Developing Inclusive and Sustainable Economic and Financial … · 2016. 5. 18. · Developing Inclusive and Sustainable Economic and Financial Systems Cite this chapter as: El-Karanshawy

Developing Inclusive and Sustainable Economic and Financial Systems

Volume 3

Islamic banking and finance – Essays on corporate finance, efficiency and product development

Editorial BoardDr. Hatem A. El-Karanshawy Dr. Azmi Omar Dr. Tariqullah Khan Dr. Salman Syed AliDr. Hylmun Izhar Wijdan Tariq Karim GinenaBahnaz Al Quradaghi

ISBN: 978-9927-118-23-4

Cover design: Natacha Fares

Copyright © 2015 The Authors

iii

CONTENTS

Foreword v

Acknowledgments vii

Preface ix

Introduction xi

PART 1: ISLAMIC CAPITAL MARKETS AND CORPORATE FINANCE

Chapter 1 Valuation of Islamic debt instruments, the Sukuk: Lessons for market developmentMohamed Ariff and Meysam Safari 1

Chapter 2 The impact of Islamic debt on company valueFitriya Fauzi, Stuart Locke, Abdul Basyith and Muhammad Idris 19

Chapter 3 Islamic financing and bank characteristics in a dual banking system: Evidence from MalaysiaMuhamed Zulkhibri 37

Chapter 4 Leverage risk, financial crisis, and stock returns: A comparison among Islamic, conventional, and socially responsible stocksVaishnavi Bhatt and Jahangir Sultan 47

Chapter 5 Is Shariah-compliant investment universally sustainable? A comparative studyMehdi Sadeghi 81

PART 2: ISLAMIC BANKING – EFFICIENCY, PROFITABILITY & INTERNATIONAL PERSPECTIVES

Chapter 6 The nexus between economic freedom and Islamic bank performance: Empirical evidence from the MENA banking sectorsFadzlan Sufian, Muhamed Zulkhibri and Abdul Majid 93

Chapter 7 Efficiency of performance of banks in the Gulf region before, during and after crises (financial and political)Abdel Latef Anouze 111

Chapter 8 The relationship between Islamic bank efficiency and stock market performance: Evidence from GCC countriesSamir Srairi, Imen Kouki and Nizar Harrathi 125

Chapter 9 Conventional banks versus Islamic banks: What makes the difference?Huseyin Aytug and Huseyin Ozturk 137

PART 3: PRODUCT DEVELOPMENT IN ISLAMIC FINANCE

Chapter 10 Estimating expected returns on Mudaraba time deposits of Islamic banksZeynep Topaloglu Calkan 151

Chapter 11 Self-adjusting profit sharing ratios for Musharakah financingVolker Nienhaus 161

Chapter 12 Indexing government debt to GDP: A risk sharing mechanism for government financing in Muslim countriesSyed Aun R. Rizvi and Shaista Arshad 173

Chapter 13 Concept and mathematics of Islamic valuation and financial engineeringNadi Serhan Aydin, FRM and Martin Rainer 181


Cite this chapter as: El-Karanshawy H A (2015). Foreword. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

ForewordHatem A. El-KaranshawyFounding Dean, Qatar Faculty of Islamic Studies, Hamad bin Khalifa University, Qatar Foundation, Doha

The International Conference on Islamic Economics and Finance (ICIEF) is the leading academic conference in the discipline organized by the International Association for Islamic Economics (IAIE) in collaboration with other key stakeholders, including the Islamic Research and Training Institute, Islamic Development Bank. It is the pioneering international conference on Islamic economics organized first in Makkah Al Mukaramah, Kingdom of Saudi Arabia, in 1976 under the auspices of King Abdulaziz University and has since been held in numerous locations around the world. The conference as such has contributed immensely to the promotion of Islamic economics and finance. Since 2011, the Qatar Faculty of Islamic Studies (QFIS), of Hamad bin Khalifa University, Qatar Foundation, has also become a key partner in organizing the conference.

The global economy continues to face the perennial problems of poverty, persistent youth unemployment, excessive inequalities of income and wealth, high levels of inflation, large macroeconomic and budgetary imbalances, exorbitant debt-servicing burdens, inadequate and aging public utilities and infrastructure, skyrocketing energy prices, and growing food insecurity. The reoccurring regional and global financial crises further intensify and magnify these problems, particularly for the underprivileged segments of the world population. As a result, many countries are at the risk of failing to achieve by 2015 the Millennium Development Goals (MDGs) set by the United Nations. Hence the achievement of an inclusive and sustainable economic and financial system has remained highly illusive.

The ICIEF presents an excellent opportunity for those interested in Islamic economics and finance to present their research and contribute to the development of an inclusive and sustainable global economic and financial

system. It is through such a setting that thoughts can be debated with the objective of advancing knowledge creation, facilitating policymaking and promoting genuine innovation for the industry and the markets. Disseminating research presented at ICIEF to the greatest number of researchers interested in the topic is important. It not only advances the discourse, but also grants those who did not have the privilege of attending the conference to partake in the discussion.

To this end, this series of five volumes (two in Arabic to follow) presents the proceedings of 8th and 9th conferences, which were held in Doha and Istanbul respectively in 2011 and 2013. Each volume focuses on a particular sub-theme within the broader theme of Developing Inclusive and Sustainable Economic and Financial Systems.

The volumes are as follows:

Volume 1: Access to Finance – Essays on Zakah, Awqaf and Microfinance

Volume 2: Islamic Economics and Social Justice – Essays on Theory and Policy

Volume 3: Islamic Banking and Finance – Essays on Corporate Finance, Efficiency, and Product Development

Volume 4: Ethics, Governance, and Regulation in Islamic Finance

Volume 5: Financial Stability and Risk Management in Islamic Financial Institutions

We hope that this academic endeavor in partnership with the Bloomsbury Qatar Foundation Publishing will benefit the Islamic economics and finance community and policy makers and that it will promote further academic study of the discipline.


Cite this chapter as: Khan T (2015). Acknowledgements. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

AcknowledgementsTariqullah KhanPresident, International Association for Islamic Economics

At the International Association for Islamic Economics (IAIE), we are grateful to acknowledge the unprecedented success of the 8th and 9th International Conferences on Islamic Economics and Finance, which were respectively organized in the Qatar National Convention Centre, Doha, December 19–21, 2011, and in the WoW Convention Centre Istanbul, September 9–10, 2013. We greatly appreciate the financial, academic and logistic support provided by the Qatar Faculty of Islamic Studies, Hamad bin Khalifa University at Qatar Foundation; Islamic Research and Training Institute at the Islamic Development Bank; and the Statistical, Economic and Social Research and Training Centre for Islamic Countries.

We offer our sincere thanks to the sponsors of the 8th International Conference on Islamic Economics and Finance in Doha. Without their partnership and generous contributions, the conference would not have been possible. In addition to the Qatar Foundation and the Islamic Development Bank, other sponsors included: Qatar Central Bank (QCB), Qatar Financial Centre Authority (QFCA), Qatar National Research Fund (QNRF), Qatar National Bank, Qatar Islamic Bank, Qatar International Islamic Bank, Masraf Al Rayan, and Qatar Airways.

We owe our deepest gratitude to the highly-esteemed panel of reviewers who volunteered to dedicate their time and energy in reviewing all the thousands of abstracts and papers that were submitted to the conferences. The reviewers of the English papers and abstracts included: Abdallah Zouache, Abdel Latef Anouze, Abdelaziz Chazi, Abdul Azim Islahi, Abdullah Turkistani, Abdulrahim AlSaati, Ahmet Tabako lu, Anowar Zahid, Asad Zaman, Asyraf Dusuki, Ercument Aksak, Evren Tok, Habib Ahmed, Hafas Furqani, Hafsa Orhan Astrom, Haider Ala Hamoudi, Hossein Askari, Humayon Dar, Ibrahim Warde, Iraj Toutounchian, Jahangir Sultan, John Presley, Kabir Hassan, Karim Ginena, Kazem Yavari, Kenan Bagci, Mabid Al-Jarhi, Maliah Sulaiman, Marwan Izzeldin, Masooda Bano, Masudul Alam Choudhury, Mehdi Sadeghi, Mehmet Asutay, Moazzam

Farooq, Mohamad Akram Laldin, Mohamad Aslam Haneef, Mohamed Ariff Syed Mohamed, Mohammed Benbouziane, Mohammed El-Komi, Monzer Kahf, Muhammad Syukri Salleh, Murat Çizakça, Nabil Dabour, Nafis Alam, Nasim Shirazi, Nazim Zaman, Necdet Sensoy, Nejatullah Siddiqi, Rifki Ismal, Rodney Wilson, Ruhaya Atan, Sabur Mollah, Salman Syed Ali, Savas Alpay, Sayyid Tahir, Serap Oguz Gonulal, Shamim Siddiqui, Shinsuke Nagaoka, Simon Archer, Tariqullah Khan, Toseef Azid, Turan Erol, Usamah Ahmed Uthman, Volker Nienhaus, Wafica Ghoul, Wijdan Tariq, Zamir Iqbal, Zarinah Hamid, Zeynep Topaloglu Calkan, Zubair Hasan, and Zulkifli Hasan. The reviewers of the Arabic papers and abstracts included Abdelrahman Elzahi, Abdulazeem Abozaid, Abdullah Turkistani, Abdulrahim Alsaati, Ahmed Belouafi, Ali Al-Quradaghi, Aly Khorshid, Anas Zarqa, Bahnaz Al-Quradaghi, Layachi Feddad, Mabid Al-Jarhi, Mohammed El-Gamal, Nabil Dabour, Ridha Saadallah, Sami Al-Suwailem, Seif El-Din Taj El-Din, Shehab Marzban and Usamah A. Uthman.

The primary objective of the conferences is to further the frontiers of knowledge in the area of Islamic economics and finance. Without the hard work and creativity of the researchers who shared their work with us, the pool of knowledge generated in the form of the conference papers and presentations would not have been possible. We thank all the authors who submitted their abstracts and papers to the two conferences.

The IAIE has always endeavored to publish most of the significant research papers contributed to its conferences. Currently the selected papers of the 8th and 9th conference are being published in five volumes under the common theme of Developing Inclusive and Sustainable Economic and Financial Systems. On behalf of the Editorial Board we acknowledge that the partnership with the Bloomsbury Qatar Foundation Publishing in this regard will be highly beneficial in disseminating research output and in promoting the academic cause.


Cite this chapter as: Omar M A (2015). Preface. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

PrefaceDr. Moh’d Azmi OmarDirector General, Islamic Research & Training Institute (IRTI)

Research in the area of Islamic banking and finance has mushroomed since the new millennium. As the Islamic banking industry continues to grow at impressive rates, researchers have pondered on fundamental questions relating to whether Islamic banks are unique from an economic standpoint. Islamic banks have been compared to their conventional counterparts on various measures such as efficiency and profitability. Capital market products such as Islamic mutual funds have also been the subject of recent research. In order to achieve favorable outcomes for the industry, it is important to empirically explore and determine the economic characteristics of Islamic banking and finance. This may inform how we may go about determining the usefulness of a company to issue Islamic debt, for example, and the impact that such an issuance would have on the stock performance of the company. In most parts of the world, Islamic banks operate in a dual banking system together with conventional banks. Researchers may be curious to explore whether operating in a dual banking system affects the type of Islamic banks that exist in the economy and the kind of products that they offer. In light of the recent LIBOR scandals, debates surrounding benchmarking of Islamic products will recapture the attention of researchers; hence, product development in Islamic banking will also remain an important area. With the increasing availability of data, empirical research in Islamic banking is expected to continue to inform industry stakeholders. Understanding the Islamic banking phenomenon from an economic perspective with the support of empirical evidence may

help to align the practices of Islamic banks to its originally intended objectives of offering a just and equitable form of financing based on sound ethical principles.

This volume titled Islamic Banking and Finance – Essays on Corporate Finance, Efficiency, and Product Development aims to bring together papers on these themes. The volume consists of selected papers from the 8th International Conference on Islamic Economics and Finance held in Doha during 19–21 December 2011 and from the 9th International Conference on Islamic Economics and Finance held in Istanbul during 9–11 September 2013. The papers are presented here in their original form as presented at the conferences, with changes limited to copyediting and correcting typographical errors. The conferences were organized by the Center for Islamic Economics and Finance, Qatar Faculty of Islamic Studies (QFIS), Hamad bin Khalifa University; Islamic Research and Training Institute (IRTI), Islamic Development Bank (IDB); International Association for Islamic Economics and Finance (IAIE); and Statistical, Economic, and Social Research and Training Centre for Islamic Countries (SESRIC).

We have selected the papers in this volume to reflect the diversity of research in the field of Islamic banking and finance. We hope that the papers that have been selected here will help to inform and motivate researchers to conduct rigorous empirical research to explore questions that remained to be answered.


Cite this chapter as: Tariq W (2015). Islamic banking and finance – Essays on corporate finance, efficiency and product development: An introduction to the issues and papers. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Islamic banking and finance – Essays on corporate finance, efficiency and product development: An introduction to the issues and papers

Wijdan TariqSenior Researcher, Center for Islamic Economics and Finance, Qatar Faculty of Islamic Studies, Hamad bin Khalifa University

Academic interest in Islamic banking and finance has grown over the years. The appeal of Islamic banking to the Muslim world is decades old. However, curiosity from the wider academic community has increased particularly in the aftermath of the global financial crisis and ongoing Eurozone crises. In recent years, prominent academics from conventional economics have either written papers on Islamic finance or commented in conferences on the viability of the Islamic finance proposition (e.g., Beck, et al., 2013; Abedifar, et al., 2013; Zaheer, et al., 2013; Ongena and Sendeniz-Yüncü, 2011; Rogoff, 2011a; Rogoff, 2011b; Roubini, 2013, Baele, et al., 2014) This is a welcome development and may help add more rigor and diversity to the Islamic finance discourse.

Due to the increasing availability of data, empirical work in Islamic finance is on the rise. The selection of papers in this volume reflects this empirical trend. The majority of the existing literature attempts to answer questions of a comparative nature. Islamic banks are often compared to conventional banks on various measures. Which banks are more stable, more profitable, and more efficient? A more fundamental question that some academics have asked is: Are Islamic banks unique to begin with?

Efficiency and profitability comparisons are perhaps the most common type of study in the field.1 Generally speaking, the more robust studies seem to show that there were few differences between Islamic and conventional banks before the crisis. During the crisis, however, evidence seems to suggest that Islamic banks were less profitable. The size of the bank is an important control variable in such comparative studies. Studies have shown that small Islamic banks appear to perform better than small conventional banks, but large conventional banks appear to perform better than large Islamic banks. Furthermore, recent anecdotal evidence appears to suggest that Islamic banks may face challenges in liquidity management, hitherto an area of lesser concern (Ali, 2013). Product development will thus remain a key topic for research in Islamic banking for reasons related to risk management as well as increasing competitive advantage.

While organizing these conferences and selecting the papers for this volume series, a deliberate attempt was

made at highlighting policy-relevance of research where possible. The first volume covers issues related to access to finance and human development including essays on zakah, awqaf, and microfinance. The second volume is a selection of academic and policy-relevant papers on economic thought from an Islamic perspective with discussions on topics such as fiscal and monetary policy, among others. The third volume (this volume) presents a wealth of empirical evidence on various issues related to Islamic banks such as profitability, efficiency, product development, and Islamic corporate financing. The fourth volume touches upon ethical, governance, and regulatory issues in Islamic finance. The fifth volume is about financial stability and risk management in Islamic financial institutions.

In the following sections, I briefly introduce the papers in this volume. Part 1 includes a selection of papers on Islamic capital markets and corporate finance. Part 2 contains empirical papers on Islamic banking – topics such as efficiency and profitability comparisons between Islamic and conventional banks. Part 3 comprises papers on product development in Islamic banks using innovative approaches.

1. Islamic capital markets and corporate financeThe first part of this volume is a collection of papers on Islamic capital markets and corporate finance. Islamic capital markets refer to those capital market products that comply with Islamic finance principles. These investment products can be categorized into two broad asset classes: Islamic equity markets and Islamic debt markets.

The Islamic equity markets are stocks and mutual funds that pass certain financial and nonfinancial screening criteria that render them to be permissible for investment by Islamic institutional and individual investors. Financial screening of stocks is generally based on Shariah principles that discourage debt and prohibit interest-based transactions. In identifying a stock as being Shariah-compliant, a common financial screen would be, for example, the debt-to-total asset ratio. If this ratio exceeds 33% for a company, its stock may be deemed impermissible for investment. Due to the nature

Tariq

xii Islamic banking and finance – Essays on corporate finance, efficiency and product development

of modern enterprise, interest income may be tolerated up to a certain percentage, e.g., as long as interest income is less than 5% of all income. Nonfinancial screens are based on the underlying business that the company is engaged in. All industries are considered permissible as long as there is no contravening of basic Shariah principles regarding trade. If a company is engaged in the trading of goods and services deemed impermissible by Shariah, then its stock would also be impermissible. Examples of such goods and services that Shariah scholars would prohibit include alcohol, gambling, and profane/violent entertainment. Since the introduction of Dow Jones Islamic Indices, a number of similar indices have been developed in different countries. Each Islamic stock index would have slightly different financial screening criteria depending on which Shariah scholars are on the respective Shariah boards.

The second asset class in the Islamic capital markets is the Islamic debt markets, which largely constitute products known as sukuk. In theory, sukuk are meant to be securitized representations of undivided shares in an underlying asset or service. In practice, however, the majority of sukuk prospectuses are drafted by lawyers with the aim of replicating bond structures. Effectively, the majority of the over $600 billion sukuk that have been issued can be considered as the Islamic equivalent of bonds.

Conventional corporate finance research is devoted mainly to the study of valuation of securities, portfolio management, capital structure of firms, dividend policy, and market inefficiencies. A significant source of corporate financing is lending by banks; a large theoretical and empirical literature exists on this topic as well. In recent years, research in the vein of corporate Islamic financing has grown with the increasing availability of reliable data.

The first paper in this section is about the valuation of sukuk. Ariff and Safari first explore whether sukuk are equivalent to bonds. Using a unique dataset from the Bondstream database, they compare sukuk and bonds issued by the same issuers with the same rating in the same market. They find evidence suggesting that bonds and sukuk are priced differently and that sukuk yields are not determined by bond yields. Their paper is also among the very few papers that attempts to propose valuation models for pricing sukuk.

The second paper by Fauzi, Locke, Basyith and Idris is an empirical analysis that explores the impact of Islamic debt (sukuk) on the value of the issuing company. They find evidence in support of the tradeoff theory of capital structure, which posits that companies actively decide between equity or debt issuance by assessing costs and benefits. In their sample, they found evidence that suggests that the issuance of sukuk was positively associated with an improvement in the financial performance of the firm, perhaps due to the associated tax benefits. They conduct further analysis on whether subsequent sukuk issuances by the same firm also positively affect financial performance, with mixed results. This paper is relevant to readers interested in theories on capital structure and their application in Islamic finance

The third paper by Zulkhibri of the Islamic Development Bank focuses on bank financing as a source of capital. He

explores the relationship between bank financing, financing rate and bank-specific characteristics in Malaysia, a country whose financial system operates on a dual-banking system. Using Bankscope data, the result of his pooled panel estimations is that Islamic banks financing behavior is dependent on the characteristics of the banks, such as its size, liquidity, and capital. This is consistent with the behavior of conventional banks. The lending behaviors of both types of banks do not differ significantly with respect to interest rates. The author also echoes sentiments of purist Islamic economists that Islamic banks should move away from debt-based instruments and towards profit-loss-sharing in order to differentiate themselves from conventional banks.

The fourth paper by Bhatt and Sultan adds a leverage risk factor to the Fama and French multifactor asset pricing model in order to investigate whether Islamic stocks are less sensitive to leverage than other stocks such as conventional stocks and socially responsible stocks. In widely cited papers, academics have reported anomalies in their results on asset pricing models (Fama and French (1992); Lakonishok, Shleifer and Vishny (1994); Kothari, Shanken and Sloan (1995)). Despite the methodological challenges of studying asset pricing models, such as concerns about data-snooping, sample selection biases, irrational behavior of market participants, and the empirical difficulties of distinguishing between competing hypotheses, this area of research remains quite popular. The application of such factor models to Shariah-compliant stocks is a nascent area of research. In Bhatt and Sultan’s paper, readers will find a valuable reference on this topic. The authors find that Shariah-compliant stocks exhibit lower risk premiums to traditional risk factors, but that they are also sensitive to the leverage risk factor. This last result may have implications for asset management practices and will also be of interest to researchers in this field.

The final paper in this section by Sadeghi, explores a number of issues surrounding Shariah-compliant stocks. He first compares Shariah-compliant sustainable stocks and conventional sustainable stocks, and then, using event study methodology, Sadeghi investigates the market reaction to the addition and deletion of a stock from a Shariah-compliant index. It is suggested that readers should regard the results in this conference paper simply as preliminary findings. Further research is needed before drawing any conclusions.

2. Islamic banking – Efficiency, profitability and international perspectivesThe first paper in this section by Sufian and Zulkhibri presents evidence on the relationship between economic freedom and the performance of Islamic banks. Economic freedom in this paper broadly refers to the extent of private ownership of resources as compared to government control. Measurements of economic freedom are taken from the Heritage Foundation and include measures of business freedom, trade freedom, investment freedom, etc. The results of the study are mixed. The authors present evidence of a positive association between financial/business freedom and the profitability of Islamic banks in the MENA region. However, monetary freedom is negatively associated with the performance of Islamic banks, lending

Eds. Hatem A. El-Karanshawy et al. xiii


support to the benefits of government intervention. The authors also offer some preliminary policy evidence based on the results of their panel regressions.

Anouze’s paper is in application of two nonparametric techniques to measure the efficiency and performance of Islamic banks, namely: Data Envelopment Analysis (DEA) and the Classification and Regression Tree (CART). The use of DEA to analyze efficiency of Islamic banks is widely applied. It essentially involves grouping data into inputs and outputs, where input variables are the resources to be minimized and output variables are the resources and products/services to be maximized to ensure a high relative efficiency score. On the other hand, there are only a handful of papers that use the CART technique to analyze Islamic banks. This data mining technique can be used to descriptively represent and explore data in a decision-tree type of format that makes it relatively simple to understand. Combining these two techniques allows this paper to make a useful contribution to the literature. Anouze analyzes the period between 1997–2007 to explore the performance of Islamic and conventional banks during the period of geo-political crises as well as the onset of the financial crises.2 Overall, Anouze identifies 15 factors that future researchers can consider as being important in predicting efficient banks.

The third paper by Srairi, Kouki, and Harrathi explores the relationship between Islamic bank efficiency and stock market performance for 25 Islamic banks in the GCC region using Data Envelopment Analysis (DEA) during the period 2003–2009. They find evidence that suggests that pure technical efficiency is positively associated with stock returns while changes in scale efficiency have no association with stock performance. Pure technical efficiency here refers to the ability of managers to efficiently utilize the resources of the bank. Scale efficiency refers to the proportional reduction in input usage if the bank can operate at optimal scale in terms of outputs produced. So in other words, the authors find evidence that suggests that stock returns respond positively to improvements in managerial efficiency. The authors also find evidence suggestive of increasing technical efficiency over the years in the sample of Islamic banks.

Aytug and Ozturk’s paper is an empirical study investigating the differences between Islamic banks and conventional banks in Turkey between 2003 and 2011. They specifically investigate whether being Islamic has any unique effect on profitability ratios. One of the distinguishing features of this paper is the application of the propensity score matching method to estimate the average treatment effect of being an Islamic bank. This statistical matching method can overcome the concerns of self-selection bias that may occur when using more conventional techniques such as the Ordinary Least Squares (OLS) and Generalized Method of Moments (GMM) estimators. For the sake of comparison, however, the authors do also use the OLS method. According to the regression results, there is an insignificant relationship between being Islamic and profitability if profitability is measured using the Return on Assets (ROA) and Return on Equity (ROE) ratios. When using the Net Interest Margin (NIM) ratio as a measure for profitability, the results suggest that Islamic banks may have a negative relationship with NIM. The results of the

propensity score matching generally confirm the findings of the regression results. More research in this area will be required to uncover the distinctiveness of using NIM as a profitability ratio for Islamic banks as opposed to ROA and ROE, and whether such differences may be driving the results in this study.

3. Product development in Islamic financeThe first paper in this section by Calkan is an application of simple mathematical techniques to enhance the competitiveness of Islamic banking products. The mudharabah-based deposits of Islamic banks – Profit Sharing Investment Accounts (PSIAs) – have an inherent competitive disadvantage compared to conventional time deposits because profit-rates are not known in advance, which may deter some customers. Calkan proposes that while Shariah scholars prohibit the fixing of profit rates in mudharabah transactions in advance, there is no prohibition on reliably estimating the expected profit rates and disclosing this to the customer. Through simulation techniques, the author provides results suggesting that Islamic banks may be able to estimate the PSIA profit rates within a 95% confidence interval. This may potentially be beneficial to banks by increasing their competitive advantage and improving their fund management; hence, this academic research paper has direct practical implications for the industry.

Nienhaus’s paper is an innovative proposal to structure a sukuk product based on musharakah with a self-adjusting profit sharing ratio. He promotes the use of profit-loss sharing (PLS) products, which are viewed by some Islamic economists as the ideal form of Islamic financing. However, the uptake of these products has been met with resistance from lenders due to the adverse selection and moral hazard problems that PLS financing arrangements pose. Nienhaus’s proposal of self-adjusting profit sharing ratios is an attempt to overcome the information problems in standard musharakah financing arrangements. Interestingly, the concept is based upon existing Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI) standards that allow for parties in a musharakah arrangement to adjust profit-sharing ratios, if mutually agreeable. Nienhaus puts forward a formula (which can be viewed as a customizable template) that would automatically adjust the profit-sharing ratio as new information emerges on the profitability of the venture. This method, he argues, would be less costly than ex-post renegotiations regarding the profit-sharing ratios, because in ex-post renegotiations, each party would strive to protect their gains/losses at the expense of the other party. On the other hand, the ex-ante model of self-adjusting ratios proposed by Nienhaus links the profit-sharing ratio to the actual performance of the venture, so the adjustment is non-discretionary. Prior research has been limited to theoretical treatment of the topic (Ahmed, 2002); hence, readers would welcome the numerical example that Nienhaus provides.3 In short, PLS financing arrangements do have a role to play in the financial system particularly in the financing of Small and Medium Enterprises (SMEs), which have been shown to be disproportionally affected by economic downturns. Researchers would be keen to build upon the model provided here to develop more feasible PLS financing products for Islamic banks to use.

Tariq

xiv Islamic banking and finance – Essays on corporate finance, efficiency and product development

The third paper by Rizvi and Arshad is a proposal to introduce Gross Domestic Product (GDP)-linked government securities. The authors also attempt to test the performance/benefits of their proposal using simple simulations on four countries, as well as using descriptive statistics and correlation analysis. Their focus is on the introduction of a GDP-linked sukuk for government financing. The idea is that indebted countries may face economic shocks that reduce their ability to repay international debt. If the rate of repayment is linked to the country’s GDP as opposed to an external benchmark such as the London Interbank Offered Rate (LIBOR), this would ensure greater risk-sharing between the sovereign borrower and international lenders. In times of economic downturns, for example, the repayments on a GDP-linked sukuk would be lower. The authors provide an overview of GDP-linked sukuk. They present some Shariah issues that may arise in structuring such products. They provide simple simulations to show the potential benefits for countries that adopt GDP-linked sukuk. Finally, the authors realize that investors/lenders would be reluctant to partake in greater risk sharing; hence, they discuss the viability of such products from the perspective of lenders. More work would be needed to ensure the viability of these products; in particular, the Shariah issues that would arise from a musharakah-based GDP-linked sukuk merits a lengthier discussion than the authors have provided. Nonetheless, the paper makes a reasonable contribution to the growing literature on GDP-linked securities.

The final paper in this section is by Aydin and Reyner. The paper makes a contribution to the Islamic financial engineering literature. Unlike previous studies, this paper uses a rigorous mathematical framework of risk-neutral valuation and arbitrage-free contract prices. The authors argue that the reference to risk-free zero bonds in conventional finance is unrealistic. Instead, the authors propose the use of a commodity basket-linked currency, which they describe in detail in the paper. The authors compare valuation of a conventional forward contract with a forward contract from an Islamic perspective and discuss the challenges of using a commodity basket in the latter case, particularly because their forward prices may not be available. Stochastic processes may be useful in this instance for estimation purposes. In short, readers interested in a financial mathematics and probability theory treatment of Islamic financial engineering will find this paper a valuable starting point. The use of financial mathematics in Islamic finance literature is scarce indeed; this fact alone makes this paper noteworthy.

3. ConclusionThe papers presented in this volume represent significant contributions to the body of research in Islamic banking, and suggest fruitful areas of future research. Looking ahead, the lack of standardized sukuk contracts and the absence of liquidity in these markets calls for further research in financial engineering and regulatory frameworks. There are continuing calls for Islamic banks to differentiate themselves from conventional banks in order to reduce their dependence on markup-based debt financing; this has been a contentious issue that has been ongoing for decades and that will remain a topic for researchers and practitioners to

ponder in the spirit of academic and intellectual discourse. It is expected that new large-scale evidence on the effects of high levels of debt in the economy (including household debt) will help to inform this debate. The topic of Shariah equity screens will continue to evolve and there is expected to be a convergence between Shariah-compliant and socially responsible investing. In addition to prevailing negative stock screening approaches, perhaps future research will explore the viability of positive-screening approaches and the role that ‘impact investing’ has to play in Islamic equity markets. Questions on how policy-makers can promote economic freedom and protect shareholders and creditors rights in a way that makes it attractive for Islamic banks in particular to do business, will continue to be explored. Increasing the competitive advantage of Islamic banks can be achieved perhaps through product development that increases information transparency and caters to the needs of customers. Finally, attempts at using alternative pricing benchmarks for Islamic securities that tie more closely to the real economy, such as GDP-linked securities, remains a fruitful area for researchers.

Notes1. For an excellent overview of the empirical literature, I

recommend readers to refer to the chapter by Ongena and Zaheer (2013).

2. It can be argued, however, that limiting the period until 2007 may not sufficiently capture the effects of the financial crisis.

3. Diaw et al (2011) also provide hypothetical examples for their GDP-linked sukuk based on the ijarah.

ReferencesAbedifar P, Molyneux P, Tarazi A. (2013) Risk in Islamic

Banking. Review of Finance. forthcoming.

Ahmed H. (2002) Incentive-Compatible Profit-Sharing Contracts: A Theoretical Treatment. In: Munawar I, Llewellyn DT (Eds.). Islamic Banking and Finance – New Perspectives on Profit-Sharing and Risk. Edward Elgar. Cheltenham. 40–54.

Ali SS. (2013) State of Liquidity Management in Islamic Financial Institutions. Islamic Economic Studies. 21(1):63–98.

Baele L, Farooq M, Ongena S. (2014) Of Religion and Redemption: Evidence from Default on Islamic Loans. Journal of Banking & Finance. 44:141–159.

Beck T, Demirgüç-Kunt A, Merrouche O. (2013) Islamic vs. Conventional Banking: Business Model, Efficiency and Stability. Journal of Banking and Finance. 37:433–447.

Diaw A, Bacha O, Lahsasna A. (2011) Public Sector Funding and Debt Management: A Case for GDP-Linked Sukuk. Paper presented at the 8th International Conference on Islamic Economics and Finance, 19–21 December 2011, Doha, Qatar.

Fama E, French K. (1992) The Cross-Section of Expected Stock Returns. Journal of Finance. 47(2):427–465.

Kothari S, Shanken J, Sloan R. (1995) Another Look at the Cross-Section of Expected Stock Returns. Journal of Finance. 50(1):185–224.

Eds. Hatem A. El-Karanshawy et al. xv


Lakonishok J, Shleifer A, Vishny R. (1994) Contrarian Investment, Extrapolation, and Risk. Journal of Finance. 49(5):1541–1578.

Ongena S, Sendeniz-Yüncü I. (2011) Which Firms Engage Small, Foreign, or State Banks? And Who Goes Islamic? Evidence from Turkey. Journal of Banking and Finance. 35:3213–3224.

Ongena S, Zaheer S. (2013) Chapter 4: Empirical evidence. In: Islamic Finance in Europe, Occasional paper series. No. 146, June. European Central Bank.

Rogoff K. (2011a). Global Imbalances Without Tears. Project Syndicate. Available at: http://www.project-syndicate.org /commentary/global-imbalances-without-tears.

Rogoff K. (2011b) What Imbalances After the Crisis? Speech at International Symposium of the Banque de France –

Regulation in the Face of Global Imbalances. Banque de France, 4 March 2011. Paris, France. Available at: https://www.banque-france.fr/fileadmin/user_upload/banque_de_france/Economie-et-Statistiques/La_recherche/GB/Session1-Rogoff.pdf.

Roubini N. (2013) There is a Lot Islamic Finance Can Teach Us. Blog. Available at: http://www.roubiniblog.com/2013/12/roubini-there-is-lot-islamic-finance.html.

Zaheer S, Ongena S, Van Wijnbergen S. (2013) The Transmission of Monetary Policy Through Conventional and Islamic Banks. International Journal of Central Banking. 8(5):175–224.


Cite this chapter as: Ariff M, Safari M (2015). Valuation of Islamic debt instruments, the Sukuk: Lessons for market development. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Valuation of Islamic debt instruments, the Sukuk: Lessons for market developmentMohamed Ariff1, Meysam Safari2

1Faculty of Business, Bond University, Gold Coast, Australia, T: (617)-5595–2296, E: [email protected] University, Selangor, Malaysia, T: (60)-17271–4676, E: [email protected]

Abstract - A key issue for the fast growing Islamic debt market is whether sukuk instruments are equivalent to conventional bonds as is practiced today by the market operators. This major research question is explored here using a large traded data set on sukuk matched with a sample of conventional bonds issued by the same issuers with same risk-rating and traded in the same market. Matched samples of sukuk and conventional bonds traded over seven years are used in our analyses, using statistical and causality tests. The results suggest that sukuk instruments are priced significantly differently and that their yields are not Granger-caused by conventional security yields or vice versa. This empirical finding does not support the market’s current practices based on the assumption that sukuk are like normal bonds. In addition, we describe the basic contract specifications, core cash flow structures in order to develop and suggest valuation models for three selected sukuk types. The major implication of these findings is that there is much more research needed to first document the independent nature of sukuk market behavior and the need for a re-examination of current market practices.

Keywords: Sukuk, bond, yield curve, yield to maturity, Islamic finance, fixed income securities, securitization

1. IntroductionSukuk is the plural form of sakk, which in Arabic means legal instrument, deed, or cheque.1 It was used in pre-Islamic era as withdrawal certificate (a form of cheque) on deposits in financial firm or authorized merchants. Later, these certificates became instrument for trading. And then it became debt instruments traded among willing holders of already issued debt just as is a conventional bond.

Being a relatively new product, there is no consensus on the exact definition of sukuk. Some of international regulatory bodies such as AAOIF, attempted to define sukuk. Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI, 2004) defines sukuk as “Certificates of equal value representing, after closing subscription, receipt of the value of the certificates and putting it to use as planned, common title to shares and rights in tangible assets, usufructs and services, or equity of a given project or equity of a special investment activity.” International Islamic Financial Market (IIFM, 2010) defined sukuk as a commercial paper that provides an investor with ownership in an underlying asset. It is asset-backed trust certificate evidencing ownership of an asset or its usufruct (earnings or fruits). It has a stable income and complies with the principle of Shariah. Unlike conventional bonds, sukuk needs to have an underlying

tangible asset transaction either in ownership or in a master lease agreement.

Islamic Financial Services Board (IFSB, 2009) definition of Sukuk is “Sukuk (plural of sakk), frequently referred to as “Islamic bonds,” are certificates with each sakk representing a proportional undivided ownership right in tangible assets, or a pool of predominantly tangible assets, or a business venture (such as a Mudarabah). These assets may be in a specific project or investment activity in accordance with Shariah rules and principles.”

2. Sukuk securities are not the same as bondsThis section provides empirical evidences that sukuk securities are different from conventional bonds. The empirical evidence is first shown as yield curves; second, paired sample mean test between pairs of sukuk and bonds issued by same issuer and for the same maturity period; and third, Granger causality test between abovementioned security pairs. Tests were conducted using data on aggregated monthly yield to maturity (YTM) data2 from Malaysian securities over August 2005 to April 2013 (total of 93 observations for each security).

Ariff and Safari

2 Islamic banking and finance – Essays on corporate finance, efficiency and product development

Descriptive statisticsDescriptive statistics for various sukuk securities and conventional bonds are presented in Table 1. The statistics suggest that the mean yield of sukuk securities for all types of issuers and for all forms of maturities is 3.92 percent. The yields vary within a minimum of 2.91 (3 months maturity Treasury sukuk securities) and the maximum of 5.66 (sukuk securities issued by AAA rated corporations with 20-year maturity). On the other hand, the mean yield of conventional bonds of all types of issuers and all maturities is 3.91 percent. These vary between a minimum of 2.90 (3 months maturity Treasury bills) and the maximum of 5.60 (conventional bonds issued by AAA rated corporate with 20-year maturity). At the issuer level, AAA rated corporate issue of sukuk securities yielded 4.29 percent while the mean of Treasury bill yields is 3.55 percent. On the other hand, the highest conventional mean yield for AAA rated corporate issuers is 4.31 percent while the lowest mean yield for conventional bills and notes issued is by the Government of Malaysia with 3.51 percent.

Yield curvesThe result data are presented in this section, as yield curves in two plots. Yield curve is the relation between the cost of borrowing and the time to maturity of a debt security for a given issuer for a given rating quality. Yield curves for sukuk securities and conventional bonds issued by government and corporations are plotted as in Figure 1 A and Figure 1B. The plots are presented as YTM of (i) conventional against (ii) sukuk issues in two plots. The two issuer types are of increasingly higher risk rating with sovereign being the lowest risk – therefore with the lowest yields – on the one end, and the AAA corporate issues with higher yield at the other end.

As Figure 1(A) suggests, the yields of Government Islamic Issues (GII) are higher than those of conventional bonds issued by the same issuer (Malaysian Government Securities, or MGS). The difference between sukuk yield and conventional bond yield is larger as maturities increase from 2 years to 15 years. The maximum difference between the yields of sukuk securities and those of conventional bonds for this category is with 3 years maturities. The difference is 8.28 basis points. On average, there is a 4.04 basis points difference between yields of sukuk securities and conventional bonds. The total outstanding value of sukuk securities issued by Malaysian government as at April 2013 was RM 155.9 billion (US$ 55 billion).

Multiplying yield difference and market size indicates that the Malaysian government needs to pay an extra RM 6.3 billion per year to investors holding sukuk securities compared to the amount conventional issues of same term and quality. This means that the sukuk investors earn RM 6.3 billion higher return compared to the investors in conventional bond market. Obviously sukuk yields are systematically higher.

Figure 1(B) shows a plot of the yields of securities issued by AAA rated corporate issuers. Yields of sukuk securities are less than yields of conventional bonds for maturities of less

than 10 years; it is more for periods beyond 10 years. The maximum difference between the yields of sukuk securities and the conventional bonds issued by corporate issuers with maturities less than 10 years is for those with 2-year maturity with a -7.28 basis points. However, the maximum amount for securities with maturities longer than 10-year term is +6.38 basis points for securities with 20-year maturity. Long-dated sukuk securities are perceived by the market as being more risky, thereby attracting higher yields. Long dated sukuk are perhaps more risky given the risk of greater uncertainty beyond 10 years. It is puzzle why the same firms issuing short-dated securities provide a safer investment. Obviously again, the yield differences are systematic.

On average, there is a -2.44 basis points difference between yields of sukuk securities and conventional bonds. The total outstanding value of sukuk securities issued by Malaysian AAA Corporate issuers in April 2013 was RM 64.48 billion (US$ 22 billion). Multiplying yield difference and the market size indicates that the AAA corporate issuers would save RM 1.57 billion (US$ 0.53 billion) per year on their sukuk securities.

Comparison of yields of Sukuk securities and conventional bondsResults of the paired sample t-tests are summarized in Table 2 (panels A and B) on the equality of means. Out of the 20 tested pairs of mean yields of sukuk and conventional bonds, 19 cases (95 percent) of all pairs showed significant differences in their yields to maturities. In 15 cases, the null hypotheses are rejected at 0.01 significance levels. Thus, one can conclude that the yields of sukuk securities differ from the case of conventional bonds, although the issuer and the issue tenure are the same. Table 2 A is a summary of the statistics pertaining to the mean yield of sukuk and conventional bonds.

As the t-statistics suggests, the mean yield of sukuk securities and conventional bonds are significantly different for all issues by Government. The difference between the means

of these are positive, indicating that sukuk securities tend to yield more than conventional bonds issued by the Government of Malaysia, ceteris paribus. Thus, the market associates higher risks (for reasons, we still do not know) to sukuk structures rather than conventional structures.

Table 2B is a summary of the statistics for sukuk securities and conventional bonds issued by AAA rated corporations. For these securities, the mean yields of sukuk securities and conventional bonds are significantly different in all cases except in the cases of 10-year maturity. The differences are negative for securities with tenure 7 year or less, while, for securities with 10 years maturity or more, the difference is positive. This means the mean of yield of sukuk securities issued by AAA rated corporations is lower than the yield of conventional bonds for issues with 7 years or less maturity. For the securities with long-term maturities (10 years and more) the mean yield of sukuk securities is higher than the conventional bonds.

Eds. Hatem A. El-Karanshawy et al. 3

Valuation of Islamic debt instruments, the Sukuk: Lessons for market development

Tabl

e 1.

Des

crip

tive

sta

tist

ics

of S

ukuk

vs.

Con

vent

iona

l bon

ds.

Issu

erTe

nure

Mea

nM

edia

nSt

d. D

evR

ange

Min

Max

Mea

nM

edia

nSt

d. D

evR

ange

Min

Max

Con

vent

iona

lSu

kuk

Gov

ernm

ent

3-M

onth

2.90

32.

890.

519

1.84

1.82

3.66

2.91

832.

890.

538

1.96

1.82

3.78

6-M

onth

2.94

22.

920.

527

1.98

1.85

3.83

2.95

442.

920.

544

2.04

1.85

3.89

1-Ye

ar3.

015

2.99

0.52

52.

061.

923.

983.

0280

2.99

0.53

02.

111.

974.

082-

Year

3.17

3.1

0.44

72.

12.

24.

33.

2039

3.11

0.43

62

2.3

4.3

3-Ye

ar3.

315

3.21

0.38

02.

132.

374.

53.

3976

3.31

0.34

01.

842.

634.

475-

Year

3.56

43.

50.

318

1.8

2.78

4.58

3.63

593.

630.

310

1.8

2.85

4.65

7-Ye

ar3.

759

3.74

0.31

91.

842.

914.

753.

8003

3.75

0.31

21.

793

4.79

10-Y

ear

3.94

33.

950.

391

1.93

3.09

5.02

3.99

673.

990.

358

1.81

3.17

4.98

15-Y

ear

4.18

64.

130.

401

1.77

3.35

5.12

4.22

854.

180.

374

1.67

3.45

5.12

20-Y

ear

4.35

74.

30.

393

1.58

3.6

5.18

4.39

464.

320.

372

1.5

3.68

5.18

Mea

n3.

515

3.47

30.

422

1.90

2.58

4.49

3.55

63.

509

0.41

11.

852.

674.

45C

orpo

rate

3-M

onth

3.33

943.

330.

453

2.17

2.28

4.45

3.31

333.

290.

455

2.17

2.24

4.41

6-M

onth

3.43

333.

390.

412

1.9

2.6

4.5

3.40

093.

350.

405

1.9

2.56

4.46

1-Ye

ar3.

6178

3.52

0.37

51.

593.

14.

693.

5534

3.47

0.35

91.

613.

044.

652-

Year

3.84

543.

70.

359

1.61

3.32

4.93

3.77

263.

670.

324

1.56

3.3

4.86

3-Ye

ar4.

0665

3.97

0.36

61.

613.

625.

234.

0016

3.93

0.32

21.

53.

625.

125-

Year

4.35

764.

270.

385

1.57

3.87

5.44

4.30

824.

240.

352

1.56

3.84

5.4

7-Ye

ar4.

6380

4.57

0.42

01.

74.

035.

734.

5942

4.53

0.40

51.

694

5.69

10-Y

ear

4.95

294.

930.

477

1.85

4.21

6.06

4.94

334.

930.

483

1.84

4.18

6.02

15-Y

ear

5.30

065.

310.

494

1.98

4.41

6.39

5.35

595.

30.

580

2.15

4.38

6.53

20-Y

ear

5.60

135.

60.

521

2.12

4.61

6.73

5.66

515.

660.

573

2.11

4.58

6.69

Mea

n4.

315

4.25

90.

429

1.81

3.60

5.41

4.29

14.

237

0.42

51.

803.

575.

38

Ariff and Safari


Figure 1. Yield curve for Sukuk securities vs. Conventional bonds.

Yield curve of government issued securities Yield curve of corporate issued securities

2.8

3

3.2

3.4

3.6

3.8

4

4.2

4.4

A B

Government Conventional IssuesGovernment Sukuk Issues

Corporate BondsCorporate Sukuk

3.2000

3.7000

4.2000

4.7000

5.2000

5.7000

0 5 10 15 20 0 5 10 15 20

Table 2A. Paired samples T-test results: Government.

Tenure Sukuk ConvΔ (Sukuk - Conv) t-Stat

3M 2.9183 2.9031 0.0152 2.789***6M 2.9544 2.9424 0.0120 2.461**1Y 3.0280 3.0147 0.0132 2.419**2Y 3.2039 3.1699 0.0340 4.789***3Y 3.3976 3.3148 0.0828 7.686***5Y 3.6359 3.5641 0.0718 8.465***7Y 3.8003 3.7588 0.0415 6.418***10Y 3.9967 3.9433 0.0533 7.037***15Y 4.2285 4.1865 0.0420 5.887***20Y 4.3946 4.3567 0.0380 5.043***

**, ***: Significant at 0.05, and 0.01 significance levels, respectively.

Table 2B. Paired samples T-test results: Corporate issues.

Tenure Sukuk ConvΔ (Sukuk -Conv) t-Stat

3M 3.3133 3.3394 -0.0260 -6.173***6M 3.4009 3.4333 -0.0325 -7.483***1Y 3.5534 3.6178 -0.0644 -7.360***2Y 3.7726 3.8454 -0.0728 -6.789***3Y 4.0016 4.0665 -0.0648 -6.047***5Y 4.3082 4.3576 -0.0495 -5.685***7Y 4.5942 4.6380 -0.0438 -7.318***10Y 4.9433 4.9529 -0.0096 -0.95315Y 5.3559 5.3006 0.0553 2.414**20Y 5.6651 5.6013 0.0638 2.571**

**, ***: Significant at 0.05, and 0.01 significance levels, respectively.

Granger causality test of yields of Sukuk and conventional bondsThe previous section showed that the mean yield of sukuk is statistically different from yield of conventional bonds. Since each pair of securities issued by the same issuer for the same period of time and for same rating, it is expected that the correlation between yields of these securities may be high. This may be a valid reason for a hypothetical argument that this difference arises from an unidentified causal relationship. One may wish to test if changes in yield of one type of security may cause change in the other series? In other words, one may want to test for Granger Causality (Granger, 1969) between yields of sukuk securities and those of conventional bonds to identify if some causal link exists.

In order to test this, two Granger causality tests were conducted on each pair of securities. First test is on the

change in yield of sukuk to verify if it causes a change in yield of conventional bonds. Second test is on the change in yield of conventional bonds to verify if it causes a change in yield of sukuk. The latter test is on the yields of conventional bonds Granger causing yields of sukuk. Results of pair-wise Granger causality test on each pair is presented in Table 3.

The first null hypothesis tested was “yield of sukuk security does not Granger cause the yield of conventional bond counterparts.” As the statistics in Table 3 suggest, out of 20 pairs of securities tested, the null hypothesis are rejected in only 6 pairs at the 0.10 significance level. Yields of sukuk securities Granger cause yield of conventional bonds in only 6 out of 20 (or 30 percent) pairs. This indicates that one may not generally conclude that yield of sukuk securities Granger cause the yields of conventional bonds. Results also



Table 3. Pair-wise granger causality tests with lags = 2.

Sukuk security does not Granger cause conventional bond

Conventional bond does not Granger cause sukuk security

Issuer Maturity F-Statistic Prob F-Statistic Prob

Government 3M 0.190 0.8266 0.8972 0.41156M 0.7611 0.4702 1.2937 0.27951Y 2.5477** 0.0842 1.5399 0.22022Y 3.7064** 0.0286 3.7080** 0.02853Y 2.4141* 0.0955 5.1703*** 0.00765Y 3.6773** 0.0293 0.8004 0.45257Y 0.6627 0.5180 0.0314 0.969010Y 0.1797 0.8358 0.2506 0.778815Y 0.1263 0.8815 0.8322 0.432520Y 0.2575 0.7736 1.1982 0.3067

Corporate 3M 0.0684 0.9339 1.6904 0.19056M 0.9525 0.3898 3.3457** 0.03991Y 3.1337** 0.486 5.6886*** 0.00482Y 3.0097* 0.0545 1.6613 0.19593Y 0.8961 0.4119 0.7585 0.47145Y 1.0998 0.3376 0.3039 0.73877Y 1.7843 0.1724 1.2941 1.279410Y 0.0696 0.9327 0.7896 0.457315Y 0.6228 0.5389 0.5011 0.607820Y 0.9677 0.3840 0.9878 0.3766

Note: *, **, ***: significant at 0.10, 0.05, and 0.01 significance levels, respectively.

show that yields of Government sukuk (1, 2, 3 and 5 years) and those of AAA rated corporate (1 and 2 years) issues Granger cause their conventional bond yields. Results do not show a consistent pattern in terms of issuer or maturity of the security for having a Granger causal effect.

The second test conducted was to check for the presence of Granger causal relation between conventional bonds and sukuk. The null hypothesis tested is “yield of conventional bonds does not Granger cause the yield of sukuk security.” Out of the 20 pairs of securities tested, the null hypothesis is rejected in 4 pairs at the 0.05 significance levels. This indicates that one may not generally conclude that the conventional bond yields Granger cause sukuk security yields. These results show that yield of conventional bonds issued by Government (2 and 3 years) and AAA rated corporate (6 months and 1 year) Granger cause their sukuk counterparts. Results do not show consistent pattern in terms of issuer or maturity of the security for having a Granger causal effect.

Finally, as in Table 3, bi-directional Granger causality (as expressed in: Enders, 1995, Hossain, 2005) between yield of sukuk and yield of conventional bonds is observable in 3 out of 20 (or 15 percent) pairs. In other words, in 3 pairs of securities, both null hypotheses are significantly rejected, or, yield of sukuk Granger cause yield of conventional bonds and the other way around. This may signal that both variables are Granger caused by a third variable yet to be explored. Results show that yield of sukuk and conventional bonds have bi-directional Granger causal relation in securities issued by Government (2 and 3 years) and AAA rated corporate

(1 year). In summary, it is reasonable to conclude that, with few exceptions, there is no causal relationship between sukuk and conventional bonds. This is the second statistically supported evidence to affirm an argument that the two types of debt instruments are not the same. This conclusion has important implication for market operation, valuation practices, risk estimation and regulatory rule setting. These are challenges to be addressed in future research.

3. Structuring sukuk contractsThis section reviews the structure of three major types of sukuk securities namely Ijarah, Musharakah, and Mudarabah.

Ijarah sukukIjarah, which means “to give something on rent” (Lewis and Algaoud, 2001), is the reward or recompense that proceeds from a rental contract between two parties, where the lessor (the owner of the asset) leases capital asset to the lessee (the user of the asset) (Gait and Worthington, 2007). There is a tendency toward lease financing (Ijarah) in Islamic banking sector, since it promises higher yields than in trade finance (Murabahah) and it also has longer financing horizon, which is an important feature for business investments (Daryanani, 2008). In order to be permissible under the Shariah, the Ijarah contract should satisfy some conditions. The primary requirement is that the lessor must be the real owner and in possession of the asset to be leased under contract. As a result, the lessor should solely bear all risks and uncertainties associated

Ariff and Safari


to the asset and be responsible for all damage, repair, insurance, and depreciation of the asset (Khan and Bhatti, 2008).

It could be inferred that charging rental payment is not allowed until the lessee actually receives the possession of the asset and shall pay the rental only as long as it is in usable condition. Moreover, in case of manufacturing defects, which are beyond the lessee’s control, the lessor is responsible. However, the lessee is responsible for the proper upkeep and maintenance of the leased asset. The intention of posing such restriction in Ijarah contract by Shariah is to protect both parties to the contract by reducing the uncertainty and ambiguity from the agreement (Wilson, 2004). In addition to that, both lessor and lessee should be clear on purpose of Ijarah and the usage of assets, moreover, the Ijarah purpose must comply with Shariah (Al-Omar and Abdel-Haq, 1996).

There are two forms of leasing contracts, or Ijarah, in Islamic finance. Ijarah, or direct leasing contract, is the case where the lessee uses the capital asset owned by the lessor, with his/her permission, for a specific period of time for a monthly or annual rent. The owner assumes ownership title during the whole contract period, and the owner should performs the ownership responsibilities such as insurance (Zaher and Hassan, 2001). In this Ijarah contract possession of asset should be transferred back to the owner after the contract matures. In other words, in pure Ijarah contracts, there is no option to transfer the ownership of the asset at maturity.

Ijarah wa Iqtina, or hire purchase, is the case of contract where the basic intention is transferring the ownership after completing the leasing period. Ijarah wa Iqtina is popularly practiced when Islamic bank purchases equipment or some other capital asset based on the request of an individual or institutional customer and then rents it to the customer for a certain fixed rent. On the other hand, the customer promises to purchase the equipment or asset within a specified period to transfer the ownership from the Islamic bank to the customer (Al-Jarhi and Iqbal, 2001). However, it should be noted that the lease contract is completely separate and independent from the contract of purchase of residuals, which has to be valued on a market-basis and cannot be fixed in advance.

The purchase contract should be an optional, non-binding contract because the quality and the market price of the asset at the end of the lease period are unclear (Chapra, 1998). One other approach is the case where the ownership is gradually transferred to the customer. In this case, and in addition to the regular rental payment, the customer shall pay installments of the value of the asset in order to reduce the ownership share of the lessor in the asset until the ownership is fully transferred to the lessee (Metwally, 2006). Ijarah wa Iqtina, having strong agreement among Shariah scholars, is widely used in the real estate, retail, industry, and manufacturing sectors (Iqbal, 1998, Warde, 2000).

Ijarah sukuk is based on the Ijarah contract. In order to issue Ijarah sukuk, the originator, who primarily owns the assets, sells the assets to a Special Purpose Vehicle (SPV), which is typically a company in an offshore tax-free site.

Then the SPV leases back the assets to the issuer at a specific predetermined rental fee and then the SPV securitize the ownership in the assets by issuing sukuk certificates to the public investors (Lewis, 2007). These sukuk certificates represent an undividable share in the ownership of the assets, which entitle the sukuk holders to distribution of the rental payments on the underlying assets. However, the rental payment could be fixed or floating for the whole period, dependent on the leasing contract between the SPV and originator. Since these sukuk certificates represent ownership in real assets, they can be traded in a secondary market.

The SPV manages the cash flows of the sukuk contract by receiving periodic rentals and installments from the originator and then disbursing the cash flows to the sukuk holders (Aseambankers, 2005). SPV also manages disbursement of lump sum maturity payments. At the maturity of a sukuk contract, the SPV no longer has a role and consequently will cease to exist. However, the Ijarah sukuk is typically issued for periods longer than five years and could be considered as long-term debt certificates. This may raise the issue of SPV’s default risk, so, the investors typically receive a direct guarantee from the issuer’s guarantee of the SPV obligations (Wilson, 2008). This guarantee also includes the obligation by the issuer to repurchase the asset from the SPV at the end of the Ijarah contract at the original sale price.

Wilson (2008) suggest that SPV does not have any of the risks associated with banks due to SPV’s nature. In other words, SPV is bankruptcy remote. If the issuer faces the bankruptcy, the creditors to the issuer cannot claim the assets held by the SPV or otherwise interfere with the rights of the sukuk-holders with respect to the underlying assets (Gurgey and Keki, 2008). As a result, SPV would be attractive to both issuers and investors, and this may justify the relatively high legal establishment costs. Figure 2 is a contract specification.

Kamali (2007) claims that the fixed and predetermined nature of rental cash flow introduces additional risk because the Ijarah sukuk holders receive steady income,

Figure 2. Ijarah sukuk structure – adapted from Wilson (2008), Abdul Majid (2003), El-Gamal (2007), Bose and McGee (2008).



which is more risk averse than is the case of common stocks. However, he mentioned general market conditions, price movements of real assets, ability of the lessee to pay the rental or installments, maintenance and insurance cost are sources of risks to the Ijarah sukuk. He concluded that because of these risk factors, the expected return on some of Ijarah sukuk may not be precisely predetermined and fixed. Thus, the fixed rental may only represent a maximum that is subject to some possible deductions.

The major criticism of Ijarah sukuk is that the return is variable or floating in most cases. Moreover, this variable rate, sometimes for simplification reasons, is mostly benchmarked or “pegged” to an interest-based index such as the London Interbank Offered Rate (LIBOR) for US$ based sukuk and in local currencies the local rates. Usmani (2002) criticized this practice by associating riba to this form of Ijarah sukuk practice. Shariah scholars suggest the usage of other non-interest benchmarks for pricing and evaluation purposes. In order to overcome the riba issue, government sukuk could be assessed by macroeconomic indicators and corporate sukuk could be assessed based on the company performance indicators.

Musharakah sukukIqbal and Molyneux (2005) defined Musharakah as “an arrangement where two or more parties establish a joint commercial enterprise and all contribute capital as well as labor and management as a general rule.” In contrast to Mudarabah contract, Musharakah investors have the right to participate in management of the business partnership, however, this right is entrusted to each investor (Shinsuke, 2007). It could be argued that Musharakah contract may require establishment of a partnership or company, where Musharakah contract parties are the participants and owners (Wilson, 2004).

Musharakah sukuk securities could be issued based on such financing concept. Musharakah type of equity finance demands that both a profit-sharing ratio and length of the joint venture agreement is decided in advance. Similar to Mudarabah, loss is shared in proportion to the capital contribution unless the loss is proven to be due to negligence of one party (Daryanani, 2008). Therefore, all profits and losses generated from the Musharakah are shared among the parties on the basis of the pre-agreed ratio. As a result, Musharakah is basically suitable for financing private or public companies as well as projects as also practiced by Islamic banks, where it is typically performed through joint ventures between banks and business firm for a certain operation (Gait and Worthington, 2007).

Musharakah, due to its nature and advantages in providing equal (but proportionate) benefits for all parties, has the support of all scholars and is valid under Shariah principles. However, El-Gamal (2000) suggests that parties to Musharakah usually need the help of legal expert to ensure that any potential Riba or Gharar is carefully avoided. On the discussion about Musharakah contracts, Chapra (1998) concluded that “The only requirement of the Shariah would be justice, which would imply that the proportional shares of partners in profit must reflect the contribution made to the business by their capital, skill, time, management ability, goodwill and contacts. Anything otherwise would not only

shatter one of the most important pillars of the Islamic value system, but also lead to dissatisfaction and conflict among the partners and destabilize the partnership. The losses must, however, be shared in proportion to capital contribution and the stipulation of any other proportion would be ultra vires and unenforceable.”

Lewis and Algaoud (2001) suggest two ways to structure a Musharakah contact. However, both types are based on the same general concept of Musharakah, where parties (capital owner and entrepreneur) are ensured an equitable share in the profit or loss on pre-agreed terms. The difference lies in the pre-agreed sharing ratio. In the first method, this pre-agreed ratio is pre-fixed and remains constant for the whole period of the contract while in the second type, the ratio is declining. The diminishing Musharakah contract is preferred by some financiers since it allows release of their capital from the investment by reducing its equity share each year and receiving periodic profits based on the remaining balance. On the other hand, the equity share of the entrepreneur increases over time to the extent that he/she becomes the sole owner of the firm.

Sukuk based on diminishing Musharakah are gaining momentum since they enable Islamic banks or Shariah-compliant investment companies to provide up-front investment funding to the issuer. In this regard, both parties establish a Special Purpose Vehicle (SPV) to administer the sukuk. In order to issue a diminishing Musharakah sukuk, the issuer transfers the ownership of an asset to the SPV to enter the partnership agreement. The investors enter the agreement by paying cash. Therefore, both the investors and the issuer are equity partners in the SPV. However, the investors share in the SPV diminishes over time as the issuer pays installments to investors to repurchase their respective shares in the asset. These installment payments plus the issuer’s rental payments for use of asset (asset’s generated income) so the contract becomes a Musharakah sukuk with cash flow stream for sukuk-holders. In fixed-ratio Musharakah sukuk, the cash flow stream for the sukuk-holder is only from the income generated from the asset and not the installment part. The structure of diminishing Musharakah sukuk is depicted in the Figure 3.

Figure 3. Diminishing musharakah sukuk structure (Wilson, 2008).

Ariff and Safari


Flexibility in payment schedule and amounts has made diminishing Musharakah sukuk more convenient to use. However, it should be highlighted that all arrangements should be agreed upon ex-ante by all parties to the SPV. The payments are usually monthly or quarterly, but not necessary in equal amounts (Wilson, 2008). Smaller installments could be made during the initial period of the sukuk, with most of the asset value or SPV capital remaining with the investors, but the amount of the installment payments could increase in a linear fashion, or according to some predetermined formula. As the issuer’s share in the asset increases through the buy-back process, the periodic rental might be expected to decrease due to a decline in remaining share. However, this does not necessarily have to be the case, especially if there is capital appreciation in the value of the asset. In other words, when installment and rental payments are aggregated, they might be constant, diminishing or increasing over time, provided both parties agree to the formula used and the documentation is transparent.

4. Sukuk cash flowsThis section reviews cash flow patterns of two major types of sukuk securities namely Ijarah sukuk and Musharakah sukuk.

Ijarah sukukIjarah sukuk can have various types of payback structures. In the simplest form, Ijarah sukuk payback could be fixed promised regular payments and not a predetermined promised maturity payment. The formal Ijarah contract does not have the option for parties to transfer the ownership of the asset at the end of the period. Thus, at the end of an Ijarah contract, the asset should be returned to the owner (capital owner or the SPV). In order to transfer the ownership back to the issuer at the maturity, one should use Ijarah wa Iqtina (lease and purchase) contract. Ijarah wa Iqtina sukuk is form of Ijarah contract where the ownership of the asset will be transferred to lessee (issuer) at the maturity of the sukuk. However, the maturity payment is not determined at the issuance time of sukuk. The valuation of the asset in this case should be conducted at the maturity time, when the market value of the asset is revealed and maturity payment is set to be equal to that.

Ijarah sukuk has fixed and predetermined rental payment (rewards) and a market valued maturity payment. However, in practice, the maturity value of the property is, sometimes, fixed and predetermined to both sides of the contract. Therefore, the cash flow pattern of Ijarah sukuk would be similar to that of a coupon-bearing conventional bond such securities are depicted in Figure 4. The valuation of this form of sukuk is similar to a straight bond because of the similarity of cash flows pattern and the tradability. Thus, a sukuk could be priced by using the Equation 1.

PR

r

M

r

Rr r

M

rt

N

t N N=+( )

++( )

= -+( )

++(=

∑1 1 1

11

1 1 ))N (1)

In Equation 1, P is the price of sukuk security, R is the amount of periodical promised payment, M is the amount

Figure 4. Cash flows pattern for sukuk with fixed promised regular payments and promised maturity payment.

Figure 5. Growing promised regular payments pattern with predetermined promised maturity payment.

of predetermined maturity payment, r is the discount rate, and N is the number of time periods remaining to maturity.

Another variation of cash flows of Ijarah sukuk happens when the periodic payments follow a growth pattern. In Ijarah sukuk, maturity payment (M) should be based on market value (hence, a priori undetermined), while the Ri, the amount of promised regular payment (rental fees) in period i is following a predetermined constant growth model with growth rate of g. However, in practice, maturity payment of Ijarah sukuk is predetermined and mentioned in contract documents. Therefore, its cash flow is similar to the one in Figure 5.

The cash flow pattern of sukuk consists of a growing annuity of promised regular payments and a promised maturity payment. Thus, using the formula for calculating the present value of an annuity, one can formulate the price of a sukuk security as Equation 2.

PR

r

M

r

Rr g

grt

N

t N

N

. –=+( )

++( )

=-( )

++

=

∑1

1

1 11

11

++( )M

rN

1 (2)

In Equation 2, P is the price of sukuk, r is the discount rate, N is the number of periods to maturity, g is the growth rate of promised regular payments, and R1 is the amount of first promised regular payment. It is assumed that the promised payments are growing at a constant rate of g, thus, R2 = R1(1 + g).

Musharakah SukukThe simplest form of cash flow generated by a Musharakah sukuk security is obtained from those securities that only pay a lump sum amount of cash in a certain and predetermined point of time in future at maturity. There is no cash payment to investors prior to maturity as shown in Figure 6. The amount of maturity payment, however, is not determined and should be based on the performance of the underlying investment. However, as in most practice cases, the maturity payment is fixed and predetermined.



This form of cash flow is the same as conventional discount bond cash flow or a bond with balloon payment. Thus the same valuation process is applicable. The valuation of these forms of sukuk could be performed using the conventional pricing approach as follow. The current price of sukuk is the maturity payment (face value) discounted to the present time, similar to discount bonds.

PMr T t( )

=+ -1

(3)

In Equation 3, P is the price of sukuk, M is maturity payment (face value), T is maturity date, t is time, and r is the discount rate. It should be noted that r should not be based on any interest bearing benchmark.

Diminishing Musharakah sukuk may possess payoff structure in a manner that it only pays some promised regular payments at certain periods of time with zero maturity payment. Amount of promised regular payments are fixed and predetermined. Cash flow pattern of such security is depicted in Figure 7.

Cash flow represented in Figure 7 is identical to a constant annuity. Thus, in order to evaluate the price of a diminishing Musharakah sukuk, present value of annuity is applicable. This will result into Equation 4.

PR

r

Rr rt

N

t N=+( )

= -+( )

=

∑1 1

11

1 (4)

In Equation 4, P is the price of diminishing Musharakah sukuk security, R is the amount of periodical promised payment, r is the discount rate, and N is the number of time periods remaining to maturity.

In diminishing Musharakah sukuk, amount of promised regular payment in period i is following a predetermined constant negative growth (declining) model with growth rate of g, which is a negative number. In other words, the cash flow stream of regular payments declines to zero at maturity (M is also equal to zero). The cash flow diagram is presented in Figure 8.

Thus, price of a diminishing Musharakah sukuk can be formulated as Equation 5.

PR

r

Rr g

grt

N

t

N

–.=+( )

=-( )

++

=

∑1

1

11

11

(5)

In Equation 5, P is the price of diminishing Musharakah sukuk, r is the discount rate, N is the number of periods to maturity, g is the negative growth rate of promised regular payments, and R1 is the amount of first promised regular payment. It is assumed that the promised regular payments are declining at a constant rate of g, thus, R2 = R1 (1 + g) < R1.

5. Sukuk world marketsThe sukuk market has grown rapidly in recent years. Emergence of more than 250 Takaful (Islamic Insurance) companies, 350 Islamic equity funds, and 370 Islamic banks worldwide has made a great demand for sukuk. Khan and Bhatti (2008) highlighted that sukuk constitute about 85 percent of the Middle Eastern capital market, US$13bn of them have been issued there with an average growth rate of over 45 percent during 2002–2007. The Middle East and Asian regions will primarily rely on sukuk to meet their US$1.5 trillion infrastructure needs over the next ten years (John, 2007) is the kind of statements one reads in commentaries.

McKenzie (2008) investigated the sukuk market statistics and reported that, while sovereign sukuk issues by Bahrain and Malaysia played an initial role in establishing the market, about 80 percent of issues between 2001 and 2006 have been corporate issues. However as reported in McKenzie (2010), most sukuk issuers are government or quasi-government organizations in 2009. In addition to that, he found that the most important market for corporate sukuk issues totaling US$44 billion over the period of 2001–2006 has been for infrastructure finance, with issuance of US$17billion, 39 percent of the total. The next largest markets were for financial services (18 percent) and energy (6 percent). McKenzie (2009) claims that there is a growing appetite and demand for investment in sukuk that goes well beyond Islamic investors amongst those investors wishing to gain exposure to diverse but high quality assets. The IFSL Islamic Finance (2012) reported that the sukuk issuance has increased significantly in past three years to the level of US$84.4billion in 2011. The amount of new sukuk issues in last ten years is shown in plots in Figure 9.

According to International Islamic Financial Market (IIFM) 2010 report, the sukuk market size grew to over US$136billion in mid-2009. As at the end of 2010, the total asset value of publicly-listed sukuk was more than US$ 197 billion in 13 markets (Alvi, et al., 2011). Mass media reports suggest a much higher asset value by including

Figure 6. Zero-promised regular payment cash flow pattern of Musharakah Sukuk securities.

Figure 7. Fixed promise regular payment cash flows pattern of Sukuk without promised maturity payment.

Figure 8. Declining promised payments cash flows pattern.

Ariff and Safari


privately-issued sukuk in a number major financial centers such as Zurich, London, Frankfurt, Singapore and others. Hence, a figure of US$ 840 billion is suggested in Ariff et al. (2012). Currently, sukuk are offered in specialized exchanges such as the Labuan Exchange in Malaysia, the Third market in Vienna, the Dubai International Finance Exchange, and the London Stock Exchange (Asad, 2009b). Governments and regulators in a variety of countries have recognized the important role that sukuk can play in capital markets and have been giving priority to developing their countries as sukuk centers (Abd Razak and Abdul Karim, 2008).

6. Sukuk regulationsAll Islamic financial and banking interactions, similar to their conventional counterparts, are subject to regulations of professional authorities. Moreover, in order to be recognized as an “Islamic” transaction, sukuk products must follow some extra procedures required by Islamic authorities (Iqbal, 1999). Shariah regulations governing sukuk, similar to other Islamic finance and banking practices in general, are dictated from three sources; international organizations, local authorities, or in-house Shariah boards.

Some major Islamic financial institutions have their own in-house Shariah Supervisory Boards (SSB). This is

mostly practiced among Islamic investment companies and banks. In the process of issuing sukuk, according to Liquidity Management Center of Bahrain (2008), Shariah advisors should study proposed sukuk structures and suggest a Shariah structure which fulfills the set economic aims sought in the issue. In addition, they should work closely with legal counsel of the issuer and the arranger (investment banker) to ensure that the legal documents are in line with Shariah requirements. Finally, they should Issue a fatwa, which is an opinion, as to the compliance of the sukuk with Islamic principles—similar to the opinion tax authorities provide on a tax question—on the whole sukuk deal before it can be put into circulation.

Islamic financial institutions are subject to rules and regulations of the local Shariah authority. Some countries such as Malaysia have set up their own Shariah Advisory Council (SAC) at the national level, which oversees the consistent application across similar situations in financial interactions. These councils are part of the Securities Commission or also part of the Central Bank of the country at national level. At the international level (the Organization of Islamic Countries) there is a Shariah Council in Saudi Arabia.

In a global perspective, there are few international organizations that attempt to regulate and screen the conduct of sukuk issuance and trade. Among these international organizations, AAOIFI, IFSB, and IIFM are the most influential ones (DIFC, 2009). Although these organizations try to base their rulings on Shariah principles, there are occasions that sukuk based on their guidelines have variations in formation. Not being national bodies, these rules cannot be imposed, so remain voluntary. Therefore, Siddiqui (2008) highlighted that more communication between these organizations will bridge the differences existing between sukuk contracts.

The decisions regarding permissibility of each sukuk contract, as mentioned above, is made by expert Muslim scholars who are appointed to the Shariah Board. Number of these experts are estimated to fall between 100 and 200, worldwide (Asad, 2009a). This indicates the urgent need for training of expert Muslim scholars to sit on Shariah boards in Islamic financial institutions. In order to address this shortfall, some Islamic institutes such as ISRA3 and INCEIF4 are planning to design special programs for training certified Shariah scholars. However, there are some other institutions that are currently providing short time courses on this topic.5

AAOIFIThe Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI) in its website6 introduces itself as “an Islamic international autonomous non-profit corporate body that prepares accounting, auditing, governance, ethical and Shariah standards for Islamic financial institutions and the industry.” AAOIFI was established in accordance with the Agreement of Association, which was signed by Islamic financial institutions in 1990 in Algiers. Then, it was registered in 1991 in Bahrain. As an independent international organization, AAOIFI is supported by institutional members (more than 200 members from 45 countries).

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Figure 9. Amount of worldwide Sukuk issues (Source: IFSL, Zawya Sukuk monitor).

69%

11%

5%

4%

3% 3%2% 3%

Malaysia

Qatar

UAE

Indonesia

S. Arabia

Bahrain

Pakistan

Others

Figure 10. Sukuk issuance by country, 2011 (Source: IFSL, Zawya Sukuk monitor).



Members include central banks, Islamic financial institutions, and other participants from the international Islamic banking and finance industry, worldwide. Esta blishment of AAOIFI represented a shift from the authority of independent Shariah supervisory boards set up in individual Islamic banking and finance operations, to a centralized model for the dissemination of standards, procedures and best practices (Gambling et al., 1993; Maurer, 2002, Pomeranz, 1997)

In May 2003, AAOIFI issued Shariah standard FAS 17 titled “Investment sukuk.” In this standard, AAOIFI issued standard for 14 different types of sukuk, where some of these sukuk are classified as tradable and others are classified as non-tradable based on the type and characteristics of the issued sukuk (AlBuolayan, 2006). AAOIFI and its Shariah Board chairman, Shaikh Muhammad Taqi Usmani, have special attention to sukuk because it is one of the most favored Islamic financing instruments and there is a possibility of variation in contract formation by practitioner. Shaikh Usmani has given some comments on the practice of sukuk over time. For instance, he has highlighted that “… since Ijarah sukuk represent the pro rata ownership of their holders in the tangible assets of the fund, and not the liquid amounts or debts, they are fully negotiable and can be sold and purchased in the secondary market. Anyone who purchases these sukuk replaces the sellers in the pro rata ownership of the relevant assets and all the rights and obligations of the original subscriber are passed on to him. The price of these sukuk will be determined on the basis of market forces, and are normally based on their profitability” (Usmani, 2001).

However, his most cited and debated comment regarding sukuk is his statement in November 2007 which declared that some 85 percent of outstanding sukuk had failed the Shariah-compliance test on the basis that they were ‘asset-based’ rather than ‘asset-backed’ with the guaranteed return of the face value of the sukuk on maturity and in the absence of a transfer in asset ownership to sukuk holders (Usmani, 2007). Since then, the juridical validity of sukuk became suspect (Hasan, 2010, Alsayyed and Malik, 2010). His reasons for such declaration were in brief as under.

• There have been cases where the assets in the sukuk were the shares of companies that do not confer true ownership but which merely offer to sukuk holders a right to returns.

• Most sukuk issued are identical to conventional bonds with regard to the distribution of profits from their enterprises at fixed percentage bench-marked on interest rates. The legal presumption regarding sukuk is that no fixed rate of profit or the refund of capital can be guaranteed.

• Virtually, all sukuk issues guarantee the return of the principal to holders at maturity (just as in conventional bonds) through a binding promise from either the issuer or the manager to repurchase the assets at the stated price regardless of their true or market value at maturity.

Later on, in February 2008, AAOIFI issued a guidance statement on accounting for investments and amendment in FAS 17 (AAOIFI, 2008). Summary of important issues raised in this guideline are:

a) Sukuk issuances have to be backed by real assets, the ownership of which has to be legally transferred to sukuk holders in order to be tradable

b) Sukuk must not represent receivables or debts, except in the case of a trading or financial entity selling all its assets or a portfolio with a standing financial obligation, in which, some debts owing by third parties, incidental to physical assets or usufruct, are unintentionally included

c) The manager of the sukuk is prohibited from extending “loans” to make up for the shortfall in the return on the assets, whether acting as a mudarib (investment manager), or sharik (partner) or wakil (agent)

d) Guarantees to repurchase the assets at nominal value upon maturity with the exception of Ijarah sukuk structures are also prohibited

e) Closer scrutiny of documentation and subsequent execution of the transaction is required by Shariah Supervisory Boards

Maurer (2010) investigated this controversial issue and concluded that for some, Usmani’s salvo was a long overdue and much needed corrective to what they saw as the excesses of sukuk issuances and structured financing vehicles that came very close to mimicking conventional bonds. To others, it was an overreaction, born of impatience with the pace of development of Islamic financial institutions and markets, and an unrealistic appraisal of what Islamic finance can actually accomplish in a globally interconnected and interdependent world.

IFSBThe Islamic Financial Services Board (IFSB), which is based in Kuala Lumpur, was established in 2002 and started operations in early 2003. It serves as an international standard-setting body of regulatory and supervisory agencies that have vested interest in ensuring the soundness and stability of the Islamic financial services industry, which is defined broadly to include banking, capital market and insurance. In advancing this mission, the IFSB promotes the development of a prudent and transparent Islamic financial services industry through introducing new or adapting existing international standards consistent with Shariah principles and recommend them for adoption. To this end, the work of the IFSB complements that of the Basel Committee on Banking Supervision, International Organization of Securities Commissions and the International Association of Insurance Supervisors.

Drafting of standards in IFSB is done on a task force working group method. The IFSB council appoints members of technical committee which are responsible for advising the council on technical issues within its terms of reference. Islamic Development Bank (IDB) Shariah Supervisory Board is responsible for Shariah supervision of IFSB’s standards. IFSB has issued three standards that affect the issuance, trading, or investing in sukuk:

• IFSB-1: Guiding principles of risk management for institutions (other than insurance institutions) offering only Islamic financial services, issued in December 2005. This guideline also includes the various risk elements affecting institutions offering investment certificates such as sukuk and operational

Ariff and Safari


consideration regarding them. This guideline offers a general perspective toward risk sources and risk management and is not specific for sukuk.

• IFSB-2: Capital adequacy standard for institutions (other than insurance institutions) offering only Islamic financial services, issued in December 2005. This standard overviews various Islamic contracts (some of which are underlying contracts of sukuk) and provide capital requirement for each.

• IFSB-7: Capital adequacy requirements for sukuk, securitizations and real estate investment, issued in January 2009. The first part of this guideline investigates the sukuk and securitization of it, sukuk structures, operational requirements pertaining to sukuk, treatment for regulatory capital purposes of sukuk and securitization exposures, and treatment of credit risk exposures of sukuk.

IIFMThe International Islamic Financial Market (IIFM), located in Bahrain, is a global standardization body for the Islamic capital and money market segment of the Islamic financial services industry. Its primary focus lies in the standardization of Islamic products, documentation and related processes. IIFM was founded with the collective efforts of Central Bank of Bahrain, Bank Indonesia, Central Bank of Sudan, Labuan Financial Services Authority (Malaysia), Ministry of Finance (Brunei Darussalam) and Islamic Development Bank (a multilateral institution based in Saudi Arabia). Besides the founding members, IIFM is supported by its permanent members, namely State Bank of Pakistan and Dubai International Financial Centre Authority (UAE). IIFM is further supported by a number of regional and international financial institutions as well as other market participants as its members. IIFM activities are under supervision of its Shariah Advisory Panel, which currently has ten members. Focus of IIFM’s work is on Islamic capital and money markets. Presently, IIFM has no specific standard or guideline pertaining to sukuk. However, it has released two reports on sukuk in 2010 and 2011. In these reports, IIFM investigated the current sukuk market from an international as well as domestic perspective. It also investigated various sukuk structures international issues. Moreover, it has studied some of sukuk issues as case study.

Sukuk ratingRating is an evaluation of a corporate or municipal bond’s relative safety from an investment standpoint. In conventional sense, it scrutinizes the issuer’s ability to repay principal and make interest payments. Then, a grade (the most controversial part of rating process) is given to the bond that indicates its credit quality. Private independent international rating companies such as Standard & Poor’s, Moody’s, and Fitch, or domestic rating agencies like RAM and MARC of Malaysia, provide these evaluations of issuer’s financial strength, or the ability to pay a bond’s principal and interest in a timely fashion. As a result, bonds are rated in a range from AAA or Aaa (the highest), to C or D, which represents a company that has already defaulted. Each rating company has its own definition and methodology for rating and own set of rating ranges.

The introduction of sukuk rating in the 90 s represents another critical milestone in the development of sukuk market. Bond ratings are principally designed to arrive at a reasoned judgment on credit risk via a careful analysis of the critical issues surrounding a specific debt on the issuer (Mohd Asri, 2004).

From the global capital markets point of view, sukuk can be rated just like any conventional bond, and can be traded as such, as well (Maurer, 2010). In general, rating agencies have the same criteria for corporate bonds rating. The criteria incorporate issue structure (repayment schedule and debt types), business risk analysis, financial risk analysis, management, ownership and other qualitative factors (Mohd Asri, 2004). However, realizing the uniqueness and types of sukuk, the rating methodology should be different to that of conventional bonds rating (Jalil, 2005). Rosly (2007) argued that sukuk structures falls in 2 categories:

• Asset-Backed sukuk, for which ratings are dependent on a risk analysis of the asset. However, investors hold rights to underlying assets through SPV and not directly; hence, sukuk performance is driven by assets and not linked to the originator.

• Unsecured sukuk, for which ratings are primarily dependent on the riskiness of the sponsor, originator, or the borrower.

Therefore, similar to conventional bonds, risk elements affecting sukuk should be thoroughly investigated by rating agencies. Among risk elements, Rosly (2007) mentioned that credit risk is the most critical one. Other factors he highlighted are currency risk (for international issues), tax risk, and reserve funds. In contrast to highly sensitive conventional bonds, sukuk are less sensitive to interest rate. Zurich based investment bank Credit Suisse believes investment in Islamic Finance and Banking products are not expose to interest rates since Islam prohibits charging interest and sukuk securities are unaffected to the credit crisis in the international finance and banking industry (Farook, 2009). Tariq (2004) summarized risk characteristics of each type of sukuk structures which is depicted in Table 4.

Security commissionSecurity Commissions in each country is a statutory body that investigates the conduct of financial markets and has enforcement power. Among all of their responsibilities, some are related to origination and exchange of sukuk securities. These responsibilities are supervising exchanges, clearing houses, and central depositors; approving authority for corporate bond issues (including sukuk securities); and regulating all matters relating to securities and futures contracts. Hence, one may assume that the all sukuk securities issued require permission and approval by Security Commission of that particular jurisdiction. Beyond the regulatory obligations, they also act as a strategic policy making body on financial markets.

Among these incentives,7 some directly affect the issuance of sukuk securities in Malaysia. Issuers of sukuk securities in Malaysia are offered tax deduction on expenses incurred in due course of issuance of Wakalah, Murabahah, Bai



Tabl

e 4.

Su

mm

ary

of r

isk

char

acte

rist

ics

of su

kuk

stru

ctur

es –

sou

rce:

(Ta

riq,

200

4).

Type

of S

ukuk

Des

crip

tion

of S

ukuk

st

ruct

ure

Cre

dit R

isk

Rat

e of

ret

urn

(Int

eres

t rat

e ri

sk)

FX r

isk

Pric

e ri

skO

ther

ris

ks

Zero

cou

pon

Suku

kIs

tisn

a, M

urab

ahah

de

bt c

erti

ficat

es –

no

n-tr

adab

le

Uni

que

basi

s of

cr

edit

ris

ks e

xist

, (K

han

and

A

hmed

, 200

1)

Very

hig

h du

e to

fix

ed ra

te,

rem

ains

for t

he

enti

re m

atur

ity

of

the

issu

e

If a

ll ot

her c

ondi

tion

s

are

sim

ilar,

FX

risk

w

ill b

e th

e sa

me

for

all c

ases

of S

ukuk

. H

owev

er, t

hose

Su

kuk

whi

ch a

re

liqui

d or

whi

ch a

re

rela

tive

ly s

hort

term

in

nat

ure

will

be

le

ss e

xpos

ed.

The

com

posi

tion

of

ass

ets

in th

e po

ol

will

als

o co

ntri

bute

to

the

FX r

isk

in

diff

eren

t way

s.

Hen

ce th

is c

an b

e

very

use

ful t

ool t

o

over

com

e th

e FX

ri

sk b

y di

vers

ifyin

g

the

pool

in d

iffer

ent

curr

enci

es.

Rel

ated

to th

e

unde

rlyi

ng

com

mod

itie

s pr

ices

an

d as

sets

in re

lati

on

to th

e m

arke

t pri

ces.

Ija

rah

Suku

k is

mos

t ex

pose

d to

this

as

th

e va

lues

of t

he

unde

rlyi

ng a

sset

s

may

dep

reci

ate

fa

ster

as

com

pare

d

to m

arke

t pri

ces.

M

aint

enan

ce o

f the

as

sets

will

pla

y an

im

port

ant p

art i

n th

is

proc

ess.

Liq

uidi

ty

of th

e Su

kuk

will

als

o

play

an

impo

rtan

t par

t in

the

risk

. Sal

am is

al

so e

xpos

ed to

pri

ce

risk

s. H

owev

er,

thro

ugh

para

llel

cont

ract

s th

ese

risk

s

can

be o

verc

ome

Liqu

idit

y R

isk

is s

erio

us

as fa

r as

the

non-

trad

able

Su

kuk

are

conc

erne

d.

Bus

ines

s ri

sk o

f the

is

suer

is a

ris

k un

derl

ying

Su

kuk

as c

ompa

red

to

trad

itio

nal fi

xed

inco

mes

. Sh

aria

h co

mpl

ianc

e ri

sk

is a

noth

er o

ne u

niqu

e in

ca

se o

f Suk

uk.

Infr

astr

uctu

re r

igid

itie

s,

lack

of e

ffici

ent

inst

itut

iona

l sup

port

in

crea

ses

the

risk

of

Suku

k as

com

pare

d

to tr

adit

iona

l fixe

d

inco

mes

, (Su

ndar

araj

an

and

Erri

co, 2

002)

Fixe

d R

ate

Ijara

h Su

kuk

Secu

riti

zed

Ijara

h,

cert

ifica

te h

olde

r ow

ns p

art o

f ass

et

or u

sufr

ucts

and

ear

ns

fixed

rent

– tr

adab

le

Def

ault

on

rent

pa

ymen

t, fi

xed

ra

te m

akes

cre

dit

risk

mor

e se

riou

s

Very

hig

h du

e to

fix

ed ra

te,

rem

ains

for t

he

enti

re m

atur

ity

of

the

issu

eFl

oati

ng

Rat

e Ija

rah

Suku

kC

erti

ficat

e ho

lder

ow

ns p

art o

f ass

et

or u

sufr

ucts

and

ea

rns

float

ing

rent

in

dexe

d to

mar

ket

benc

hmar

k su

ch a

s

LIBO

R –

trad

able

Def

ault

on

rent

pa

ymen

t, fl

oati

ng

rate

mak

es d

efau

lt

risk

less

er s

erio

us –

se

e pr

evio

us c

ase

Exis

ts o

nly

wit

hin

th

e ti

me

of th

e

float

ing

peri

od

norm

ally

6

mon

ths

Fixe

d ra

te H

ybri

d/

Pool

ed S

ukuk

Secu

riti

zed

pool

of

asse

ts; d

ebts

mus

t no

t be

mor

e th

an

49%

, floa

ting

rate

po

ssib

ility

exi

sts

–

trad

able

Cre

dit r

isk

of d

ebt

part

of p

ool,

de

faul

t on

rent

s,

fixed

rate

mak

es

cred

it r

isk

seri

ous

Very

hig

h du

e to

fix

ed ra

te,

rem

ains

for t

he

enti

re m

atur

ity

of

the

issu

e

Mus

hara

kah

Te

rm F

inan

ce

Suku

k (M

TFS)

Med

ium

term

re

deem

able

M

usha

raka

h

cert

ifica

te b

ased

on

dim

inis

hing

M

usha

raka

h –

tr

adab

le a

s w

ell

as re

deem

able

Mus

hara

kah

has

hi

gh d

efau

lt r

isk

(K

han

and

Ahm

ed,

2001

), h

owev

er,

MTF

S co

uld

be

base

d on

the

st

reng

th o

f the

en

tire

bal

ance

she

et

Sim

ilar t

o flo

atin

g

rate

, but

, uni

que

be

caus

e th

e ra

te

is n

ot in

dexe

d

wit

h a

benc

hmar

k

like

LIBO

R,

henc

e le

ast

expo

sed

to

this

ris

kSa

lam

Suk

ukSe

curi

tize

d sa

lam

, fix

ed-r

ate

and

no

n-tr

adab

le

Uni

que

cred

it r

isk

(K

han

and

Ahm

ed,

2001

)

Very

hig

h du

e

to fi

xed

rate

Ariff and Safari


Bithaman Ajil, Musharakah, Mudarabah, Ijarah, and Istisna sukuk securities until 2015. Moreover, the Malaysian Security Commission (SC) has recognized SPV as a channel to transfer funds and hence, exempted them from income taxes. In addition to that, the issuer is permitted to tax deduct the cost of issuance of sukuk incurred by SPV Company. On the other hand, sukuk investors are receiving some incentives such as tax exemption on profits earned from investing in sukuk securities (except for the profits due to convertible loan stocks).

To add to the sukuk market flavor, SC offers intermediaries income tax exemption for qualified institutions “in respect of statutory income derived from regulated activity of dealing in securities and advising on corporate finance relating to arranging, underwriting and distributing non-ringgit sukuk originating from Malaysia which are issued or guaranteed by the Government or approved by the SC until the year of assessment 2014.”

7. Institutional developmentsSimilar to conventional practices of accounting, there is no unique standard for Islamic accounting and finance. However, in a similar fashion to conventional counterparts, there are attempts to develop one. AAOIFI, which gained respect of many Islamic countries and institutions, has developed standards for accounting statements. In the section 6.2 of the “Statement of Financial Accounting,” AAOIFI states that the following is the main focus of financial accounting in Islam: “on the fair reporting of the entity’s financial position and results of its operations, in a manner that would reveal what is halal (permissible) and haram (forbidden). Moreover, they highlighted one of the key objectives of financial reports: Information about the Islamic bank’s compliance with the Islamic Shariah and its objectives and to establish such compliance; and information establishing the separation of prohibited earnings and expenditures, if any, which occurred, and of the manner in which these were disposed of.”

Such an objective is not merely the same as the objectives set by the IASB (International Accounting Standards Board) in IFRS (International Financial Reporting Standards). In the 2001 “Framework for the Preparation and Presentation of Financial Statements,” IASB mentioned that the objective of financial statements is to “provide information about the financial position, performance and changes in financial position of an entity that is useful to a wide range of users in making economic decisions.” Hence, one may conclude that these two objectives are not necessarily the same.

Amin (2011) examined the implication of such difference in objectives and showed that there would be an actual difference in outcomes. He claimed that due to the difference in standards, particularly with respect to the presence of SPV in the structuring of sukuk securities, there would be a different outcome in financial reports. He argues that the root cause of such difference lies in the different interpretation of the role of the SPV, especially the fact that AAOIFI requires the ownership transfer of assets to it, while the IFRS considers sukuk transaction purely as a financing transaction and does not require the transferring of title of assets in the balance sheet statement.

Amin (2011) suggests that as there are more attempts to develop an internationally accepted accounting standards by IASB, Islamic accounting should follow their lead and focus more on requiring such standards to adhere to the demands of the Muslim world. He suggests that IFRS may include Islamic requirements in the standards by mandating practitioners to provide such information as footnote disclosures or in other formats.

Former issues indicate that there are potential conflicts between the outcome of following AAOIFI standards and IFRS standards in practice. Hence, firms may face a dilemma in adopting the proper accounting standards. Since issuance of sukuk securities is one of the many accounting and finance practices of firms, they may generally follow the IFRS standards, which are more comprehensive. In addition to that, audit firms may prefer conducting audit service for firms who follow a more acceptable standard (IFRS) rather than a less common one (AAOIFI). This might result in higher fees for auditing firms that adopt AAOIFI standards and make them less favored. Moreover, majority of accountant practitioners are educated with IAS methods and standards. The number of well-versed accountants in Islamic regulations is still lower than the demand; hence, many firms ought to hire conventional accountants.

These issues have hindered the acceptance and adoption of AAOIFI standards by many firms. Therefore, similar to Amin’s (2011) recommendation, we suggest that the AAOIFI should focus more on collaboration with international standard setting bodies and negotiate with them to incorporate and embed the demands of Muslim practitioners as well as investors in the internationally recognized standards such as IFRS.

8. Future developmentsPro-market regulatory framework is a key feature for long-term growth and sustainability of any financial market development. Presently, the contracts are arranged and organized individually and hence, the cost of issuance of sukuk securities is higher compared to conventional bonds. Lack of standardized sukuk contracts results in similar costly structuring of contracts. Standardization of contracts strengthens the convenience of securitization process, and will boost the cooperation among regulatory bodies. This will lead to one-stop servicing of primary market making, thus lessening the burdens of issuance process.

Market deepening activities should be undertaken to improve the lack of liquidity in the market. Majority of sukuk markets suffers from illiquidity to the extent that 70 percent or more are not traded as at 2011 has never been publicly traded since the issuance until their maturity. Introduction of discount houses as the intermediary between the exchange and the custormers would improve liquidity.

Market broadening activities are also required to further develop sukuk markets. Introducing more securities targeting at specific needs of customers would help introduce a broader list of securities. For example, the sukuk market has 6 traded ones although conceptually



there are 14 different types. These securities, though combination of two or more types of existing contracts into one structure, would result in new a security to address the unique needs of the issuer. More efforts should be put into financial engineering of sukuk securities.

Finally, education is another critical area that needs more attention in order to ensure the continuity and sustainability of sukuk markets. Currently, education and training efforts across the world is highly un-coordinated. Standardization of education and training on Islamic finance in general and in particularly sukuk securities is absent. There is no accreditation for different programs offered to the public. Although an international accreditation body for education programs is ideal, national accrediting bodies may serve the need for the time being. Existence of accreditation for Islamic finance programs may not only ensure the quality of the programs, but also is vehicle for standardization of these programs.

Although some limited number institutions (like INCEIF in Malaysia) offer specialized programs, majority of universities do not offer Islamic finance programs. Underlying reasons for shortage should be further investigated. Some potential reasons might be lack of enough syllabuses, or even lack of demand for such programs. However, universities may collaborate and share their resources to develop “market intelligence” database. Such database can be used to identify the potential programs, courses and syllabus. For instance, especial programs may be developed for training of Shariah board members.

Notes1 During the 3rd century AD, financial firms in Persia

(currently known as Iran) and other territories in the Persian Sassanid Dynasty issued letters of credit known as “chak.” KHARAZMI, A. A.-A. M. B. A. (1895) Maf t h al-Ool m. IN VLOTEN, G. V. (Ed. Leiden. In post-Islamic Arabic documents this word has transformed into “sakk” FLOOR, W. (1990) AK (legal document, testament, money draft, check). IN YARSHATER, E. (Ed. Encyclopedia Iranica Online Edition. New York., Columbia University.

2 Data is collected from Bondstream database, a product of Bond Pricing Agency Malaysia. Sample is limited to Malaysia due to unavailability of parallel markets for sukuk and bonds in other countries. Moreover, Malaysia accommodate almost two third of world sukuk issues.

3 International Shariah Research Academy for Islamic Finance, based in Kuala Lumpur, Malaysia.

4 International Centre for Education in Islamic Finance, based in Kuala Lumpur, Malaysia.

5 Besides the ISRA and INCEIF which offer long term programs, we can mention, among all, Dr. Kahf Institute, REDmoney Islamic Finance training institute, and UK Islamic Finance Council.

6 http://www.aaoifi.com/aaoifi/TheOrganization/Over-view/tabid/62/language/en-US/Default.aspx accessed on 9/4/2013.

7 For complete list of such incentives refer to the website of Security Commission of Malaysia http://www.sc.com.my/main.asp?pageid=593&menuid=322&newsid=&linkid=&type=(Accessed on 11/4/2013)

ReferencesAAOIFI (2004) Shariah Standards for Financial Institutions.

Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI).

AAOIFI (2008) Guidance Statement on Accounting for Investments and Amendment in FAS 17. Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI).

Abd Razak, D. & Abdul Karim, M.A. (2008) Development of Islamic Finance in Malaysia: A Conceptual Paper. 8th Global Conference on Business & Economics. Florence, Italy.

Abdul Majid, A.R. (2003) Development of Liquidity Management Instruments: Challenges and Opportunities. International Conference on Islamic Banking: Risk Management, Regulation and Supervision. Jakarta –Indonesia.

Al-Jarhi, M.A. & Iqbal, M. (2001) Islamic Banking: Answers to Some Frequently Asked Questions Jeddah, Saudi Arabia, Islamic Research and Training Institute, Islamic Development Bank.

Al-Omar, F. & Abdel-Haq, M. (1996) Islamic Banking: Theory, Practice & Challenges, Karachi, Oxford University Press.

Albuolayan, A. (2006) Rapid Surge in Sukuk Market. ArabNews, Jeddah, Saudi Arabia. Jeddah, Saudi Arabia, Saudi Research & Publishing Co.

Alsayyed, N.A. & Malik, F. (2010) Sukukization: Islamic Economic Risk Factors in Shariah View. Working Paper of International Shari’ah Research Academy for Islamic Finance (ISRA). Kuala Lumpur, Malaysia, International Shari’ah Research Academy for Islamic Finance (ISRA).

Alvi, I.A., Mohammed, A.R., Khan, G.Z., Nasser, U.M., Naseer, B. & Khan, M.S. (2010) Sukuk Report, A Comprehensive Study of The International Sukuk Market. In Alvi, I.A. (Ed. International Islamic financial market. 1 ed. Manama, Bahrain, International Islamic financial market.

Alvi, I.A., Mohammed, A.R. & Naseer, U.M. (2011) Sukuk Report, A Comprehensive Study of The International Sukuk Market. In Alvi, I. A. (Ed. International Islamic Financial Market. 2 ed. Manama, Bahrain, IIFM.

Amin, M. (2011) Accounting for sukuk under IFRS and AAOIFI accounting standards. In Jaffer, S. (Ed.) Global Growth, Opportunities and Challenges in the Sukuk Market. Euromoney Trading Inc.

Asad, S.I.-U.-D. (2009a) The business of shariah advice The DAWN newspaper, Karachi, Pakistan. Karachi, Pakistan, Pakistan Herald Publication (Pvt) Limited.

Asad, S.I.-U.-D. (2009b) An overview of the sukuk market. Jang Newspaper. Karachi, Pakistan, Jang Group of Newspapers.

Aseambankers (2005) Capitalising on Opportunities in the Sukuk Industry. Aseambankers Malaysia Berhad. Kuala Lumpur, Aseambankers.

Bose, S., SR. & McGee, R.W. (2008) Islamic Investment Funds: An Analysis of Risks and Returns. SSRN eLibrary.

Chapra, M.U. (1998) The Major Modes of Islamic Finance. 6th Intensive Orientation Course on “Islamic Economics, Banking and Finance.” the Islamic Foundation, Leicester, U.K.

Ariff and Safari


Daryanani, N. (2008) A Deeper Understanding on the Prohibition of Riba. Management. Nottingham, University of Nottingham, UK.

DIFC (2009) Dubai International Financial Centre Sukuk Guidebook, Dubai, UAE.

El-Gamal, M.A. (2000) A Basic Guide to Contemporary Islamic Banking and Finance. Plainfield, Indiana, Islamic Society of North America.

El-Gamal, M.A. (2007) Mutuality as an antidote to rent-seeking Shariah arbitrage in Islamic finance. Thunderbird International Business Review, 49, 187–202.

Enders, W. (1995) Applied Econometrics Time Series, New York, John. Wiley & Sons.

Farook, R. (2009) Global Financial Crisis Unthinkable Under Islamic Banking Principles. Sunday Observer. Sri Lanka, The Associated Newspapers of Ceylon Ltd.

Floor, W. (1990) AK (legal document, testament, money draft, check). In Yarshater, E. (Ed. Encyclopedia Iranica Online Edition. New York., Columbia University.

Gait, A.H. & Worthington, A.C. (2007) A Primer on Islamic Finance: Definitions, Sources, Principles and Methods. Working Papers Series of University of Wollongong. Wollongong, Australia, University of Wollongong.

Gambling, T., Jones, R. & Karim, R.A.A. (1993) Credible Organizations: Self-Regulation V. External Standard-Setting in Islamic Banks and British Charities. Financial Accountability & Management, 9, 195–207.

Granger, C.W.J. (1969) Investigating Causal Relations by Econometric Models and Cross-spectral Methods. Econometrica, 37, 424–438.

Gurgey, U. & Keki, E. (2008) Sukuk in Turkey. International Financial Law Review, 27, 111–111.

Hasan, Z. (2010) Islamic Finance: What Does It Change, What It Does Not – The Structure – Objectives Mismatch and Its Consequences. Working Paper. Kuala Lumpur, Malaysia, International Centre for Education in Islamic Finance (INCEIF).

Hossain, A. (2005) Granger-Causality Between Inflation, Money Growth, Currency Devaluation and Economic Growth in Indonesia, 1951–2002. International Journal of Applied Econometrics and Quantitative Studies, 2, 23.

IFSB (2009) Capital Adequacy Requirements for Sukuk, Securitizations and Real Estate Investment. In Board, I.F.S. (Ed. IFSB. Kuala Lumpur, Malaysia, Islamic Financial Services Board.

IFSL (2012) Islamic Finance. IN Mckenzie, D. (Ed. Finaicial Market Series. London, International Financial Services London.

Iqbal, M. (1998) Islamic banking. In Kahf, M. (Ed.) Lessons in Islamic Economics. Jeddah, Islamic Research and Training Institute.

Iqbal, M. & Molyneux, P. (2005) Thirty years of Islamic banking: History, Performance and Prospects, Houndmills, Basingstoke, Hampshire; New York, Palgrave Macmillan.

Iqbal, Z. (1999) Financial engineering in Islamic finance. Thunderbird International Business Review, 41, 541–559.

Jalil, A. (2005) Islamic Bonds Issues: The Malaysian Experience. In T. Ramanayanah, Mohamed Sulaiman, Hasnah Harun, Ruhaini Ali, Aizzat Mohd. Nasurdin, Nabsiah Abdul Wahid & Intan Osman (Eds.) The 6th Asian Academy of Management Conference, 9–11 December 2005. Casuarina Ipoh, Perak, Malaysia, Asian Academy of Management.

John, I. (2007) Sukuks key to meeting $1.5tr infrastructure needs of ME, Asia. Khaleej Times. Dubai, Galadari Printing and Publishing.

Kamali, M.H. (2007) A Shari’ah Analysis of Issues in Islamic Leasing. J.KAU: Islamic Econ, 20, 3–22.

Khan, M.M. & Bhatti, M. I. (2008) Development in Islamic banking: a financial risk-allocation approach. The Journal of Risk Finance, 9, 40–51.

Khan, T. & Ahmed, H. (2001) Risk Management: An Analysis of Issues in the Islamic Financial Industry, Jeddah, Saudi Arabia, Islamic Development Bank – Islamic Research and Training Institute.

Kharazmi, A.A.-A.M.B.A. (1895) Maf t h al-Ool m. In Vloten, G.V. (Ed.) Leiden.

Lewis, M.K. (2007) Islamic Banking in Theory and Practice. Monash Business Review, 3, 1–8.

Lewis, M.K. & Algaoud, L.M. (2001) Islamic banking, Cheltenham, UK, Edward Elgar.

Liquidity Management Center (2008) The Guide to Sukuk Market. Bahrain, November 2008.

Maurer, B. (2002) Anthropological and Accounting Knowledge in Islamic Banking and Finance: Rethinking Critical Accounts. The Journal of the Royal Anthropological Institute, 8, 645–667.

Maurer, B. (2010) Form versus substance: AAOIFI projects and Islamic fundamentals in the case of sukuk. Journal of Islamic Accounting and Business Research, 1, 32–41.

McKenzie, D. (2008) Islamic Finance 2008. International Financial Services London, Islamic Finance Working Group, 8.

McKenzie, D. (2009) Islamic Finance 2009. International Financial Services London (IFSL), 8.

McKenzie, D. (2010) Islamic Finance 2010. International Financial Services London (IFSL), 8.

Metwally, M.M. (2006) Economic Consequences of Applying Islamic Principles in Muslim Societies. Journal of Islamic Banking and Finance, 23, 11–33.

Mohd Asri, N. (2004) The Effect of Islamic Private Debt Securities Rating Changes on Firm’s Common Stock Returns. The Journal of Muamalat and Islamic Finance Research, 1, 25–38.

Pomeranz, F. (1997) The Accounting and Auditing Organization for Islamic Financial Institutions: An Important Regulatory Debut. International Journal of Accounting, Auditing and Taxation, 6, 123–30.



Rosly, S.A. (2007) Islamic Capital Market: Shariah Stocks, Sukuk, i-Reits, Islamic unit trust funds. LOFSA-INCEIF Workshop on Islamic Finance, 14th-15thNovember 2007, Labuan, Malaysia. Labuan, Malaysia, International Center For Education in Islamic Finance (INCEIF).

Shinsuke, N. (2007) Beyond the Theoretical Dichotomy in Islamic Finance: Analytical Reflections on Mur ba ah Contracts and Islamic Debt Securities. Kyoto Bulletin of Islamic Area Studies,, 1, 72–91.

Siddiqui, R. (2008) Contributing to the Development of the Islamic Capital Market: The Dow Jones Citigroup® Sukuk Index. Dow Jones Islamic Market Indexes NewsLetter, December, 1–2.

Sundararajan, V. & Errico, L. (2002) Islamic financial institutions and products in the global financial system: Key issues in risk management and challenges ahead. Risk Management in an Islamic Financial System, September 1, 2002, Tehran, Iran. Tehran, Iran Iran Banking Institute, Central Bank of the Islamic Republic of Iran.

Tariq, A.A. (2004) Managing Financial Risks of Sukuk Structures. School of Business and Economics. Leicestershire, UK, Loughborough University.

Usmani, M.T. (2001) Principles Of Shari’ah Governing Islamic Investment Funds.

Usmani, M.T. (2002) An Introduction to Islamic Finance, The Hague, Kluwer Law International.

Usmani, M.T. (2007) Sukuk and their Contemporary Applications. Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI), 14.

Warde, I. (2000) Islamic Finance in the Global Economy, Edinburgh Edinburgh University Press.

Wilson, R. (2004) Overview of the sukuk market. In Adam, N.J. & Thomas, A. (Eds.) Islamic Bonds: Your Guide to Issuing, Structuring and Investing in Sukuk. University of Durham, United Kingdom, Euromoney Institutional Investor PLC.

Wilson, R. (2008) Innovation in the structuring of Islamic sukuk securities. Humanomics, 24, 170–181.

Zaher, T.S. & Hassan, M.K. (2001) A Comparative Literature Survey of Islamic Finance and Banking. Financial Markets, Institutions & Instruments, 10, 155–199.

AcknowledgmentAn earlier version of this paper was presented at the 23rd Global Finance Conference, Chicago, USA in May 2012 at De Paul University. Authors would like to appreciate the feedbacks received from reviewers and participants of that conference. This research was funded by Khazanah Nasional Bhd., Malaysia, and the Maybank Endowed chair professorship at the University Putra Malaysia. The remaining errors are solely those of the authors.


Cite this chapter as: Fauzi F, Locke S, Basyith A, Idris M (2015). The impact of Islamic debt on company value. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

The impact of Islamic debt on company valueFitriya Fauzi1, Stuart Locke2, Abdul Basyith, Muhammad Idris1Faculty of Economics, University of Muhammadiyah, Palembang, Indonesia, Email: [email protected] Management School, The University Waikato

Abstract - This study uses micro-econometric analysis to examine the impact of Islamic debt on firm value and firm financial performance by observing Malaysian firms. A number of significant contributions to corporate finance arise from this research in relation to Islamic debt instruments and firm financial performance. First, it provides evidence of the Islamic debt impact on firm value and firm financial performance. Second, and very importantly it provides new insights, adding substantially to the very few studies that have been conducted on these types of instruments.

The choice of model employed is specified according to its diagnostic testing results for non-normality, heteroskedasticity, multicollinearity, endogeneity and linearity in. A test is conducted to confirm that there are no outliers in the data set prior to the diagnostic testing. Poolability and co-integration testing are also included. Based on the diagnostic results, data are analysed using the dynamic panel generalised method of moment (GMM using a quarterly balanced panel of 80 Malaysian firms issuing Islamic debt which spans from 2000 to 2009. This method is employed to investigate the impact of Islamic debt issues on firm value and/or firm financial performance.

The result reveals that Islamic debt has a significant positive impact on company value and firm financial performance. It also confirms that trade-off theory holds well in the Malaysian context for Islamic debt financing. Furthermore, the coefficient for Islamic debt is higher than the coefficient for non-Islamic debt, suggesting that the Islamic debt provides a higher contribution to firm value and to the improvement of firms’ financial performance compared to non-Islamic debt.

Keywords: Islamic debt, firm value

1. IntroductionModes of financing are a part of capital structure, and the relationship between capital structure and firm value is important and continues to be debated in the literature. An unleveraged firm can be seen as an all equity firm, whereas a leveraged firm is made up of ownership equity and debt. A firm’s debt to equity ratio provides a measure of the leverage or gearing. The influence of debt to equity on the value of a firm is the subject of multiple articles, including the seminal work of Modigliani and Miller (M&M). Modigliani and Miller achieved notoriety with their proof that capital structure made no different to firm value. Subsequent relaxation of the M&M assumptions suggested that debt to equity choice does impact on firm value. More recent research considering bankruptcy costs has contributed to empirical estimates of optimal leverage. The M&M argument is that, in a perfect market, how a firm is financed is irrelevant to its value. However, even though

it is widely accepted that the world is not made up of perfect markets, the issue of optimal D/E ratio remains contentious and unresolved.

Moreover, there are considerable studies that have investigated the impact of conventional debt on firm value, and those studies have generated few theories which are implemented up to now. However, so far there is no investigation of the financial performance and value consequences of using Islamic debt as opposed to non-Islamic debt (conventional debt). The lack of prior research into the financial aspect of Islamic debt does not reflect a lack of concern for such matters, but rather is a reflection of the newness of the topic. Therefore, the extent to which prior research is applicable to Islamic debt, particularly in emerging market, requires analysis, and this study is useful to supplement existing studies in this field and serves as a reference for studies in the future.

Fauzi et al.


The literature relating to Islamic debts has predominantly focused on the legal aspects of Islamic law, concept, basic requirements and the validity of how the debts are conducted in Islamic finance as general Islamic debt (Cakir, 2007; Mirakhor, 1996; Ashhari, 2009; Somolo, 2009; Tariq, 2007; Wilson, 2008). So far, researchers have been unable to find research that looks at the effects of Islamic debts on the value of the company in international contexts. Haneef (2009) discusses the history of Sukuk, explaining how it has evolved from an asset backed structure, where Sukuk holders have ownership rights over the underlying asset, to an asset based structure, where Sukuk holders rank paripassu with unsecured creditors. Other scholars (Abd.Sukor, 2008; Al-Amine, 2001; Juan, 2008; Kamali, 2007; Mohd Yatim, 2009; Mokhtar, 2009; Al-Amine, n.d.; Al Amine, 2008; Usmani, 1999, n.d.; Vishwanath, 2009; Wilson, 2008, n.d.; Yean, n.d.) also discuss the structure and the regulation of the Sukuk market in relation to Shariah perspective and Shariah compliancy. Therefore, this study attempts to examine the impact of Islamic debt on company value.

The rest of this paper is organised as follows. Section two presents the significance emergence of Islamic debt in the fast growing form of financing in emerging and mature markets. Section three provides literature review followed by section four, five and six which present the methodology, analysis and conclusion.

2. Significant emergence of Islamic debtThe Islamic financial and economic system has existed since the time of the prophet Muhammad SAW. During that time, buying and selling, and savings and loans activities were not as extensive as they are now. However, the principle remains the same; no interest charged and no non halal products and activities permitted. The interest system is not used at all because it is forbidden by Allah SWT. The banning was declared in the Quran and the Hadith.

Islamic debt, known as Sukuk, has evolved to become a significant part of corporate capital trading in the secondary market. The Accounting and Auditing Organization of Islamic Financial Institutions (AAOIFI) has also defined Islamic debt as certificates of equal value representing undivided shares in the ownership of tangible assets, usufruct and services or (in the ownership of) assets of the particular projects or any specified investment activity. Investment of Sukuk should be distinguished from common shares and bonds. While shares represent the ownership of a company as a whole and are for an indefinite period, Sukuk represent specified assets and are for a given period of time. Sukuk, unlike bonds, carry returns based on cash flow originating from the assets on the basis of which they are issued (Ayub, 2007; p. 392).

Islamic debt includes no periodic interest payments and provides a different cash flow profile when compared with non-Islamic debt instruments for borrowing companies and lenders. There is a socio-religious dimension relating to major principles that underlie all business transactions under Islamic law. All business transactions must adhere the teaching of the Islamic foundation, which is the Quran and Sunnah. There are at least four major prohibitions in Islamic business transactions. The first is the prohibition of

riba, known as adding any interest payments to a loan or other financing contract. The second is the prohibition from gharar and maisir, known as uncertainty and gambling; so transactions embodying these attributes will be considered invalid. The third is the prohibition of non-halal business transactions, such as alcohol, gambling and any other things that are prohibited and considered as non-halal. The fourth is the general prohibition of contracts that fail to meet the highest Shariah standards (Ayub, 2007).

The development of Sukuk is supported by many factors, including the development of Islamic banking (takaful) and an increasing demand for Islamic products in the debt market. The development of Sukuk with its associated types of structure has given rise to much discussion and debate among scholars of Islamic law. The uniqueness of Islamic debt compared to non-Islamic debt is that Islamic debt offers a secure investment based on the principle of rent and profit sharing without legalised interest system. It is constituted by pure motive of cooperation based on Islamic law. How the market prices this security in term of the yield curve and how risk pricing is embedded in the value and performance of the firms.

Recent innovations in Islamic finance have changed the dynamics of the Islamic finance industry, especially in the debt markets. Sukuk became increasingly popular as companies sought to raise funds by offering corporate Sukuk. It has become significant for raising funds in the international capital markets through Islamic Shariah. Increases in this market have been strong all over the world, especially in Malaysia, UAE and Saudi Arabia. In 1996 total Sukuk issued was USD 0.05b rising to USD 15.5b by the end of 2008. The most significant increase occurred in 2007 with more than 130 issues valued at USD 34.3b. The trend is apparent in Table 1 which shows a rapid expansion by value through to the financial crisis in 2008. Malaysia accounts for 43.7% of Sukuk issues followed by UAE with 30.1% and Saudi Arabia representing 10.4%. The size of offering by country for 2009 is shown in Table 2.

In the early years of Sukuk’s emergence as a financial instrument, murabahah and istisna were the most significant forms of issuance, accounting for 62.5% and 19.5% respectively. This changed between 2002 and 2007 when musyarakah and ijarah become the largest type of issue, accounting for 36.3% and 28.3% of the total market. In 2008 to 2009 the ranking reversed with ijarah and musyarakah accounting for 43.4% and 20.8% respectively as reflected in Table 3.

The increase in the issue size from year to year indicates that this market was gradually developing. It became lucrative for both the Sukuk issuer and the Sukuk holder, receiving increased support in the form of market surveillance and regulation, and from market participants.

The evolution of Sukuk structure is presented in Figure 1. The evolving of the original structure is due to the needs of this market for its product development. At the beginning of the emergence of Sukuk is debt based Sukuk. Murabahah Sukuk is one of the debt-based forms. The second stage of the evolution is asset-based Sukuk. One of the forms of this structure is Ijarah Sukuk. The last stage


The impact of Islamic debt on company value

Table 1. Global Sukuk issuance by year.

Year Value USD billion Number

1996 0.05 11997 0.9 11998 – –1999 0.2 42001 1.6 162002 2.9 232003 4.2 322004 3.5 502005 7.8 962006 19.5 992007 34.3 1302008 15.5 174

Source: ZawyaSukuk Monitor, 2009

0

1996

1997

1998

1999

2001

2002

2003

2004

2005

2006

2007

2008

20

40

60

80

100

120

140

160

180

200

Value USD billion

Number

Table 2. Global Sukuk issuance by country in 2009.

Country Value USD Billion Value in%

Malaysia 31.5 43.67Bahrain 5.2 7.21Indonesia 0.3 0.42UAE 21.7 30.08Pakistan 1.5 2.08Brunei Darussalam 0.7 0.97Kuwait 1.8 2.50Saudi Arabia 7.5 10.40Qatar 1.3 1.80UK 0.2 0.28Sudan 0.13 0.18USA 0.16 0.22Germany 0.14 0.19Total 72.13 100

Source: ZawyaSukuk Monitor, 2009

Value USD Billion

Value in %

0

5

10

15

20

25

30

35

40

45

50

Mala

ysia

Bahrai

n

Indo

nesia

UAE

Pakist

an

Brune

i Daru

ssalam

Kuwait

Saudi

Arabia

Qatar

UKSud

anUSA

German

y

of the evolution is equity-based Sukuk or partnership-based Sukuk. The forms of this structure are musyarakah and istisna’.

3. Literature reviewCapital structure is widely discussed in the finance literature. The mixture of debt to equity in the financial structure of companies and whether it will impact upon financial performance risk and valuation is the subject of theoretical and empirical studies. In general, capital structure theories are classified into three categories; first, the zero impact hypotheses or the Modigliani and Miller theory; second, the positive impact hypotheses or trade-off theory; third, the negative impact hypotheses or pecking order theory.

Modigliani and Miller theoryModigliani and Miller (1958) argue that capital structure is irrelevant, thus the total cash flows a company makes for all investors (debt holders and shareholders) are the same

regardless of capital structure. On the other hand Jensen and Meckling (1976) states that the amount of leverage in a firm’s capital structure is associated with its performance.

Furthermore, several researchers have conducted numerous studies which aim to examine the relationship between capital structure and firm performance. However, until now the evidence regarding this study is contradictory and mixed. (Ebaid, 2009; Ni & Yu, 2008; Phillips, 2004) find consistent results with Modigliani and Miller (M&M) theorem. On the other hand, (Abor, 2007; Bhabra, Liu & Tirtiroglu., 2008) find inconsistent results with M&M theorem. The focus on country effect, for example, developed and emerging markets has been done in some of the studies (Bhabra et al., 2008; Ebaid, 2009). Other studies document a focus on firm size, such as large, medium or small companies (Abor, 2007).

The Modigliani and Miller theory (1958) assumes that a capital market is perfect (no transaction or bankruptcy costs, perfect information), individuals and corporations can borrow at the same rate, and no taxes. It does not

Fauzi et al.


Figure 1. The evolution of Sukuk structure.

The evolution of Sukuk

1. Debt based Sukuk

4. Hybrid/mixed Sukuk

3. Equity based Sukuk

2. Asset based Sukuk

Table 3. Global Sukuk issuance by structure type.

YearType of structures

Value USD billion

Value in%

Phase I Murabahah 1.6 62.5(1996–2001) Al Salaam 0.16 6.3 Istisna 0.5 19.5 Ijarah 0.25 9.8 Mudarabah 0.05 2.0 Musyarakah – – Al Istithmar – – Hybrid – – Other – –

Total 2.56 100.0

Phase II Murabahah 4.9 6.8(2002–2007) Al Salaam 1.9 2.6

Istisnaa 4.1 5.7Ijarah 20.5 28.3Mudarabah 8 11.0Musyarakah 26.3 36.3Al Istithmar 2.9 4.0Hybrid 2.8 3.9Other 1 1.4

Total 72.4 100.0

Phase III Murabahah 4 12.8(2008–2009) Al Salaam 0.05 0.2

Istisnaa 0.08 0.3Ijarah 13.6 43.4Mudarabah 2.5 8.0Musyarakah 6.5 20.8Al Istithmar 3.5 11.2Hybrid 0.075 0.2Al Wakalah 1 3.2

Total 31.305 100.0

Source: ZawyaSukuk Monitor, 2009.

matter if the firm’s capital is raised by issuing stock or debt, or what the firm’s dividend policy is. Therefore, the M&M theory concludes that capital structure is irrelevant.

Modigliani and Miller made two propositions under these conditions. Their first proposition was that the value of a leveraged firm is the same as the value of an unleveraged firm. Their second proposition was that the expected return on equity is positively related to leverage because the risk to equity holders increases with leverage. These two propositions stand on the assumption that taxes are ignored, and bankruptcy cost and other agency costs were not considered.

When taxes were taken into account, they had two propositions as well. First, they state that the value of the firm is positively related to leverage. It means that corporate leverage lowers tax payments because corporations can deduct interest payments but not dividend payments. Secondly, they state that the cost of equity rises with leverage because the risk to equity rises with leverage. These propositions assume that firms have a capital

structure almost entirely composed of debt. But in the real world, firms cannot stand only with debt or a hundred percent leverage because an increase in debt will increase bankruptcy cost and agency cost. Consequently, it means that no optimal capital structure exists.

Trade-off theoryAfter the seminal work of M&M, little research has been done to explore capital structure in which some assumptions were proposed; trade off theory and pecking order theory. The trade-off theory derived from the models based on taxes and agency cost. Modigliani and Miller (1963), DeAngelo and Masulis (1980) and Jensen and Meckling (1976) suggest the firm has an optimal capital structure by offsetting the advantages of debt and the cost of debt. Therefore, trade off theory refers to the idea that a company chooses how much debt finance and how much equity finance to use by balancing the costs and benefits. It states that there is an advantage to financing with debt, the tax benefits of debt, and tax benefits to be had, but there is also a cost to financing with debt, the costs of financial distress including bankruptcy costs, and agency costs. This theory suggests that there is a positive relationship between debt level and firm performance. Moreover, the implication of this trade off theory is that firms have target leverage and they adjust their leverage toward the target over time. In addition, Harris and Raviv (1990) imply that higher leverage can be expected to be associated with larger firm value, higher debt level relative to expected income, and lower probability of reorganization following default.

The empirical relevance of the trade-off theory has often been questioned. Some research has been conducted to investigate this theory and the results from various contexts are mixed and inconclusive. The evidence does indicate there are likely to be differences attributable to firm size, country and the maturity of the respective capital market.

Pecking order theoryPecking order theory was developed by Myers and Majluf (1984). Myers and Majluf (1984) consider firms must issue common stock to raise cash to undertake a valuable investment opportunity. Management is assumed to know more about the firm’s value than potential investors. Investors interpret the firm’s actions rationally.



An equilibrium mode1 of the issue-investment decision is developed under these assumptions. The model shows that firms may refuse to issue stock, and therefore may pass up valuable investment opportunities. The model suggests explanations for several aspects of corporate financing behaviour, including the tendency to rely on internal sources of funds, and to prefer debt to equity if external financing is required.

In addition, Frank and Goyal (2003) state that capital structure is acquired in accordance with the priority of the firm in which internal funding is preferable and external funding is less preferable. If it is needed, firms could use external funding from the lowest risk debt. Therefore, pecking order theory refers to the idea that companies prefer to use their sources of financing from internal financing to equity. If external financing is required, firms issue the safest security first. That is, they start with secure debt, then perhaps equity as a last choice. In addition, issue costs are least for internal funds, low for debt and highest for equity. There is also the negative signaling to the stock market associated with issuing equity, positive signaling associated with debt, and asymmetric information between managers and investors. This theory suggests that there is a negative relationship between debt level and firm performance. Therefore, the implication of this pecking order theory is that firms prefer to depend on internal sources of funds and prefer debt to equity if external financing is required. Thus, a firm’s leverage is not driven by the trade-off theory, but rather by results of the firm’s attempts to mitigate signalling effect and information asymmetry.

The majority of studies have been conducted in mature markets with some based on developing markets including Asia, Africa and the Middle East. Previous studies also consider firm size, investigating whether large corporations and small and medium sized corporations behave differently. The results are mixed and inconclusive. In addition, previous studies examined different institutional structure (Booth, Aivazian, Demirguc-Kunt & Maksimovic., 2001); different governance mechanisms (Wiwattanakantang, 1999); different market power and firms investment (Eriotis, Frangouli & Neokosmides., 2002); different regional risk (Zeitun & Tian, 2007); different firm characteristics, ownership structure and industry membership (Bhabra et al., 2008). In comparison with the abundance of studies on the relationship between capital structure and firm performance and the determinant factors of capital structure in conventional debt, only a few studies focus on Islamic debt. This study will provide evidence about how Sukuk impacts upon financial performance and corporate value in markets.

Furthermore, the study of leverage impact on firm value has become an important aspect of capital structure theory. Myers (1984) claims that the firm value depends on the debt ratio, similarly, many studies have focused on the impact of the debt level and the debt type on a firm’s financial performance (Ebaid, 2009; Ghosh & Cai, 1999; Hatfield, Cheng & Davidson, 1994; Coleman, 2007; Talberg, Winge, Frydenberg & Westgaard, 2008). However, few recent studies investigate the impact of the Islamic debt type on a firm’s value and financial performance. This study attempts to investigate the impact of Islamic debt on company value and a firm’s financial performance.

4. Methodology

DataThe data for this study were obtained from the Islamic Finance Information Service (IFIS) database. The sampling period is 2000 to 2009, which is ten years and, this study used quarterly data. This quarterly data is important since the issuance of Islamic debt for every firm is in different quarters. Initially, this study proposed to investigate the debt choice impact on a company value and firm’s financial performance using Malaysian firms as a sample. Further, this study notes that 227 companies from Malaysia issued Islamic debt from 2000 to 2009. From those 227 Malaysian companies, 106 companies are public companies, and 121 companies are limited companies. From the 106 public companies, 31 companies have been excluded from the data list because of the unavailability of their financial statement data. In addition, the sample of Islamic debt offering must have data availability on the size of the offering, the maturity length, the history of the issuance, and other accounting data information.

For panel data analysis, the availability of ten years’ worth of data is required, particularly quarterly data. To mitigate the problem of missing values, this study uses multiple imputations by including the weighted value to compensate the missing value excluded in the model (Raghunathan, 2004).

VariablesFirm value and firm financial performance are dependent variables. Each of the performance indicators measures a different aspect of performance. Tobin’s Q is used as a firm value indicator,and it is considered as a market reflection of the firm’s activities and performances. Return on Asset (ROA) and Return on Equity (ROE) metrics are used as a firm financial performance indicator. ROE measures the performance from the perspective of the equity holders; meanwhile ROA measures the asset productivity and operating profit margin. It is important to note that none of these measures truly reflect the complete picture by themselves but have to be seen in conjunction with other metrics.

The proportion of the Islamic debt to non-Islamic debt, the proportion of Islamic debt, the frequency of Islamic debt issuance and the type of Islamic debt issued are used as independent variables. The proportion of Islamic debt is calculated as the total Islamic debt divided by the total assets/or total Islamic debt divided by the total Islamic debt plus total equity. The proportion of non-Islamic debt is calculated as the total of non-Islamic debt divided by the total assets/or total non-Islamic debt divided by the total non-Islamic debt plus total equity.

For the proportion of Islamic debt, the frequency of Islamic debt issued and the type of Islamic debt, dummy variables are employed. To avoid too many parameters and to find the unique least square estimates for the model, this study uses only n-1 dummy; therefore, the baseline is chosen for every set of the specifications. In addition, n-1 dummy may mitigate the problem of multicollienarity among the regressors (Baltagi, 2005). The choice of a baseline category

Fauzi et al.


is essentially arbitrary, for this study fits precisely with all regression models regardless of which category is selected for this role. The value and meaning of the individual dummy-variable coefficients δ1, δ2, ζ1, ζ1, η1 and η2 depend, however, on which category is chosen as the baseline.

The first group dummy is aimed at examining the effect of the Islamic debt proportion on each company, and three categories are set: first, a company having an Islamic debt proportion below the average of the Islamic debt proportion; second, a company having an average Islamic debt proportion; and third, a company having an Islamic debt proportion higher than the average of the Islamic debt proportion. The first category will be set as “1” if companies have a below average Islamic debt proportion; otherwise it is set equal to “0”. The second category will be set as “1” if companies have an average Islamic debt proportion; otherwise it is set equal to “0”. The third category will be set as “0” if companies have a higher than average Islamic debt proportion. The baseline category is used for this dummy if the company has a higher than average proportion Islamic debt. The average of the Islamic debt proportion is 8.06%, which is calculated by the total of the Islamic debt proportion over the total number of companies.

The second group dummy is aimed at examining the effect of the Islamic debt issuance frequency on each company, and three categories are set: first, a company issuing an Islamic debt only once; second, a company issuing an Islamic debt for the second time; and third, a company issuing an Islamic debt more than twice. The first category will be set as “1” if companies issue an Islamic debt only once; otherwise it is set equal to “0”. The second category will be set as “1” if companies issue an Islamic debt for the second time; otherwise it is set equal to “0”. The third category will be set as “0” if companies issue an Islamic debt more than twice. The baseline category is used for this dummy if the company has more than twice of the Islamic debt issuance.

The third group dummy is aimed at examining the effect of the Islamic debt type on each company, and three categories are set: first, a company issuing a debt-based type of Islamic debt; second, a company issuing an asset-based type of Islamic debt; and third, a company issuing an equity-based type of Islamic debt. The first category will be set as “1” if companies issue a debt-based type; otherwise it is set equal to “0”. The second category will be set as “1” if companies issue an asset-based type; otherwise it is set equal to “0”. The third category will be set as “0” if companies issue an equity-based type. The baseline category is used for this dummy if the company issues an equity-based type of the Islamic debt.

Prior research suggests that the performance of each firm may differ according to their size, because larger firms have greater economies of scale in the transaction costs associated with long term debt, which may influence the results and inferences (Ramaswamy, 2001; Frank & Goyal, 2003; Coleman, 2007; Jermias, 2008; Ebaid, 2009). In addition, larger firms have less potential of bankruptcy cost; therefore, firm size should be positively related to the borrowing capacity (Krishnan & Moyer, 1997). This study uses a natural logarithm of the total assets as a proxy for firm size as the control variable (Naceur & Goaied, 2002;

Akhtar, 2005; Zeitun & Tian, 2007; Talberg et al., 2008). The natural logarithm is applied for the firm size variable owing to the skewness and kurtosis problem. Further, natural logarithm ensures that the actual regressor has less statistical noise in the regression model, and moderates the effects of the large size of the firm.

Model specificationThis study uses the panel data method which allows the unobservable heterogeneity for each observation in the sample to be eliminated and multicollinearity among variables to be alleviated. Unobservable heterogeneity might result in spurious correlations with the dependent variables, which would bias the coefficient obtained (Baltagi, 2005). Before proceeding to the model specification, diagnostic testing of normality, heteroskedasticity, multicollinearity, and autocorrelation, was conducted to determine the appropriate method used in this study. The specification testing results are provided in Tables 4 and 5.

The heteroskedasticity result is 462.99 with p-value 0.0000, suggesting that there is a heteroskedaticity problem. Therefore, this problem needs to be catered to obtain efficient and unbiased results. The skewness and kurtosis results are 45.72 with p-value 0.000 and 4.25 with p-value 0.0392, suggesting that non-normal distribution, thus this non-normal distribution has to be treated. Therefore, outliers’ checking is conducted prior to data analysing.

The multicollinearity result is 10.9100 with p-value 0.000, suggesting no multicollinearity problem among the explanatory variables. Before proceeding to the endogeneity test and linearity test results, a brief conclusion made is that heteroskedasticity and non-normality problems have to be

Table 4. Summary of the specification testing results.

Tests p-value

Heteroskedasticity 462.99* 0.0000Skewness 45.72* 0.0000Kurtosis 4.25* 0.0392Multicollinearity 10.91* 0.0000Linearity 2.9055* 0.0000Endogeneity Endogeneity exist

*Sig. at 1% significance level.

Table 5. The DWH test for endogeneity of regressors.

Variables Tobin’s Q ROA ROE

Islamic Debt Proportion

0.0820(0.7745)

6.9848*(0.0083)

0.0035(0.9526)

Non-Islamic Debt Proportion

37.4723*(0.0000)

0.0036(0.9523)

14.2038*(0.0002)

*Sig. at 1% significance level.



treated. The linearity test result for group 1 is 2.905506 with p-value 0.0000, which rejects the null hypothesis of nonlinearity. Similar to the result for group 1, the linearity test result for group 2 is -4.659 with p-value 0.000, which supports the linear model. The endogeneity test result reveals that the regressors in the model present endogeneity.

Supported by numerous previous studies, by the assumptions above, by the endogeneity tests and by the linearity tests, the Generalised Method of Moments is appropriate as it corrects for heteroskedasticity, the endogeneity problems and reduces multicollinearity, hence improving the efficiency of the estimates.In conclusion, according to specification testing results, a linear dynamic panel GMM is employed.

Before constructing the dynamic panel GMM model, the equation below is a starting point for this study to establish if the debt choice has an impact on a firm’s value and firm’s financial performance. A model for the regression of Islamic debt, non-Islamic debt, the proportion of Islamic debt, the frequency of Islamic debt issuance, and the type of the Islamic debt issued is then:

a b b δ δ ζζ η η

= + + + + ++ + + +

y X X K K NN Z Z u

it i i i i i i i

i i i it

1 1 2 2 1 1 2 2 1 1

2 2 1 1 2 2 (1)

µ λ= + +u vit i t it 002.eps (2)

i = 1, …, N; t = 1, …, T,

where yi is firm’s value and/or firm’s financial performance. Xi1 is Islamic debt, Xi2 is non-Islamic debt and, Xi3 is firm size. K is the dummy proportion for Islamic debt, N is the dummy frequency for Islamic debt and Z is the dummy Islamic debt type. mi denotes the unobservable individual effect, lt denotes the unobservable time effect, and vit is the remainder stochastic disturbance term. This model describes three parallel regression planes, which can differ in their intercepts. Hereafter, the X, K, N, Z will be referred as Xit (set of regressors):

y x u N and t Tit i it it= + + = … = …a β ′ , , , , ,i 1 13.eps (3)

where xit is a K × 1 vector of regressors, β is a K × 1 vector of parameters to be estimated, and ai represents time-invariant individual nuisance parameters. Under the null hypothesis, uit is assumed to be independent and identically distributed (i.i.d.) over periods and across cross-sectional units.

The GMM equation model (Blundell & Bond, 1998) is specified as follows (Cameron & Trivedi, 2010):

a b η e- = - + ′ + +- -y y y X( 1)it it it it i it1 1 (4)

where yit is the Tobin’s Q at time t for firm i, Xit is a set of regressors, ηi is an unobserved firm-specific effect and eit is a stochastic error. Moving to the right yit−1, it is to obtain:

a b η e= + ′ + +-y y X .it it it i it1 (5)

Taking into account the first difference, one can elide the unobserved firm-specific effect:

a b e e- = - + ′ - + -- - - - -y y y y X X( ) ( ) ( )it it it it it it it it1 1 2 1 1 (6)

where X includes lag performance for explanatory varia bles, yit−1 as well as dependent variables. The first-differencing eliminates potential bias that arises from unobservable heterogeneity. After first-differencing, GMM estimation uses lagged values as instruments for Xit - Xit−1:

Supposing that the regressors are predetermined, it is possible to obtain consistent estimates of coefficients performing a GMM estimator that exploits the following orthogonality conditions;

e e( )- = ≥ = …- -E y s t T0 for 2 and 3, ,it s it it 1

(7)

E X s t Tit s it it- --( ) = ≥ = …e e 1 0 2 3, ,for and (8)

Then, the instrumental variable estimation in the first difference model is:

y y y X

t p T

,

1, ,it it p it p it it1 1

Δγ γ b eΔ = +…+ Δ + Δ ′ + Δ= + …

- - (9)

where Δ is the first difference operator. The first difference in equations and levels using their past levels/first differences are used for the instrumented variables. Further, a test of overidentifiying restrictions is necessary to test the validity of overidentifying instruments in an overidentified model to identify that the parameters of the model are estimated using optimal GMM. This test is called Hansen’s test, and the null hypothesis is that all instruments are valid. At last, weak instruments testing is done to identify whether the instrument is weak, and the overidentified model is used because the model has only one endogenous regressor that is overidentified (Cameron & Trivedi, 2010; p. 191–199).

5. AnalysisThe sample used consists of 80 listed firms issuing Islamic debt for the period of 2000 to 2009. Therefore, there are approximately 3,200 observations used. Table 6 provides the descriptive statistics used in this study. The table depicts the number of observations, mean, standard deviation, minimum and maximum value of each variable. The dependent variables are Tobin’s Q, ROA and ROE,and each of these dependent variables is regressed toward its explanatory variables.This study divides all explanatory variables into four categories. The first category is the debt structure used by the firm. The second category is the frequency of Islamic debt issuance. The third category is the Islamic debt proportion issued. The fourth category is the Islamic debt type issued. Firm size and year of Islamic debt issued are used as control variables.

The mean value for Tobin’s Q is 0.1679 with a range of -1.6600 to 1.9938, suggesting that most of the firms experienced low firm performance based on the market measure. A low Tobin’s Q may indicate that the stock is

Fauzi et al.


undervalued. Theoretically, stock being undervalued is likely to happen in a firm which has a stable earning history, a historically consistent return on equity and a higher earnings growth rate compared to the market average. Apparently, this seems to be consistent with the sample used for this group, in which the majority of firms are large firms (see the mean value of firm size, which suggests that most of the firms are big firms).

The mean value for ROA is 0.0925 with a range of 0.0100 to 0.1526. Though the mean value of ROA is considerably small, this positive value indicates that the firms in the sample create shareholder value over the sampling period. This positive value also indicates an effective utilisation of firm assets in generating an operating surplus in the business. This lower value of ROA may indicate that the firms are asset-intensive firms. If so, they thus require more money to be invested into the business to continue generating earnings. According to a common rule, ROA below 5% indicates asset-heavy firms (for example; manufacturing, railroads, telecommunication providers, car manufacturers, etc); meanwhile ROA above 20%

indicates asset-light firms (for example, agency firms, software firms, advertising firms, etc). The ROA is approximately 9% which may indicate that the majority of the firms used in the sample are asset-heavy firms, and represent a variety of sectors. These are a few examples of the firms used in the sample: Esso Malaysia is one of the biggest fuel providers in Malaysia, Hubline is one of the biggest shipping service providers, Kinsteel is one of the largest steel millers, Kuala Kepong is the largest rubber plantation and manufacturer, and Zecon is a construction, infrastructure, toll concession and property development company.

The mean value for ROE is 0.0156 with a range of 0.0021 to 0.2292, suggesting that most of the firms experienced low firm performance based on accounting measures. However, the positive value indicates that the firms in the sample create shareholder value and operating efficiency is positively translated into benefits to the owners. Furthermore, the lower value of ROE may indicate that the majority of the firms require more capital invested as discussed in point two, where it is noted that the majority

Table 6. Descriptive statistics.

Variables Obs. Mean Std. Dev. Min Max

Dependent variablesTobin’s Q 80 0.1679 0.2129 -1.6600 1.9938ROA 80 0.0925 0.0004 0.0100 0.1526ROE 80 0.0156 0.0352 0.0021 0.2292Explanatory variables the debt structure of the firmIslamic Debt Proportion 80 0.0806 0.0847 0.0102 0.4576Non-Islamic Debt Proportion 80 0.2174 0.1725 0.0598 0.8732The frequency of Islamic debt issuanceFirst Issuance 80 0.0000 0.0000 0.0000 0.0000Second Issuance 80 0.1316 0.3381 0.0000 1.0000More Than two Issuance 80 0.4211 0.4938 0.0000 1.0000The proportion of Islamic debt issuedIslamic Debt Below Average 80 0.8813 0.3235 0.0000 1.0000Islamic Debt Average 80 0.0000 0.0000 0.0000 0.0000Islamic Debt Above Average 80 0.1164 0.3208 0.0000 1.0000The type of Islamic debt issued 80 Debt Type of Islamic Debt 80 0.0000 0.0000 0.0000 0.0000Asset Type of Islamic Debt 80 0.1053 0.3069 0.0000 1.0000Equity Type of Islamic Debt 80 0.1316 0.3381 0.0000 1.0000Control Variables Size effectFirm Size 80 6.0388 0.7254 4.6032 8.4924Year effectYear 2001 80 0.0119 0.1081 0.0000 1.0000Year 2003 80 0.0625 0.2440 0.0000 1.0000Year 2004 80 0.1563 0.3660 0.0000 1.0000Year 2005 80 0.3438 0.4787 0.0000 1.0000Year 2006 80 0.1719 0.3803 0.0000 1.0000Year 2007 80 0.1406 0.3504 0.0000 1.0000Year 2008 80 0.0938 0.2938 0.0000 1.0000Year 2009 80 0.0313 0.1754 0.0000 1.0000



of the firms are asset-heavy. Therefore, the lower value of ROE does not mean that they have lower performance. Moreover, those asset-heavy firms have less competition as the entry barrier is high. This can be said to be one of the competitive advantages of these firms.

The mean value for Islamic debt proportion is 0.0806 with a range of 0.0102 to 0.4576, indicating that most of the firms issued small amounts of Islamic debt. This may be due to the fact that this Islamic debt is traded in the thin trading, moreover, some of the Islamic debt type certificates cannot be traded in the stock exchange due to its Islamic law issue. The mean value for non-Islamic debt proportion is 0.2174 with a range of 0.0598 to 0.8732, indicating that most of the firms are not highly leveraged. This also suggests that the majority of the firms are less risky since excessive debt can lead to greater interest payments and principal repayment burden.

First issuance is used as a baseline category for the frequency of Islamic debt issuance, and it takes the value of zero. The mean value for the second issuance of Islamic debt is 0.1316 with a range of 0.0000 to 1.0000, suggesting that only 13.16% of the firms issued Islamic debt for the second time. The mean value for more than two issuance is 0.4211 with a range of 0.0000 to 1.0000, suggesting that most of the firms issued Islamic debt more than twice.

The average of the Islamic debt proportion is 8.06%, which is calculated by the total of the Islamic debt proportion over the total firms in the sample, and thus, this 8.06% average value is used as the average category. The mean value for Islamic debt below average is 0.8831 with a range of 0.0000 to 1.000, suggesting that most of the firms issued Islamic debt no greater than 10% (below the average). Islamic debt average is used as a baseline category for the proportion of Islamic debt issued, and it takes the value of zero. The mean value of Islamic debt above average is 0.1164 with a range of 0.0000 to 1.0000, suggesting that only a few firms issued Islamic debt greater than the average. This may be due to the fact that excessive debt issued might increase the probability of default. Therefore, the issuers have to assess the trade-off between the Islamic debt and any other potential risks arising as a result of this debt.

Debt type is used as a baseline category for the Islamic debt type and it takes the value of zero. The mean value for asset type of Islamic debt is 0.1053 with a range of 0.0000 to 1.0000, suggesting that only 10.53% of the firms in the sample issued this type of Islamic debt. The mean value for the equity type of Islamic debt is 0.1316 with a range of 0.0000 to 1.0000, suggesting that only 13.16% of the firms in the sample issued this type of Islamic debt.

The mean value for firm size is 6.0388 with a range of 4.6032 to 8.4924, suggesting that most of the firms are big firms (see explanation on point two). During the sampling period 2000 to 2009, Islamic debt is only issued during these eight years – 2001, 2003 to 2009. Islamic debt is mostly issued in 2005 which accounted for 34.38%. The mean value for 2001, 2003, 2004, 2005, 2006, 2007, 2008 and 2009 are 1.19%, 6.25%, 15.63%, 34.38%, 17.19%, 14.06%, 9.38% and 3.13% respectively from the total sample.

Table 7 provides a pairwise correlation matrix of the explanatory variables. The highest correlation is between the Islamic debt proportion and Tobin’s Q, which counts for 0.4379 (p-value 0.0000) and this value is significant. The second highest correlation is between the proportion of Islamic debt and the Islamic debt above average, which counts for 0.5979 (p-value 0.0000). The third highest correlation is between the proportion of Islamic debt and the Islamic debt below average, which counts for -0.5965 (p-value 0.0000). The rest of the correlation coefficient is less than 0.5, and it is considered as a low correlation between the explanatory variables, thus, giving less cause for concern about the multicollinearity problem.

Table 8 presents the dynamic GMM panel regression results. There are three regression equations, and there are four explanatory variable categories.

The debt structure of the firm and Tobin’s Q, ROA and ROEThe coefficient of Islamic debt and non-Islamic debt are a positive and significant, indicating that these two variables have a positive effect on a firm’s financial performance. Both variables are statistically significant at a 1% level. This finding can be better explained by trade-off theory. According to prior literature, a firm has an optimal capital structure by offsetting the advantages of debt and the cost of debt (Modigliani & Miller, 1963; DeAngelo & Masulis, 1980; Jensen & Meckling, 1976; Haris & Raviv, 1990; Frank & Goyal, 2003), and this theory apparently can also be applied to Islamic debt. Trade-off theory refers to the idea that a company chooses how much debt finance and how much equity finance to use by balancing the costs and benefits. It states that there is an advantage to financing with debt and the tax benefits of debt, and fortunately Islamic debt is exempted from the taxes. Moreover, the use of leverage is one way to improve firm performance (Champion, 1999), and firms prefer debt financing because they anticipate a higher return (Hadlock & James, 2002). Furthermore, this finding is in line with Krishnan and Moyer (1997) and Abor (2005) who find a positive relationship between capital structure choice and firm financial performance in developing countries. In particular, Krishnan and Moyer (1997) include Malaysia as one of the sample in their study.

The positive result for Islamic debt coefficient obtained supports the trade-off theory, which was derived from the models based on taxes and agency cost. From the point of view of internal management, having Islamic debt in their debt structure brings more pressure to the management as Islamic debt is more expensive compared to non-Islamic debt, hence, improving the firm’s efficiency is important to maximise asset utilisation due to the Islamic debt obtained. At the end, this action leads to improvement in the firm’s performance. Moreover, debt may reduce agency costs by reducing cash flows available for expropriation and investments in negative net present value projects (Harris & Raviv, 1990; Jensen, 1986), as does Islamic debt. Furthermore, compared to equity issues, the issue of debt will not dilute the managers’ equity holdings as a proportion of total equity, but further enhance the alignment of interests (Fleming, Heaney & McCosker, 2005). In addition, though conventional debt and Islamic

Fauzi et al.


Tabl

e 7.

Pai

rwis

e co

rrel

atio

n m

atri

x fo

r exp

lana

tory

var

iabl

es.

Vari

able

sTo

bin’

s Q

RO

AR

OE

Isla

mic

Deb

t Pr

opor

tion

Non

-Isl

amic

D

ebt P

ropo

rtio

nFi

rst

Issu

ance

Tobi

n’s

Q1.

0000

RO

A0.

1525

***

(0.0

000)

1.00

00

RO

E0.

0382

**

(0.0

353)

0.18

29**

* (0

.000

0)1.

0000

Isla

mic

Deb

t Pro

port

ion

0.43

79**

* (0

.000

0)0.

0393

**

(0.0

304)

0.10

06**

* (0

.000

0)1.

0000

Non

-Isl

amic

Deb

t Pro

port

ion

0.09

97**

* (0

.000

0)0.

0528

***

(0.0

036)

0.04

66**

* (0

.010

1)-0

.277

3***

(0

.000

0)1.

0000

Firs

t Iss

uanc

e0.

0000

(0

.000

0)0.

0000

(0

.000

0)0.

0000

(0

.000

0)0.

0000

(0

.000

0)0.

0000

(0

.000

0)0.

0000

(0

.000

0)Se

cond

Issu

ance

-0.0

681*

**

(0.0

002)

0.02

85

(0.1

162)

0.00

3 (0

.866

6)-0

.084

7***

(0

.000

0)-0

.091

9***

(0

.000

0)0.

0000

(0

.000

0)M

ore

than

2 Is

suan

ce-0

.069

9***

(0

.000

1)0.

0346

**(0

.056

5)0.

0920

***

(0.0

000)

-0.0

334*

(0

.065

9)-0

.078

0***

(0

.000

0)0.

0000

(0

.000

0)Is

lam

ic D

ebt B

elow

Ave

rage

-0.3

119*

**(0

.000

0)0.

0780

***

(0.0

000)

0.10

48**

*(0

.000

0)-0

.596

5***

(0.0

000)

0.07

88**

*(0

.000

0)0.

0000

(0.0

000)

Isla

mic

Deb

t Ave

rage

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)0.

0000

(0.0

000)

Isla

mic

Deb

t Abo

ve A

vera

ge0.

3119

***

(0.0

000)

0.08

13**

* (0

.000

0)0.

1059

***

(0.0

000)

0.59

79**

* (0

.000

0)-0

.081

7***

(0

.000

0)0.

0000

(0

.000

0)D

ebt T

ype

of Is

lam

ic D

ebt

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)0.

0000

(0.0

000)

Ass

et T

ype

of Is

lam

ic D

ebt

-0.0

556*

**(0

.002

1)0.

0228

(0.2

080)

0.00

87(0

.630

4)-0

.040

0**

(0.0

274)

-0.0

587*

**(0

.001

2)0.

0000

(0.0

000)

Equi

ty T

ype

of Is

lam

ic D

ebt

0.00

45(0

.805

6)0.

0073

(0.6

892)

0.10

06**

*(0

.000

0)0.

1094

***

(0.0

000)

0.20

26**

*(0

.000

0)0.

0000

(0.0

000)

Firm

Siz

e-0

.207

9***

(0.0

000)

0.00

82(0

.687

8)0.

1971

***

(0.0

000)

-0.1

369*

**(0

.000

0)0.

4492

***

(0.0

000)

0.00

00(0

.000

0)Ye

ar 2

001

-0.0

056

(0.7

587)

0.00

00(1

.000

0)-0

.048

3***

(0.0

078)

0.18

82**

*(0

.000

0)0.

1362

***

(0.0

000)

0.00

00(0

.000

0)Ye

ar 2

003

0.05

73**

*(0

.001

6)0.

0000

(1.0

000)

-0.0

292

(0.1

074)

0.06

11**

*(0

.000

7)0.

0971

***

(0.0

000)

0.00

00(0

.000

0)Ye

ar 2

004

0.03

48**

(0.0

547)

-0.1

001*

**(0

.000

0)-0

.088

4***

(0.0

000)

0.02

71(0

.134

9)-0

.110

6***

(0.0

000)

0.00

00(0

.000

0)Ye

ar 2

005

0.20

12**

*(0

.000

0)0.

0097

(0.5

919)

-0.0

105

(0.5

624)

0.30

00**

*(0

.000

0)-0

.090

1***

(0.0

000)

0.00

00(0

.000

0)Ye

ar 2

006

0.06

89**

*(0

.000

1)0.

0000

(1.0

000)

0.00

37(0

.836

8)0.

0496

***

(0.0

062)

-0.0

866*

**(0

.000

0)0.

0000

(0.0

000)

Year

200

70.

0078

(0.6

681)

0.03

55**

(0.0

501)

0.05

42**

*(0

.002

8)0.

0237

(0.1

917)

0.06

24**

*(0

.000

6)0.

0000

(0.0

000)

Year

200

8-0

.036

1**

(0.0

466)

0.00

00(1

.000

0)0.

0732

***

(0.0

001)

0.01

53(0

.398

)0.

0330

***

(0.0

690)

0.00

00(0

.000

0)Ye

ar 2

009

-0.0

688*

**(0

.000

1)0.

0000

(1.0

000)

0.00

10(0

.957

2)-0

.017

7(0

.330

4)-0

.014

8(0

.413

5)0.

0000

(0.0

000)



Tabl

e 7.

(Co

ntin

ued)

Seco

nd

Issu

ance

Mor

e Th

an

2 Is

suan

ceIs

lam

ic D

ebt

Bel

ow A

vera

geIs

lam

ic D

ebt

Abo

ve A

vera

geD

ebt T

ype

of

Isla

mic

Deb

tA

sset

Typ

e of

Is

lam

ic D

ebt

Seco

nd Is

suan

ce1.

0000

Mor

e th

an 2

Issu

ance

-0.3

320*

**(0

.000

0)1.

0000

Isla

mic

Deb

t Bel

ow A

vera

ge0.

1188

***

(0.0

000)

-0.0

309*

(0.0

886)

1.00

00

Isla

mic

Deb

t Ave

rage

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)Is

lam

ic D

ebt A

bove

Ave

rage

-0.1

170*

**(0

.000

0)0.

0290

(0.1

103)

-0.9

890*

**(0

.000

0)1.

0000

Deb

t Typ

e of

Isla

mic

Deb

t0.

0000

(0.0

000)

0.00

00(0

.000

0)0.

0000

(0.0

000)

0.00

00(0

.000

0)1.

0000

Ass

et T

ype

of Is

lam

ic D

ebt

-0.1

335*

**(0

.000

0)0.

0548

***

(0.0

025)

0.01

99(0

.273

2)-0

.020

9(0

.248

7)0.

0000

(0.0

000)

1.00

00

Equi

ty T

ype

of Is

lam

ic D

ebt

-0.0

364*

*(0

.045

0)-0

.016

60.

0075

(0.6

785)

-0.0

109

(0.5

496)

0.00

00(0

.000

0)-0

.133

5***

(0.0

000)

(0.3

603)

Firm

Siz

e0.

1010

***

(0.0

000)

-0.0

076

(0.7

086)

0.20

83**

*(0

.000

0)-0

.206

0***

(0.0

000)

0.00

00(0

.000

0)-0

.036

3*(0

.074

0)Ye

ar 2

001

-0.0

426*

**(0

.018

8)-0

.093

4***

(0.0

000)

-0.1

008*

**(0

.000

0)0.

1025

***

(0.0

000)

0.00

00(0

.000

0)-0

.037

5**

(0.0

384)

Year

200

30.

0088

(0.6

293)

0.00

00(0

.000

0)-0

.118

9***

(0.0

000)

0.12

18**

*(0

.000

0)0.

0000

(0.0

000)

-0.0

900*

**(0

.000

0)Ye

ar 2

004

0.04

58**

*(0

.011

5)-0

.092

6***

(0.0

000)

-0.0

096

(0.5

979)

0.01

19(0

.513

2)0.

0000

(0.0

000)

-0.0

144

(0.4

266)

Year

200

50.

0449

***

(0.0

132)

-0.0

038

(0.8

321)

-0.3

430*

**(0

.000

0)0.

3406

***

(0.0

000)

0.00

00(0

.000

0)-0

.089

7***

(0.0

000)

Year

200

6-0

.101

1***

(0.0

000)

0.17

30**

*(0

.000

0)-0

.076

1***

(0.0

000)

0.07

44**

*(0

.000

0)0.

0000

(0.0

000)

-0.0

186

(0.3

065)

Year

200

70.

0231

(0.2

020)

0.07

33*

(0.0

001)

-0.0

665*

**(0

.000

2)0.

0591

***

(0.0

011)

0.00

00(0

.000

0)0.

1052

***

(0.0

000)

Year

200

80.

0046

(0.8

011)

-0.0

772*

**(0

.000

0)-0

.010

2(0

.573

7)0.

0037

(0.8

405)

0.00

00(0

.000

0)0.

0168

(0.3

550)

Year

200

9-0

.020

0(0

.270

4)0.

0082

(0.6

508)

0.01

89(0

.298

7)-0

.018

6(0

.304

0)0.

0000

(0.0

000)

0.06

61**

*(0

.000

3)

Fauzi et al.


Tabl

e 7.

(Co

ntin

ued)

Equi

ty T

ype

of

Isla

mic

Deb

tFi

rm S

ize

Year

200

1Ye

ar 2

004

Year

200

5Ye

ar 2

006

Equi

ty T

ype

of Is

lam

ic D

ebt

1.00

00

Firm

Siz

e0.

2151

***

(0.0

000)

1.00

00

Year

200

1-0

.042

6***

(0.0

188)

0.00

85(0

.674

6)1.

0000

Year

200

3-0

.102

2***

(0.0

000)

0.01

67(0

.410

7)-0

.028

7(0

.113

1)Ye

ar 2

004

-0.1

200*

**(0

.000

0)-0

.102

0***

(0.0

000)

-0.0

338*

(0.0

627)

1.00

00

Year

200

5-0

.061

8***

(0.0

007)

-0.0

172

(0.3

959)

-0.0

474*

**(0

.008

9)-0

.133

5***

(0.0

000)

1.00

00

Year

200

6-0

.101

1***

(0.0

000)

-0.0

618*

**(0

.002

3)-0

.028

4(0

.117

1)-0

.080

1***

(0.0

000)

-0.1

124*

**(0

.000

0)1.

0000

Year

200

70.

0781

***

(0.0

000)

0.09

17**

*(0

.000

0)-0

.024

4(0

.178

4)-0

.068

8***

(0.0

001)

-0.0

966*

**(0

.000

0)-0

.057

9***

(0.0

014)

Year

200

80.

2361

***

0.12

15**

*(0

.000

0)-0

.015

0(0

.408

5)-0

.042

2***

(0.0

198)

-0.0

593*

**(0

.001

1)-0

.035

6**

(0.0

499)

(0.0

000)

Year

200

9-0

.020

0(0

.270

4)0.

0244

(0.2

299)

-0.0

056

(0.7

566)

-0.0

158

(0.3

826)

-0.0

222

(0.2

202)

-0.0

133

(0.4

623)

Year

200

7Ye

ar 2

008

Year

200

9

Year

200

71.

0000

Year

200

8-0

.030

5*(0

.092

2)1.

0000

Year

200

9-0

.011

5(0

.527

8)-0

.007

0(0

.698

1)1.

0000

***S

ig. a

t 1%

sig

nific

ance

leve

l, **

Sig.

at 5

% s

igni

fican

ce le

vel a

nd *

Sig.

at 1

0% s

igni

fican

ce le

vel.



Table 8. Summary of the regression result.

Variables

Dynamic GMM

Tobin’s Q ROA ROE

L1 0.5742***(0.0174)

0.1220***(0.0041)

0.1547***(0.0015)

Constant 0.3180***(0.0988)

8.7500***(0.0410)

0.8463**(0.4264)

The debt structure of the firmIslamic Debt Proportion 0.7441***

(0.0723)0.0159***(0.0025)

0.0276**(0.2540)

Non-Islamic Debt Proportion

0.5372***(0.0159)

0.0177***(0.0015)

0.3357***(0.0544)

The frequency of Islamic debt issuanceD1_First Issuance (Omitted) (Omitted) (Omitted)D2_Second Issuance

-0.0961*(0.0951)

-0.0774***(0.0140)

-0.7018***(0.2332)

D3_More Than 2 Issuance 0.2090***(0.0562)

0.0323***(0.0026)

1.2480***(0.1135)

The proportion of Islamic debt issuedD1_ID Below Average 0.0006*

(0.0129)0.0181***(0.0010)

0.1319*(0.1146)

D2_ID Average (Omitted) (Omitted) (Omitted)D3_ID Above Average

-0.0248**(0.0111)

-0.0036***(0.0010)

-0.0135**(0.1266)

The type of Islamic debt issuedD1_Debt Type of ID

(Omitted)

(Omitted)

(Omitted)

D2_Asset Type of ID

0.0423(0.1076)

0.0250***(0.0074)

0.3337(0.2695)

D3_Equity Type of ID 0.1764*(0.1049)

0.0346***(0.0027)

1.0458***(0.0532)

Control Variables Size effectFirm Size -0.0844***

(0.0142)-0.0040***(0.0005)

-0.2354***(0.0654)

Year effectYear 2001 -1.5848*

(0.9510)-0.0169*(0.0097)

-0.5066***(0.1019)

Year 2003 0.0458***(0.0098)

0.0092***(0.0024)

-0.0786*(0.0433)

Year 2004 0.0130(0.0109)

0.0173***(0.0019)

0.3969***(0.0657)

Year 2005 0.0171***(0.0048)

0.0058***(0.0014)

0.0981***(0.0159)

Year 2006 -0.0346***(0.0072)

-0.0018***(0.0003)

-0.0392***(0.0078)

Year 2007 0.1160***(0.0177)

0.0080***(0.0020)

-0.0889***(0.0298)

Year 2008 0.0397*(0.0236)

0.0017***(0.0006)

0.8061***(0.1146)

Year 2009 0.0587**(0.0294)

0.0048***(0.0019)

0.7980**(2.8289)

J-StatisticsChi2

21514.850.0000

78177.740.0000

1.27e+060.0000

Fauzi et al.


debt are fundamentally different, they perform similarly in a competitive market as these two instruments are affected by the same factors (Kraciska & Nowak, 2012).

Although both debt types have a positive impact, the coefficient for Islamic debt is higher than the coefficient for non-Islamic debt (only for Tobins’ Q), suggesting that the Islamic debt provides a higher contribution to the improvement of firms’ financial performance compared to non-Islamic debt. Furthermore, it can be concluded that when Islamic debt is chosen as a tool of firm financing, (1) the markets react positively to firm performance, thus this positive reaction might lead to the stock becoming overvalued;(2) Islamic debt not only improves the effectiveness of the firm’s management s in managing their assets to generate profits, but it also improves the operating efficiency of the total business; (3) firms are effective in managing their operation efficiency which in the end contributes to the owners’ wealth because ROE measures the performance from the perspective of the equity-holders. There are a few reasons for this significant positive contribution of Islamic debt issuance. First, Islamic debt is claimed and advertised as a secure investment due to its structure. Second, Islamic debt is given a special privilege such as stamp duty and exempted tax for both issuers and investors. Third, Islamic debt is guaranteed by the special purpose vehicle (SPV); in case of default the Islamic debt holders may recourse the assets underlying the Islamic debt. Fourth, though there were a few cases of default in Middle East, those cases have no impact on the investors’ perspective, as some investors investing in Islamic debt only do so only to comply with the religious matter. Fifth, the majority of investors are non-Muslim, with an increasing presence of foreign investors (PricewaterhouseCoopers Malaysia, 2008). Sixth, the Islamic debt issuance contributes to an increase in the issuer’s stock returns (Nagano, nd.).

Moreover, from the issuers’ perspective, there are benefits issuing Islamic securities, in particular, Islamic debt. The key benefits are tax incentives, value proposition and regulatory process. First, for tax incentives, the issuers are exempted from stamp duty, tax deductible of issuance cost, and the special purpose vehicle (SPV) is exempted from tax, and tax neutrality. Second, for value proposition, there is a wider investors’ base, Islamic debt is attractively priced due to the strong demand, there is strong structuring expertise in the Islamic finance industry, and Islamic debt enhances the issuers’ profile. Third, in terms of regulatory process, the process facilitates the issuance process, the rating of Islamic debt is automatically approved for AAA-rated for Islamic debt issued in domestic (Malaysian) currency and A-rated Islamic debt issued in foreign currency, any amendment to terms of approved Islamic debt need only to inform the Securities Commission, and exchangeable Islamic debt is exempted from rating. From the point of view of shareholders, the usage of debt increases their wealth, and because of this, markets believe that Islamic debt positively contributes to the firm performance. Moreover, Islamic debt issuance contributes to an increase in the issuers’ total factor productivity (Nagano, n.d.).

Furthermore, the positive result may be due to the stabilised nature of the Malaysian financial system which has evolved in line with the changing structure of the economy. The changes in the economic structure and financial system

in turn have had an important influence in shaping the increasing complexity and sophisticated nature of its capital market along with the implementation of regulations, and these changes support firms to operate more effectively and efficiently, increasing the confidence of markets. Moreover, a more diversified financial system, in particular, the rapid growth of the Malaysian Islamic Capital Market and the Malaysian debt market, has increased the alternative sources of financing available to corporations.

These key benefits supports the theory that the choice of capital structure may help mitigate agency costs (Jensen & Meckling, 1976). According to the agency costs theory, high leverage or a low equity/asset ratio reduces the agency costs of outside equity and increases firm value by constraining or encouraging managers to act more in the interests of shareholders. Moreover, corporate debt has a disciplining effect on management, since it serves to reduce the free cash flow and therefore minimises management’s discretionary spending.

Overall, the finding for ROA and ROE are similar to Tobin’s Q which also supports the trade-off theory (Modigliani & Miller, 1963; DeAngelo & Masulis, 1980; Jensen & Meckling, 1976; Haris & Raviv, 1990; Frank & Goyal, 2003), and this theory apparently can also be applied to Islamic debt.

The frequency of Islamic debt issuance and Tobin’s Q, ROA and ROEThe coefficient for first issuance of Islamic debt is a positive and significant at 1% level of significance (but for Tobins’ Q which significant at 10% level of significance), suggesting that the first issuance of Islamic debt affects higher firm performance. This also indicates that (1) the markets react positively to the issuance of Islamic debt when it is first introduced to the market; (2) the firm effectively utilises its assets to generate profits for the shareholders, and additional debt, in particular Islamic debt, pushes the management to perform better. There are several factors that might contribute to this positive finding. First, the managers of the firms are compelled to put more effort into generating more profits. Because some of Islamic debt is in the form of partnership (profit and loss sharing agreement), Islamic debt tends to place greater pressure on the managers to manage the firms effectively. Second, there is a broad-based coordination of government policies which resulted in a comprehensive public policy that supports growth and innovation in the Islamic financial market, in particular, Islamic debt. Third, the importance of government intervention, such as tax incentives and required ratings improves issuers’ and investors’ confidence. Fourth, the rapid growth of Islamic finance signifies that Islamic debt has moved from the pioneering stage to being an established financing instrument that serves as a commercially viable and effective tool for mobilising investment assets to finance productive economic activities. Fifth, in the beginning of the Islamic finance initiation, Islamic debt offered those competitiveness features, particularly cost effectiveness, secureness and efficiency. As such, the market had high expectations of this new instrument, the upshot was that Islamic debt brought more pressure on managers to manage their firms effectively in order to meet market expectations. Sixth, apart from being well-regulated by various standards and guidelines, Malaysia is also the only



country that makes it compulsory for all tradable corporate debt securities to be rated to enhance investors’ confidence and to assist in the investment decision-making process. Another distinguishing factor for the Malaysian Islamic debt market is the establishment of a centralised, national level Shariah supervisory board, which ensures that every Islamic debt issued in Malaysia, is in full compliance with the Shariah. All these factors provide sufficient protection to investors in the Islamic debt and conventional debt markets.

However, the coefficient for the second issuance of Islamic debt is a negative and significant at 5% level of significance, suggesting that the issuance of Islamic debt for a second time lowers firm performance. This negative finding is similar to the study by Godlewski et al. (2010), which suggest that Islamic debt expansion has a detrimental effect on firm value. This negative finding may indicate that (1) either the management of the firms have loosened their control because of overconfidence from the first successful issuance of Islamic debt or that the management have expropriated the firms’ previous profits; (2) the markets have experienced, observed and learnt from the first Islamic debt issuance, leading underconfidence in the markets over this second issuance, which in turn may affect the share price of those firms issuing Islamic debt; (3) low credit rating of firms issuing Islamic debt as this is associated with high risk. The gap between the first and the second Islamic debt issuance ranges between two to six years. Presumably, in that time period, investors observed the firm’s performance, their Islamic debt rating, the market conditions such as the frequency of default cases of Islamic debt. In Malaysia, cases of Islamic debt default were few and it is something that raises concern on the investors’ protection because a default occurs due to the breach of any binding obligations under the original terms of the agreement between the issuer and the Sukuk holders. Thus this factor may contribute to the negative result.

Furthermore, debt is also a source of information which indicates the firm’s current condition that investors can use to monitor and evaluate major operating decisions of the firm in two ways. Firstly, the mere ability of the firm to make its contractual payments to debt-holders provides information. Secondly, in the event that the organisation fails to make the payments, their ways to resolve the matter either through informal negotiation or formal bankruptcy proceedings will disseminate considerable information to the investors (Harris & Raviv, 1990). In sum, the negative relationship of the second issuance of Islamic debt and its firm’s performance is probably either a result of the previous firm performance in meeting their obligation of payment or a result of inefficient utilisation of their firms’s assets.

Fortunately, the coefficient for more than two issuance of Islamic debt is a positive and significant at 1% level of significance, suggesting the issuance of Islamic debt for more than two improves a firm’s financial performance. This may indicate that after having a few experiences in issuing Islamic debt, the issuance of Islamic debt later on impacts positively on firm performance. This may be caused by the fact that (1) the debt-holders of Islamic debt closely monitor the management of the firm to ensure that the firm can generate profits and distribute a periodic

stream of cash flow over time. Thus, Islamic debt also reduces the agency problem within the company and hence increases firm value. (2) That as the industry grows, it is more apparent that there is more demand by non-Muslim investors and issuers to play a role in the industry. Here in Malaysia, for instance, there is just as strong a demand for Shariah compliant products among non-Muslims as there is among Muslims (PricewaterhouseCoopers Malaysia, 2008).From the view point of markets, this may indicate that the markets have learnt through several issuances of Islamic debt and therefore they have greater confidence in subsequent issuances compared to the second issuance of Islamic debt. However, investors are irrational according to the behavioural finance theory. Their decision may be influenced by the magnitude issue, their bias selection and the lucky event issue.

The proportion of Islamic debt issued and Tobin’s QThe coefficients for the proportion of Islamic debt below the average and at the average are a positive and significant varies at 10% and 1% level of significance. These positive and significant results may be caused by internal and external factors. In terms of internal factors, the proportion of Islamic debt issued at a certain level stimulates the management to work effectively. For external factors, there are two views; first from the markets’ view, second from the view of government support. From the markets’ view, the proportion of a certain level of Islamic debt may be considered as tax exempted stimulation as the profits derived from Islamic debt are exempted from the taxes. Furthermore, the markets have confidence over the assets/projects underlying the Islamic debt contract which may bring profits in future; therefore, this market confidence affects their stock price. With regards to government support, the Malaysian government has provided an interesting model to promote the co-existence of an ethical and societal-based finance through issuing a few regulations that appeal to Muslim and non-Muslim investors; hence these regulations issued can assure the credibility of this instrument. Furthermore, the regulating body has taken vital steps to develop a facilitative regulatory framework, to create a large pool of players, to introduce a comprehensive range of innovative and competitive Islamic financial product and services, and to ensure sufficient depth to facilitate liquidity management, hence creating market confidence.

Though debt reduces the agency costs of free cash flow by reducing the cash flow available for spending at the discretion of managers (Jensen, 1986), an increased leverage also has costs; as leverage increases the risk of default also increases. This theory supports the result for Islamic debt above the average which is a negative and significant at 1% level of significance. This finding suggests that the greater the proportion of Islamic debt issued, the lower the firm performance. This result is similar to the empirical result for non-Islamic debt, in that the proportion of debt at a certain level may hamper firm performance as an additional incurrence of debt gives no guarantee that firm performance will be higher. This is mainly because as the leverage increases, so does the risk of default, which provides a greater incentive for lenders to monitor the firm. Though it is claimed that Islamic debt is more secure than

Fauzi et al.


the conventional debt, this result finds no support for that claim. On the contrary, this finding supports the notion that as the leverage increases, the probability of default also increases, and Islamic debt is no exception to this rule. Overall, the result for Islamic debt proportion Tobins’ Q, ROA and ROE has similarity.

The type of Islamic debt and Tobin’s QThe coefficients for the debt-type and equity-type are a positive and significant at 1% and 10% level of significance for Tobins’ Q and ROE. While all types of Islamic debt are a positive and significant at 1% level of significance for ROA. Though the finding for ROA is slightly different than for Tobin’s Q, this result does not impair on the Tobin’s Q result, as it is common for different methods of calculation to give different results. The finding suggests that debt-types and equity-types affect higher firm performance. The result supports the notion that certain types of debts have a different impact on shareholders’ wealth (Mikkelson & Partch, 1986); hence, this finding can also be applied to Islamic debt.

Furthermore, the finding can be explained by the different Islamic debt structure. This is important since the structure determines the obligation of the originator/issuers. There is typically a requirement that on maturity of the Islamic debt or upon an event of default, the originator has a purchase obligation to repurchase the assets which enables the Special Purpose Vehicle (SPV) to redeem the outstanding certificates and repay the Sukuk holders. In this regard, the rights of Sukuk holders in the event of default will vary depending on whether the Sukuk structure is an asset-based or an asset-backed structure. The positive result for debt-based and equity-based Sukuk may be caused by their structure. The assumptions that may be raised is that debt-based and equity-based are in the structure of asset-backed Sukuk, and asset-based is in the structure of asset-based Sukuk. Thus, the rights of the Sukuk-holders depend on the structure of Islamic debt. For example, in the case of Sukuk ijarah, if the Sukuk is asset-backed, this allows the holders to liquidate the underlying asset in the event of default to recover most of their investments. On the other hand, if the Sukuk is asset-based, this only represents beneficial ownership on the underlying asset and it restricts the holders’ rights in the event of a default.

The coefficient for firm size is a negative and significant for all four regression equations, suggesting that bigger firms which having Islamic debt in their debt structure have a lower firm performance. The negative result may be due to the fact that bigger firms are already well-stabilised in terms of cash flows and profits because of their well-stabilised capital structure; hence changing its capital structure with a new unproven instrument may endanger the firm’s credibility and ability to maintain their stable cash flows and profits. This notion leads to the markets’ perspective on the firms’ capability in the future; the markets may have lower confidence and in turn, this affects the stock price of the firms.

Apart from year 2004, all the years reveal a significant result. All the years (2003, 2004, 2005, 2007, 2008, and 2009) have a positive coefficient except year 2001 and year 2006. Malaysia, with its economic strength, supportive government policies, educated workforce, developed

infrastructure, vibrant business environment and quality of life, has always been an attractive market for foreign investors. Therefore, the coefficients for year 2003, 2004 and 2005 are supported.

Despite the challenging global economy, Malaysia has continued to pursue liberalisation, enhancing the entrepreneurial and investment environments. The economy scores above the world average in many of the ten economic freedoms (World Bank, 2011). The trade regime is relatively open despite lingering non-tariff barriers. However, corruption and a judicial system that remains vulnerable to political influence pose significant challenges to economic freedom. 2001 and 2006 were two years which yielded a negative and significant impact. The first, 2001, may be due to the global economic slowdown overall. Significantly, though, a general election was held in 2003 and again in 2008, revealing a pattern in which there is a two year gap between this political event and a year yielding a negative and significant impact. This may indicate that before the general election, the political situation in Malaysia heats up, which affects the market players.

The Malaysian economy has been surprisingly resilient in spite of the global slowdown in 2007.Malaysia has only felt a minor impact from the slowing US economy, but emerging challenges in the form of soaring food prices and the persistent rise in global oil prices are weighing down heavily on economic prospects. Furthermore, to avoid the fiscal deficit, the government announced a revamp in oil subsidies, pushing up the price of petrol diesel, which has adverse implications for inflation and economic growth. However, in 2008 and 2009, the business confidence index increased as it indicates by the rise of sales and production, higher export sales, higher capacity utilisation, higher domestic demands and higher capital investment. The gross domestic product growth was sustained at a certain targeted level. This growth was driven by high commodity prices, strong private consumption and steady investment, and supported by fiscal spending. The business condition index would be a better indicator of current economic activity as it relies on firm-level information. Therefore, the positive and significant coefficients for year 2007, 2008 and 2009 are supported.

6. ConclusionIn sum, the findings for all three categories of explanatory variables, along with their control variables for all metrics (Tobins Q, ROA and ROE), are only slightly different in their coefficient value. Almost all the coefficient signs and significance values reveal the same direction and a similar significance value. The coefficients for Islamic debt is higher than the coefficient for non-Islamic debt and, overall, the findings for Tobin’s Q, ROA, ROE and EVA support the trade-off theory (Modigliani & Miller, 1963; DeAngelo & Masulis, 1980; Jensen & Meckling, 1976; Haris & Raviv, 1990; Frank & Goyal, 2003) and this theory apparently can also be applied to Islamic debt.

ReferencesAbd. Sukor, M.E., Muhamad, R, Gunawa, A.Y. (2008)

Malaysian Sukuk: Issues in accounting standard. Shariah Journal, 16(1), 63–74.



Abor, J. (2005) The effect of capital structure on profitability: An empirical analysis of listed firms in Ghana. Journal of Risk Finance, 6, 438–447.

Abor, J. (2007) Debt policy and performance of SMEs: Evidence from Ghanaian and South African firms. The Journal of Risk Finance, 8(4), 364–379.

Al Amine, M.A.B. (2008) Sukuk market: Innovations and challenges. Islamic Economic Studies, 15(2), 1–22.

Al Amine, M.A.B. (n.d.). The Islamic bonds market: Possibilities and challenges. International Journal of Islamic Financial Services, 3(1).

Al-Amine, M.A.B.M. (2001) Istisna’ and Its application in Islamic banking. Arab Law Quarterly, 16(1), 22–48.

Ashhari, M.Z., Chun, L.S., Nassir, A. Md. (2009) Conventional vs Islamic bond announcements: The effects on shareholders’ wealth. International Journal of Business and Management, 4(6).

Ayub, M. (2007) Understanding Islamic Finance. England: John Wiley & Sons, Ltd.

Baltagi, B.H. (2005) Econometric Analysis of Panel Data (3rd

ed.). West Sussex, England: John Wiley & Son Ltd.

Bhabra, H., Liu, T., Tirtiroglu, D. (2008) Capital Structure Choice in a Nascent Market: Evidence from Listed Firms in China. Financial Management, 37(2), 341.

Blundell, R., Bond, S. (1998) Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87, 115–143.

Booth, L., Aivazian, V., Demirguc-Kunt, A., Maksimovic, V. (2001) Capital Structures in Developing Countries. Journal of Finance, 56(1), 87–130.

Cakir, S., Raei, F. (2007) Sukuk vs. Eurobonds: Is there a difference in value-at-risk, IMF Working Paper (Vol. WP/07/237): International Monetary Fund.

Cameron, A.C., Trivedi, P.K. (2010) Microeconometrics using stata (Rev. Ed.).College Station, Texas: Stata Press.

Champion, D. (1999) Finance: The joy of leverage. Harvard Business Review, 77(4), 19–22.

Coleman, A.K. (2007) The impact of capital structure on the performance of microfinance institutions. The Journal of Risk Finance, 8(1), 56–71.

DeAngelo, H. & Masulis, R.W. (1980) Optimal capital structure under corporate and personal taxation. Journal of Financial Economics, 8(1), 3–29.

Ebaid, I.E.S. (2009) The impact of capital-structure choice on firm performance: empirical evidence from Egypt. The Journal of Risk Finance, 10(5), 477–487.

Eriotis, N.P., Frangouli, Z., Neokosmides, Z.V. (2002) Profit margin and capital structure: An empirical relationship. The Journal of Applied Business Research, 18(2), 85–88.

Fleming, G., Heaney, R., McCosker, R. (2005) Agency costs and ownership structure in Australia. Pacific-Basin Finance Journal, 13, 29–52.

Frank, M.Z., Goyal, V.K. (2003) Testing the pecking order theory of capital structure.Journal of Financial Economics, 67(2), 217–248.

Ghosh, A., Cai, F. (1999) Capital structure: New evidence of optimality and pecking order theory. American Business Review, 17(1): 32–38.

Hadlock, C.J., James, C.M. (2002) Do banks provide financial slack? Journal of Finance, 57, 1383–1420.

Haneef, R. (2009) From asset-backed to asset-light structures: The intricate history of Sukuk. ISRA International Journal of Islamic Finance, 1(1).

Harris, M., Raviv, A. (1990) Capital structure and the informational role of debt. The Journal of Finance, 45(2), 321–349.

Hatfield, G.B., Cheng, L.T.W., Davidson, W.N. (1994) The determination of optimal capital structure: The effect of firm and industry debt ratios on market value. Journal of Financial and Strategic Decisions, 7(3):1–14.

Jensen, M.C. (1986) Agency costs of free cash flow, corporate finance and takeovers. American Economic Review, 76(2), 323–329.

Jensen, M.C., Meckling, W.H. (1976) Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3(4), 305–360.

Jermias, J. (2008) The relative influence of competitive intensity and business strategy on the relationship between financial leverage and performance. The British Accounting Review, 40(1), 71–88.

Juan, S. (2008) Prospects and challenges for developing corporate Sukuk and bond markets: International Monetary Fund. International Journal of Islamic and Middle Eastern Finance and Management, 1(1), 20–30.

Kamali, M.H. (2007) A shari‘ah analysis of issues in Islamic leasing. Islamic Economic, 20(1), 3–22.

Kraciska, O., Nowak, S. (2012) What’s in it for me? A primer differences between Islamic and conventional finance in Malaysia. IMF Working Paper, 151.

Krishnan, V.S., Moyer, R.C. (1997) Performance, capital structure and home country: An analysis of Asian corporations. Global Finance Journal, 8(1), 129–143.

Mikkelson, W.H., Partch, M.M. (1986) Valuation effects of security offerings and the issuance process. The Journal of Financial Economics, 15, 31–60.

Mirakhor, A. (1996) Cost of capital and investment in a non-interest economy. Islamic Economic Studies, 4(1), 35–46.

Modigliani, F., Miller, M. (1963) The cost of capital, corporation finance and the theory of investment. American Economic Review, 48, p.261–297.

Modigliani, F., Miller, M.H. (1958) The cost of capital, corporation finance and the theory of investment. American Economic Review, 48(3), 261.

Mohd Yatim, M.N., Shah, M. (2009) Sukuk (Islamic bond): A crucial financial instrument for securitisation of debt for the debt-holders in shari’ah-compliant capital market. International Journal of Business and Management.

Mokhtar, S. (2009) A synthesis of Shariah issues and market challenges in the application of wa’d in equity-

Fauzi et al.


based Sukuk. ISRA International Journal of Islamic Finance, 1(139–145).

Myers, S., C., Majluf, N., S. (1984) Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13, 187–221.

Myers, S.C. (1984) The capital structure puzzle. Journal of Finance, 39, 575–592.

Naceur, S.B., Goaied, M. (2002) The relationship between dividend policy, financial structure, profitability and firm value. Applied Financial Economics, 12; 843–849

Nagano, M. (n.d.). Islamic finance and the theory of capital structure. Discussion Papers in Economics, Society of Economics Nagoya City University, 51.

Ni, J., Yu, M. (2008) Testing the Pecking-Order Theory: Evidence from Chinese Listed Companies. The Chinese Economy, 41(1), 97–113.

Phillips, P.A., Sipahioglub, M.A. (2004) Performance implications of capital structure: evidence from quoted UK organisations with hotel interests. Service Industries Journal, 24(5), 31–51.

PricewaterhouseCoopers Malaysia.(2008) Malaysia, Asia’s Islamic Finance hub (Kuala Lumpur). Retrieved from: http://www.pwc.com/my/en/publications/islamic-finance-hub.jhtml

Raghunathan, T.E. (2004) What do we do with missing data? Some options for analysis of incomplete data. Annual Reviews Public Health, 25, 99–117.

Ramaswamy, K. (2001) Organisational ownership, competitive intensity and firm performance: An

empirical study of the Indian Manufacturing sectors. Strategic Management Journal, 22, 989–998.

Somolo, E. (2009) Islamic Pricing Benchmark. ISRA Inter-national Journal of Islamic Finance, 1(1).

Talberg, M., Winge, C., Frydenberg, S. & Westgaard, S. (2008) Capital structure across industries. International Journal of the Economics of Business, 15(12), 181–200.

Tariq, A.A., Dar, H. (2007) Risks of Sukuk structures: implications for resource mobilization. Thunderbird International Business Review, 49(2), 203–233.

Usmani, M.T. (1999) The Concept of Musharakah and Its Application as an Islamic Method of Financing. Arab Law Quarterly, 14(3), 203–220.

Vishwanath, S.R., Azmi, S. (2009) An overview of Islamic Sukuk bonds. The Journal of Structured Finance, 14(4), 58–67.

Wilson, R. (2008) Innovation in the structuring of Islamic Sukuk securities. Humanomics, 24(3), 170–181.

Wilson, R. (n.d.). Islamic Bonds: Your Guide to Issuing, Structuring and investing in Sukuk.

Wiwattanakantang, Y. (1999) An empirical study on the determinants of the capital structure of Thai firms. Pacific-Basin Finance Journal, 7(3–4), 371–403.

World Bank. (2011).

Yean, T.W. (n.d.). Sukuk: Issues and the way forward.

Zeitun, R., Tian, G.G. (2007) Capital structure and corporate performance: Evidence from Jordan. Australasian Accounting Business and Finance Journal, 1(4): 40–61.


Cite this chapter as: Zulkhibri M (2015). Islamic financing and bank characteristics in a dual banking system: Evidence from Malaysia. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Islamic financing and bank characteristics in a dual banking system: Evidence from MalaysiaMuhamed ZulkhibriEconomic Research and Policy Department, Islamic Development Bank, P.O Box 5925, Jeddah 21432, Saudi Arabia, E-mail: [email protected]

Abstract - An understanding of financing behaviour explains the performance of Islamic banks as an alternative to the conventional finance model as suppliers of capital for businesses and entrepreneurs. In the Islamic banking system, banks are suppliers of capital and not lenders unlike that stipulated in a traditional banking system. To date, Islamic banks have become the major source of capital in Malaysia and lending behaviour is an important policy variable. In this context, the paper examines the relationship between bank financing, bank financing rate and bank-specific characteristics in a dual banking system. The evidence suggests that the bank-specific characteristics are important for Islamic banking financing behaviour. The Islamic banks’ financing behaviour is consistent with behaviour of conventional banks in that the bank lending operates through banks depending on the size, and level of liquidity and capital. The findings also suggest that that there is no significant difference between Islamic bank financing and conventional bank lending behaviour with respect to interest rates.

Keywords: Islamic banks, base financing rate, bank financing, panel regression analysis

1. IntroductionAn Islamic bank is a deposit-taking institution, which includes all functions currently known as banking activities. The bank mobilizes funds on the basis of mudaraba (profit-sharing) or wakalah (as an agent charging a fixed fee for managing funds), which form part of its liabilities, while financing on a profit-and-loss sharing (PLS) basis or through the purchase of goods (on cash) and sale (on credit) or other trading, leasing and manufacturing activities, form part of the assets. Apart from demand deposits, which are treated as interest-free loans from the clients to the bank and are guaranteed to be repaid in full, it plays the role of an investment manager for the owners of deposits, akin to a universal bank.

In contrast, conventional banks are understood in a generic sense as financial intermediaries. Their main task is to provide indirect finance, in contrast to direct finance through financial markets. Banks are understood to channel surplus funds from the household sector into the corporate sector facing a deficit, as they invest more than internal or direct finance. Banks play a vital role in the economy enabling more productive investment than would be possible merely on the basis of profits and financial market funds. Hence, an asymmetric information and

understanding of the functioning of banks must lead to the conclusion that Islamic banks are not viable.

The distinguishing features of Islamic banks are the prohibition of charging or paying interest, the impermissibility of demanding collateral and, to a small extent, compulsory charitable spending (Khan and Mirakhor 1992). Profit has to be generated merely by primary and secondary modes of Islamic finance (Chapra 2000). Primary modes include profit-sharing arrangements such as mudaraba (partnership) and musharakah (equity participation), while secondary modes are essentially mark-up pricing or leasing arrangements.

Many scholars also argue that Islamic banking resembles conventional banking schemes. This includes the claim that the majority of Islamic lending has a debt-like character (Aggarwal and Youssef 1996). El-Gamal (2005) has concluded that Islamic finance is primarily a form of rent-seeking legal arbitrage and simply seeks to replicate the operations of conventional financial instruments. However, some researchers have argued that Islamic financial institutions have huge potential over the conventional banking model as an alternative finance to absorb macro-financial shocks and promote economic growth (Dridi and Hasan 2010; Mills and Presley 1999).

Zulkhibri


In term of deposit, Islamic banks use mainly the risk-sharing PLS instruments, while in financing, most Islamic banks rely on debt-like instruments (mark-up financing and a guaranteed profit margin) that are based on deferred obligation contracts. Moreover, conventional interest rates (the London Interbank Offered Rate (LIBOR) or a domestic equivalent) will always be a benchmark for Islamic banks’ mark-up. As a result, in the case of such debt-like instruments, the pricing of Islamic financing is not a function of real economic activity but is based on a pre-determined interest rate plus a credit risk premium.

The objective of the paper is to investigate the determinants of Islamic financing while taking into consideration bank-specific characteristics. Understanding this behaviour indicates how efficiently Islamic banks perform their roles as suppliers of capital for businesses and entrepreneurs. However, little is known about the determinants of bank financing which operate alongside conventional banks in the dual banking system. Moreover, due to the fact that the rate of return on retail PLS accounts closely follows interest rates offered by conventional banks in Malaysia (Chong and Liu 2009; Cervik and Charap 2011), the paper employs a panel-pooled regression methodology by investigating the cross-sectional differences in the way that Islamic banks respond to base financing rates across bank-specific characteristics.

2. Overview of Malaysian Islamic financial industryMalaysia’s Islamic finance industry has been in existence for over 30 years. The enactment of the Islamic Banking Act 1983 enabled the country’s first Islamic Bank to be established. Malaysia’s overall strategy in the development of Islamic banking can be summarized under four pillars:

1. A full-fledged Islamic banking system operating on a parallel basis with a full-fledged conventional system (dual banking system).

2. A step-by-step approach, in the context of an overall long term strategy.

3. A comprehensive set of Islamic banking legislation and a common Shariah Supervising Council for all Islamic banking institutions.

4. A practical and open-minded approach in developing Islamic financial interests.

An important feature in the implementation of Islamic banking in Malaysia and creating a viable Islamic banking system is that three basic elements have been adopted:

1. A large number of instruments and range of different types of financial instruments must be available to meet the different needs of different investors and borrowers.

2. A large number of institutions with an adequate number of different types of institutions participating in the Islamic banking system to provide depth to the Islamic banking system.

3. An Islamic interbank market to support an efficient and effective system linking the system to the institutions and the instruments.

As it can be observed in Figure 1, the share of Islamic assets in the overall banking system is growing significantly, from around 7% in 2006 to 20% in 2012. As at the end of 2012, the country’s Islamic banking system had accumulated a total of RM119 billion in assets, or about 20% of the total assets of the banking sector, which is RM0.6 trillion. To date, Malaysia has 16 Islamic Banks, which comprises nine local Islamic Banks and seven foreign Islamic Banks. Figure 2 shows the composition of the Islamic financing modes. It shows that

87% 86% 82% 80% 79% 78% 77%

7% 8% 14% 16% 17% 18% 20%

6% 6% 4% 4% 4% 4% 3%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2006 2007 2008 2009 2010 2011 2012

Commercial Bank Islamic Bank Investment Bank

Source: Bank Negara Malaysia (2013)

Figure 1. Total assets: Islamic and conventional banks.


Islamic financing and bank characteristics in a dual banking system: Evidence from Malaysia

financing and conventional lending rates on loans in Malaysia. Between 2009 and 2012, the correlation of base lending rate of conventional banks and the Islamic base financing rate was about 76 percent. Accordingly, despite the fact that conventional and Shariah-compliant Islamic banks operate in different banking environments, it is surprising that the Islamic base financing rate closely tracks interest rates offered by conventional banks in Malaysia.

the Bai Bithaman Ajil and Ijarah Thumma Al Bai dominate the composition with 32% and 23%, respectively, over the period 2006–2012. This dominant trend of using mark-up and debt-like instruments in Islamic financing practices support some of the arguments that Islamic banking is akin to conventional banking in practical terms.

As shown in Figure 3, the data reveal a high degree of correlation between the base financing rate on retail


41% 37% 33% 32% 34% 32% 32%

29%30%

30% 29% 27%26% 23%

21% 19% 16% 15% 18% 20% 20%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2006 2007 2008 2009 2010 2011 2012

Bai Bithaman Ajil Ijarah Ijarah Thumma Al-Bai Murabahah

Musyarakah Mudharabah Istisna' Others

Figure 2. Composition of Islamic financing modes.

Figure 3. Islamic rate of return and conventional average lending rate.


Correlation=0.76

Conventional

Islamic

4.50

2009 2010 2011 2012

5.00

5.50

6.00

6.50

7.00

Zulkhibri


3. Literature reviewIslamic banking is different from conventional banking from a theoretical perspective because interest (riba) is prohibited in Islam (rate of return on deposits cannot be fixed by the bank and interest cannot be charged on loans). The prohibition of interest is combined with the common belief that banks channel funds towards productive investment, which makes Islamic banking and Western economic theory inconsistent with each other. A unique feature of Islamic banking is the PLS paradigm, which is largely based on the mudaraba (profit-sharing) and musharakah (equity participation) concepts of Islamic contracting.

The concepts of Islamic finance in using the rate of returns as a replacement for interest can be divided into two strands of argument. The idealist literature attempts to look at the key concepts of Islamic finance such as PLS, money, interest and profit from an ideal perspective. A pre-determined return to the lender, dependent on the borrowing period and independent of the borrower’s uncertainty, is not permissible under Islamic banking. This means that the ideal and most ‘Islamic’ form of each concept should be accepted as valid. Much of the literature on Islamic banking and finance in the 1960s and the theoretical studies on Islamic banking fall under this category.

Another line of argument based on maslaha-oriented literature would be at the extreme end of the continuum. According to this view, riba should not be interpreted in a simplistic fashion as modem bank interest. Any interest-based bank could theoretically be an Islamic bank provided that the Islamic ideals of justice, equity, fairness, non-exploitation were its guiding principles; humane terms of providing finance to those ‘needing’ them were practised; and, it provided one way of helping the economically disadvantaged classes of society to raise their standard of living. Nonetheless, it can be seen from this that there has been a gradual shift from the idealist position to a more pragmatic, mark-up based and less risky version.

In the conventional literature, the interest rate has long been recognized not only by classical and neo-classical economists, but also by contemporary economists as one of the factors that determine the level of savings in the economy, and that the interest rate has a positive relationship with savings. However, Haron (2001) found similar positive relationship behaviour for the profit rate declared by Islamic banks. In other words, Islamic bank customers are guided by the profit maximization theory since there is no pre-determined rate of return involved in the Islamic banking system. Since depositors at Islamic banks possess similar attitudes to those at the conventional banks, the interest rate will continue to have an influence on the operations of Islamic banks.

On a similar note, while funding activities are carried out mainly through the participatory PLS model, it is well-established in the literature that Islamic banks follow their conventional counterparts in creating assets through non-PLS, debt-like instruments with a pre-determined, fixed rate of return; in line with the findings of Beck et al. (2010), there are “few significant differences in business orientation, efficiency, asset quality or stability” between conventional and retail Islamic banks. As a result, given the implicit link to interest rates on the asset side of the balance sheet, PLS rates of return follow conventional bank deposit rates.

More recent literature on Islamic finance tries to establish the difference between Islamic rates of return and conventional banks interest rates based on empirical assessments. Cervik and Charap (2011) compare the empirical behaviour of conventional bank deposit rates and the rate of returns on retail Islamic PLS investment accounts in Malaysia and Turkey. The findings show that conventional bank deposit rates and PLS rate of returns exhibit co-integration in the long-run, and that conventional bank deposit rates cause returns on PLS accounts. Moreover, the time-varying volatility of conventional bank deposit rates and PLS returns is correlated and is statistically significant.

Such correlations have been observed in other studies. In the case of Malaysia, Chong and Liu (2009), for example, find that retail Islamic deposit rates mimic the behaviour of conventional interest rates. The study shows that only a small portion of Islamic bank financing is strictly PLS-based, and that Islamic deposits are not interest-free, but are very much pegged to conventional deposits. The findings also suggest that the Islamic resurgence worldwide drives the rapid growth in Islamic banking rather than the advantages of the PLS paradigm, implying that regulations similar to those of conventional banks should be applied for the Islamic bank.

Similarly, Kasri and Kassim (2009) examined the relationship between investment deposits and rate of return, including interest rate for Islamic banks in Indonesia over the period 2000 to 2007. Using a vector autoregressive model (VAR) model, the study reveals that the mudaraba investment deposit in the Islamic banks are co-integrated with return of the Islamic deposit, interest rate of the conventional banks’ deposit, number of Islamic banks’ branches, and national income in the long-run. The finding also suggests that rate of return and interest rate move in tandem, indicating that Islamic banks in Indonesia are exposed to benchmark risk and rate of return risk.

In practice, the main explanation of the similarity between Islamic bank profit rate and conventional banks can be attributed to the differences in perceptions of riskiness (theoretically and practically) at the institutional and systemic level, particularly on the asset side. In addition, Islamic banks lose on the grounds of liquidity, assets and liabilities concentrations and operational efficiency, whereas they tend to win in the field of profitability. Nevertheless, Islamic banking could provide a further guarantee, albeit still marginal, against systemic risks in certain emerging financial markets.

4. Data and estimation methodologyThe study employs a panel of annual bank level data of all Islamic banks operating in the Malaysia covering the period 2006–2012. The financial statements of Islamic banks operating in Malaysian Islamic banking sectors are collected from the Bankscope database of Bureau van Dijk’s company. The macroeconomic variables: consumer price index, real gross domestic product, Islamic base financing rate, and monetary policy rate are taken from various issues of Quarterly Statistical Bulletin published by Central Bank of Malaysia.

Table 1 reports the basic descriptive statistics for the sample:



Using a panel of information for 17 individual banks, we initially estimate the benchmark model for Islamic bank financing to allow for asymmetric response to bank characteristics and monetary policy. This has usually been done by introducing interaction terms between Islamic base financing rate and bank discriminatory variables. Beside these variables, we control for economic activities and consumer prices, which allow us to control for demand-side effects on Islamic bank financing. By combining time series of cross-sectional observation, panel data give more informative data, more variability, less co-linearity among the variables, more degrees of freedom and more efficiency (Gujarati and Sangeetha 2007). The panel-data estimation method of both pooled-regression and fixed-effect model is preferred. Fixed-effects specification is mainly used to account for time-invariant unobservable heterogeneity that is potentially correlated with the dependent variable. To test for estimation robustness of the models, we employ random-effect estimations and use all diagnostic tests to validate the models. Our baseline model specification is as follows:

Δ ΔFIN BFR SIZE LIQUIDITY

CAPI

it i j t jj

l

it it= + + +

+

-=

- -∑m β γ ω

ϕ

10 1 0 1

0 TTAL BFR SIZE

BFR LIQUIDITY

it jj

l

t j it

jj

l

t j i

-=

- -

=-

+

+

∑

∑

11

1

1

γ

ω

Δ

Δ

*

* tt

j t j itj

l

j tj

l

j t

BFR CAPITAL

GDP PRICES

-

- -=

-=

-

+

+ +

∑

∑

1

11

11

ϕ

κ l

Δ

Δ Δ

*

111j

l

i i t=

∑ + +υ e , (1)

where ∆ is the first-difference operator, FIN is the Islamic banks financing, GDP is the logarithm of real GDP, PRICES is the logarithm of consumer price index, BFR is the Islamic financing rate, SIZE, LIQUIDITY and CAPITAL are the bank size, liquidity and capitalisation respectively. The subscript i denotes banks where i = 1, …, N; t denotes time where t = 2006–2012; ni denotes individual bank effects and ei,t denotes error-term.

The choice of bank-specific characteristics is based on the theoretical assumption that a certain type of bank is expected to be more responsive to financing shocks since it operates in a dual banking system, and these characteristics are widely used in the empirical literature. Following Gambarcota (2005), the three measures for bank characteristics of size (SIZE), liquidity (LIQUIDITY) and capitalisation (CAPITAL) are defined as follows:

Size:

∑= - =SIZE A

A

Nln

lnit it

iti

N

T

1

s

Liquidity:

LIQUIDITYLAA

LA A

NTit

it

it

it iti

N

Tt

T

= -

=

=

∑∑/

/1

1ps

Capitalisation:

s

CAPITALKA

K A

NTit

it

it

it iti

N

Tt

T

= -

=

=

∑∑/

/1

1

Bank size (SIZE) is measured by the logarithm of total assets (A). Relatively, banks with a smaller size may face higher constraints in raising external funds, thus forcing them to reduce their lending (Kasyhap and Stein 1995, 2000). Liquidity (LIQUIDITY) is measured by the ratio of liquid assets (cash and short-term funds) to total assets (LA). More liquid banks can draw down on their liquid assets to shield their financing portfolios and are less likely to cut back on financing in the face of rising cost or rate of return. Capitalisation (CAPITAL) is measured by the ratio of capital and reserve to total assets (K). Since raising bank capital is costly, the bank tends to adjust the lending behaviour to meet the required level of capital. In the face of a rising rate of return, a bank’s cost of financing rises while the remuneration of bank assets remains the same. Hence, the financing of highly-leveraged banks is expected to be more responsive to changes in the rate of return than the financing of well-capitalised banks (Kishan and Opiela 2006).

All three criteria are normalised with respect to their average (NT) across all the banks in the respective sample in order to get indicators that sum to zero over all observations. For the Eqation (1), the average of the interaction term (∆BFR*SIZE, ∆BFR*LIQUIDITY and ∆BFR*CAPITAL) is, therefore, zero and the parameters are directly interpretable as the overall Islamic rate of return effect on Islamic bank financing. To remove the upward trend in the case of size (reflecting the fact that size is measured in nominal terms), or the overall mean in the case of liquidity and capitalisation, the bank characteristic variables are defined as deviations from their cross-sectional means at each time period.

The assumption is that small, less liquid and poorly capitalised banks react more strongly to changes in base financing rate. This would correspond to a significant

Table 1. Descriptive statistics.

Variable# of Obs. Mean

Std. Dev. Min Max

Financing 69 14.15 1.28 7.41 16.57Assets 71 14.91 0.95 11.83 16.96Liquidity 70 0.08 0.04 -0.02 0.20Capital 66 0.09 0.04 -0.02 0.24GDP 119 20.32 0.08 20.17 20.42Base Financing Rate

119 4.60 0.05 4.51 4.66

Overnight Rate 119 2.92 0.46 2.00 3.50Prices 119 3.29 0.24 2.93 3.70

Note: Financing, Assets, GDP and Prices are in logarithmic forms; Liquidity is defined as a ratio of liquid assets (cash and short-term funds) to total assets; Capital is defined as ratio of capital and reserve to total assets. Source: Author’s own computation from Bankscope.

Zulkhibri


positive coefficient for the interaction terms ∆BFR*SIZE, ∆ BFR*LIQUIDITY, and ∆BFR*CAPITAL, and means that banks with these characteristics reduce their financing growth rate more strongly in response to a restrictive shock of base financing rate than do larger, more liquid and well-capitalised banks.

Since Islamic banks operate in the dual banking system, conventional interest rates may influence the Islamic bank financing behaviour. Equation (1) may represent the overall effect of Islamic bank financing without monetary policy. Huang (2003) argues that, under the conventional system, changes in interest rates have a larger effect on bank loans supplies because banks’ ability to insulate their financing supplies from changes in monetary policy will be restricted, in particular, during periods of tight monetary conditions. We try to test this hypothesis for Islamic financing behaviour by including monetary policy rate, where MP is monetary policy shock proxy by overnight policy rate in Equation (2) and estimate the following model:

Δ

Δ

FIN

BFR SIZE LIQUIDITY

CAPI

it

i j t jj

l

it it= + + +

+

-=

- -∑m β γ δ

ϕ

10 1 0 1

0 TTAL BFR SIZE

BFR LIQUIDITY

it jj

l

t j it

jj

l

t j i

-=

- -

=-

+

+

∑

∑

11

1

1

γ

ω

Δ

Δ

*

* tt j t j itj

l

t jj

l

j tj

l

BFR CAPITAL

MP GDP

- - -=

-=

-=

+

+ +

∑

∑

1 11

11

1

ϕ

ζ κ

Δ

Δ

*

∑∑ ∑+ + +-=

l υ ej tj

l

i i tPRICESΔ 11

,

(2)

5. Empirical resultsTable 2 reports the results for our benchmark model of Islamic bank financing, while Table 3 to Table 4 report the results from fixed-effect and random-effect. The direct impact of changes in the base financing rate on bank financing is negative and significant. The coefficients for base financing rate range from 1.78 to 5.47, which means that an increase of base financing rate by one percentage point leads to a decrease in the bank financing in the range of between 1.7% to 5.5%. The result of our benchmark models in line with the basic theoretical prediction is similar to the bank lending channel of conventional banks (Ehrmann et al. 2003). Since the Islamic rate of return implicitly tracks interest rates offered by conventional banks (Chong and Liu 2009), the results also explain that the reduction in Islamic bank deposits may not be completely substituted by other forms of financing in order to continue to meet financing demand, thus leading to a reduction in Islamic bank financing. The results for fixed-effect and random-effect provide similar observations, albeit with a lower impact of base financing rate on bank financing between -0.21 to -1.76. The estimated regression equations for all models explain the behaviour of financing in the range of 30% - 97%. All diagnostic tests confirm the good fit of the models.

The results from Table 2 to Table 4 also show the importance of bank-specific characteristics with respect to

the bank lending behaviour. The variable of SIZE is positive and highly significant for all models. In the fixed-effect and random-effect model, SIZE is positive and significant, ranging from 0.51 to 1.39. Larger banks might be more efficient due to scale economies, while the theoretical and empirical literature on the relationship between size and stability is ambiguous (Beck et al. 2013). This suggests that size is an important factor characterising the banks’ financing reaction with large banks being expected to minimise cost. This finding is also consistent with Fadzlan and Zulkhibri (2009), who suggest that larger financial institutions in Malaysia attain a higher level of technical efficiency in their operations and exhibit an inverted U-shape behaviour.

In the case of the liquidity characteristic, the results show that the coefficient of LIQUIDITY is positively associated and highly significant with bank financing, and is between 1.16 and 9.47. Only banks that have a larger share of liquid assets, or that are bigger, are able to shield their lending relationships. This evidence points to the fact that Islamic banks are able to protect their financing portfolios by drawing down on their liquid assets and are, therefore, less likely to cut on financing, whereas the latter have better access to external finance due to their size in order to retain their preferred liquidity ratio. This finding also implies that, in periods of rising base financing rate, a borrower from a less liquid bank, on average, tends to suffer from a sharp decline in financing more than does a customer of a more liquid bank. The result is in line with the findings by Brooks (2007) that liquidity is the main determinant explaining credit supply in Turkey.

Looking at the coefficient of capitalisation, CAPITAL, it appears that bank financing is positively associated with bank capitalisation, or the bank capital structure. The results suggest that market participants may perceive highly capitalised banks as being less risky (Kishan and Opiela 2000). Consequently, it should be more expensive for poorly-capitalised banks to finance externally. Such poorly-capitalised banks try to avoid the cost of falling below the regulatory minimum capital requirements or the increased risk of violating the capital requirement by holding capital buffers and asset buffers (short-term risk-weighted assets rather than customer financing) that can be liquidated if the bank runs into problems with the capital requirement. The more short-term risk-weighted assets (other than customer loans) the bank holds on its balance sheet (i.e., the higher the bank’s asset buffer), the lower the risk of violating the capital requirements will be. The short-term risk-weighted assets will soon be liquid, thereby reducing the capital requirement in the near future. Also, the higher the bank’s capital buffer, the lower the risk of violating the capital requirement will be.

The macroeconomic variables included in the bank financing models control for the demand-side effect, and only the real GDP growth variable is significant in the equation, where it has a positive coefficient. The response of credit to economic activity is consistent with expectation. The facts that the coefficient of real GDP is significant may imply that the economic activities are taken into account in financing decisions in an important way. On the other hand,



the price variable is negatively related to bank financing, but is insignificant. The rise in inflation may be associated with the variability of the inflation rate and will generate uncertainty about the future return on investments. This, in turn, discourages firms from undertaking investments which, consequently, reduces their financing demand. However, the price variable is insignificant for the results of all regressions.

Due to the potential interrelations between bank financing and conventional interest rate, all bank financing models are run with overnight policy rate (ON). The coefficients in all regressions are negatively related to bank financing and vary within a reasonable magnitude (0.05 to 0.52), but are broadly lower than the base financing rate. The results of these regressions suggest that the reaction of banks to changes in interest rates remains the same as the change base financing rate and is robust in terms of a different type of econometric

specifications. This finding broadly supports the findings by Chong and Liu (2009), Cervik and Charap (2011), and Beck et al. (2013) that there is no significant difference in bank financing behaviour with respect to interest rates. Furthermore, Kasri and Kassim (2009) confirm that the conventional interest rate is one of the determinants for saving deposits in Indonesia. This evidence explains why the bulk of Islamic bank financing is based on the mark-up principle and is very debt-like in nature (i.e. murabaha and ijarah), rather than using the principle of PLS. Despite their operations being different from those of conventional banks, Islamic banks seem to face asymmetric information, severe adverse selection and moral hazard problems similar to those of their counterparts in their attempts to provide funds to entrepreneurs. However, the use of debt-like instruments is a rational response on the part of Islamic banks to informational asymmetries in the environments in which they operate.

Table 2. OLS estimation: Islamic bank financing and characteristics models.

ΔFINANCEit

Bank-specific characteristics

SIZE SIZE LIQUIDITY LIQUIDITY CAPITAL CAPITAL

ΔBFRt -1.841*** -1.787*** -3.156*** -3.509*** -4.363*** -5.478***(0.165) (0.259) 0.272 0.516 (0.155) (0.681)

Bank-specific char.SIZE 1.110*** 1.113*** – – – –

(0.050) (0.052)LIQUIDITY 2.282*** 1.811*** – –

– – (1.827) (1.924)CAPITAL – – – – 3.065*** 2.887***

(0.884) (0.873)Impact of BFRΔBFRxSIZE -0.145*** -0.146*** – – – –

(0.021) (0.021)ΔBFRxLIQUIDITY – – -4.591*** -4.489*** – –

(0.657) (0.671)ΔBFRxCAPITAL – – – – -1.806*** -1.032***

(0.550) (0.542)Impact of MPΔONt – 0.053*** – 0.052*** – 1.237***

(0.194) (0.194) (0.737)Control VariablesΔGDPt 1.495* 1.142* 6.227* 8.363* 8.636*v 9.778**

(0.635) (0.710) (4.329) (4.969) (2.115) (7.256)ΔPRICESt -1.311 -1.828 -6.205 -9.292 -2.899 -2.277

(4.105) (4.561) (9.318) 10.106 (16.869) (18.852)R-square 0.848 0.848 0.332 0.341 0.771 0.664

Notes: Standard errors are reported in parenthesis. Standard errors and covariances are White-heteroskedasticity-consistent. The subscript i denotes banks and the subscript t denotes time, where t = 2006–2012. Bank-specific characteristics: SIZE is defined as logarithm of total asset; LIQUIDITY is defined as ratio of liquid asset (interbank deposits and securities); CAPITAL is defined as capital and reserves to total assets; BFRt is the base financing rate; MPt is the overnight interest rate; GDPt is the logarithm of real GDP; PRICESt is the logarithm of consumer price index; *significant at 10%, **significant at 5%, and ***significant at 1%.

Zulkhibri


Table 3. Fixed effect estimation: Islamic bank financing and characteristics models.

ΔFINANCEit



ΔBFRt -0.208*** -0.294*** -0.470*** -1.764*** -0.820*** -0.339***(0.128) (0.194) (0.415) (0.516) (0.171) (0.186)

Bank-specific char.SIZE 0.595*** 0.508** – – – –

(0.036) (0.554)LIQUIDITY 1.792*** 1.161** – –

– – (1.894) (0.815)CAPITAL – – – – 6.067** 8.815**

(2.354) (2.526)Impact of BFRΔBFRxSIZE -0.222** -0.021** – – – –

(0.010) (0.159)ΔBFRxLIQUIDITY – – -1.605** -1.458* – –

(0.511) (1.108)ΔBFRxCAPITAL – – – – -3.011** -3.671**

(1.599) (1.335)Impact of MPΔONt – -0.043*** – -0.528* – -0.368**

(0.017) (0.317) (0.169)Control VariablesΔGDPt 1.193* 1.538* 5.170* 5.021* 5.386** 8.222**

(0.746) (1.207) (3.116) (3.410) (2.494) (4.319)ΔPRICESt -0.844 -0.357 -0.361 -0.305 -0.980 -0.229

(2.197) (2.385) (7.153) (7.446) (7.967) (9.035)Constant 7.588*** 8.784 10.902*** 19.164*** 18.058*** 17.516***

(0.357) (7.917) (0.937) (16.003) (2.494) (2.553)

Firm Dummy Yes Yes Yes Yes Yes YesR-square 0.972 0.972 0.738 0.751 0.735 0.743F-statistic 72.29 66.77 5.989 5.856 5.231 4.995

Notes: Standard errors are reported in parenthesis. Standard errors and covariances are White-heteroskedasticity-consistent. The subscript i denotes banks and the subscript t denotes time, where t = 2006–2012. Bank-specific characteristics: SIZE is defined as logarithm of total asset; LIQUIDITY is defined as ratio of liquid asset (interbank deposits and securities); CAPITAL is defined as capital and reserves to total assets; BFRt is the base financing rate; MPt is the overnight interest rate; GDPt is the logarithm of real GDP; PRICESt is the logarithm of consumer price index; *significant at 10%, **significant at 5%, and ***significant at 1%.

To analyze further the impact of banks reducing their bank financing in response to a change in base financing rate with respect to specific characteristics, we have interacted the base financing rate variable with bank size (∆BFR*SIZE), liquidity (∆BFR*LIQUIDITY) and capitalization (∆BFR*CAPITAL); this is to investigate the economic arguments that there is a unique role for Islamic banks in the dual banking system, and the importance of heterogeneity among banks. Tables 2, 3 and 4 also report the results of the base financing rate with respect to bank-specific characteristics for bank financing using a fixed-effect and random-effect model. The estimates of bank-specific characteristics coefficients provide interesting results. The estimate coefficients of BFR*LIQUIDITY, BFR*CAPITAL and BFR*SIZE consistently show a positive sign, and are highly

significant at the conventional level. These results suggest that banks’ lending and their ability to obtain other sources of funding are factors that are affected indirectly through bank-specific characteristics.

In summary, we find strong evidence of the asymmetric adjustment of bank financing through bank-specific characteristics, as reported in the literature for conventional banks and in line with the arguments of Kashyap and Stein (1995, 2000) and Kishan and Opiela (2000). Moreover, banks react differently to the base financing rate depending on their own specific characteristics; a bank with higher capitalisation, in particular, is expected to increase financing more than a bank with greater size and liquidity. Furthermore, since an Islamic bank is operating in a dual



Table 4. Random effect estimation: Islamic bank financing and characteristics models.

ΔFINANCEit



ΔBFRt -0.368*** -0.324*** -0.452*** -0.334*** -0.377*** -0.371***(0.028) (0.025) (0.214) (0.215) (0.142) (0.105)

Bank-specific char.SIZE 0.618*** 1.399*** – – – –

(0.035) (0.496)LIQUIDITY 9.471*** 4.639** – –

– – (1.805) (2.858)CAPITAL – – – – 7.307*** 8.521***

(2.284) (2.331)Impact of BFRΔBFRxSIZE -0.221** -0.244** – – – –

(0.010) (0.143)ΔBFRxLIQUIDITY – – -1.539*** -0.189*** – –

(0.102) 0.075ΔBFRxCAPITAL – – – – -3.406*** -3.667**

(1.129) (1.777)Impact of MPΔONt – -0.041** – -0.472** – -0.323***

(0.021) (0.231) (0.036)Control VariablesΔGDPt 1.222* 1.423* 5.365* 9.664* 6.131* 8.530*

(0.842) (1.112) (4.104) (4.380) (3.517) (5.198)ΔPRICESt -0.657 -1.394 -0.698 -0.808 -3.444 -5.773

(2.191) (0.502) (7.112) (7.406) (2.066) (8.947)Constant 7.101*** 4.052 10.873*** 13.830 17.810*** 17.412

(0.346) (7.067) (0.861) (14.852) (2.297) (2.340)

Firm Dummy Yes Yes Yes Yes Yes YesR-square 0.908 0.904 0.394 0.413 0.400 0.409F-statistic 93.58 72.83 6.256 5.515 5.455 4.968

Notes: Standard errors are reported in parenthesis. Standard errors and covariances are White-heteroskedasticity-consistent. The subscript i denotes banks and the subscript t denotes time, where t = 2006–2012. Bank-specific characteristics: SIZE is defined as logarithm of total asset; LIQUIDITY is defined as ratio of liquid asset (interbank deposits and securities); CAPITAL is defined as capital and reserves to total assets; BFRt is the base financing rate; MPt is the overnight interest rate; GDPt is the logarithm of real GDP; PRICESt is the logarithm of consumer price index. *significant at 10%, **significant at 5%, and ***significant at 1%.

banking system, the asymmetric information problems faced by Islamic banks are expected to affect the ability to protect the financing lines from policy-induced reductions in deposits, and result in Islamic bank financing behaviour.

6. ConclusionA significant number of empirical studies have explored the bank lending behaviour of conventional banks over the past decades, while studies on Islamic bank financing behaviour remain scarce due to lack of bank-level data. Under the dual banking system, to the extent that financial constraints vary with banks’ ability to access other sources of financing, the implication is that Islamic bank financing responses to the bank financing rate and the conventional interest rate are

contingent on observable bank-specific characteristics. Understanding this mechanism is crucially important, in the context that Islamic banks have increasingly played a dominant role in the Malaysian financial system.

This paper analyses the importance of bank-specific characteristics with respect to Islamic bank financing in Malaysia. The results obtained from pooled panel estimation allow us to draw several significant conclusions about Islamic bank financing behaviour in Malaysia within a dual banking system. The evidence gathered in this study suggests that bank-specific characteristics are important for Islamic banking financing behaviour. The financing behaviour of Islamic banks is consistent with the behaviour of conventional banks in that the bank lending operates

Zulkhibri


according to banks’ size, and level of liquidity and capital (Goldniuk 2006). The results of these regressions also suggest that the reaction of Islamic banks financing to changes in interest rates is the same as for conventional banks, and are robust to different types of econometric specifications.

Many problems and challenges relating to Islamic instruments, financial markets, and regulations must be addressed and resolved. A complete Islamic financial system with its identifiable instruments and markets is still at a relatively early stage of evolution. The functioning of Islamic banks should rapidly differentiate itself from conventional banking. Due to the existence of moral hazards and adverse selection in the industry, an Islamic bank is not able to provide a full-fledged alternative finance to conventional finance. Moreover, an Islamic bank does not develop in the path that was envisioned by the Islamic scholars (Saeed 1996). One of the drawbacks is the low level of participation in PLS arrangements, which seems contradictory to the essential concept of Islamic banking. In practice, it would beneficial for Islamic banks to stop replicating the conventional banking models that concentrate mainly on debt-based instruments and mark-up models, but instead to move over to the PLS model.

ReferencesAggarwal R, Yousef T. “Islamic Banks and Investment

Financing.” Journal of Money, Credit, and Banking, 32(1), 2000. 93–120.

Bank Negara Malaysia, Quarterly Bulletins, Kuala Lumpur.

Beck T, Demirguc-Kunt A, Merrouche O. “Islamic vs. Conventional Banking: Business Model, Efficiency and Stability.” Journal Banking and Finance, 37(2), 2013. 433–447.

Brooks PK. “Does the Bank Lending Channel of Monetary Transmission Work in Turkey?” IMF Working Paper 07/272, Washington DC: International Monetary Fund. 2007.

Cervik S, Charap J. “The Behavior of Conventional and Islamic Bank Deposit Returns in Malaysia and Turkey.” IMF Working Paper, No. WP/11/156, Washington DC: International Monetary Fund, 2011.

Chong B, M-H Liu. “Islamic Banking: Interest-Free or Interest-Based?” Pacific- Basin Finance Journal, 17, 2009.124–144.

Chapra MU. “The Future of Economics: An Islamic Perspective.” The Islamic Foundation, Leicester, UK, 2000.

Dridi J, Hasan M. “Have Islamic Banks Been Impacted Differently than Conventional Banks During the Recent Global Crisis?” IMF Working Paper, No. WP/10/201, Washington DC: International Monetary Fund, 2010.

El-Gamal M. “Mutuality as an Antidote to Rent-Seeking Sharia-Arbitrage in Islamic Finance”. http://www.ruf.rice.edu/~elgamal/. 2005.

Ehrmann M, Gambacorta L, Martinez-Pages J, Sevestre P, Worms A. “Financial Systems and the Role of Bank in Monetary Transmission in the Euro Area”. In I. Angeloni, A. Kashyap, and B. Mojon, (eds.), Monetary Transmission in the Euro Area: A Study by the Eurosystem Monetary Transmission Network, Cambridge University Press, 2003. 235–269.

Fadzlan S, Zulkhibri M. “Post-Crisis Productivity Change in Non-Bank Financial Institutions: Efficiency Increase or Technological Progress?”Journal of Transnational Management, 14(2), 2009. 124–154.

Haron S. “Islamic banking and finance.” Leading issues in Islamic Banking and Finance, 20 (1), 2001. 17–32.

Huang Z. “Evidence of Bank Lending Channel in the U.K.” Journal of Banking and Finance, 27, 2003. 491–510.

Gambacorta L. “Inside the Bank Lending Channel.” European Economic Review, 49, 2005. 1737–1759.

Gujarati DN, Sangeetha S. “Basic Econometric.” 4th Edition, McGraw-Hill Education Books Ltd., India, 2007.

Golodniuk I. “Evidence on the Bank-Lending Channel in Ukraine.” Research in International Business and Finance, 20, 2006. 180–99.

Kasri RA, Kassim, S. “Empirical Determinants of Saving of the Islamic Banks in Indonesia.” Journal of King Abdul Aziz University, Islamic Economics, 22(2), 2009.

Khan M, Mirakhor A. “Islamic Banking”. The New Palgrave Dictionary of Money and Finance, London: Macmillan Press, 1992.

Kashyap A, Stein J. “The Impact of Monetary Policy on Bank Balance Sheets.” Carnegie Rochester Conference Series on Public Policy, 1995. 51–195.

Kashyap A, Stein J. “What Do a Million Observations Say About the Transmission Mechanism of Monetary Policy.” American Economic Review, 90(3), 2000. 407–428.

Kishan P, Opelia T. “Bank Capital and Loan Asymmetry in the Transmission of Monetary Policy.” Journal of Banking and Finance, 30, 2006. 259–285.

Khan M. “Time Value of Money and Discounting in Islamic Perspective.” Review of Islamic Economics, Vol. 1, No. 2, 1991. 35–45.

Mills P, Presley J. “Islamic Finance: Theory and Practice.” London, UK: Macmillan, 1999.

Saeed A. “Islamic Banking and Interest: A Study of the Prohibition of Riba’ and its Contemporary Interpretation.” Leiden: EJ Brill, 1996.


Cite this chapter as: Bhatt V, Sultan J (2015). Leverage risk, financial crisis, and stock returns: A comparison among Islamic, conventional, and socially responsible stocks. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Leverage risk, financial crisis, and stock returns: A comparison among Islamic, conventional, and socially responsible stocksVaishnavi Bhatt1, Jahangir Sultan2

1Ramaiah Institute of Management Studies, Banglaore, India, Email: [email protected] Hughey Center for Financial Services, Bentley University, Email: [email protected]

We thank Marcia Cornet, Vishwanathan Iyer, Kartik Raman, and seminar participants at Bentley University for their useful suggestions. We thank Dow Jones Indexes for providing us with information on proprietary indexes. We remain responsible for all remaining errors. Please do not quote without permission.

Abstract - According to the financial press, firms with low leverage have lower distress risk due to their reduced exposure to the credit market, especially during credit crises. Compared to their conventional and socially responsible (SRI) counterparts, Shariah compliant (SC) stocks are low-leverage stocks. Our hypothesis is that SC firms would be less sensitive to leverage risk and thus would be ideal for wealth preservation during declining market environment. We find that the leverage risk factor performs consistently across various categories of firms and its impact is more pronounced during the recent financial crisis. However, we also find that compared to the conventional stocks, SC stocks are also quite sensitive to the leverage factor. In contrast, the SRI class of stocks has the least sensitivity to leverage risk factor, suggesting they can be attractive for wealth preservation during credit crises.

Keywords: asset pricing, leverage, returns, financial crisis

JEL Codes: G00, G01, G10, G11, G12, G32

1. IntroductionAs the divine code of law, the Shariah is a code of conduct that guides business transactions for the Muslims and are based on the Quran and the edicts of Prophet Muhammad (pbuh). Hence, the guidelines set forth in the Shariah become imperative to every Muslim and govern all aspects of life, whether they may be of personal, social, political, economic or financial nature. Shariah compliant (SC) stocks1 are low-leverage stocks with high asset backing, compared to their conventional and socially responsible (SRI) counterparts. It is a widely held belief that firms with low leverage have lower distress risk due to their reduced exposure to the credit market. Naturally, these firms are capable of promoting flight to safety, especially in a declining market environment.

In this paper, we examine if SC stocks have lower sensitivity to economy wide leverage risk. To this extent, we create a new leverage risk factor (LEV) on the basis of firm-specific financial leverage (total debt over assets)2. The risk factor LEV (defined as the return on high leverage stocks minus

the return on low leverage stocks) is a non-diversifiable risk premium and therefore should be included in any multifactor asset pricing model. The evidence that high leverage requires higher risk premium can be indicative of the notion that high leverage can be value destructive, especially when equity prices are falling in a persistent fashion. The fact that Islamic stocks may have lower credit market exposure is important for wealth preservation during both good and bad times. Milly and Sultan (2009) report that Islamic stocks listed globally have outperformed conventional stocks and SRI stocks during the 2007–2009 economic crisis3. It would be interesting to examine how these stocks respond to the traditional risk factors (such as market risk premium, size, and value) as well as the leverage risk factor. If indeed Islamic stocks have lower sensitivity to the leverage risk factor, it would be indicative of their attractiveness for wealth preservation when investors are looking for safer assets.

In this paper, using a sample of 3704 globally traded stocks for the period January 2000- April 2009, we construct a risk factor based on firm-specific leverage find that the inclusion

Bhatt and Sultan


of the leverage risk factor leads to a weakening of the significance of the traditional FF variables. Furthermore, we show that, in comparison to the traditional FF factors, the economic and statistical significance of the leverage risk factor is high, especially during the financial crisis. We also demonstrate that the leverage risk factor contributes to the systematic risk of a firm and represents the underlying macroeconomic fundamentals. Finally, we show that compared to the conventional stocks, SC stocks display substantially lower risk premium to traditional risk factors. We also find that similar to the conventional stocks, Islamic stocks are also sensitive to the leverage factor, thus leading us to suggest that a leveraged based screening of Islamic stocks may not be ideal for wealth preservation especially during a credit crises. An investor must search for other redeemable characteristics in Islamic stocks that can help preserve equity value during falling equity prices.

The remainder of the paper is as follows. In Section II, we review the link between leverage and stock returns. In Section III, we discuss the recent financial crisis to motivate the empirical model. In Section IV we offer empirical results, and the final section concludes the paper.

2. Review of literatureA detailed analysis of the sensitivity of SC stocks to the leverage risk is tricky. In the first place, one must demonstrate that, in the context of a multifactor asset pricing model, the previous risk factors are incapable of capturing economy wide leverage risk. Once a reliable risk factor is constructed, a researcher can proceed to the next stage to investigate whether such risk factor is significant in an asset pricing model. Finally, the analysis can proceed to examine if there are differences in the way different categories of firms respond to this newly created risk factor.

Consider the following multifactor asset pricing model (Fama-French (1992))

b b b b e- = + - + + +r r r r R R( )t ft mt ft t SMB t HML t0 1 2 , 3 , (1)

shows that excess return on a portfolio (rt – rft) is explained by the sensitivity of its return to three factors: the excess return on a broad market portfolio (rmt – rft); the difference between the return on a portfolio of small stocks and the return on a portfolio of large stocks (SMB, small minus big); and the difference between the return on a portfolio of high-book-to-market stocks and the return on a portfolio of low-book-to-market stocks (HML, high minus low).

Our analysis thus leads us to first address an important question which has largely been ignored in the literature. Fama and French (1992) note that SMB (return on a portfolio of small firms minus the return on a portfolio of large firms) and HML (return on a portfolio of high book to market firms minus return on a portfolio of low book to market firms) are statistically important in explaining the cross-section of equity returns. Subsequent work by academics and practitioners has sought to verify the effects of these factors (FF factors, from hereafter) on cross-section of equity returns (for example, see Fama and French (1993, 1995, and 1998), Liew and Vassalou (2000), Davis, Fama and French (2000), Sivaprasad and Muradoglu (2009),

and Vassalou and Xing (2004)). A common finding in the literature is that value stocks earn a premium over growth stocks. Similarly there is evidence that small sized stocks earn a premium over big stocks.

These so-called empirical anomalies continue to generate controversies in the literature. For instance, are value and size premiums caused by the underlying risk factors of firms falling within these categories? Similarly, the notion of whether value and size premiums reflect incorrect extrapolation of past earnings growth by the market and subsequent correction of the mispricing errors, continues to receive attention in the literature (see Eom and Park (2008) for a recent survey).

How well do FF risk factors capture financial distress risk? Fama and French (1992) note that the combination of book to market and size describes the cross-section of average stock returns and absorb the apparent roles of other variables like leverage and E/P. The authors note that the SMB and HML factors are correlated with leverage and, therefore, adequately represent financial distress. The ability of the traditional FF factors to directly capture leverage risk is critical for asset management, especially when leverage risk becomes a source of systemic risk in the economy. The implication for an investor facing such catastrophic shocks is simple. If size and value based strategies do not perform consistently well across good and bad times, the rationale behind such investing strategy is at risk.

However, Fama and French (1992 and 1993) deal with the market leverage (assets over market value of equity) and the book leverage (assets over book value of equity), which may not directly capture the sensitivity of the firms to economy wide leverage risk4. In particular, the debt market exposure of a firm is a major determinant of the distress risk that may not be directly captured by the FF factors. Furthermore, to the extent that excessive leveraging and major credit events can lead to correlated defaults, we may find that the debt market exposure is monotonically increasing in financial leverage. In essence, the resulting credit crisis produces contagion-like effects with leverage risk as being the primary catalyst. According to Fama and French (1996), if default risk becomes correlated across firms, market participants, especially workers in distressed firms, tend to avoid all distressed firms in general. We believe that this presents an ideal opportunity for volatility spillover among firms in the economy, with the extent of spillover monotonically rising in leverage.

Surprisingly, very few studies have empirically examined the role of leverage risk factor in asset pricing. Chan and Chen (1991) examine the effects of financial leverage (book value of debt and preferred stock over market value of equity) on stock returns and find a positive relationship. Unfortunately, their analysis does not investigate if factor loadings on the financial leverage can subsume the effects of HML and SMB. As Fama and French (1992) write, “It would be interesting to check whether loadings on their distress factors absorb the size and book-to-market equity effects in average returns documented here.” Ferguson and Shockley (20003) write, “... a three-factor empirical model that includes factors based on relative leverage and relative distress should outperform the Fama and French (1993) three-factor model in the cross section”.


A comparison among Islamic, conventional, and socially responsible stocks

An investigation into this topic is timely given the recent financial crisis when economy-wide leverage played a key role in exacerbating the risk exposure especially for the leveraged5 financial and non-financial firms. As the subprime crisis deepened, coupled with escalating liquidity crisis, the credit market virtually dried up, limiting access to funds. The TED spread (difference between the interest rates on Eurodollar loans and short-term U.S. T-bill) rose in July 2007, then spiked even higher in September 2008, reaching as high as 4.65% on October 10, 2008. While the impact was felt mostly by the hedge funds, insurance agencies, banks, and firms directly involved in construction business and mortgage lending, the effects of the liquidity crisis also had affected the non-financial firms as well. Thus, the financial crisis in 2007–2008 had a devastating contagion-like effect on credit risk, with leverage risk acting as the centrepiece. An analysis of the Islamic stocks and their conventional counterparts is critical from the point of view of academic as well as the practitioner community. If Islamic stocks have lower sensitivity to the leverage risk factor, then these stocks would be ideal for wealth management, especially during financial crises.

There are several studies on the relationship between leverage and stock returns. See Chou, Ko, and Lin (2010) for a recent survey. In one strand of the literature, leverage is positively related to stock returns, especially for weak firms with poor investment opportunities. Accordingly, as debt increases the risk exposure of such firms, investors demand a premium. Sivaprasad and Muradoglu (2009) find that leverage has a significant positive relation with stock returns. Gomes and Schmid (2009) show that equity returns are increasing in market leverage. Ho, Strange and Piesse (2008) conduct a similar study for the Hong Kong stock exchange and conclude that market leverage (Assets/Market value of equity) exhibit a significant conditional relationship with the stock returns. Bhandari (1988) performs cross sectional regressions between monthly average returns and the leverage ratios for the period 1948–1979 and finds that the debt equity ratio has a positive effect on stock returns. Ferguson and Shockley (2003) include relative leverage (D/E) and relative distress risk, based on Altman’s Z score. They find that their model performs better than the three factor FF model in explaining stock returns. On similar lines, Chou et al (2010) propose an augmented five factor model which incorporates both FF factors as well as Ferguson and Shockley factors and demonstrate that this augmented five factor model explains most of the asset pricing anomalies.

In contrast, there are several studies that offer rationales for supporting a negative relationship between financial leverage and stock returns. The debt-overhang theory (Meyers, 1977) provides a convenient framework to suggest why leverage reduces equity return. Accordingly, as leverage increases, the distress risk increases, and shareholders pass up positive NPV projects. As a result, the stock price decreases, reflecting underinvestment in successful projects and a decline in firm value (Meyers (1977)). Other explanations include firms substituting debt for equity especially during economic crisis when the cost of equity financing is higher than the cost of debt financing (Dimitrov and Jain (2006)); managerial preference for equity over debt because high debt payments can reduce equity returns, especially when firms do not take advantage of growth opportunities (Lang, et al (1995)); the benefit of external disciplining mechanism

of debt financing (Jensen and Meckling (1976) and Fama and Jensen (1983)); and a reduction of the manager’s ability to waste free cash (Jensen (1986)). Overall, these studies imply that debt reduces agency costs and managerial waste, improves disclosure, and thus reduces equity risk premium. As a result, leverage is decreasing in stock returns.

The previous discussion suggests that the leverage risk factor is important for asset pricing models. Our focus in this paper is to examine the extent to which the well-known anomalies (size and book to market effects) are resolved by directly adding leverage as a systematic risk factor. Leverage risk becomes fundamental risk especially when firms’ exposure to the debt market becomes pervasive and correlated across the economy. Fama and French (1996) recognize that investors avoid financially distressed firms because distress risk is correlated across the economy. We suggest that when leverage risk becomes correlated across the economy, it has a contagion-like effect on firms in general, especially those with high exposure to the debt market. To this extent, while size and book to market factors are correlated with the leverage of the firm, they may not adequately capture the firm’s direct exposure to the economy wide systemic risk due to excessive leverage. Finally, to the extent that Islamic stocks tend to have low leverage and are involved only in permissible economic activities under the guidelines of the Quran and Sunnah, may have reduced exposure to interest rate volatility. This simple and powerful proposition has not been fully addressed in the literature. If Islamic stocks continue to act like their conventional counterparts, it only goes to reaffirm the harmful effect of riba as firms take on more debt.

Our suggestion is consistent with the anecdotal evidence from the recent financial crisis when leverage risk became one of the primary drivers of the global economic crisis. There was plenty of evidence of such systemic risk in the recent financial crisis: debt markets such as the commercial paper market, the repo market, and short-term bank borrowing virtually dried up. Altogether, increased leverage of firms, especially of hedge funds, insurance agencies, banks, and mortgage companies, coupled with a liquidity crisis, took a heavy toll on the global economy.

In the next section, we discuss the link between leverage risk factor and selected macroeconomic variables such as the industrial production, unemployment, inflation, credit spread and term spread. Our intent is to draw inferences on the effects of the leverage risk factor on stock returns across various time periods.

3. Leverage risk and the financial crisis—contemporary evidenceIn 2004, the US Securities and Exchange commission granted a waiver of the international standards of maximum accounting leverage ratio6 (which was about 12) for five major securities firms – Goldman Sachs, Merrill Lynch, Morgan Stanley, Lehman Brothers and Bear Sterns.7 Subsequently, many of the investment banks boosted their leverage ratios to as high as 30. Mortgage giants Freddie Mac and Fannie Mae had leverage levels close to 60 to 1 (2008 data), which can be very lucrative if the asset prices rise, but is disastrous when asset prices fall. A recent report8 cites excessively high leverage ratios prevailing in

Bhatt and Sultan


the housing market and the underlying mortgage backed securities as the culprit behind the credit crisis. Towards the end of the year 2009, the global economy was afflicted with excessive indebtedness which adversely affected the worldwide economy. For example, average household sector debt increased 141 per cent of disposable income in the United States and 177 percent in the United Kingdom. Furthermore, the best known banks in the US and Europe had their leverage (assets/equity) rising to forty, sixty or even hundred times the size of their equity capital.9

There is a broad consensus that increased leverage affects stock returns during the financial crisis. According to the popular press10, under normal circumstances where stock prices deviate from their underlying fundamentals, prices tend to bounce back to their intrinsic values, thereby restoring the efficiency of the equity markets. However, during a prolonged crisis, price discovery process takes longer, and stocks move away from their intrinsic values for a longer period of time. In addition, when investors are pessimistic about the financial markets, they may miss out on profitable arbitrage opportunities as prices move. In fact, due to the significant mispricing in the market, the US subprime crisis caused share prices of various US and European banks to fall and exerted immense pressure on these banks in the form of deteriorating profit margins.11

From a balance sheet perspective, companies reduce their leverage ratios either by selling off their assets (thereby restructuring their balance sheets) or by issuing new shares. Both of these strategies have different implications on the expected returns from the investor perspective. According to James Lee, Vice Chairman of JP Morgan,12 in spite of the efforts by the financial sector to augment their capital levels to as high as $300 billion firms have not been able to bring down the leverage to pre-crisis levels.

While many financial institutions and asset managers have been deleveraging since 2008, the process might eventually diminish the ability of these institutions to produce attractive returns, especially when they are unable to grow their balance sheets. In such circumstances, as financing gets costlier, firms focus on augmenting their capital level rather than investing it. In this process—“The big get bigger and the rest get smaller”13—has a direct impact on the stock returns of these firms. In other words, higher leverage levels increase the risk exposure of the firms and present higher growth opportunities, which should lead to higher stock returns. In contrast, lower leverage levels shrink the balance sheet of the firm and also reduce their competitiveness, having a negative impact on the shareholder value and stock returns.

Leverage risk during the financial crisis has macroeconomic implications. Notwithstanding the de-leveraging efforts of banks and other financial institutions, as of November 6, 2009, banks in particular exhibited 40 to 1 leverage (assets over equity capital). Similarly, the deleveraging efforts undertaken by many governments have also led to adopting restrictive monetary policy, resulting in higher interest rates. However, analysts argue that increasing interest rates and withdrawing funds from the financial system may cause the economy to exacerbate the effects of the credit risk. It has also been forecasted14 that deliberate attempts by the governments to deleverage will lead to lower wages in developed countries and a permanent unemployment

of 15% to 25%. Such macroeconomic instability has the potential to push investors away from the stock and bond markets. Furthermore, an increase in the perceived risk in the financial markets would prompt investors requiring a higher risk premium, which directly affects the expected returns on these stocks. So, deleveraging could have negative effects and is expected to reduce productivity. Overall, leverage affects expected returns not as a firm specific variable but as a systematic risk factor.

4. Empirical resultsOur initial sample includes weekly data for approximately 4000 stocks from 55 countries from January 2000 to April 2009. Our sample includes both financials (banks, S&Ls, credit unions, mortgage financing companies, real estate firms, and insurance companies) and non-financial firms. Since financial firms, especially banks and insurance firms, operate with high leverage, we will also separate financials from the aggregate sample to examine if financials stocks have different sensitivity to the risk factors.

We eliminate stocks having negative book to market equity from the sample in the construction of the risk factors.15 Also, the number of stocks each year used in the construction of factors varies depending on the availability of data for the corresponding year. This eliminates the problem of survivorship bias in the sample. The data for the weekly stock returns are extracted from Datastream, while the data related to economic fundamentals like size, book to market equity and leverage are extracted from FactSet. Stock returns are in US dollar terms and are based upon log relatives of weekly stock prices. The Dow Jones Global Index is used as the market benchmark, and the US risk-free rate is used as a proxy for global risk free rate16. We use previous year-end fundamentals to form portfolios for each successive year; the rationale behind this is that investors use information contained in the balance sheets and financial statements to predict future returns. Investors are assumed to follow a buy and hold policy with annual portfolio rebalancing.

Construction of risk factorsWe sort all stocks in the sample by size, book to market and leverage and categorize them in 3 groups (top 30%, middle 40%, and bottom 30%). Using the independent sorting procedure we construct value weighted portfolios formed by the intersection of three portfolios based on size, three portfolios based on book to market equity and three portfolios based on leverage (Debt/Assets). In all, we have 3*3*3 = 27 portfolios. The returns on these annually rebalanced portfolios create the dependent variable. In addition to the XMKT (market risk premium), the FF factors are: SMB (size mimicking portfolio constructed each week by taking the simple average of the returns on small sized portfolios minus returns on big sized portfolios), HML (book to market mimicking portfolios constructed each week by taking the simple average of the returns on high book to market portfolios minus the returns on low book to market portfolios) and LEV (leverage mimicking portfolios constructed each week by taking simple average of the returns on high leveraged portfolios minus the returns on low leverage portfolios).

Table 1 reports the number of stocks used for the construction of factors and portfolios each year which varies depending on the availability of data and meeting specific requirements



(for e.g. positive book to market equity). The correlation matrices for the sample across the three periods are reported in Tables 2A-2C. For the aggregate period, there is a positive correlation of .45 between HML and LEV, which is expected since both these factors closely represent the distress risk of the firm. In Table 2B, for the non-crisis period, the correlation coefficients are as follows: 0.407 (LEV, HML), -.1451 (LEV, XMKT), and .205 (LEV, SMB). In Table 2C, for the crisis period, there are some interesting changes. For example, the correlation between LEV and HML increases further, and the correlation between LEV and XMKT actually turns positive. Finally, the correlation between LEV and SMB actually turns negative. Table 3 reports summary statistics for the risk factors, XMKT, SMB, HML and LEV.

Table 1 A. Number of stocks.

Year No. of stocks

2000 37452001 43442002 43782003 43792004 43962005 44032006 43992007 43992008 44062009 4391

Table 2A. Correlation matrix of the explanatory factors for—all stocks (Jan 2000–April 2009).

XMKT SMB LEV HML

XMKT 1.000 -0.087 0.031 -0.057SMB -0.087 1.000 -0.030 0.086LEV 0.031 -0.030 1.000 0.451HML -0.057 0.086 0.451 1.000

Table 2B. Correlation matrix of the explanatory factors—all stocks (Jan 2000–June 2007).

XMKT SMB LEV HML

XMKT 1.000 -0.151 -0.145 -0.213SMB -0.151 1.000 0.205 0.283LEV -0.145 0.205 1.000 0.407HML -0.213 0.283 0.407 1.000

Table 2C. Correlation matrix of the explanatory factors—all stocks (July 2007–April 2009).

XMKT SMB LEV HML

XMKT 1.000 -0.007 0.222 0.143SMB -0.007 1.000 -0.448 -0.301LEV 0.222 -0.448 1.000 0.523HML 0.143 -0.301 0.523 1.000

Table 3. Descriptive statistics of the returns on Market factor, SMB, HML, and LEV factors.

XMKT SMB LEV HML

Mean -0.002 0.002 0.000 0.001Median 0.001 0.002 0.000 0.001Maximum 0.115 0.054 0.045 0.053Minimum -0.221 -0.063 -0.035 -0.049Std. Dev. 0.027 0.017 0.008 0.013Skewness -1.398 -0.272 -0.077 0.127Kurtosis 13.623 4.883 6.630 6.155

XMKT is defined as rm-rf where rf is the return on the risk free asset and rm is the return on the market portfolio. SMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. HML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. LEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”.

Macroeconomic variables and factor loadingsIn this section we demonstrate that the FF and leverage risk factors have macroeconomic implications. Several studies have shown that macroeconomic variables predict expected returns on stocks and bonds. See for example, Abel (1999), Fama (1981), Elton, et. al. (2001), Vassolou (2003) and Petkova (2006) and references therein. These studies show a significant positive relationship between the excess market returns and indicators of economic growth. We extend this analysis and test for the relationship between selected macroeconomic variables and returns on SMB, HML, and LEV factors.

We choose the following variables to represent the world economic environment for these globally traded stocks: growth rate in industrial production (world), unemployment rate (world), inflation (U.S.), credit spread (U.S.) and term spread (U.S.) during our sample period. Credit spread is defined as the difference in the weekly yield on Moody’s AAA corporate bonds and 1 year maturity government Treasury notes. Term spread is calculated as the difference between the weekly yield on 1 year treasury notes and 3 month treasury bills. The source of the data is the FRED database at the St. Louis Federal Reserve. Monthly data for industrial production for the world and unemployment rates have been obtained from the database of IHS Global Insights (http://www.ihsglobalinsight.com/EconomicFinancialData). Monthly inflation rates for the U.S. are obtained from the website www.Inflationdata.com.

Table 4 represents the results for multivariate regressions of each of the macroeconomic variables on lagged excess market returns and returns on SMB, HML and LEV. We use three lags to extract the maximum information content of these factors. The regressions are estimated at following periodicity: monthly for inflation, industrial production

Bhatt and Sultan


growth rate, unemployment rate and weekly for credit spread and term spread. The regression model is:

b b bb b e

= + - ++ + +

- -

- -

Y r r RR R

( )kt m ft t i t i SMB

t i HML t i LEV t

0 1 2 ,

3 , 4 , (2)

where, Ykt represents each of the following macroeconomic variables: monthly percentage change in industrial production growth rate, inflation and unemployment rate, weekly credit spread and term spread, and i represents the number of lagged terms 1 to 3 to reduce serial correlation. All macroeconomic variables have been tested for unit root and those with unit root have been differenced once to induce stationarity.

Panel A represents the coefficients and t-statistics for each of the above macroeconomic variables regressed against one lag of the independent variables. The results indicate that the leverage risk factor affects the unemployment rate and inflation (at 5% level of significance) while remains insignificant for term spread, industrial production and credit spread. In Panel B, unemployment exhibits significant sensitivity to LEV lagged one period and inflation shows significant sensitivity to LEV lagged two periods. Panel C also shows significant factor loadings on LEV for all the macroeconomic variables at different lag lengths. With respect to the other factors, SMB shows significant factor loadings for industrial production and term spread (at the first and second lags) while the impact of excess market returns on these macroeconomic variables seems to be weak. Note that HML seems to have limited ability to predict these economic variables at the first and the second lags but tends to exhibit a significant impact on these variables at the third lag (significant for unemployment and inflation). Overall, the results emphasize that the “leverage risk factor” is a systematic risk factor, though its effects on macroeconomic variables are not uniform.

These results have powerful implications for the US economy bouncing back from a severe financial crisis. First, researchers argue that “deeper the decline in GDP, peak to trough, the more rapid the post recession rebound.” A recent report17 suggests that this is the case only if there is a significant increase in the private sector liabilities. According to the report, a 0.3% drop in employment rate requires the real GDP growth higher than 3%, which in turn requires a 5% rise in the private sector liabilities, and subsequently, has a significant impact on the level of industrial production. In fact, we have witnessed slow moving recoveries following the 1980, 1991 and 2001 recessions, with the slowness being attributed to low levels of private liabilities during these periods. This supports the positive relationship between the leverage risk factor and industrial production and a consistently negative relationship between unemployment rate and the leverage risk factor which has been documented in the earlier section.

Next, the credit spread is a representative of firm’s default risk. A high credit spread indicates stringent credit markets and higher risk levels. However, the last quarter century witnessed some of the major developments in finance, for e.g. “securitisation” and introduction of “structured products” which generate cash flows from underlying

pool of assets like mortgages or credit card receivables (commonly known as collateralized debt obligations). Investors relied on the major credit rating agencies like Moody’s and Standard & Poor’s for the acceptance of these products. The collapse of the subprime lending sector and the resulting credit crisis in 2007 and 2008 exposed a colossal failure of the credit rating agencies; which also paved the way for a near-complete closure of markets for these products. In a nutshell, the credit spreads did not reflect the true economic risk underlying the corporate debt, hence it is difficult to establish a true empirical relationship between the leverage risk factor and the credit spread variables.

Notwithstanding the previous discussion, we find a positive relationship between LEV and the credit spread (see Table 4, Panel C). This is consistent with the evidence that the firms hit hard by the credit crisis were those that relied heavily on debt to finance growth like Home Depot, Toyota Motor and FedEx.18 Stock prices of these firms, including investment banks like Citigroup and UBS AG, plummeted during the recent market meltdown. Bear Stearns and American Home Mortgage are notable examples of firms which were coerced to sell their holdings at far below their book values. In general, there was a continuous re-pricing of risk in the stock market and stock prices plummeted. In contrast, the US treasury yields were falling due to flight to safety, while the rates on mortgage debts failed to decline at the same pace. This resulted in higher credit spreads because mortgage debts were most risky and demanded a premium over Treasury bonds. Thus, the positive relationship between leverage risk factor and the credit spread as seen in Table 4 is plausible since both the variables are representative of the increased exposure of the firm to distress risk caused by over leveraging.

Finally, when inflation is uncertain, investors demand inflation risk premium. Inflation induces volatility in the returns on debt and hence there is a leverage risk premium. Whether the relationship between inflation and leverage risk premium is positive or negative depends on the interaction between inflation, taxes (corporate tax and personal tax), expected return on assets, and the amount of debt used in the project. According to Armitrage (2005), as inflation increases, the real tax adjusted weighted average cost of capital decreases because higher inflation alleviates the corporate taxes on the firms’ real profits and increases the tax advantage on debt. However in the presence of personal taxes, higher inflation causes an increase in the tax rates on real returns to debt. This increases the leverage risk of the firms which are heavily dependent on debt and thus demand a premium over firms which rely less on external debt. For our sample, we find mixed evidence (positive and negative) of the relationship between inflation and the risk factors (See Table 4).

Explaining cross-section of returnsIn this section, we present our regression results by including leverage factor as a systematic risk factor. First, we test for the significance of the FF risk factors. Next we add LEV to the regression model to compare results across three periods: January 2000 –April 2009 (aggregate), January 2000 – June 2007 (non-crisis), and July 2007 – April 2009 (crisis)19. To check on the robustness of these results, we will



Table 4. Multivariate regressions of macroeconomic variables conditional on factor returns during the aggregate period.

The following regression is estimated to demonstrate the link between Fama-French factors and economic variables:

b b b b b e- + + + +- - - -Y = + r r R R R( )kt m ft t i t i SMB t i HML t i LEV t0 1 2 , 3 , 4 ,

where Ykt represents each of these macroeconomic variables (monthly Industrial production growth rates, monthly unemployment rate, monthly data for percentage change in inflation rates and weekly data for credit spread and term spread) for the combined period (January 2000 to April 2009). i represents the number of lagged terms 1 to 3. rf is the return on the risk free asset and rm is the return on the market portfolio. rf is the return on the risk free asset and rm is the return on the market portfolio. RSMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. RHML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. RLEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”.

Panel A

Industrial Unemployment rate Credit Spread Term Spread% change in inflation rate

Coeff. t-stats Coeff. t-stats Coeff. t-stats Coeff. t-stats Coeff. t-stats

XMKT t-1 -3.241 -0.307 0.007 0.044 -0.0001 -2.081** 0.0000 -1.660* 1.847 -0.172SMB t-1 -41.724 -2.249** 0.338 1.110 0.0001 1.326 -0.0001 -2.965** 1.532 1.365HML t-1 -29.364 -1.090 0.729 1.721* 0.0001 0.681 -0.0001 -1.393 -6.337 0.959LEV t-1 62.315 1.540 -1.667 -2.302** -0.0003 -1.404 0.0001 1.348 5.998 -2.207**R-square 0.105 0.114** 0.030** 0.028** 0.152**

Panel B



XMKT t-1 -2.582 -0.205 0.016 0.098 -0.0001 -1.619 0.0000 -1.501 2.308 1.707SMB t-1 -47.422 -3.096** 0.438 1.786* 0.0001 1.928** -0.0001 -2.992** 1.172 0.792HML t-1 -20.096 -0.794 0.578 1.588 0.0001 0.507 -0.0001 -1.802* -5.263 -1.897*LEV t-1 24.896 0.903 -0.910 -1.840* -0.0002 -0.960 0.0001 1.362 3.367 0.814XMKTt-2 13.845 2.104** -0.350 -2.413** -0.0001 -3.493** 0.0000 -0.388 0.361 0.454SMB t-2 -23.054 -1.497 0.087 0.305 0.0001 0.835 0.0000 -0.519 0.187 0.123HML t-2 -21.410 -0.897 0.447 1.018 0.0000 -0.183 -0.0002 -2.547** 0.306 0.165LEV t-2 61.648 1.553 -1.043 -1.412 -0.0001 -0.687 0.0001 0.951 5.379 2.274**R-square 0.224 0.219 0.054 0.042 0.186

Panel C



XMKT t-1 -5.246 -0.410 0.090 0.535 -0.0001 -1.651* 0.0000 -1.402 2.403 1.678*SMB t-1 -48.352 -3.397 0.480 2.004** 0.0001 1.835* -0.0001 -2.961 0.803 0.537HML t-1 -7.562 -0.297 0.230 0.579 0.0001 0.540 -0.0001 -1.733 -4.378 -1.353LEV t-1 44.872 1.699 -1.224 -2.597** -0.0002 -1.038 0.0001 1.267 3.673 0.859XMKTt-2 18.343 2.842 -0.458 -3.220** -0.0001 -4.013** 0.0000 -0.505 0.739 0.991SMB t-2 -25.884 -2.343 0.196 0.941 0.0000 0.603 0.0000 -0.585 0.067 0.049HML t-2 -6.985 -0.384 0.270 0.946 0.0000 -0.320 -0.0001 -2.321 0.110 0.053LEV t-2 1.522 0.045 -0.043 -0.084 -0.0001 -0.787 0.0001 0.812 6.470 2.284**XMKTt-3 16.418 2.282 -0.476 -3.596** 0.0000 -0.529 0.0000 0.473 0.308 0.338SMB t-3 -25.738 -1.619 0.087 0.316 0.0000 0.354 0.0000 0.225 0.393 0.214HML t-3 2.303 0.149 -0.631 -2.107** -0.0001 -0.923 0.0000 0.738 5.296 2.655**LEV t-3 2.740** -1.109 -2.042** 0.0003 1.953* 0.0000 0.153 -2.493 -0.861R-square 0.430 0.412 0.066 0.044 0.215

(*indicates significant at 10% level, ** indicates significance at 5% level)

Bhatt and Sultan


further classify firms into two groups: financial and non-financial. Financial firms include all financial institution as well as real estate and mortgage firms. The popular adage is that leverage is a two-way sword. It magnifies returns in an up market and magnifies losses in a down market. Finally, we test our main hypothesis that Islamic stocks would be less sensitive to the leverage risk factor than conventional and socially responsible stocks. Our primary rationale is that low leverage of Islamic stocks would lessen the interest rate exposure of these firms.

We use the following firm-specific GARCH model:

r r r rt ft mt ft- = + - + ++ +

R RR

t SMB t HML

t LEV t

β β β ββ e

0 1 2 3

4

( ) , ,

, (3)

t t tN e ψ σ| ~ ( , ),-120 (4)

σ a e δ σt i

q

t i ij

p

t j2

1

2

1

= + +=

-=

-∑ ∑Wi

(5)

where rt – rft in the mean equation is the weekly excess return on asset i, rft is the weekly risk free rate (US T-bill), rmt – rft is the market risk premium (XMKT), and SMB, HML and LEV are Fama-French factors and the leverage risk factor, defined earlier. The variance equation (5) models the conditional variance as a GARCH(p,q) process where p and q denote the lag length. W is the intercept term, a is the ARCH term and δ is the GARCH term. a and δ terms are expected to be positive and significant determinants of the conditional variance of changes in the excess return. The primary reason for using the GARCH model is that preliminary diagnostics suggest that the weekly excess returns have time varying variance with volatility clustering and fat tails. The GARCH models are estimated using the Bollerslev-Wooldridge (1992) corrections to deal with excess kurtosis. As noted earlier, standard t-statistics based inferences in the presence of excess kurtosis in the residuals are asymptotically invalid because standard errors are biased downward, leading to false acceptances.

Factor loadings at the firm levelWe test the above model at both the firm and portfolio level for all 3,707 financial and non-financial firms. Each week from January 2000 to April 2009 we run cross sectional regressions of weekly excess stock returns on XMKT, SMB, and HML factors. Next, we add LEV to test for its significance in addition to the market factor and the Fama-French factors. For robustness check, we test for the partial F-statistics of LEV to see whether this additional factor contributes significantly in explaining the cross section of expected returns (in addition to the market factor and the traditional FF factors).

Tables 5 exhibit the summary of the impact of XMKT, SMB, HML and LEV factors on the returns of firm and portfolios. Model 1 is the traditional FF case and Model 2 includes the LEV factor in addition to the FF factors. As shown, we have 3707stocks in the sample. Note that in Table 5 and subsequent tables, we only include regression results that are significant at least at the 5% level. In Panel A, the results show that for the aggregate period (2000–2009), in 3,304 instances XMKT is positive. A positive sign for the

XMKT is consistent with the single factor CAPM model. The distribution of the SMB is about half positive and half negative. The HML is positive in 1,539 and negative in 216 cases. When LEV is added to the model (Model 2), we find that, there is a .24% increase in the number of cases ((3312/3304)-1) where XMKT is significant. With the addition of LEV, there is a 2.54% increase in the number of cases where SMB is significant. Surprisingly, the number of cases HML is positive and significant drops by 16.58%. Finally, in 2,208 instances, LEV is positive, though in 125 instances it is negative.

The results (Panel B) for the non-crisis period (2000-June 2007) are similar. The number of cases where the factors is significant changed as follows: .39% (XMKT), 1.83% (SMB), and -2.05% (HML). With regard to positive and negative impact of the factors on stock returns, there are some changes compared to the aggregate period (Model 2). For example, SMB, the number of negative cases is now 680, representing a 40% decline from the previous model. In contrast, HML, now has 539 instances for which the coefficients are negative, indicating a 17.43% increase from the previous value. Finally, we have 862 instances of positive and 149 cases of negative coefficients for LEV. It appears that, compared to Model 2 (aggregate period), there is a large number of instances the regression coefficients are insignificant. Altogether, the number of significant cases drops by 56%, suggesting that the LEV factor is able to capture systemic risk in the economy across good and bad times quite well.

However, the contribution of LEV in capturing leverage risk is evident when we estimate firm-specific regressions for the crisis period (Panel C). During July 2007-April 2009, compared to the non-crisis period, there is a 201.07% increase in the number of the cases where LEV is significant. This increase is indicative of several stylized facts during the escalating financial crisis afflicting the global economy. It appears that the credit crisis had a contagion-like effect, impacting firms across all spectrums of leverage. In essence, firms were hard hit especially when access to the debt market was severely limited because of reluctance among financial institutions to lend. The results suggest that for 3,038 firms, the sensitivity to LEV is positive and significant. Only in 66 cases the variable has negative coefficients. Compared to the non-crisis period, the addition of LEV during the credit crisis leads to a change in the number of significant cases for the remaining factors: XMKT (-84.91%), SMB (-14.97%), and HML (-27.54%). In particular, the number of negative coefficients for HML is higher than the positive ones, indicating that during the recent credit crisis, a value based investment strategy would have earned investors negative risk premium. Again, it supports the notion that the HML may not have been a good proxy for the distress risk during this period.

Factor loadings at the portfolio levelTable 5 also the highlights portfolio-specific regressions (Panels D-F) for the three periods. Based on the intersection of these three factors, we have 27 portfolios with annual rebalancing. The results reconfirm our earlier finding that the addition of LEV weakens the significance of the traditional FF factors. For the aggregate period (Panel D), we find that LEV is significant and positive for 19 out of 27 portfolios. The number drops to 17 when we estimate the



Tabl

e 5.

Sum

mar

y of

resu

lts

show

ing

the

num

ber o

f sto

cks

and

port

folio

s w

hich

sho

wed

sig

nific

ant s

ensi

tivi

ties

to X

MK

T, S

MB,

HM

L, a

nd L

EV fa

ctor

s.

M

odel

1

Mod

el 2

rr

r

rR

R

Nit

ftm

tft

tSM

Bt

HM

Lt

tt

-=

+-

++

+

-ββ

ββ

e

eψ

01

23

10(

)

|~

(,

,,

σσ

σa

eδ

σ

t

ti

iq

ti

ijp

tj

2

2

1

2

1

),

=+

+=

-=

-∑

∑W

rr

r

rR

RR

itft

mt

ftt

SMB

tH

ML

tLE

Vt

t-=

+-

++

++

ββ

ββ

βe

e0

12

34

()

|,

,,

ψψσ

σa

eδ

σ

tt

ti

iq

ti

ijp

tj

N

-

=-

=-

=+

+∑

∑1

2

2

1

2

1

0~

(,

),

W

whe

re, r

i is

the

retu

rn o

n po

rtfo

lio i;

r f is

the

retu

rn o

n th

e ri

sk fr

ee a

sset

and

r m is

the

retu

rn o

n th

e m

arke

t por

tfol

io. R

SMB i

s th

e re

turn

on

the

size

mim

icki

ng p

ortf

olio

con

stru

cted

by

taki

ng

the

sim

ple

aver

age

of th

e re

turn

s ea

ch w

eek

of a

ll “s

mal

l” p

ortf

olio

s m

inus

“bi

g” p

ortf

olio

s. R

HM

L is

the

retu

rn o

n bo

ok to

mar

ket m

imic

king

por

tfol

io c

onst

ruct

ed b

y ta

king

the

sim

ple

aver

age

of th

e re

turn

s ea

ch w

eek

of a

ll “h

igh

BE/M

E” p

ortf

olio

s m

inus

“lo

w B

E/M

E” p

ortf

olio

s. R

LEV i

s th

e re

turn

on

leve

rage

mim

icki

ng p

ortf

olio

s co

nstr

ucte

d by

taki

ng th

e si

mpl

e av

erag

e of

the

retu

rns

each

wee

k of

all

“hig

h le

vera

ge”

port

folio

s m

inus

“lo

w le

vera

ge p

ortf

olio

s”. A

ll in

dica

ted

coeffi

cien

ts w

ith

(*)

are

sign

ifica

nt a

t lea

st a

t the

5%

leve

l of s

igni

fican

ce.

All

stoc

ks: 3

707

All

Port

folio

s: 2

7

Pane

l A: A

ggre

gate

Per

iod

Pane

l D: A

ggre

gate

Per

iod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

3304

1208

1539

Mod

el 1

Posi

tive

2713

18N

egat

ive

011

1421

6

Neg

ativ

e0

93

Tota

l33

0423

2217

55

Tota

l27

2221

Mod

el 2

Posi

tive

3312

1234

1005

2208

Mod

el 2

Posi

tive

2712

1719

Neg

ativ

e0

1147

459

125

Neg

ativ

e0

96

1

Tota

l33

1223

8114

6423

33To

tal

2721

2320

% c

hang

e in

si

gnifi

canc

e (b

y m

odel

)0.

24%

2.54

%-1

6.58

%

% c

hang

e in

si

gnifi

canc

e (b

y m

odel

)0.

00%

-4.5

5%9.

52%

Pane

l B: N

on-c

risi

s Pe

riod

Pane

l E: N

on-c

risi

s Pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

3570

1461

1244

0M

odel

1Po

siti

ve27

1518

Neg

ativ

e0

612

417

N

egat

ive

09

3

Tota

l35

7020

7316

61

Tota

l27

2421

Mod

el 2

Posi

tive

3584

1431

1088

882

Mod

el 2

Posi

tive

2715

1817

Neg

ativ

e0

680

539

149

Neg

ativ

e0

95

3

Tota

l35

8421

1116

2710

31To

tal

2724

2320

%ch

ange

in

sign

ifica

nce

(by

mod

el)

0.39

%1.

83%

-2.0

5%

%ch

ange

in

sign

ifica

nce

(by

mod

el)

0.00

%0.

00%

9.52

%

(Con

tinu

ed)

Bhatt and Sultan


Tabl

e 5.

(Co

ntin

ued)

Pane

l C: C

risi

s pe

riod

Pane

l F: C

risi

s pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

439

401

1330

0M

odel

1Po

siti

ve4

016

Neg

ativ

e11

515

6496

N

egat

ive

017

0

Tota

l55

419

6514

26

Tota

l4

1716

Mod

el 2

Posi

tive

157

854

439

3038

Mod

el 2

Posi

tive

07

127

Neg

ativ

e38

494

174

066

Neg

ativ

e0

712

0

Tota

l54

117

9511

7931

04To

tal

014

1327

%ch

ange

in

sign

ifica

nce

(by

mod

el)

-2.3

5%-8

.65%

-17.

32%

%

chan

ge in

si

gnifi

canc

e (b

y m

odel

)0.

00%

-17.

65%

-18.

75%

% c

hang

e in

si

gnifi

canc

e (b

y pe

riod

)-8

4.91

%-1

4.97

%-2

7.54

%20

1.07

%%

chan

ge in

si

gnifi

canc

e (b

y pe

riod

)-1

00.0

0%-4

1.67

%-4

3.48

%35

.00%

model for the non-crisis period (Panel E). In contrast, we find that in all cases, LEV is positive and significant during the crisis period (Panel F). We also note that, in comparison to the aggregate period, there is a 100% reduction in the number of cases XMKT is significant during the crisis period. For the remaining variables, percentage change in significance is as follows: SMB (-41.67%) and for HML (-43.48%), indicating an across the board weakening of the FF factors during the financial distress. In contrast, there is a 35% increase in the number of instances where LEV is positive and significant.

Overall, the FF factors seem to lose their significance when LEV as a systemic risk is included in the model. In particular, during a financial crisis period, sensitivity to LEV at the firm and portfolio level suggests that the traditional FF factors may not be adequately capturing the effects of economy-wide distress arising from excess leverage. Therefore, sensitivity to this systemic risk translates into additional risk premium that is not adequately captured by the FF risk factors.

Portfolio-specific regression results across aggregate, non-crisis and crisis periods are provided next in Tables 6–8 to highlight the magnitude of the coefficients and to check for robustness of adding LEV. In Table 6, there are several stylized facts for the aggregate period. First, there is a noticeable increase in the adjusted R2 when LEV is added as an explanatory variable. Second, as reported earlier, with the addition of LEV in Model 2, the traditional FF factors tend to lose their statistical significance. Finally, we find that in many instances, the coefficient of HML actually turns negative.

In Table 7, we report the results for the non-crisis period and the results indicate that while the variable LEV is an important explanatory power, its addition makes only marginal impact on Model 2. There is an increase in the adjusted R2 but by a small margin. In contrast, during the crisis period (Table 8), the addition of the LEV makes a substantial contribution to the overall forecast ability of Model 2. The adjusted R2 increases by a considerable margin. In addition, the size of the coefficients across the 27 portfolios is large, ranging from 1.56% (portfolio #7) to 4% (portfolio #18). The magnitude of the coefficients indicates the heightened sensitivity of firms to the economic distress during the period. As indicated earlier, the results indicate that our leverage risk factor performs quite well in representing systemic risk in the global economy. Also note that the number of negative significant coefficients for HML increases considerably (from 5 to 12)20 which suggests that the value based investment strategy may not with falling equity prices. In contrast, LEV has a positive relationship with stock returns for all 27 portfolios, suggesting that investors demand a premium for investing in high leverage portfolios during the credit crisis.

The negative effects of leverage risk factor on stock returnsIn a number of cases (see Tables 5–8), leverage risk has a negative effect on stock returns, which is consistent with several existing studies. For example, Penman et al. (2007) decompose the book to price ratio of a firm into two components. The first component is the enterprise book to price (which represents the operating risk of the firm),



Table 6. Factor loadings of all firms for the aggregate period (January 2000 to April 2009).

Model 1 Model 2

r r r r R R

N

( )

| ~ (0, ),it ft mt ft t SMB t HML t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 ,

12

2

1

2

1

∑ ∑

b b b b ee ψ σ

σ a e δ σ

- = + - + + +

= W + +

-

=-

=-

r r r r R R R

N

( )

| ~ (0, ),it ft mt ft t SMB t HML t LEV t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 , 4 ,

12

2

1

2

1

∑ ∑

b b b b b ee ψ σ

σ a e δ σ

- = + - + + + +

= W + +

-

=-

=-

where, ri is the return on portfolio i; rf is the return on the risk free asset and rm is the return on the market portfolio. RSMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. RHML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. RLEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”. All indicated coefficients are significant at 5% level of significance. GARCH models are estimated using the Bollerslev-Wooldridge corrections to the standard errors. Model 1 excludes LEV. Model 2 includes LEV. Coefficients of the GARCH variance equations are not reported to conserve space. They are available upon request.

Aggregate Period

Portfolio

Model 1 Model 2

Intercept MKT SMB HMLAdj.

R-square Intercept MKT SMB HML LEVAdj.

R-square

1 0.002* 0.794* 0.388* -0.235* -0.015 0.002* 0.794* 0.389* -0.231* -0.023 -0.0212 0.002* 0.663* 0.344* -0.047 0.000 0.002* 0.642* 0.301* -0.193* 0.722* 0.1033 0.001 0.742* 0.598* 0.053 -0.004 0.001 0.727* 0.560* -0.254* 1.303* 0.1694 0.002 0.557* 0.390* 0.261* 0.014 0.001 0.547* 0.364* 0.248* 0.145 0.0425 0.000 0.661* 0.546* 0.431* -0.001 0.000 0.665* 0.521* 0.353* 0.396* 0.0646 0.000 0.723* 0.565* 0.565* 0.024 0.000 0.701* 0.510* 0.444* 0.457* 0.1037 0.000 0.622* 0.640* 0.744* 0.055 -0.001 0.639* 0.671* 0.753* -0.118 0.0218 0.000 0.765* 0.731* 0.812* 0.022 0.000 0.767* 0.720* 0.740* 0.306* 0.0599 0.000 0.840* 0.842* 0.923* 0.080 0.000 0.862* 0.890* 0.725* 0.741* 0.14110 0.000 0.896* 0.053 -0.264* -0.041 0.000 0.900* 0.044 -0.289* 0.141 -0.02811 0.001 0.690* 0.062 -0.072 0.009 0.001 0.700* 0.023 -0.169* 0.453* 0.07912 0.000 0.718* 0.009 0.016 -0.021 0.000 0.733* -0.050 -0.155 0.807* 0.10113 0.000 0.630* 0.083 0.372* -0.014 0.000 0.628* 0.050 0.306* 0.337* 0.03414 0.001 0.631* 0.052 0.340* 0.042 0.001 0.648* -0.003 0.214* 0.548* 0.12915 0.001 0.632* 0.138* 0.462* 0.032 0.001 0.634* 0.089 0.306* 0.659* 0.15116 0.000 0.630* 0.226* 0.892* 0.108 0.000 0.628* 0.222* 0.874* 0.053 0.11217 0.001 0.749* 0.250* 0.684* 0.051 0.001 0.752* 0.203* 0.590* 0.515* 0.13218 0.001 0.810* 0.188* 0.872* 0.086 0.001 0.814* 0.130* 0.716* 0.939* 0.21519 0.000 0.772* -0.470* -0.288* 0.198 0.000 0.775* -0.467* -0.274* -0.100 0.18920 0.000 0.656* -0.396* 0.035 0.094 0.000 0.651* -0.406* -0.022* 0.409* 0.14821 0.000 0.613* -0.379* 0.082 0.122 0.000 0.622* -0.358* -0.075 0.650* 0.20222 0.001 0.698* -0.387* 0.255* 0.064 0.001 0.671* -0.414* 0.180* 0.312 0.10423 0.001 0.721* -0.283* 0.352* 0.073 0.001 0.718* -0.338* 0.261* 0.599* 0.15824 0.001 0.671* -0.253* 0.505* 0.103 0.001 0.667* -0.262* 0.357* 0.543* 0.18425 0.002* 0.804* -0.839* 0.856* 0.297 0.002* 0.806* -0.848* 0.937* -0.430* 0.27726 0.000 0.773* -0.321* 0.901* 0.179 0.000 0.769* -0.290* 0.670* 1.035* 0.25527 0.002* 0.846* -0.486* 0.880* 0.255 0.002* 0.857* -0.562 0.688* 1.271* 0.401

measured as the ratio of book value of operating assets to their market value. The second component is the financial leverage component (which represents the financing risk of the firm), measured as the ratio of market value of debt to market value of equity. The authors find that enterprise book to price ratio has a significant positive relationship with the expected stock returns while the “leverage” component of book to price ratio has negative relationship

with the expected stock returns. Johnson (2004) documents a negative relationship between leverage and cross section of expected returns after controlling for firm specific characteristics like volatility. See Arditti (1967), Dimitrov and Jain (2006) for similar results. In particular, Dimitrov and Jain (2006) note that during economic distress, raising equity is costlier than debt (e.g., bank financing or line of credit), so firms would prefer to increase leverage. So,

Bhatt and Sultan


Table 7. Factor loadings of all firms for non-crisis period (January 2000 to June 2007).

Model 1 Model 2

∑ ∑

b b b b ee ψ σ

σ a e δ σ

- - + + +

= W + +

-

=-

=-

r r = + r r R R

| N( ,

( )

~ 0 ),it ft mt ft t SMB t HML t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 ,

12

2

1

2

1 ∑ ∑

b b b b b ee ψ σ

σ a e δ σ

- = + - + + + +

= W + +

-

=-

=-

r r r r R R R

N

( )

| ~ (0, ),it ft mt ft t SMB t HML t LEV t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 , 4 ,

12

2

1

2

1

where, ri is the return on portfolio i; rf is the return on the risk free asset and rm is the return on the market portfolio. RSMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. RHML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. RLEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”. All indicated coefficients with (*) are significant at 5% level of significance. GARCH models are estimated using the Bollerslev-Wooldridge corrections to the standard errors. Model 1 excludes LEV. Model 2 includes LEV. Coefficients of the GARCH variance equations are not reported to conserve space. They are available upon request.

Portfolio

Non-crisis Period

Model 1 Model 2

Intercept XMKT SMB HMLAdj.

R-square Intercept XMKT SMB HML LEVAdj.

R-square

1 0.002* 0.856 0.449* -0.259* 0.473 0.002* 0.851* 0.454* -0.224* -0.195 0.473 2 0.002* 0.778 0.457* -0.139 0.320 0.002* 0.783* 0.447* -0.204* 0.343 0.328 3 0.001 0.822 0.678* 0.005 0.378 0.001 0.846* 0.646* -0.253* 1.119* 0.419 4 0.001 0.673 0.397* 0.278* 0.394 0.001 0.674* 0.396* 0.277* 0.005 0.392 5 0.000 0.709 0.565* 0.435* 0.440 0.000 0.722* 0.540* 0.378* 0.313* 0.449 6 0.000 0.781 0.618* 0.562* 0.433 0.000 0.777* 0.572* 0.479* 0.311* 0.450 7 -0.001 0.785 0.686* 0.712* 0.532 -0.001 0.777* 0.699* 0.772* -0.221* 0.531 8 0.000 0.807 0.764* 0.817* 0.501 0.000 0.810* 0.760* 0.766* 0.212* 0.502 9 0.000 0.878 0.882* 0.925* 0.471 0.000 0.899* 0.916* 0.775* 0.637 0.47510 -0.001 0.973 0.142* -0.367* 0.466 -0.001 0.974* 0.142* -0.369* 0.008 0.46411 0.000 0.743 0.069 -0.073 0.446 0.000 0.764* 0.045 -0.143 0.322* 0.45612 0.000 0.791 0.081 -0.070 0.324 0.000 0.812* 0.027 -0.178 0.661* 0.36113 0.000 0.699 0.127* 0.346* 0.383 0.000 0.703* 0.106* 0.310* 0.213* 0.38614 0.001 0.687 0.077 0.346* 0.405 0.001 0.715* 0.049 0.256* 0.429* 0.42215 0.001 0.662 0.176* 0.436* 0.400 0.001 0.682* 0.119* 0.338* 0.559* 0.43716 0.000 0.701 0.294* 0.846* 0.482 0.000 0.700* 0.300* 0.870* -0.084 0.48117 0.001 0.781 0.298* 0.668* 0.469 0.001 0.797* 0.271* 0.599* 0.392* 0.48418 0.001 0.852 0.233* 0.834* 0.446 0.001 0.886* 0.171* 0.729* 0.832* 0.47319 0.000 0.865 -0.360* -0.354* 0.542 0.000 0.929* -0.349* -0.415* -0.430* 0.55220 0.000 0.693 -0.367* 0.040 0.502 0.000 0.694* -0.374* -0.003 0.314* 0.51121 0.000 0.664 -0.308* 0.037 0.437 0.000 0.682* -0.304* -0.085 0.539* 0.46122 0.001 0.823 -0.335* 0.255* 0.420 0.001 0.814* -0.345* 0.221* 0.170 0.42223 0.001 0.769 -0.226* 0.324* 0.478 0.001 0.779* -0.280* 0.266* 0.523* 0.50424 0.001 0.715 -0.205* 0.498* 0.453 0.001 0.721* -0.215* 0.392* 0.447* 0.48025 0.002* 0.883 -0.747* 0.812* 0.401 0.002* 0.867* -0.753 0.930* -0.699* 0.44126 0.000 0.951 -0.496* 0.940* 0.363 0.000 1.014* -0.478 0.707* 0.666* 0.36327 0.002* 0.891 -0.416* 0.846* 0.395 0.002* 0.921* -0.450 0.677* 0.992* 0.456

falling equity returns during economic distress and rising leverage support the empirical finding that leverage and return on equity may be negatively correlated.

Managerial preference for debt over equity financing is also related to the value of the firm and its future prospects. Lang et al. (1995) find a negative relationship between financial leverage and future growth of a firm. The authors

emphasize that the negative relationship between leverage and growth is more visible for firms with a low Tobin’s q since these firms are characterised by negligible growth opportunities not recognised by the capital markets. The study further rationalises that managers of firms with considerably lucrative growth opportunities generally do not opt for a high leverage21 because high interest payments on debt tend to erode the profitability of the firm which



Table 8. Factor loadings of all firms during crisis period (July 2007 to April 2009).

Model 1 Model 2

r r r r R R

Nit ft mt ft t SMB t HML t

t t

- = + - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( )

| ~ ( ,, ,

σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R

|it ft mt ft t SMB t HML t LEV t

t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N( ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ),

W


Portfolio

Crisis period

Model 1 Model 2

Intercept XMKT SMB HMLAdj.

R-square Intercept XMKT SMB HML LEVAdj.

R-square

1 -0.003 0.075 -0.320 0.161 -0.034 -0.001 -0.034 0.264 -0.942* 2.931* 0.4102 -0.005 0.084 -0.164 0.221 -0.040 -0.002 0.020 0.461* -0.974* 3.193* 0.4443 -0.003 0.077 -0.326 0.498 -0.005 0.000 -0.028 0.393* -0.831* 3.633* 0.5494 -0.003 0.087 -0.079 0.608* -0.010 -0.003 0.038 0.238 -0.180 1.784* 0.3125 -0.003 0.145 -0.151 0.711* 0.012 -0.003 0.029 0.385* -0.364 2.610* 0.4506 -0.004 0.167* 0.060 0.757* 0.004 -0.004 0.064 0.598* -0.196 2.643* 0.4277 -0.004 0.078 0.129 0.782* 0.053 -0.003 -0.033 0.436* 0.177 1.563* 0.3118 -0.002 0.196* -0.081 0.855* 0.053 -0.001 0.019 0.544* -0.087 2.736* 0.4739 -0.002 0.112 0.077 1.347* 0.136 0.000 0.018 0.748* -0.080 3.851* 0.58710 -0.003 0.086 -0.520* 0.361 0.026 -0.001 0.063 0.022 -0.927* 3.042* 0.45811 -0.004 0.037 -0.619* 0.522 0.053 -0.003 -0.007 -0.128 -0.910* 3.149* 0.50612 -0.004 0.109 -0.791* 0.463 0.105 -0.003 0.063 -0.247 -0.859* 3.078* 0.53813 -0.003 0.100 -0.370* 0.702* 0.059 -0.003 0.057 0.031 -0.274 2.140* 0.38514 -0.003 0.127 -0.728* 0.724* 0.132 -0.002 0.039 -0.119 -0.442* 3.061* 0.55215 -0.004 0.095 -0.586* 0.908* 0.130 -0.003 0.034 -0.016 -0.490* 3.328* 0.56116 -0.003 0.098 -0.275 0.871* 0.134 -0.003 0.028 0.114 0.096 1.895* 0.37317 -0.003 0.031 -0.556* 1.278* 0.175 -0.002 0.012 0.020 0.039 3.287* 0.60918 -0.005 0.051 -0.863* 1.541* 0.264 -0.003 0.045 -0.134 -0.112 4.033* 0.65719 -0.002 0.173* -1.069* -0.026 0.182 -0.002 0.028 -0.554* -0.877* 2.388* 0.52220 -0.001 0.119 -0.964* -0.011 0.113 -0.002 0.000 -0.349 -1.069* 2.744* 0.47921 -0.003 0.194* -1.017* 0.092 0.117 -0.002 0.044 -0.475* -0.853* 2.596* 0.53122 -0.002 0.146 -0.903* 0.278 0.089 -0.001 -0.042 -0.480* -0.416 2.083* 0.35623 -0.002 0.160 -1.064* 0.445 0.178 -0.002 0.009 -0.485* -0.600* 2.883* 0.53424 -0.004 0.157 -0.897* 0.749* 0.199 -0.003 0.024 -0.386 -0.413 2.875* 0.56125 -0.004 0.196 -1.176* 0.887* 0.374 -0.002 0.129 -0.735* -0.004 2.702* 0.58126 -0.004 0.043 -1.124* 1.435* 0.386 -0.003 0.048 -0.587* 0.333 2.972* 0.68027 -0.005 -0.008 -1.782* 1.765* 0.454 -0.002 -0.048 -1.153* 0.486* 3.693* 0.771

prevents the firm from utilizing the benefits of these growth opportunities. Hence, a negative relationship between leverage and growth seems rational, which implies a negative relationship between leverage and stock returns.22

The negative effect of leverage on return on equity is also consistent from a corporate governance perspective. Jensen and Meckling (1976) suggest that increased debt levels

have direct implications on the cash flow of the company by enforcing regular interest payments on debt which controls managerial expropriation. Fama and Jensen (1983) explain that increased debt levels adds to the default risk of the firm and affects the manager’s reputations adversely in case the firm defaults on its interest payments or debt. This imposes a constraint on manager expropriation and leads to better corporate disclosures. In addition, Jensen

Bhatt and Sultan


(1986) suggests that leverage increasing transactions such as LBOs, new debt issues (bonds), and stock repurchase reduce the manager’s access to free cash, thus reducing their waste. He further suggests that debt reduces the agency cost. This implies that as leveraging increases, external monitoring increases, and managerial efficiency is expected to rise. Furthermore, this may be imply that as firms become efficient, shareholders demand less risk premium for leverage, and as a result, stock prices fall with higher leverage.

Consistent with the above discussion, there are also more instances of significant negative coefficients for LEV during the non-crisis period (which was a period with profitable investment opportunities in the market). For example as reported in Table 5, there are 149 cases of negative coefficients on LEV (Model 2, Panel B) during the non-crisis period, but the number reduces to 66 during the crisis period (Panel C). At the portfolio level (Panel F), compared to the non-crisis period (Panel E), the number of negative coefficients for LEV reduces from 3 to 0.

Leverage and investment strategyThese results have investment implications that suggest investing in highly leveraged firms. However, an investor needs to decide between excessively high leverage level and the negative effects of leverage on financial distress (Luoma and Spiller (2002)). See Bris and Koskinen (2002) for further evidences. A recent report23 elaborates that the regular interest payments on debt for those companies which fund their investments through debt tend to erode the cash flow levels of the company by adding to the operating expenses of the firm. The flip side of the argument is that a firm with highly profitable growth opportunities and with a strong cash flow position would still earn a higher return on equity since they yield high profit margins. The report claims that a period of economic recovery is characterized by a strong economic momentum which bolsters earnings potentials of levered firms. The rationale behind this is that debt is cheaper for firms with promising growth prospects, and such they perform at the peak levels when debt is easily available24. The economic recovery in 2003 provides strong evidence to this fact when the federal funds rate was approximately 1.25%, which in turn stimulated economic growth to jump from 1% to 7%. During this period, levered companies, high yield bonds and bank loans yielded attractive returns.

These results do not suggest that as efficiency increases, stock price decreases. Rather, as firms become more efficient, debt becomes cheaper and such companies can afford to have high debt levels in their capital structure (thereby decreasing the overall cost of capital) without increasing their credit risk. Due to lower risk levels, investors do not need additional compensation for excessive leverage as in the case of firms which are not efficient. Also, in efficient markets, due to strong corporate governance principles and better disclosures, the probability of insider information is reduced and information of the company is quickly reflected in the stock prices. Hence there is no scope for mispricing or arbitrage opportunities; so returns fall.

Leverage risk of financial and real estate firmsWe perform additional robustness tests by separating the financial stocks in the sample from the non-financial

stocks25. Such an examination is critical because it removes industry-specific effects of the credit crisis since the effects may not have been uniformly distributed among financial and non-financial firms. Financial firms included in the sample include banks, S&Ls, credit unions, mortgage financing companies, real estate firms, and insurance companies. Clearly, these firms bore the brunt of the credit crisis due to over speculation, deregulation, and over leveraging. We re-construct FF and LEV factors and estimate firm26 and portfolio-specific regressions using two separate samples of firms: the first sample with 645 financial stocks and the second sample with 2,975 non-financial stocks.

A summary of the regression results are reported in Panels A-F, Table 9. In Panels A-C, there is evidence that the leverage risk factor performs well across the three periods, especially during the crisis period. The results support the hypothesis that the addition of LEV weakens the significance of the traditional FF factors. For the aggregate period (Panel A), LEV is significant and positive for 17 out of 27 portfolios. During the non-crisis period (Panel B), the significance of LEV drops, we now have 14 positive and 4 negative instances. During the crisis period (Panel C), in 23 out of 27 cases, LEV is positive and significant. Note that, in comparison to the non-crisis period, there is a -85.19% change in the number of cases XMKT is significant during the crisis period. For the remaining risk factors, the change in significance is as follows: SMB (−5.26%) and HML (−59%), suggesting a weakening of the FF factors during the financial distress. In contrast, there is an increase of 27.78% in the number of instances where LEV is positive and significant during the crisis period.

In Panels D-F, we report a summary of statistically significant results for the non-financial firms in the sample. We confirm our previous findings that the addition of LEV weakens the significance of the traditional FF factors considerably. For the aggregate period (Panel D), LEV is positive in 16 out of 27 portfolios. During the non-crisis period (Panel E), in 14 instances LEV has positive and significant coefficients, and in 4 instances the coefficients are negative and significant. Similar to our earlier findings, in 27 out of 27 cases, LEV is positive and significant during the crisis period (Panel F). In comparison to the non-crisis period, there is a -100% change in the number of cases where XMKT is significant during the crisis period. For the SMB, the number of significant cases changes by -39.13% and for HML, the number of significant cases changes by -57%, confirming the fact that the power of the FF factors weakens during the financial distress. In contrast, there is a 50% increase in the number of instances where LEV is positive during the crisis period.

Details of these portfolio-specific regressions across aggregate, non-crisis and crisis periods are provided in Tables 10–15 to demonstrate the contribution of the LEV on a case by case basis. Tables 10–12 report the results for the financial stocks while Tables 13–15 report the results for the non-financial stocks. In these tables we also report the adjusted R2 for Model 1 (without LEV) and Model 2 (with LEV) and the results confirm our earlier results. In Table 10, first, there is a noticeable increase in the adjusted R2 when LEV is added as an explanatory variable, indicating increased forecasting power of Model 2 during the aggregate period. Second, as reported earlier, with the



Tabl

e 9.

Sum

mar

y of

fact

or lo

adin

gs fo

r fina

ncia

l and

non

fina

ncia

l sto

ck p

ortf

olio

s.

M

odel

1

Mod

el 2

rr

=

+

rr

RR

|N

(,

itft

mt

ftt

SMB

tH

ML

t

tt

--

++

+

-

ββ

ββ

e

eψ

01

23

10

()

,,

~σσ

σa

eδ

σ

t

ti

iq

ti

ijp

tj

2

2

1

2

1

),

=+

+=

-=

-∑

∑W

rr

=

+

rr

RR

|N

(,

itft

mt

ftt

SMB

tH

ML

t

tt

--

++

+

-

ββ

ββ

e

eψ

01

23

10

()

,,

~σσ

σa

eδ

σ

t

ti

iq

ti

ijp

tj

2

2

1

2

1

),

=+

+=

-=

-∑

∑W

whe

re, r

i is

the

retu

rn o

n po

rtfo

lio i;

r f is

the

retu

rn o

n th

e ri

sk fr

ee a

sset

and

r m is

the

retu

rn o

n th

e m

arke

t por

tfol

io. R

SMB i

s th

e re

turn

on

the

size

mim

icki

ng p

ortf

olio

con

stru

cted

by

taki

ng

the

sim

ple

aver

age

of th

e re

turn

s ea

ch w

eek

of a

ll “s

mal

l” p

ortf

olio

s m

inus

“bi

g” p

ortf

olio

s. R

HM

L is

the

retu

rn o

n bo

ok to

mar

ket m

imic

king

por

tfol

io c

onst

ruct

ed b

y ta

king

the

sim

ple

aver

-ag

e of

the

retu

rns

each

wee

k of

all

“hig

h BE

/ME”

por

tfol

ios

min

us “

low

BE/

ME”

por

tfol

ios.

RLE

V is

the

retu

rn o

n le

vera

ge m

imic

king

por

tfol

ios

cons

truc

ted

by ta

king

the

sim

ple

aver

age

of

the

retu

rns

each

wee

k of

all

“hig

h le

vera

ge”

port

folio

s m

inus

“lo

w le

vera

ge p

ortf

olio

s”.

Fina

ncia

l sto

ck p

ortf

olio

sN

on fi

nanc

ial s

tock

por

tfol

ios

Pane

A: A

ggre

gate

Per

iod

Pane

l D: A

ggre

gate

Per

iod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

277

15M

odel

1Po

siti

ve27

1318

Neg

ativ

e0

157

Neg

ativ

e0

83

Tota

l27

2222

Tota

l27

2121

Mod

el 2

Posi

tive

277

1517

Mod

el 2

Posi

tive

2713

1816

Neg

ativ

e0

167

0N

egat

ive

09

61

Tota

l27

2322

17To

tal

2722

2417

% c

hang

e in

si

gnifi

canc

e (b

y m

odel

)0.

00%

4.55

%0.

00%

% c

hang

e in

si

gnifi

canc

e (b

y m

odel

)0.

00%

4.76

%14

.29%

Pane

l B: N

on-c

risi

s Pe

riod

Pane

l E: N

on-c

risi

s Pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

278

150

Mod

el 1

Posi

tive

2714

18N

egat

ive

012

7N

egat

ive

09

3To

tal

2720

22To

tal

2723

21M

odel

2Po

siti

ve27

815

14M

odel

2Po

siti

ve27

1418

14N

egat

ive

011

74

Neg

ativ

e0

93

4To

tal

2719

2218

Tota

l27

2321

18

% c

hang

e in

si

gnifi

canc

e (b

y m

odel

)0.

00%

-5.0

0%0.

00%

% c

hang

e in

sig

nific

ance

(b

y m

odel

)0.

00%

0.00

%0.

00%

Bhatt and Sultan


Tabl

e 9.

(Co

ntin

ued)

Pane

l C: C

risi

s pe

riod

Pane

l F: C

risi

s pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

150

30

Mod

el 1

Posi

tive

00

6N

egat

ive

025

4N

egat

ive

06

4To

tal

1525

7To

tal

06

10M

odel

2Po

siti

ve4

06

23M

odel

2Po

siti

ve0

100

27N

egat

ive

018

30

Neg

ativ

e0

49

0To

tal

418

923

Tota

l0

149

27%

cha

nge

in s

igni

fican

ce

(by

mod

el)

-73.

33%

-28.

00%

28.5

7%%

cha

nge

in s

igni

fican

ce

(by

mod

el)

0.00

%13

3.33

%-1

0.00

%

% c

hang

e in

si

gnifi

canc

e (b

y pe

riod

)-8

5.19

%-5

.26%

-59.

09%

27.7

8%%

cha

nge

in s

igni

fican

ce

(by

peri

od)

-100

.00%

-39.

13%

-57.

14%

50.0

0%

addition of LEV in Model 2, the statistical significance of the traditional FF factors tend to weaken. Finally, we find that in many instances the coefficient of HML actually turns negative. During the non-crisis period (Table 11), the addition of LEV to the model makes only marginal impact on the forecast power of the Model 2. The adjusted the R2 changes by a small margin. In contrast, we find that during the crisis period (Table 12), the addition of the LEV makes a substantial contribution to the overall forecast ability of Model 2. The adjusted R2 increases by a substantial margin. Furthermore, the size of the coefficient for LEV across portfolios is large, similar to the results reported earlier. The magnitude of the coefficient clearly indicates an increased sensitivity of firms to the economic distress. Similar results are obtained for non-financial stocks in Tables 13–15.

Overall, our analysis shows that financial and non-financial categories of stocks have similar exposure to the debt market, despite the fact that the concept of leverage and its use varies across these two categories of firms. It again reinforces the notion that the financial crisis had a contagion-like effect on all types of firms. It also establishes the fact that our leverage risk factor is able to capture economy-wide risk from over leveraging during the financial crisis period.

Test for robustnessEarlier, we reported that the correlation between LEV and HML as follows: .45 (aggregate period), .40 (non-crisis period), and .51 (crisis period). These correlations may be viewed as high, raising a criticism that LEV factor may be collinear with HML, and as such, rendering the effects of HML insignificant in majority of the cases. We argue that while HML and LEV are both balance sheet variables and are therefore should be correlated, they do not represent similar risk factor in the economy. That is to say that HML does not represent LEV and LEV does not represent HML. Both are capturing economy wide risk. However, while HML is supposed to be capturing distress risk (Fama and French 1993)), it does not adequately capture systemic risk of firms when their exposure to the debt market rises due to economy-wide problems with over leveraging. To this extent, LEV adds unique information to the model and does a good job in capturing systemic risk related to the over exposure of firms to the debt market.

While a correlation of .51 may not be indicative of multicollinearity, it is important to examine if these results are robust to such statistical artefact. In the present context, we do not find a glaring evidence of multicollinearity because it would have been reflected in high F-statistics with insignificant t-statistics for the estimated coefficients.

Despite the fact that multicollinearity is not an issue, we decided to estimate the partial F-statistics to check the robustness of these results to multicollinearity. The partial F-statistic determines the incremental explanatory power of adding additional variables to the basic model. In the present context, a significant partial F statistic (critical value is 3.32 at the 1% significance level) provides justification for adding LEV to the model containing the traditional FF factors.

Table 16 reports the partial F-statistics (across all the three groups - all stock portfolios, financial stocks only portfolios,



Table 10. Factor loadings for financial and non-financial firms for the aggregate period (January 2000 to April 2009).

Model 1 Model 2

∑ ∑

b b b b ee ψ σ

σ a e δ σ

- - + + +

= W + +

-

=-

=-

r r = + r r R R

| N( ,

( )

~ 0 ),it ft mt ft t SMB t HML t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 ,

12

2

1

2

1 ∑ ∑

b b b b b ee ψ σ

σ a e δ σ

- - + + + +

= W + +

-

=-

=-

r r = + r r R R R

| N( ,

( )

~ 0 ),it ft mt ft t SMB t HML t LEV t

t t t

t ii

q

t i ij

p

t j

0 1 2 , 3 , 4 ,

12

2

1

2

1


Portfolio

Aggregate Period

Model 1 Model 2



R-square

1 0.002* 0.709* -0.026 -0.189* -0.060 0.002* 0.709* -0.026 -0.189* -0.001 -0.0632 0.001 0.379* 0.514* -0.395* -0.098 0.001 0.263* 0.431* -0.277* 0.857* 0.1133 0.003* 0.783* 0.138* -0.253* 0.060 0.002 0.716* 0.163* -0.189* 1.112* 0.3484 0.001 0.485* 0.151* 0.337* 0.023 0.001 0.478* 0.159* 0.333* 0.149 0.0435 0.000 0.473* 0.069 0.203* -0.015 0.000 0.441* 0.066 0.256* 0.419* 0.1496 0.003* 0.463* 0.082* 0.188* 0.038 0.003* 0.423* 0.085* 0.192* 0.462* 0.2047 0.000 0.383* 0.173* 0.822* 0.121 0.000 0.409* 0.175* 0.827* -0.105 0.0948 0.002 0.796* 0.211* 0.687* -0.052 0.001 0.702* 0.193* 0.753* 0.544* 0.1219 0.002* 0.756* 0.399* 0.783* -0.002 0.001 0.673* 0.368* 0.797* 0.802* 0.26710 0.002 0.546* -0.315* -0.245* 0.178 0.002 0.535* -0.317* -0.230* 0.112 0.20911 0.001 0.552* -0.190* -0.125* 0.088 0.001 0.505* -0.220* -0.126* 0.332* 0.21412 0.002* 0.520* -0.154* -0.302* 0.109 0.001 0.491* -0.139* -0.247* 0.809* 0.35913 0.001 0.503* -0.120* 0.312* 0.133 0.001 0.501* -0.122* 0.312* 0.030 0.14014 0.001 0.683* -0.117* 0.164* -0.022 0.001 0.633* -0.127* 0.202* 0.384* 0.12315 0.002* 0.369* -0.085 0.085 0.071 0.002* 0.356* -0.092* 0.092 0.489* 0.26216 0.000 0.413* -0.064 0.718* 0.187 0.000 0.423* -0.068 0.713* -0.071 0.17517 0.003* 0.713* -0.074 0.627* 0.072 0.002* 0.643* -0.108 0.707* 0.602* 0.24118 0.002* 0.414* -0.112* 0.482* 0.142 0.002* 0.371* -0.177* 0.584* 0.737* 0.40819 0.001 0.668* -0.591* -0.067 0.316 0.001 0.653* -0.598* -0.070 0.119 0.34020 0.001* 0.593* -0.582* -0.171* 0.341 0.001 0.564* -0.605* -0.158* 0.169 0.39121 0.001 0.743* -0.598* 0.007 0.298 0.001 0.706* -0.609* 0.010 0.374* 0.39222 0.001 0.670* -0.911* 0.035 0.498 0.001 0.661* -0.905* 0.037 0.108* 0.51023 0.002* 0.693* -0.458* 0.051 0.270 0.001 0.664* -0.475* 0.118 0.324 0.36824 0.002* 0.740* -0.676* 0.207* 0.420 0.002 0.682* -0.771* 0.190* 0.577* 0.56625 0.001 0.609* -0.617* 0.711* 0.373 0.001 0.608* -0.617* 0.711* 0.028* 0.37526 0.001 0.627* -0.639* 0.611* 0.357 0.001 0.579* -0.646* 0.689* 0.452* 0.43927 0.001 0.679* -1.120* 0.785* 0.478 0.000 0.505* -1.087* 1.113* 1.562* 0.699

and non-financial stock only portfolios) for the aggregate, non-crisis, and crisis periods, respectively. For the combined stock portfolios, the partial F statistic is significant in 22 out of 27 portfolios during the aggregate period. During the non-crisis period, the number of cases of significant partial F statistics is reduced to 18. Similar results can be seen for financial and non-financial stock portfolios during the aggregate and the non-crisis period. However, the effect

of LEV is predominantly high during the crisis period with significant partial F statistics in 27 cases for combined and non-financial stock portfolios, and in 26 cases for financial stocks only portfolios. This supports the evidence presented earlier suggesting that compared to HML, LEV incorporates additional and unique information concerning distress risk exposure of the firms. In particular, the effect of LEV is particularly dominant during the crisis period.

Bhatt and Sultan


Table 11. Factor loadings for financial and non-financial firms for non-crisis period (January 2000 to June 2007).

Model 1 Model 2

r r = + r r R R

| ,it ft mt ft t SMB t HML t

t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~N( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W


Portfolio

Non-crisis Period

Model 1 Model 2



R-square

1 0.002* 0.757* 0.110 -0.172* 0.168 0.002* 0.784* 0.110 -0.223* -0.269* 0.1702 0.001 0.547* 0.922* -0.542* 0.143 0.001 0.521* 0.897* -0.496* 0.375* 0.1553 0.003* 0.873* 0.365* -0.325* 0.265 0.002 0.867* 0.411* -0.283* 0.673* 0.3174 0.002 0.586* 0.311* 0.296* 0.262 0.002 0.589* 0.316* 0.294* -0.035 0.2605 0.000 0.508* 0.167* 0.171* 0.182 0.000 0.499* 0.160* 0.193* 0.186* 0.1856 0.003* 0.497* 0.145* 0.208* 0.236 0.003* 0.466* 0.136* 0.208* 0.296* 0.2577 0.000 0.643* 0.362* 0.800* 0.342 -0.001 0.683* 0.443* 0.826* -0.408* 0.3338 0.001 0.842* 0.354* 0.678* 0.265 0.001 0.834* 0.312* 0.693* 0.359* 0.2749 0.002* 0.792* 0.547* 0.759* 0.307 0.002 0.753* 0.513* 0.786* 0.477* 0.34710 0.002* 0.592* -0.252* -0.282* 0.301 0.002* 0.594* -0.251* -0.286* -0.038 0.30011 0.001 0.600* -0.104* -0.143* 0.224 0.001 0.589* -0.121* -0.143* 0.126 0.23212 0.002* 0.590* -0.056 -0.311* 0.168 0.002 0.562* -0.055 -0.276* 0.411* 0.21513 0.001 0.551* -0.054 0.295* 0.299 0.001 0.557* -0.038 0.293* -0.153 0.30114 0.002* 0.444* -0.095* 0.137* 0.280 0.001 0.732* -0.045 0.149* 0.277* 0.19815 0.003* 0.362* -0.033 0.030 0.158 0.002* 0.370* -0.035 0.052 0.335* 0.19216 0.001 0.524* 0.063 0.702* 0.316 0.001 0.553* 0.065 0.666* -0.480* 0.35117 0.003* 0.778* 0.043 0.588* 0.233 0.003* 0.768* 0.020 0.624* 0.377* 0.24918 0.002* 0.417* -0.017 0.419* 0.242 0.002* 0.405* -0.020 0.452* 0.417* 0.28319 0.001 0.727* -0.493* -0.040 0.452 0.001 0.731* -0.483* -0.044 -0.096 0.44920 0.001* 0.663* -0.454* -0.224* 0.504 0.001* 0.663* -0.454* -0.224* 0.001 0.50321 0.001 0.776* -0.512* 0.003 0.522 0.001 0.765* -0.535* 0.017 0.209* 0.53722 0.001 0.745* -0.795* 0.061 0.518 0.002 0.755* -0.781* 0.056 -0.110 0.51623 0.002* 0.747* -0.367* -0.008 0.376 0.002* 0.749* -0.366* 0.009 0.088 0.37824 0.002 0.789* -0.523* 0.235* 0.442 0.002 0.773* -0.544* 0.234* 0.212* 0.46125 0.001 0.655* -0.380* 0.651* 0.236 0.001 0.668* -0.360* 0.602* -0.436* 0.25326 0.002 0.728* -0.472* 0.497* 0.272 0.002 0.726* -0.472* 0.525* 0.184 0.26827 0.002 0.780* -0.850* 0.655* 0.347 0.001 0.712* -0.862* 0.956* 1.241* 0.497

Factor loadings by types of firmsPreviously, we noted that LEV is a good proxy for distress risk across financial and non-financial stocks. We find that stocks have similar sensitivities to the leverage risk factor. In this section, we further conduct an additional robustness check to examine whether there are differences in the way various categories of firms respond to the economy wide risk factors because they are classified by stock

exchanges as meeting desired criteria for various style of investing. For instance, the Dow Jones classifies investing in certain stocks (popular household names) under broad categories such as socially responsible investing because these firms promote social, environmental, and corporate responsibility. To this extent, we consider conventional, Islamic and Socially Responsible Investing (SRI) stocks, where each group exhibits distinct characteristics27. There



Table 12. Factor loadings for financial and non-financial firms for the crisis period (July 2007 to April 2009).

Model 1 Model 2

r r = + r r R R

| N ,it ft mt ft t SMB t HML t

t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W


Portfolio

Crisis period

Model 1 Model 2



R-square

1 -0.004 -0.126 -0.577* -0.168 0.130 -0.003 -0.154 -0.283 0.051 0.529 0.2032 -0.003 -0.051 -0.644* -0.708* 0.137 0.001 -0.104 -0.002 -0.187 1.207* 0.3573 -0.005 0.596* -1.159* -1.536* 0.348 -0.002 0.266* -0.263* -0.952* 1.642* 0.6534 -0.003 0.101 -0.197 0.265 0.069 -0.003 0.048 -0.065 0.336 0.257 0.1025 -0.005 0.332* -0.794* -0.222 0.217 -0.003 0.074 -0.140 0.184 1.169* 0.4836 -0.005 0.390* -0.636* -0.394 0.136 -0.003 0.201 0.227 0.131 1.551* 0.4087 -0.005 0.092 -0.115 0.554* 0.059 -0.004 0.069 -0.027 0.619* 0.167 0.0858 -0.006* -0.034 -0.553* 0.103 0.144 -0.004 0.284* -0.184 0.241 0.813* 0.2519 -0.003 0.252 -0.724* 0.190 0.292 0.001 0.178 0.094 0.978* 1.691* 0.54210 -0.005 0.344* -0.998* -0.588 0.361 -0.003 0.238* -0.695* -0.404 0.637* 0.44911 -0.004 0.261* -1.037* -0.531 0.398 -0.003 0.140 -0.664* -0.276 0.733* 0.50612 -0.006 0.257* -1.394* -1.299* 0.499 -0.004 0.133 -0.611* -0.809* 1.446* 0.69613 -0.004 0.239* -0.568* -0.145 0.238 -0.004 0.193* -0.434* -0.083 0.228 0.27714 -0.003 0.266* -1.062* -0.332 0.350 -0.002 0.152 -0.544* 0.050 0.843* 0.46615 -0.009* 0.326* -1.256* -0.687 0.383 -0.005 0.103 -0.422* -0.132 1.616* 0.60316 -0.002 0.143 -0.338* 0.393* 0.231 -0.002 0.054 -0.161 0.442* 0.268* 0.26517 -0.004 0.203* -1.145* -0.110 0.423 -0.002 0.140 -0.600* 0.421 1.005* 0.53918 -0.007* 0.304* -1.314* 0.109 0.488 -0.006 0.087 -0.596* 0.519* 1.275* 0.67119 -0.001 0.308* -1.277* -0.542 0.469 0.000 0.227 -0.869* -0.225 0.758* 0.54320 -0.001 0.207 -1.383* -0.437 0.552 0.000 0.107 -0.913* -0.011 0.917* 0.64421 -0.006 0.264 -1.458* -1.107* 0.473 -0.003 0.118 -0.837* -0.497* 1.275* 0.62622 -0.003 0.292* -1.459* -0.528 0.614 -0.002 0.178* -1.127* -0.292 0.615* 0.66123 -0.004 0.197* -1.537* -0.452 0.625 -0.002 0.107 -0.881* 0.056 1.182* 0.73624 -0.005 0.154 -1.729* 0.039 0.704 -0.001 0.106 -1.108* 0.268 1.275* 0.80925 -0.006 0.235 -1.570* 0.070 0.647 -0.005 0.130 -1.146* 0.462 0.725* 0.67126 -0.001 0.034 -1.683* 0.908* 0.695 0.002 -0.052 -0.970* 1.247* 1.255* 0.77027 -0.008* 0.393* -2.280* 0.121 0.720 -0.005 0.161 -1.374* 0.854* 1.686* 0.826

are distinct differences among these groups with respect to the fundamentals such as size, ROA, ROE, leverage, return on capital, PE ratio, and EPS. (See Milly and Sultan (2009) for further evidences.)

We use stocks included in the Dow Jones Islamic Index (DJIM) which is a proprietary index of stocks classified as Islamic stocks by the Dow Jones Shariah Board. Because of proprietary nature of such classifications, the names of the

stocks are withheld though some of the common household names in the US may be classified as Islamic stocks because they meet the requirements set by the Dow Jones Shariah Board. On October 29, 2010, the market capitalization of the Dow Jones Islamic World Index was $20 billion with 2,369 stocks. The weights (%) for some of the major countries in the index are as follows: US (50.54), UK (6.71), Japan (5.42), Canada (5.27), Switzerland (3.45), Australia (3.26), France (2.97), India (2.5), Taiwan (2.2),

Bhatt and Sultan


Table 13. Factor loadings for the non-financial stock portfolios for aggregate period (January 2000 to April 2009).

Model 1 Model 2

r r = + r r R R


t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W


Portfolio

Aggregate Period

Model 1 Model 2



R-square

1 0.002* 0.776* 0.453* -0.425* 0.076 0.002 0.774* 0.468* -0.395* -0.196 0.0572 0.002* 0.646* 0.380* -0.097 0.011 0.002 0.630* 0.391* -0.246* 0.679* 0.0953 0.000 0.698* 0.614* 0.048 0.041 0.000 0.686* 0.548* -0.185* 1.123* 0.1664 0.001 0.584* 0.345* 0.215* 0.023 0.001 0.587* 0.342* 0.206* 0.043 0.0235 0.000 0.668* 0.533* 0.376* 0.048 0.000 0.677* 0.521* 0.336* 0.211* 0.0716 0.000 0.775* 0.570* 0.586* 0.014 0.000 0.779* 0.554* 0.551* 0.181 0.0317 0.000 0.621* 0.594* 0.690* 0.109 0.000 0.626* 0.613* 0.708* -0.101 0.0948 0.000 0.741* 0.729* 0.787* 0.075 0.000 0.748* 0.723* 0.752* 0.136 0.0839 0.000 0.838* 0.812* 0.893* 0.103 0.000 0.871* 0.818* 0.767* 0.454* 0.13010 -0.001 0.958* 0.098 -0.309* -0.054 -0.001 0.943* 0.112 -0.276* -0.161 -0.06011 0.001 0.709* 0.094 -0.144 0.025 0.001 0.727* 0.084 -0.197* 0.238* 0.04712 0.000 0.704* 0.021 -0.094 0.004 0.000 0.740* -0.007 -0.224* 0.558* 0.06413 0.000 0.682* 0.092 0.281* -0.016 0.000 0.696* 0.081 0.246* 0.189 -0.00714 0.001 0.629* 0.026 0.287* 0.036 0.001 0.673* -0.019 0.185* 0.463* 0.08815 0.000 0.665* 0.168* 0.474* 0.032 0.000 0.692* 0.125* 0.358* 0.531* 0.10016 0.000 0.686* 0.341* 0.790* 0.056 0.000 0.687* 0.331* 0.757* 0.137 0.06517 0.000 0.723* 0.270* 0.641* 0.051 0.000 0.753* 0.210* 0.528* 0.498* 0.11018 0.001 0.848* 0.253* 0.796* 0.033 0.000 0.856* 0.182* 0.699* 0.809* 0.14019 0.001 0.745* -0.459* -0.372* 0.194 0.001 0.743* -0.454* -0.347* -0.166 0.18720 0.000 0.638* -0.372* -0.008 0.054 0.000 0.651* -0.392* -0.053 0.369* 0.09821 0.000 0.589* -0.353* 0.000 0.102 0.000 0.604* -0.364* -0.131 0.628* 0.18522 0.001 0.685* -0.394* 0.224* 0.043 0.001 0.685* -0.396* 0.210* 0.049 0.04323 0.001 0.672* -0.243* 0.257* 0.031 0.001 0.682* -0.301* 0.178* 0.493* 0.09424 0.001 0.647* -0.165* 0.438* -0.003 0.000 0.659* -0.195* 0.312* 0.504* 0.06725 0.001 0.666* -1.074* 1.002* 0.236 0.001 0.672* -1.089* 1.105* -0.505* 0.23626 0.001 0.772* -0.187* 0.701* 0.054 0.000 0.767* -0.233* 0.585* 0.814* 0.14027 0.002 0.791* -0.408* 0.796* 0.125 0.001 0.794* -0.443* 0.589* 1.068* 0.241

Germany (1.73), South Korea (1.56), Brazil (1.5), Russia (1.47), China (1.39), Hong Kong (1.25), and Sweden (1.09). Among some of the traditionally Muslim majority countries, the weights are: Malaysia (.35), Kuwait (.22), Qatar (.08), UAE (.03), and Bahrain (.01).

Our selection of SC stocks is in line with the recent interest in the performance of faith based investing, with its overarching goal to promote the betterment of society,

relative to conventional investment strategies, which lack such ethical ambition. SC stocks are popular among a new class of investors that, in addition to profit motives, is also driven by their desire to live ethically and invest morally. Compared to the conventional Western financial system, Islamic finance is a newcomer to the global financial world, encompassing somewhere between $750 billion to $1 trillion of investments in firms and projects that are classified as SC. Yet, over the past few



years Islamic investments have become more competitive and consequently attractive not only to Muslim but also non-Muslim investors seeking alternative investments opportunities, which live up to high ethical as well as nominal performance standards. As a result, the number of Islamic mutual funds and exchange traded funds world-wide has increased considerably from merely 8 before 1992 to more than 300 in 2008, with an estimated

market capitalization of $300 billion and numerous traditional US financial institutions joining to partake in this development.

Similarly, the SRI class of stocks is a relative newcomer, which has gained popularity in recent years. In the early 2000s, we have seen a dramatic interest in socially responsible investing that poured billions of dollars

Table 14. Factor loadings for non-financial stocks portfolio for non-crisis period (January 2000 to June 2007)

Model 1 Model 2

r r = + r r R R


t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W


Portfolio

Non-crisis Period

Model 1 Model 2



R-square

1 0.002 0.846* 0.462* -0.390* 0.483 0.002 0.830* 0.482* -0.321* -0.364* 0.4962 0.002* 0.732* 0.415* -0.098 0.238 0.002* 0.741* 0.407* -0.164 0.360* 0.2513 0.000 0.782* 0.627* 0.115 0.347 0.000 0.821* 0.599* -0.126 0.923* 0.3714 0.001 0.673* 0.363* 0.268* 0.406 0.001 0.666* 0.366* 0.278* -0.057 0.4085 0.000 0.718* 0.518* 0.415* 0.451 0.000 0.726* 0.513* 0.394* 0.109 0.4506 0.000 0.825* 0.573* 0.613* 0.453 0.000 0.835* 0.561* 0.590* 0.130 0.4517 -0.001 0.758* 0.638* 0.673* 0.551 -0.001 0.750* 0.647* 0.719* -0.147 0.5548 0.000 0.789* 0.729* 0.815* 0.527 0.000 0.795* 0.726* 0.791* 0.092 0.5259 0.000 0.903* 0.807* 0.953* 0.471 0.000 0.939* 0.800* 0.862* 0.385* 0.46910 -0.001 1.023* 0.121 -0.323* 0.465 -0.001 0.982* 0.125 -0.243* -0.292* 0.48011 0.001 0.775* 0.057 -0.095 0.458 0.001 0.790* 0.051 -0.127 0.150 0.45712 0.000 0.764* 0.044 -0.038 0.357 0.000 0.812* 0.026 -0.125 0.401* 0.36513 0.000 0.756* 0.140* 0.277* 0.401 0.000 0.766* 0.136* 0.255* 0.111 0.39614 0.001 0.689* 0.026 0.328* 0.411 0.001 0.731* 0.000 0.257* 0.366* 0.41915 0.000 0.711* 0.160* 0.499* 0.444 0.000 0.751* 0.132* 0.408* 0.440* 0.45816 0.000 0.759* 0.368* 0.798* 0.475 0.000 0.761* 0.366* 0.791* 0.031 0.47317 0.000 0.759* 0.280* 0.657* 0.495 0.000 0.794* 0.240* 0.570* 0.394* 0.50218 0.001 0.898* 0.249* 0.801* 0.447 0.001 0.932* 0.192* 0.726* 0.710* 0.46119 0.000 0.849* -0.432* -0.375* 0.549 0.000 0.912* -0.380* -0.428* -0.402* 0.56520 0.000 0.664* -0.367* 0.013 0.475 0.000 0.676* -0.385* -0.026 0.312* 0.48621 0.000 0.639* -0.323* -0.002 0.414 -0.001 0.769* -0.400* -0.109 0.649* 0.41822 0.001 0.801* -0.370* 0.278* 0.403 0.001 0.799* -0.369* 0.290* -0.048 0.40423 0.001 0.727* -0.242* 0.308* 0.460 0.001 0.769* -0.268* 0.242* 0.396* 0.47024 0.000 0.699* -0.173* 0.472* 0.416 0.000 0.731* -0.208* 0.381* 0.422* 0.43525 0.001 0.820* -1.110* 1.026* 0.436 0.001 0.764* -1.118* 1.202* -0.918* 0.49226 0.001 0.927* -0.342* 0.879* 0.358 0.001 0.977* -0.360* 0.763* 0.552* 0.35827 0.001 0.865* -0.388* 0.849* 0.410 0.001 0.890* -0.412* 0.656* 0.885* 0.444

Bhatt and Sultan


Table 15. Factor loadings for non-financial stock portfolios during crisis period (July 2007 to April 2009).

Model 1 Model 2

r r = + r r R R| N ,

it ft mt ft t SMB t HML t

t t

- - + + +

-

β β β β ee ψ

0 1 2 3

1 0( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W


Portfolio

Crisis Period

Model 1 Model 2



R-square

1 -0.005 0.156 0.586 -1.208* -0.092 -0.004* 0.051 0.718* -1.277* 2.808* 0.3912 -0.005 0.124 0.500 -0.967* -0.087 -0.003* 0.044 0.832* -1.075* 2.866* 0.4423 -0.004 0.043 0.392 -0.652 -0.092 -0.002 0.017 0.756* -0.717* 3.192* 0.5354 -0.004 0.062 -0.008 0.459 -0.073 -0.004 0.047 0.377* -0.081 1.710* 0.2925 -0.005 0.086 0.047 0.359 -0.043 -0.005 0.049 0.537* -0.216 2.292* 0.4396 -0.006 0.143 0.427 0.191 -0.030 -0.005 0.053 0.733* -0.074 2.438* 0.4177 -0.004 0.079 0.233 0.579* 0.033 -0.003 −0.019 0.540* 0.307 1.427* 0.3158 -0.004 0.110 0.393 0.305 0.001 -0.003 −0.008 0.716* 0.141 2.412* 0.4719 -0.004 0.151 0.646 0.032 -0.040 -0.002 0.018 0.840* 0.150 3.242* 0.57510 -0.005 -0.004 -0.268 0.055 -0.066 -0.002 0.078 0.212 -0.604* 2.539* 0.41711 -0.006* -0.004 -0.549* 0.112 -0.066 -0.004 0.019 0.055 -0.769* 2.855* 0.47912 -0.005 0.047 -0.535 -0.161 -0.043 -0.003 0.070 -0.087 -0.736* 2.957* 0.50413 -0.004 0.025 -0.410 0.674* -0.073 -0.003 0.092 0.032 -0.116 1.839* 0.33214 -0.005 0.041 -0.471 0.241 -0.042 -0.003 0.034 -0.056 -0.287 2.786* 0.48715 -0.006* 0.064 -0.234 0.221 -0.045 -0.003 0.032 0.147 -0.227 2.963* 0.51516 -0.004 0.033 -0.027 0.539 -0.010 -0.002 −0.008 0.445* 0.150 1.938* 0.35917 -0.005 0.015 -0.527 0.772* -0.022 -0.002 −0.014 -0.015 0.161 3.016* 0.53318 -0.006 -0.004 -0.415 0.834* 0.004 -0.003 0.090 0.104 0.035 4.151* 0.61219 -0.004 0.177 -0.643* -0.673* -0.020 -0.002 0.054 -0.451* -0.773* 2.140* 0.44420 -0.004 0.034 -0.667* -0.626* -0.020 -0.003 −0.012 -0.255 -0.933* 2.589* 0.44521 -0.005* 0.114 -0.818* -0.241 -0.009 -0.003 0.058 -0.427* -0.616* 2.331* 0.48722 -0.003 0.118 -0.640* -0.076 -0.070 -0.003 −0.021 -0.302 -0.430 1.827* 0.23923 -0.004 0.098 -0.543 -0.309 -0.040 -0.003 0.000 -0.276 -0.420 2.608* 0.44824 -0.005 0.101 -0.468 0.036 -0.044 -0.003 0.017 -0.119 -0.201 2.566* 0.45325 -0.003 -0.033 -0.337 0.756* 0.039 -0.003 −0.013 -0.085 0.351 1.865* 0.33126 -0.006* 0.171 -1.072* 0.708* 0.010 -0.004 0.141 -0.379* 0.093 2.762* 0.48827 -0.005 -0.018 -0.622 -0.079 -0.024 -0.003 0.002 -0.546* 0.192 3.637* 0.609

into companies known for their efforts to offer ethical investments and projects that promoted environmental sustainability. In terms of the portfolio allocation and structure, Islamic and socially responsible investing (SRI) stocks exhibit strong similarities, whereas conventional stocks are not subject to any other qualitative or quantitative constraints. Although SRI funds were initially

conceived in a religious context as well, socially responsible investing has expanded to take in consideration “the so-called ‘triple bottom line’, commonly known as the ‘three P’s rule: people, planet and profit’” (Forte & Miglietta, 2007, p. 3). Most recently, assets under SRI management were estimated to have increased “from $639 billion in 1995 … to $2.71 trillion in 2007”, while “assets in all types



Table 16. Partial f-statistics testing for the significance of contribution made by the LEV factor.

Restricted Model Unrestricted Model

r r = + r r R R


t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W

where ri is the return on portfolio i; rf is the return on the risk free asset and rm is the return on the market portfolio. RSMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. RHML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. RLEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”. Partial f-statistics and the p-values test for the significance in the contribution of R-square made by the new model (which includes the LEV factor). The factors SMB, HML and LEV have been rebalanced for financial stock portfolios and non financial stock portfolios. (*) indicates significance at 5% level of significance. GARCH models are estimated using the Bollerslev-Wooldridge corrections to the standard errors. Model 1 excludes LEV. Model 2 includes LEV. Coefficients of the GARCH variance equations are not reported to conserve space. They are available upon request.

Portfolio

All stock portfolios Financial stock portfolios Non-financial stock portfolios

Aggregate period

Non-crisis period

Crisis period

Aggregate period

Non-crisis period

Crisis period

Aggregate period

Non-crisis period

Crisis period

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

1 -1.62 1.04 71.90* -0.17 2.04 9.64* -8.36 10.65* 75.57*2 56.73* 5.39* 82.82* 115.83* 6.41* 33.22* 45.46* 7.58* 90.27*3 101.36* 28.68* 116.58* 214.21* 30.94* 83.68* 72.96* 16.11* 127.92*4 14.96* -0.01 45.12* 11.11* -0.07 4.60* 1.26 1.73 49.54*5 34.46* 7.13* 75.91* 94.36* 2.16 49.55* 12.75* 0.13 81.80*6 43.57* 13.12* 70.42* 101.84* 11.67* 44.31* 9.16* -0.18 73.26*7 -15.57 0.34 36.19* -13.06 -3.84 3.66 -6.98 2.99 39.88*8 19.58* 2.22 75.98* 95.79* 5.93* 14.53* 5.07* -0.88 84.74*9 35.62* 4.19* 103.75* 178.07* 24.36* 52.41* 15.69* -0.56 137.36*10 6.96* -0.11 76.02* 19.93* 0.39 16.06* -1.71 12.03* 78.90*11 37.50* 8.58* 87.33* 78.43* 4.73* 21.61* 12.12* 0.92 99.36*12 66.64* 23.42* 89.13* 189.67* 24.14* 62.10* 31.73* 5.81* 104.67*13 25.01* 2.55 50.96* 4.82* 2.34 6.09* 5.04* -1.87 58.03*14 49.24* 12.04* 89.40* 80.92* -38.57 21.53* 28.36* 6.12* 97.85*15 68.41* 26.13* 93.21* 126.42* 17.04* 53.15* 37.70* 10.87* 109.78*16 3.29 0.71 36.76* -6.22 22.03* 5.42* 5.44* -0.52 55.13*17 46.24* 12.02* 105.33* 108.33* 9.49* 24.76* 33.00* 6.07* 112.73*18 80.45* 21.10* 108.63* 217.60* 23.24* 53.18* 61.30* 11.21* 148.07*19 -4.56* 9.18* 68.14* 18.88* -0.99 16.25* -3.45 15.24* 79.61*20 31.43* 8.18* 67.16* 41.14* 0.03 25.47* 24.38* 9.57* 79.64*21 49.63* 18.60* 83.87* 75.18* 13.06* 39.69* 50.31* 3.95* 91.88*22 22.87* 2.70 39.97* 12.65* -0.83 13.91* 1.05 1.35 39.27*23 49.86* 21.54* 72.94* 75.48* 2.52 40.51* 34.80* 8.93* 84.35*24 48.31* 21.56* 78.73* 162.88* 15.23* 52.25* 37.14* 13.95* 86.59*25 -12.15* 28.19* 47.55* 2.89* 9.85* 7.90* 0.62 43.70* 42.22*26 50.41* 1.07 87.66* 72.14* -0.97 31.74* 49.04* 1.20 88.80*27 118.71* 44.78* 131.34* 355.94* 116.64* 58.15* 74.25* 25.09* 153.14*

of socially and environmentally screened funds [… in the US] rose to $201.8 billion.” (2007 Report on Socially Responsible Investing Trends in the United States, 2008, p. ii) The premise of the “three P’s rule” is reflected in a definition of socially responsible investing, which can be found in the 2005 Report on SRI Trends in the United States released by the Social Investment Forum:

Socially responsible investing (SRI) is an investment process that considers the social and environmental consequences of investments, both positive and negative, within the context of rigorous financial analysis… It is a process of identifying and investing in companies that meet certain standards of Corporate Social Responsibility (CSR)

Bhatt and Sultan


(2004 Report on Socially Responsible Investing Trends in the United States. 10 Year Review, 2005, p. 2).

The congruence of Islamic and SRI stocks stems from the fact that both do not have profit maximization as their sole objective, but rather strive to achieve a paramount, ethical obligation and a social-utilitarian function. In the case of Islamic funds, the religious responsibilities and regulations outlined in the Shariah, take precedence over profit in order to further the establishment of a just and moral Islamic economic system and ultimately society.

In contrast, profit maximization is the dominant objective in traditional fund management. Conventional equity portfolio strategies include neither positive nor negative screens, whose purpose it is to align the portfolio with certain ethical, qualitative standards. As such, conventional funds are not subject to the qualitative screening procedures that are so imperative to Islamic and SRI funds. Additionally, Islamic funds differ from SRI and conventional ones, since their provisions incorporate quantitative screens that are directly based on ethical paradigms found in the Shariah. Furthermore, Islamic funds have to comply with certain income purification requirements, which are derived from the teachings of the Holy Quran and Sunnah.

The hypothesis tested is that high leverage increases exposure to the credit market and subsequently translates into shareholders demanding higher risk premium. Recall that Islamic stocks have low leverage, they are significantly more asset-backed than conventional firms, and are not involved in the business of speculation, production of weapons, alcohol, pork, and entertainment. More specifically, Islamic funds typically screen out companies with excessive reliance on debt, where the typical maximum level of total debt to market capitalization is set at 33 percent28.

The first step towards applying our leverage risk factor to these index classifications is to recreate the FF and LEV specific to each category of stocks. This is followed by estimating GARCH regressions at the firm and portfolio level.

Factor loadings of conventional stocksIn Table 17, we report a summary of firm and portfolio specific regressions by groups. The first panel reports the results for the firms belonging to the conventional stock category. We find that at the firm level, the inclusion of LEV produced some interesting results. Compared to the aggregate period, the number of instances where the XMKT is significant drops by 79.57% at the firm and by 100% at the portfolio level. The change in significance for SMB is as follows: 8.89% at the firm and −17.39% at the portfolio level. The results for the HML are consistent across both the firm and the portfolio level. The number of instances where HML is significant at the firm level drops by 56.13% and by 44.44% at the portfolio level. Finally, the number of instances LEV is significant increases by 231% at the firm and by 58.82% at the portfolio level. Overall, these results are qualitatively similar to the ones reported earlier and confirm our earlier finding that the inclusion of LEV subsumes the effects of the traditional FF factors to a great extent.

Factor loadings of Islamic stocksPanel B reports the results for the Islamic group of stocks. Compared to the aggregate period, there is a remarkable change in the number of cases of where XMKT is significant (−87.6% at the firm and by −92.59% at the portfolio level). The change in significance for SMB is as follows: 9.89% at the firm and −44.44% at the portfolio level. The results for the HML are again consistent across both the firm and the portfolio level. The change in statistical significance for HML is as follows: −56.44% at the firm and by −60% at the portfolio level. Finally, the number of instances LEV is significant increases by 98.9% at the firm and by 73.33% at the portfolio level. Again, our results are quite consistent with the previous results reported without the index classifications. Islamic stocks behave similar to the conventional stocks when it comes to sensitivities to economic risk factors.

Factor loadings of SRI stocksIn Panel C, we report the results for 238 stocks classified as SRI group of stocks. Compared to the previous groups, we have some unusual results. We find that, compared to the aggregate period, the number of instances where the XMKT is significant drops by 89% at the firm and by 100% at the portfolio level. For SMB, the changes in the number of significant cases are: −66% (firm-level) and −80.95% (portfolio-level). In contrast to our previous results, the number of instances where HML is significant at the firm level increases by 56.43% and by 35% at the portfolio level. Finally, the number of instances where LEV is significant drops by 11.11% the firm and by 56.25% at the portfolio level.

With respect to the effects of LEV risk factor, the results for the SRI group are quite different from the Conventional and Islamic stocks, suggesting that stocks in this category are less sensitive to the economy-wide leverage risk factor. Certainly, leverage risk for this type of firms is not unusually different but perhaps the nature of the business these firms are involved may make it less susceptive to economy wide leverage risk. It may also be possible that during the financial crisis, while socially responsible investing would have earned positive risk premium with respect the HML, SRI investors would have earned a negative risk premium when leverage was employed as a stock picking strategy. Whether SRI investing produces a lower return because these stocks are generally less sensitive to the economy wide risk factors suggests that these stocks may offer significant diversification benefits. Overall, further research along these lines would offer more clues as to why SRI stocks have negative risk premium for leverage risk.

Partial F-testTable 18 reports partial F-statistics (across all the three groups - all stock portfolios), for the aggregate, non-crisis, and crisis periods in order to test for the significance of the contribution made by LEV. For the combined stock portfolios, the partial F statistic is significant in 24 out of 27 portfolios during the aggregate period. During the non-crisis period, the number of significant partial F statistics is reduced to 17 cases. However, the effect of LEV is prominent during the crisis period with significant partial F statistics in all 27 portfolios. For the Islamic stocks, the



Tabl

e 17

. Fa

ctor

load

ings

by

type

s of

firm

s.

M

odel

1

Mod

el 2

rr

=

+

rr

RR

|N

,it

ftm

tft

tSM

Bt

HM

Lt

tt

--

++

+

-

ββ

ββ

e

eψ

01

23

10

()

,,

~(

σσ

σa

eδ

σ

t

ti

iq

ti

ijp

tj

2

2

1

2

1

),

=+

+=

-=

-∑

∑W

rr

=

+

rr

RR

R

|it

ftm

tft

tSM

Bt

HM

Lt

LEV

t

t

--

++

++

ββ

ββ

βe

e0

12

34

()

,,

,

ψψσ

σa

eδ

σ

tt

ti

iq

ti

ijp

tj

N,

-

=-

=-

=+

+∑

∑1

2

2

1

2

1

0~

(),

W

whe

re, r

i is t

he re

turn

on

port

folio

i; r f i

s the

retu

rn o

n th

e ri

sk fr

ee a

sset

and

r m is

the

retu

rn o

n th

e m

arke

t por

tfol

io. R

SMB i

s the

retu

rn o

n th

e si

ze m

imic

king

por

tfol

io c

onst

ruct

ed b

y ta

king

the

sim

ple

aver

age

of th

e re

turn

s ea

ch w

eek

of a

ll “s

mal

l” p

ortf

olio

s m

inus

“bi

g” p

ortf

olio

s. R

HM

L is

the

retu

rn o

n bo

ok to

mar

ket m

imic

king

por

tfol

io c

onst

ruct

ed b

y ta

king

the

sim

ple

aver

age

of

the

retu

rns e

ach

wee

k of

all

“hig

h BE

/ME”

por

tfol

ios m

inus

“low

BE/

ME”

por

tfol

ios.

RLE

V is t

he re

turn

on

leve

rage

mim

icki

ng p

ortf

olio

s con

stru

cted

by

taki

ng th

e si

mpl

e av

erag

e of

the

retu

rns

each

wee

k of

all

“hig

h le

vera

ge”

port

folio

s m

inus

“lo

w le

vera

ge p

ortf

olio

s”. A

ll in

dica

ted

coeffi

cien

ts w

ith

(*)

are

sign

ifica

nt a

t 5%

leve

l of s

igni

fican

ce.

PAN

EL A

Con

vent

iona

l sto

cks:

230

8C

onve

ntio

nal p

ortf

olio

s: 2

7

Agg

rega

te P

erio

dA

ggre

gate

Per

iod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

1757

810

23M

odel

1Po

siti

ve27

418

Neg

ativ

e0

1439

225

N

egat

ive

017

0

Tota

l17

5714

4712

48

Tota

l27

2118

Mod

el 2

Posi

tive

1697

879

712

08M

odel

2Po

siti

ve27

212

20

Neg

ativ

e0

1582

482

84N

egat

ive

018

30

Tota

l16

9715

9012

7912

92To

tal

2720

1520

%ch

ange

in s

igni

fican

ce

(by

mod

el)

-3.4

1%9.

88%

2.48

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

-4.7

6%-1

6.67

%

Non

-cri

sis

Peri

odN

on-c

risi

s Pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

2076

791

70

Mod

el 1

Posi

tive

277

17

Neg

ativ

e0

1108

378

N

egat

ive

014

0

Tota

l20

7611

1512

95

Tota

l27

2117

Mod

el 2

Posi

tive

2056

686

345

9M

odel

2Po

siti

ve27

715

15

Neg

ativ

e0

1152

499

80N

egat

ive

016

32

Tota

l20

5611

5813

6253

9To

tal

2723

1817

%ch

ange

in s

igni

fican

ce

(by

mod

el)

-0.9

6%3.

86%

5.17

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

9.52

%5.

88%

(Con

tinu

ed)

Bhatt and Sultan


Cri

sis

peri

odC

risi

s pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

290

262

1026

0M

odel

1Po

siti

ve10

027

Neg

ativ

e76

996

12

Neg

ativ

e0

160

Tota

l36

612

5810

38

Tota

l10

1627

Mod

el 2

Posi

tive

9252

548

017

44M

odel

2Po

siti

ve0

59

27

Neg

ativ

e32

873

611

541

Neg

ativ

e0

141

0

Tota

l42

012

6159

517

85To

tal

019

1027

%ch

ange

in s

igni

fican

ce

(by

mod

el)

14.7

5%0.

24%

-42.

68%

%ch

ange

in s

igni

fican

ce

(by

mod

el)

0.00

%18

.75%

-62.

96%

%ch

ange

in s

igni

fican

ce

(by

peri

od)

−79.

57%

8.89

%−5

6.31

%23

1.17

%%

chan

ge in

sig

nific

ance

(b

y pe

riod

)−1

00.0

0%−1

7.39

%−4

4.44

%58

.82%

PAN

EL B

Isla

mic

sto

cks:

116

1Is

lam

ic p

ortf

olio

s: 1

161

Agg

rega

te

Peri

odA

ggre

gate

Per

iod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

1004

402

252

Mod

el 1

Posi

tive

2712

12

Neg

ativ

e0

214

196

N

egat

ive

05

8

Tota

l10

0461

644

8

Tota

l27

1720

Mod

el 2

Posi

tive

1019

401

248

388

Mod

el 2

Posi

tive

2712

1210

Neg

ativ

e0

244

239

202

Neg

ativ

e0

79

4

Tota

l10

1964

548

759

0To

tal

2719

2114

%ch

ange

in s

igni

fican

ce

(by

mod

el)

1.49

%4.

71%

8.71

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

11.7

6%5.

00%

Tabl

e 17

. (C

onti

nued

)



Non

-cri

sis

Peri

odN

on-c

risi

s Pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

1087

236

259

Mod

el 1

Posi

tive

2710

13

Neg

ativ

e0

276

234

N

egat

ive

07

8

Tota

l10

8751

249

3

Tota

l27

1721

Mod

el 2

Posi

tive

1089

254

265

178

Mod

el 2

Posi

tive

2711

124

Neg

ativ

e0

254

263

280

Neg

ativ

e0

78

11

Tota

l10

8950

852

845

8To

tal

2718

2015

%ch

ange

in s

igni

fican

ce

(by

mod

el)

0.18

%-0

.78%

7.10

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

5.88

%-4

.76%

Cri

sis

peri

odC

risi

s pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

8350

819

1M

odel

1Po

siti

ve0

07

Neg

ativ

e33

4864

N

egat

ive

011

3

Tota

l11

655

625

5

Tota

l0

1110

Mod

el 2

Posi

tive

9248

913

387

3M

odel

2Po

siti

ve2

106

26

Neg

ativ

e43

6997

38N

egat

ive

00

20

Tota

l13

555

823

091

1To

tal

210

826

%ch

ange

in s

igni

fican

ce

(by

mod

el)

16.3

8%0.

36%

-9.8

0%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

-9.0

9%-2

0.00

%

%ch

ange

in s

igni

fican

ce

(by

peri

od)

-87.

60%

9.84

%-5

6.44

%98

.91%

%ch

ange

in s

igni

fican

ce

(by

peri

od)

-92.

59%

-44.

44%

-60.

00%

73.3

3%

(Con

tinu

ed)

Bhatt and Sultan


PAN

EL C

SRI s

tock

s: 2

38SR

I Por

tfol

ios:

27

Agg

rega

te P

erio

dA

ggre

gate

Per

iod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

224

8018

3M

odel

1Po

siti

ve27

920

Neg

ativ

e0

690

N

egat

ive

010

1

Tota

l22

414

918

3

Tota

l27

1921

Mod

el 2

Posi

tive

224

7919

112

Mod

el 2

Posi

tive

2711

218

Neg

ativ

e0

690

51N

egat

ive

010

29

Tota

l22

414

819

163

Tota

l27

2123

17

%ch

ange

in s

igni

fican

ce

(by

mod

el)

0.00

%-0

.67%

4.37

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

10.5

3%9.

52%

Non

-cri

sis

Peri

odN

on-c

risi

s Pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

235

7812

40

Mod

el 1

Posi

tive

279

17

Neg

ativ

e0

8211

N

egat

ive

010

3

Tota

l23

516

013

5

Tota

l27

1920

Mod

el 2

Posi

tive

237

7613

318

Mod

el 2

Posi

tive

2711

178

Neg

ativ

e0

837

27N

egat

ive

010

38

Tota

l23

715

914

045

Tota

l27

2120

16

%ch

ange

in s

igni

fican

ce

(by

mod

el)

0.85

%-0

.63%

3.70

%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

10.5

3%0.

00%

Tabl

e 17

. (C

onti

nued

)



Cri

sis

peri

odC

risi

s pe

riod

XM

KT

SMB

HM

LLE

VX

MK

TSM

BH

ML

LEV

Mod

el 1

Posi

tive

836

222

Mod

el 1

Posi

tive

03

27

Neg

ativ

e10

120

N

egat

ive

01

0

Tota

l18

4822

2

Tota

l0

427

Mod

el 2

Posi

tive

1239

219

28M

odel

2Po

siti

ve0

227

7

Neg

ativ

e14

150

12N

egat

ive

02

00

Tota

l26

5421

940

Tota

l0

427

7

%ch

ange

in s

igni

fican

ce

(by

mod

el)

44.4

4%12

.50%

-1.3

5%%

chan

ge in

sig

nific

ance

(b

y m

odel

)0.

00%

0.00

%0.

00%

%ch

ange

in s

igni

fican

ce

(by

peri

od)

-89.

03%

-66.

04%

56.4

3%-1

1.11

%%

chan

ge in

sig

nific

ance

(b

y pe

riod

)-1

00.0

0%-8

0.95

%35

.00%

-56.

25% results are quite strong. The number of cases F-statistics is

significant is 17 (aggregate period), 15 (non-crisis period), and 27 (crisis period). Finally, we find that the SRI stocks only portfolios are not sensitive to the leverage risk during the credit crisis.

Overall, the regression results suggest that while the sensitivities of the portfolio returns to the FF factors are significant during the aggregate and the non-crisis periods, there are important changes in the sign and significance of these factors during the crisis period. Their significance also weakens with the introduction of leverage as a risk factor, almost to the tune of being subsumed by the leverage risk factor. The effects of the market factor are persistent before the crisis period but surprisingly became insignificant during the crisis period. Leverage factor is consistently significant across all the periods and its effect is more prominent during the crisis period due to the greater debt exposure of the firms and higher macroeconomic risk. The results further support the conclusions drawn in the earlier tables.

Leverage risk factor for US stocksA potential shortcoming of the preceding results is due to the fact that our previous samples include stocks traded globally and may not accurately quantify the effects of the credit crisis on the US market. We therefore conduct another experiment using US stocks only. This additional exercise is carried out by excluding all non-US stocks, creating traditional Fama-French factors, and adding our newly created LEV factor to represent financial distress. In addition, we also estimate GARCH regressions to demonstrate that our risk factors represent macroeconomic shocks as well. While the results are not presented to save space, we can summarize the results as follows. We find that for predicting US industrial production, the following variables are statistically significant: SMB, HML, XMKT, HML (t-2), and LEV (t-2). For predicting US unemployment rate, variables such as SMB, HML, LEV and several interaction variables on LEV (for 2008 credit crisis period) are significant at various lag length. We find that several interaction variables using LEV are significant in predicting the credit spread and term spread. Finally, dummy variables on LEV are highly significant in affecting changes in the US inflation rate. Overall, it is safe to conclude that the risk factors contain adequate information on the US economy.

Next, GARCH regressions are estimated. For the aggregate period (Jan2000 – April 2009), out of 27 portfolios, SMB has 18 (6) positive (negative) coefficients. HML has 16 (8) positive (negative) coefficients. In contrast, for Model 2 (where we add LEV), the results are as follows: SMB has 18 positive and 4 negative coefficients, HML has 17 positive and 9 negative coefficients, and LEV has 6 positive and 17 negative coefficients. For the non-crisis period (January 2000 – June 2007), SMB has 18 (7) positive (negative) coefficients. HML has 13 (10) positive (negative) coefficients. In contrast, for Model 2, SMB has 18 positive and 7 negative coefficients, HML has 14 positive and 9 negative coefficients, and LEV has 6 positive and 18 negative coefficients. In all cases, there is a significant increase in R2 when we add LEV in the model.

Bhatt and Sultan


Table 18. Partial f-statistics testing for the significance of contribution made by the LEV factor.

Old Model New Model

r r = + r r R R


t t

- - + + +

-

β β β β e

e ψ0 1 2 3

1 0

( ) , ,

~ ( σσ

σ a e δ σ

t

t ii

q

t i ij

p

t j

2

2

1

2

1

),

= + +=

-=

-∑ ∑W

r r = + r r R R R


t

- - + + + +β β β β β e

e0 1 2 3 4( ) , , ,

ψψ σ

σ a e δ σ

t t

t ii

q

t i ij

p

t j

N ,-

=-

=-= + +∑ ∑

12

2

1

2

1

0~ ( ),

W

where, ri is the return on portfolio i; rf is the return on the risk free asset and rm is the return on the market portfolio. RSMB is the return on the size mimicking portfolio constructed by taking the simple average of the returns each week of all “small” portfolios minus “big” portfolios. RHML is the return on book to market mimicking portfolio constructed by taking the simple average of the returns each week of all “high BE/ME” portfolios minus “low BE/ME” portfolios. RLEV is the return on leverage mimicking portfolios constructed by taking the simple average of the returns each week of all “high leverage” portfolios minus “low leverage portfolios”. All indicated coefficients with (*) are significant at 5% level of significance. Partial f-statistics and the p-values test for the significance in the contribution of R-square made by the new model (which includes the LEV factor). GARCH models are estimated using the Bollerslev-Wooldridge corrections to the standard errors. Model 1 excludes LEV. Model 2 includes LEV. Coefficients of the GARCH variance equations are not reported to conserve space. They are available upon request.

Conventional Portfolios Islamic Portfolios SRI Portfolios

Aggregate period

Non-crisis period

Crisis period

Aggregate period

Non-crisis period

Crisis period

Aggregate period

Non-crisis period

Crisis period

Portfolio Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

Partial f-statistic

1 21.91* -1.27 69.32* -9.88 40.84* 20.16* 43.73* 24.27* 2.562 34.89* 5.64* 83.46* 16.99* -0.38 45.00* 0.86 0.74 -2.263 59.14* 17.21* 119.34* 92.84* 53.73* 51.58* 1.26 37.32* 1.224 4.67* -0.71 25.16* -3.19 6.05* 24.07* 29.93* 18.52* 0.125 8.03* -0.21 51.10* 7.80* 0.02 20.48* -1.26 0.02 -1.376 28.87* 4.78* 95.20* 16.85* -0.50 43.11* -3.29 20.11* -2.127 6.48* 0.03 21.11* -5.00 13.93* 17.95* 39.16* 22.08* 4.96*8 12.42* 1.32 80.78* -10.27 2.37 36.12* 3.40 0.46 -1.589 73.34* 45.25* 110.37* 11.15* 0.00 102.05* -16.20 16.59* 6.18*10 39.51* 0.63 71.80* -8.71 21.96* 38.52* 17.58* 8.50* 0.0811 37.85* 7.25* 74.64* -3.90 5.45* 34.93* 2.31 0.13 -0.4812 107.88* 39.23* 96.55* 12.21* 0.02 47.64* -3.58 13.58* 0.4213 21.39* 1.59 35.39* -9.71 6.82* 22.37* 16.78* 1.64 2.3014 44.05* 9.55* 82.74* 7.34* -0.35 35.52* 4.54* 0.87 -2.0315 57.86* 14.22* 98.65* 36.39* 10.23* 57.74* -8.10 7.48* -0.9416 5.55* 2.44 29.10* -1.82 14.17* 15.63* 16.09* -0.36 1.8717 36.48* 4.02* 109.43* 4.21* 0.32 38.63* 4.81* 1.16 -0.1618 49.86* 19.90* 103.38* 30.52* 2.30 74.60* -6.02 15.96* 4.00*19 0.84 -2.49 65.33* -15.51 40.79* 45.68* 45.41* 39.12* -0.7420 27.10* 6.44 72.09* 35.40* 6.89* 44.07* 3.41 -2.41 -1.1621 108.17* 29.41* 69.69* 26.46* 2.18 48.00* -7.28 16.48* -0.7622 -8.32 -5.24 48.33* -10.31 2.29 10.15* 22.20* 6.17* -0.5923 92.18* 24.24* 79.91* 22.92* 5.23* 39.53* 0.15 0.56 -0.6624 98.98* 29.60* 45.80* 36.11* 4.99* 72.64* -1.92 0.55 1.6625 -5.79 13.44* 20.00* 54.37* 108.31* 3.51 53.43* 71.05* 2.7826 25.65* 2.07 77.86* 19.86* 0.54 43.08* -2.53 1.10 -0.9727 258.76* 81.47* 134.57* 88.06* 47.13* 92.43* 3.03 53.30* -0.37

For the crisis period (July 2007- April 2009), SMB has 20 positive coefficients. HML has 17 (8) positive (negative) coefficients. For Model 2, SMB has 19 positive and 1 negative coefficients, HML has 17 positive and 8 negative coefficients, and LEV has 9 positive and 5 negative coefficients. As noted earlier, there is a marked improvement in the regression R2 when LEV is added.

Finally, we estimated the partial F-statistics to measure the marginal significance of LEV in the model. Similar to the results for global stocks, we find that LEV contributes to improving the overall significance of the model. In all three periods (aggregate, non-crisis and turbulent), the partial F-statistics is significant in majority of the cases.



5. ConclusionsFama and French (1993) note that the traditional FF factors SMB and HML are good proxies for the underlying distress risk of the firms. As of today, a comparison of how well SMB and HML explain stock returns across good times and bad is missing from the literature. In particular, an investigation into whether the economy wide leverage factor replicates the underlying economic fundamentals and contributes to systematic risk especially during bad times is still abstruse; and we attempt to unravel this puzzle in the present study. Our hypothesis is that, compared to conventional stocks with high leverage, we would expect SC stocks to have lower sensitivity to risk factors, as well as lower risk premium. This finding would significantly reaffirm the notion that excessive leverage and engaging in economic activities that are not consistent with the principles of Islamic transactions can destroy economic and social values, especially during falling market environment.

Using weekly data on stock returns for 3704 firms, we test for the significance of the factors constructed on the basis of size, book to market equity, and leverage. We find that the significance of the market factor is drastically reduced during the recent crisis while the explanatory power of the Fama-French factors, SMB and HML is reduced considerably. In contrast, leverage risk factor performs considerably well across all there periods, especially well during the financial crisis, in capturing systemic risk in the economy. Its addition to the model is directly correlated with the reduction of the economic and statistical significance of the traditional Fama-French factors.

The main result of this paper is that the effects of leverage risk are robust to heterogeneity of the firms in the sample. To show that, we perform cross-sectional regressions across three distinct categories of stocks i.e. Conventional, Islamic, and SRI stocks. First, as indicated in the earlier section, excess market returns play a leading role in explaining the cross section of expected returns prior to the crisis period, but the effects of the market factor consistently phased out across all the three categories of stocks during the crisis period. The effects of the leverage factor are consistently significant (except in the case of the socially responsible investing stocks) throughout; however leverage factor gains momentum during the crisis period and has a significant effect on the cross-section of expected returns on stocks and portfolios. The sensitivities of stock returns to the Fama-French factors are lower after the introduction of the leverage factor.

In a nutshell, the contribution of leverage risk to asset pricing has been quite strong. The results indicate that leverage based risk factor can explain a substantial portion of the cross-section of stock returns across financial and non-financial stocks, as well as, various categories of stocks including conventional, Islamic, and SRI stocks. These results have powerful implications for asset management using various types of stocks and also during periods of great uncertainties.

Notes 1. Sharia compliant stocks are household names in

mostly developed countries. Surprisingly, only few stocks with enough liquidity and strong balance sheet

data from the emerging and Muslim countries are included in the Dow Jones Islamic Index.

2. The leverage measure which we are using is the market value of debt to market value of assets and not book value of debt to market value of assets. Both debt to equity and debt to assets are measures of capital structure of a company reflecting the amount of fixed liabilities. The only difference being that debt to equity ratio is more specific to the overall capital used in the company while debt to assets ratio is a much broader measure.

3. The authors used weekly data to examine the relative performance of investing in three different types of stocks –conventional, Islamic, and SRI stocks. Both in sample (Jan 2000-June 2007) and out of sample (July 2007-April 2009) mean-variance optimization indicated a portfolio with Islamic stocks generated significantly larger Sharp ratios. The authors claim that a low credit market exposure of Islamic stocks was largely responsible for the relative superior performance. The results are robust even when financial and real estate companies are removed from the sample.

4. A combination of these two leverage factors produces the book to market ratio. See Fama and French (1992).

5. The crisis had a major impact in September and October 2008 when there was a huge withdrawal of $144.5 billion from the money market. Major institutions like Lehman Brothers, Bear Stearns, Merrill Lynch, Fannie Mae, Freddie Mac and AIG had to bear the brunt of high debt market exposure.

6. Accounting leverage is defined as assets/(assets-liabilities).

7. Source: http://neutralobserver.blogspot.com/2008/ 11/understanding-financial-crisis-leverage.html

8. This part of the discussion has been adapted from “Leverage 101: The Real Cause of Financial Crisis”, Sept. 25, 2008, extracted from http://seekingalpha.com/article/97299-leverage-101-the-real-cause-of-the-financial-crisis

9. Source: http://www.huffingtonpost.com/niall-ferguson/beyond-the-age-of-leverag_b_163872.html

10. Source: http://sg.biz.yahoo.com/080625/67/4hbq2.html

11. Source: http://sg.biz.yahoo.com/080625/67/4hbq2.html

12. Source: http://www.financialweek.com/article/ 20080624/REG/705337846/-1/FWDAILYALERT01

13. As stated in a research note by James Lee, vice chairman of J.P.Morgan Chase – extracted from http://www.financialweek.com/article/20080624/REG/705337846/-1/FWDAILYALERT01

14. As stated by Bob Chapman, “Upsurge of Global Leveraged Speculation: The Financial Crisis is not over”, Global Research, November 6, 2009 – extracted from http://www.globalresearch.ca/index.php?context = va&aid = 15959

15. This is consistent with the portfolio formation procedure as suggested in Fama and French (1993). However, for the purpose of firm specific analysis, we consider all stocks.

16. To avoid complications, we restrict the 3-month T-bill return to zero for the months of December 2008 and January, 2009 when intraday return on T-bills was often negative.

Bhatt and Sultan


17. Source: http://www.lombardodier.com/annexes/ 23056/23074/Investment_Strategy_Bulletin_06.10.09.pdf, “Is de-leveraging an obstacle to recovery?”

18. http://www.wikinvest.com/concept/2007_Credit_Crunch

19. We split the periods to specifically test the impact of LEV factor during the non-crisis and the crisis period. Given the fact the overleveraging leads to increased risk exposure in the economy, we believe that this part of the systematic risk was not captured fully by the traditional FF factors. This leads us to conjecture that LEV factor has more direct implications for the performance of the stocks during the recent credit crisis and hence we expect the LEV factor to exhibit stronger effects during this period.

20. See Table 5.21. According to the study, managers choose leverage on

the basis of the private information about the future growth prospects and hence, the financial health of the firm.

22. Furthermore, a firm with low leverage (having low Tobin’s q and insignificant growth opportunities) are harder hit during distress periods as compared to firms with higher leverage ratios (with major growth prospects and positive NPV projects). This also explains a negative relationship between credit spread and firm’s leverage.

23. Source: “How leverage can increase a company’s return on equity”, Putnam Spectral Funds, extracted from: http://www.putnam.com/spectrum/return-on-equity.htm

24. However the risk substantially increases with the excessive use of debt since the firm is under a pressure to service its debt on a regular basis. In addition, during economic distress, the assumption that debt is available at a lower cost may not hold true. The recent credit crisis of 2007 presents plentiful evidence where debt became costly. In fact, a firm which undertakes risky projects may not enjoy the low cost of debt because the riskiness of its operations may require the debt holders to be paid a higher interest.

25. It is believed that financial firms exhibit different characteristics as compared to non-financial firms and hence show different sensitivities to the risk factors. For instance, high leverage for a financial firm has different implication as compared to a non-financial firm with high debt levels. This further rationalizes the idea of conducting a robustness check by separating out financial firms from the sample.

26. Firm-specific regressions are not reported to conserve space. They are available on request.

27. We thank Dow Jones for providing us with the proprietary list of stocks classified as conventional, Islamic and SRI stocks.

28 Specifically, the debt ratio (short-term plus long-term debt as a percent of market capitalization) must not exceed 33%, interest income should not represent more than 5% of total revenue, the ratio of accounts receivables to total assets does not exceed 45%, and the ratio of cash and interest bearing securities to market capitalization does not exceed 33%. See Dow Jones website for more.

ReferencesAbel, AB. 1988. Stock prices under time-varying dividend

risk: An exact solution in an infinite-horizon general equilibrium model. Journal of Monetary Economics. 22:375–393.

Abel, AB. 1999. Risk premia and term premia in general equilibrium. Journal of Monetary Economics. 43:3–33.

Arditti, FD. 1967. Risk and the required return on equity. Journal of Finance, 22:19–36.

Armitrage, S. 2005. The cost of capital: Intermediate theory. Cambridge University Press.

Bhandari, LC. 1988. Debt/Equity ratio and expected common stock returns: Empirical evidence. Journal of Finance. 43:507–528.

Black, F. 1972. Capital market equilibrium with restricted borrowing. Journal of Business. 45:444–455.

Black, F. 1990. Mean reversion and consumption smoothing. Review of Financial Studies. 3:107–114.

Bris, A, Koskinen, Y. 2002. Corporate leverage and currency crises. Journal of Financial Economics. 632, 275–310.

Chan, KC, Chen, N. 1991. Structural and return characteristics of small and large firms. Journal of Finance. 46, 1467–1484.

Chou, PH, Ko KC, Lin S. 2010. Do relative leverage and relative distress really explain size and book-to-market anomalies? Journal of Financial Markets. 13, 77–100.

Cochrane, J. 2001. Asset Pricing. University Press. New Jersey.

Connor, G, Korajczyk, R. 1989. An intertemporal equilibrium beta pricing model. Review of Financial Studies. 2:373–392.

Daniel, K, Titman S. 1997. Evidence on the characteristics of cross-sectional variation in stock returns. Journal of Finance. 52, 1–33.

Davis, JL, Fama, E, French K. 2000. Characteristics, covariances, and average returns: 1929 to 1997. Journal of Finance. 55:389–406.

Dempsey, M. 2009. The Fama and French three-factor model and leverage: compatibility with Modigliani and Miller propositions. Investment Management and Financial Innovations. 6:48–53.

Dimitrov, V, Jain, P. 2006. The value relevance of changes in financial leverage. Working Paper. http://ssrn.com/abstract = 708281.

Dow Jones & Company. (2009, February 27). Dow Jones Indexes. Retrieved August 9, 2009, from Dow Jones & Company: http://www.djindexes.com/mdsidx/index.cfm?event = showIslamicStats#fund.

Dow Jones & Company. (2009, February 28). Dow Jones Sustainability Index World Components. SAM Indexes. Retrieved September 8, 2009, from Dow Jones & Company: http://www.sustainability-index.com/djsi_protected/djsi_world/components/SAM_DJSIWorld_Components.pdf.



Dow Jones & Company. (2009, March 2). Guide to the Dow Jones Islamic Market Indexes. Dow Jones & Company.

Dow Jones Sustainability World Index. SAM Indexes. (2009, February 1). Dow Jones Sustainability World Index. SAM Indexes. Retrieved September 8, 2009, from Dow Jones & Company: http://www.sustainability-index.com/djsi_pdf/publications/Factsheets/SAM_IndexesMonthly_DJSIWorld.pdf.

Elton, EJ, Gruber, MJ, Agrawal, D, Mann, C. 2001. Explaining the rate spread on corporate bonds. Journal of Finance. 247–277.

Eom, KS, Park JH. 2008. Evidence on the three-factor and characteristics models: Korea, SSRN: http://ssrn.com/abstract = 1329664.

Fama, E, French, K. 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial Economics. 33:3–56.

Fama, E, French, K. 1992. The cross-section of expected stock returns. Journal of Finance. 47:427–465.

Fama, E, French, K. 1995. Size and book-to-market factors in earnings and returns. Journal of Finance. 50:131–155.

Fama, E, French K. 1996. Multifactor explanations of asset pricing anomalies. Journal of Finance. 51: 55–84.

Fama, EF, French K. 1998. Value versus growth: The international evidence. Journal of Finance. 53:1975–1999.

Fama, EF. 1981. Stock returns, real activity, inflation and money. American Economic Review. 71, 545–565.

Fama, EF, Jensen M. 1998. Separation of ownership and control. Michael C, Jensen Foundations of Organizational Strategy, Harvard University Press.

Fama, E, Schwert, WG. 1977. Asset returns and inflation. Journal of Financial Economics. 5:115–146.

Ferguson, M, Shockley R. 2003. Equilibrium anomalies. The Journal of Finance. 58(6):2549–2580.

Gomes, JF, Schmid L. 2009. Levered returns. Forthcoming. Journal of Finance.

Jensen, M, Meckling WH. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics. 3:305–360.

Jensen, M. 1986. The agency costs of free cash flow: Corporate finance and takeovers. American Economic Review. 76:323–329.

Johnson, T, 2004. Forecast dispersion and the cross section of expected returns. Journal of Finance. 59:1957–1978.

Kandel, S, Stambaugh, RF. 1990. Expectations and volatility of consumption and asset returns. Review of Financial Studies. 3:207–232.

Lakonishok, J, Shleifer AR, Vishny. 1994. Contrarian investment, extrapolation, and risk. Journal of Finance. 49:1541–1578.

Lally, M. 2004. The Fama and French model, leverage, and the Modigliani-Miller propositions. Journal of Financial Research. 27:341–349.

Lang, L, Ofek, E, Stulz, R. 1996. Leverage, Investment and firm growth. Journal of Financial Economics. 40:3–29.

Liew, J, Vassalou, M. 2000. Can book-to-market, size and momentum be risk factors that predict economic growth. Journal of Financial Economics. 57:221–245.

Luoma, GA, Spiller, EA, Jr. 2002. Financial accounting return on investment and financial leverage. Journal of Accounting Education. 202:131–138.

Milly, M, Sultan, J. 2009. Portfolio diversification during financial crisis: An analysis of Islamic asset allocation strategy. Bentley University Working Paper.

Modigliani, F, Miller, M. 1958. The cost of capital, corporation finance and the theory of investment. American Economic Review. 48:261–297.

Penman, SS, Richardson, Irem, T. 2007. The book-to-price effect in stock returns: Accounting for Leverage. Journal of Accounting Research. 45:427–467.

Petkova, R. 2006. Do the Fama-French factors proxy for innovations in predictive variables? Journal of Finance. 61:581–612.

Report on Socially Responsible Investing Trends in the United States. 10 Year Review. Washington D.C.: Social Investment Forum. 2004.

Report on Socially Responsible Investing Trends in the United States. Washington D.C.: Social Investment Forum. 2008.

Ron, YWH, Strange, R, Jenifer, P. 2008. Corporate financial leverage and asset pricing in the Hong Kong market. International Business Review. 17(1):1–7.

Sivaprasad, S, Muradoglu, YG. 2009. An Empirical test on leverage and stock returns. University of Westminster Working Paper.

Vassalou, M, Yuhang X. 2004. Default risk in equity returns. Journal of Finance. April: 831–868.

Vassalou, M. 2003. News related to future GDP growth as a risk factor in equity returns. Journal of Financial Economics. 68:47–73.


Cite this chapter as: Sadeghi M (2015). Is Shariah-compliant investment universally sustainable? A comparative study. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Is Shariah-compliant investment universally sustainable? A comparative studyMehdi SadeghiDepartment of Applied Finance and Actuarial Studies, Macquarie University, Sydney-Australia, T: (61) 2–98508527, E: [email protected]

Abstract - The current paper reports the outcome of investigating the sustainability and efficiency of Shariah–compliant investment from the global and cross-country perspectives. Our findings, thus far, suggest that global Shariah compliant sustainable shares performed slightly better than global sustainable shares in general, during 2006-2011, and Shariah compliant shares performed substantially better than global market during the same period. The superior performances of Islamic market indexes suggest that Shariah compliant investment is more resilient and sustainable compared to their counterparts in the long term. Further evidence from our cross country study suggests that, Shariah compliant investments perform better than the market on whole in Muslim countries, and worse than the market in predominately non Muslim countries. These findings have important implications for investors, regulators, customers, and Islamic financial institutions.

Keywords: Shariah-compliant sustainable investment, Shariah-compliant investment, index addition and deletion, event study, abnormal returns, liquidity changes

JEL Classification: G14, G15

1. IntroductionThe market for Islamic financial services is growing at an impressive rate, reaffirming its position as one of the most dynamic sectors in international finance. The Islamic finance industry enjoyed a compound annual growth rate for 2006–2009 of 28%1. The current value of Shariah compliant assets managed worldwide, according to the International Monetary Fund (IMF) estimates, now tops USD 1 trillion. The value of these assets is forecasted to hit US$1.6 trillion by 2013. This growth represents a major achievement, as well as new challenges for investors, regulators, customers, and also Islamic financial institutions themselves.

The biggest share of Islamic financial belongs to the Islamic banks. The S&P2 report indicate that the assets of top 500 Islamic banks in 2008 was $639bn, and grew by 28.6 percent to $822 bn in 2009. There are also Shariah-compliant investment funds within Islamic financial system that cover a wide range of sectors including real estate, equities, infrastructure, and energy. According to Lipper data for 2010, 586 Islamic funds were in operation with $37bn of assets under management, with a bias towards

equity funds (303), mixed asset (101), money markets (77), and sukuk funds (77).

With the recent troubles in the global economy, finance industry has been looking at Islamic contracts as the possible means of preventing such meltdowns from ever materializing again. Another area that has received more in-depth media coverage is the field of sustainability. Recent changes in the world of investment have made asset owners and managers increasingly aware of the potential risk and value impact of environmental, social, and governance (ESG) factors, on an investment profile. There are arguments in financial literature in favour of both areas as safer approaches, and less vulnerable to questionable financial transactions, which may have led to the global recession beginning in 2008. These arguments have been substantiated by some empirical findings that suggest some Islamic financial institutions and companies focused on sustainability have been more resilient to financial crisis. For instance, Hasan and Dridi (2010), report that Islamic banks have been more resilient than conventional banks during recent global financial crises. This view was also corroborated by external rating agencies’ reassessment of Islamic banks’ risk, which was generally found to be

Sadeghi


more favourable than—or similar to—that of conventional banks (with the exception of UAE) (ibid). Some studies also suggest that companies with a strong commitment to sustainability have outperformed their industry averages by 17%3.

But are Islamic finance and sustainability finance compa-tible? What’s really involved in incorporating sustainability criteria and Islamic principles into investment decisions? Can they make a material difference to investment performance? We start answering these questions by highlighting similarities and differences between these two. Islamic finance and socially responsible investing (SRI) approaches have a lot in common with respect to the screening process, and criteria used for stock selection. Sustainability, on the other hand, goes above and beyond SRI by considering positive screens, promoting investment in companies with best practices. According to World Economic Forum Report (2011) “Sustainable investing is an investment approach that integrates long-term environmental, social and governance (ESG) criteria into investment and ownership decision-making with the objective of generating superior risk-adjusted financial returns”4. As the financial crisis receded into a period of uncertainty in the past two years, recognition that sustainability, corporate governance and transparency are important factors in portfolio management has emerged. This is a fundamental shift away from the ideological and political corner of SRI to the real performance of sustainability.

Some researches assert that Islamic finance holistic and dynamic perception of SRI is more effective in taking into consideration the reality and ever-changing circumstances of societies in contrast to Western humanistic theories. They conclude that corporations operation on a piety-based business paradigm acknowledge their social responsibility to their workers, managers, other corporations, customers, and society as a whole more significantly (Dusuki and Abdullah, 2007). However, regardless of their similarities, and theoretical arguments in support of one or another, sustainability and Shariah-compliant investments are assessed on the basis of long-term trends in yield, profitability, and efficiency in use of limited financial resources.

In January 2006, Dow Jones Indexes launched the world’s first Dow Jones Islamic Market Sustainability Index. This index merges Islamic investing principles and sustaina-bility criteria by combining the methodology of Dow Jones Islamic Market Indexes5 and Dow Jones Sustainability Indexes. To be included in the index, companies must be components of both the Dow Jones Islamic Market Index and the Dow Jones Sustainability World Index. Linking Shariah compliant investment performance to sustainability is, perhaps, the most effective way to highlight the importance of ESG governing factors to Islamic finance. The time series data provided by Dow Jones Indexes is an invaluable resource to help us investigate whether Islamic finance is a sustainable practice in the long term.

Current paper is a progress reports on our ongoing long term research objective of testing the efficiency and sustainability of Shariah compliant investment opportunities around the world. We have used time series data on Dow Jones

Islamic Market Index and Dow Jones Islamic Market Sustainability Indexes and their constituents to see if there is any significant difference between the performances of these indexes with Dow Jones Global Stock Market Index. We also investigate whether there is any significant change in the efficiency and liquidity of market following Islamic index addition and deletion events.

This study is important for several reasons. First, although Shariah-compliant investment is similar to SRI, an area that has already attracted a great deal of research interest, certain differences is evident in the screening procedures that make Shariah-compliant investment different. For instance, some Islamic funds do not exclude weapons manufacturers but they do exclude conventional banks, while SRI funds normally exclude weapon manufacturing firms and do not exclude banks. As another difference, concerns about environmental issues are not as important in screening Shariah-compliant companies as they are for SRI funds. Furthermore, Shariah-compliant companies are subject to certain financial ratio tests that are not relevant to conventional SRI companies6.

Second, Miller-Modigliani capital structure theory contemplates that in an imperfect capital market with corporate taxes, companies can increase their assets’ value by increasing their leverage. Given that Shariah-compliant companies are constrained by their level of borrowing, it would be interesting to investigate how this constraint can affect their value.

Third, finance theory based on the efficient market hypothesis (EMH) considers shares with identical risk and return as perfect substitutes for each other. This makes market demand for securities elastic and horizontal. Since Shariah-compliant equities are not a perfect substitute for the conventional equities, their demand may not be horizontal. This can bring about a different outcome to the study of a Shariah-compliant index revision.

Fourth, Islamic screening criteria reduce the number of available shares to invest. It is claimed by critics that the reduction of the investment universe through screening will reduce the performance. Similar counterarguments have been raised regarding sustainability criteria (Freidman, 1996). It would be interesting to investigate how this constraint can affect Shriah-compliant portfolios.

Finally, academic research on the performance of Shariah-compliant investments is rare, and to the best of our knowledge, no similar study on the impacts of the Shariah-compliant index revisions has been conducted before.

Our results, thus far suggest that global Shariah compliant sustainable shares perform worse than global Shariah compliant shares in the long term. However, they both perform better than global stock market as a whole. Further evidence from individual countries suggests that, Shariah compliant investments perform better than the market in Muslim countries, and worse than the market in predominately non Muslim world. The rest of this study is organized as follows: Section II is allocated to a short review of research background. We outline our methodology, data and hypothesis development in Section III. Empirical


Is Shariah-compliant investment universally sustainable? A comparative study

findings are discussed in Section IV. Section V articulates our conclusions, and describes the limits of our study.

2. Research background and literature review

Research backgroundWe started our study with an investigation of the market performance and liquidity of Shariah-compliant Index (SI) portfolio following its introduction by Bursa Malaysia. Malaysia has one of the largest Islamic fund markets in the world. It had 155 unit trusts and mutual funds at the end of June 2010 with a total volume of about RM22.69 billion. Our findings show that, overall, introduction of SI had a positive impact on the financial performance and the liquidity of included shares in this country7.

As time series data on Shariah compliant indexes become more readily available for other parts of Muslim world through index providers, such as Dow Jones Islamic Market Index8, we decided to extend our study to the MENA (Middle East and North Africa) market in the second stage of our study. MENA region is another important hub in Islamic finance, with large market and appropriate financial infrastructure. Constrained by the availability of times series data, we used event study methodology and the improved models of liquidity measures, first to index addition to equity markets in Qatar, Kuwait, Oman, and UAE. Our findings showed an even stronger result than for Malaysia in that market reacts positively to the introduction of Shariah compliant shares in these countries. This was reflected in short and long-term market performance and the improvement in the liquidity of shares9. One of the limitations of recent study was the small number of companies in our sample. To test the robustness of findings with larger samples, we extended our investigation to Jordan and Egypt. Our results overwhelmingly supported the robustness of our earlier findings of countries in the Gulf region10.

Overall, our research on seven markets in Islamic countries in showed that investors’ reaction to the introduction of Shariah-compliant shares is positive. This is reflected in improvement in the share price and market liquidity up to 150 days following the index addition. The positive outcome for six countries in MENA region is especially important because they were found from the data that became available by Dow Jones Indexes immediately following the start of financial crisis, suggesting that Shariah compliant investments in Islamic countries has been more resilient to financial crisis than conventional investments.

In addition to Muslim countries, Islamic finance is practiced outside the Muslim world without ties to any particular jurisdiction. Shariah compliant investments are defined according to certain norms and conditions that can be applied anywhere in the world where there is a market and people who wish to engage in financing transactions in a manner which is consistent with Shariah law. This progress is specially facilitated by a form of reverse financial engineering that reconstructs conventional financial products into Shariah compliant instruments. This innovation has significantly increased Muslims investments

in Shariah compliant companies in non-Muslim countries around the world11. In the case of equities, the differences between Shariah compliant shares and their conventional forms are even less significant, only requires screenings. This screening process is similar to the screening of Socially Responsible Investing (SRI) instruments.

In the third stage of our research we decided to investigate Shariah-compliant index addition and deletion to predominantly non-Muslim countries, starting with Australia as the first sample. Australia’s skills in complex financial engineering and experience in infrastructure, resources, property and agriculture provide her with a unique opportunity to develop Shariah-compliant investments. This country also has easy access to rapidly growing Islamic financial markets with over a billion in population to accommodate their demand12.

A through presentation of our findings on all eight countries studied so far is too long to report here. In order to show the contrasting nature of market reaction to Index addition and deletion events in predominately Muslim and non-Muslim countries, we report the report the results on two sample countries of Egypt and Australia in section IV.

Literature reviewFrom a theoretical perspective, there are two explanations for the effects of stock additions to an index: demand-based and information-based. The demand-based explanation sees index changes as information-free events. For example, Shleifer (1986), by employing the downward-sloping demand curve hypothesis, showed that the price effects following index changes are due to the demand from index tracking. These effects can be temporary or permanent. The temporary effect is explained by the price pressure hypothesis, predicting a reversal of initial price increases in the long run (Harris and Gurel, 1986). The permanent effect is explained by the imperfect-substitute hypothesis, which assumes that there would be no price reversal, as the new price reflects changes in the distribution of security holdings in equilibrium13.

Information-based explanations include the information hypothesis and the liquidity hypothesis. Unlike the demand-based explanations, information-based explanations as-sume that index changes are not information-free events. Some studies, such those by Dhillon and Johnson (1991) and Jain (1987), support the information hypothesis: they showed that the addition of a stock to the index conveys favorable news about the firm’s prospects and a permanent price increase can result following this event. Amihud and Mendelson (1986), Beneish and Whaley (1996), and Hegde and McDermott (2003) contended that the price reactions can be explained by changes in market liquidity. According to the liquidity hypothesis, the price increase at index inclusion is caused by the increased liquidity due to the greater visibility of the shares, greater interest from institutional investors, higher trading volume, and lower bid-ask spreads. Amihud and Mendelson (1986) suggested that the increase in stock liquidity is positively related to the firm’s value through a reduction in the cost of capital. Previous studies, such as Harris and Gurel (1986), and

Sadeghi


Hegde and McDermott (2003) reported liquidity increases following index additions.

The topic of Shariah-compliant index revision is important from two perspectives. First, the nature of companies’ activities and their capital structure makes them Shariah compatible in the first place. Second, changes in investors’ demand result in subsequent market price reactions, according to our earlier discussion. For example, reduction in the level of debt in the capital structure can make a company Shariah-compliant, bringing about an increase in the demand from Muslims and higher share prices if demand is not fully elastic. At the same time, the lower level of debt may move the capital structure of the company to a suboptimal level, at a higher cost of capital than in equilibrium. This may send negative signals to the market when shares are added to a Shariah-compliant index. As a result, it is possible that the interaction of opposing market forces on index revision will bring about different outcomes compared with the effects of conventional index additions. Therefore, it is not possible to predict clearly how the performance and liquidity of shares included in or excluded from the DJIM index will change, as it largely depends on how the net effects of the influential factors are revealed through our empirical investigation.

3. Data and methodologyTo determine the impact of additions to and deletions from the DJIM index, we applied several measures of both short and long-term price and liquidity performance. We applied standard event study methodology to find the initial stock price reaction of firms when an announcement of an index change was made. We also applied several liquidity measures to investigate the magnitude and direction of liquidity changes following the index revision. Data for this research has been collected through Dow Jones Indexes and Bloomberg.

Price effectOur event-study methodology calculates the abnormal returns. An abnormal return is the difference between the realized return observed from the market and the benchmark return. The return to the market portfolio is estimated via both ordinary least square (OLS) and Scholes and William (1977) procedures. The latter method is usually used when stocks do not trade at the same level of frequency as the market index and OLS may produce biased beta estimates. This problem is exacerbated for infrequently or thinly traded stocks as the sampling interval is reduced14. The advantages of these models are that they control for the effect of market movements through the market portfolio, and also allow for an individual security’s responsiveness as measured by beta. Return on the All Ordinaries index was used as a proxy for the market rate of return.

We defined the event date as the day that a stock was added to or deleted from the DJIM index. For each event, the return time series data were divided into an estimation period and an event window. The estimation time series data are used to calculate the benchmark parameters, and the event window period is used for computing prediction errors based on the estimated parameters. Abnormal returns are represented by the prediction errors. The abnormal returns

during the event windows can be interpreted as a measure of the effect of the event on the value of the firms, which is reflected in their share price.

Our event window extended from 10 days before to 25 days after the event. This asymmetric event window was chosen to examine the extended effect of excess returns in the post-event period15.

The normal returns of stocks are the expected returns if there are no events. The normal returns are estimated over a period of time outside the event window (Peterson, 1989). For applications in which the determinants of the normal return are expected to change due to the event, the estimation period can fall on both sides of the event window. This period commences 125 trading days before and ends 125 trading days after the event dates, excluding the event period of day -10 to Day 25. As a result, the estimation period consists of Day -135 to Day -11 and Day 26 to Day 150. We did not allow the event period to overlap with the estimation period, to avoid biasing the parameter estimates in the direction of the event effect.

The following section describes the event study methodology that we used in our study. MacKinlay (1997), and Kothari and Warner (2004) have provided a survey of event study methods, and we follow their papers to describe the models here.

Liquidity effectMarket liquidity is an elusive concept and difficult to measure. In this study, we use six proxies to evaluate changes in market liquidity during post-event periods, compared to the corresponding control periods. The large number of tests helps to confirm the robustness of our findings and reduces the chance of making wrong inferences.

These liquidity proxies include: 1) quoted spread, as the simple difference between bid and ask prices; 2) percentage spread, as the quoted spread normalized by the midpoint of the bid and ask prices; 3) changes in the volume of trade as the daily average of the transaction size, normalized on the average volume of trade in the control period; 4) changes in volatility, measured by the standard deviation of returns; 5) the Amivest liquidity ratio, as the average ratio of share volume to absolute return over all days with non zero returns; and 6) the proportion of zero daily returns. Zero daily return is related to trading speed because the days with zero return indicate delays or difficulties in executing an order, interrupting the continuity of trading.

In calculating the percentage bid-ask spread and change in the volume of trade, we largely follow Hegde and McDermott (2003). Changes in the volume of trade are directly related, and changes in the bid-ask spreads and volatility are inversely related to the market liquidity. It is important to note that an increase in the volume accompanied by an increase in volatility can actually impede market liquidity. The Amivest liquidity ratio is estimated according to Amihud and Mendelson (2002). This ratio measures the ability of a share to absorb changes in trading volume without any significant change in share price. Change in this variable is directly related to the liquidity. In estimating the proportion of zero daily



returns, we follow Bekaret et al. (2004), as they found it a reasonable proxy for a liquidity measure to use in their study. Change in this variable is inversely related to the market liquidity.

4. Results

Global findingsFigure 1 presents the Performance of Dow Jones Islamic Market Sustainability Index (DJIMSI), Dow Jones Islamic Market Index (DJIMI), Dow Jones Global Sustainability Index (DJGSI), and Dow Jones World Stock Index (DJWSI) during 1/1/2006 to 1/5/2011. Our findings show that Dow

Jones Islamic Market Sustainability Index out performs Dow Jones Global Sustainability Index by less than 1% during this period. However, Dow Jones Islamic Market stock Index shows a much higher return of 22% compare to Dow Jones World stock Market Index. The superior performances of Islamic Market indexes suggest that Shariah compliant investment is more resilient and sustainable compare to their counterparts within the family of Dow Jones Indexes in the long term.

Cross country findingsA through presentation of our findings for all eight countries studied so far is too long to report here. The Results reported here is only from Egypt and Australia, in order to show the contrasting nature of market reaction to Index addition and deletion in two predominately Muslim and non Muslim country.

Price effectTable 1 presents the estimated CARs for index additions in the pre- and post-event periods for Egypt. The coefficient for CARs, accumulated during the period (-10, 0), is -1.71%. However, it is not statistically significant at the conventional levels. The CARs coefficient estimated over the shorter period (-5, 0) increases to 2.41% and becomes statistically significant at the 0.05 level. CARs coefficients for Day 0 (the event day) and for (0, 5) increase further to 2.85% and 3.44%, respectively, and become highly significant at the 0.01 level. CARs for (0, 15) drop to 2.69% and remain statistically significant at the 0.01 level. CAR coefficients increase further to 6.31% during (0, 30) and remain statistically significant at the 0.01 level.

The prolonged effects of the index additions on CARs in Table 1 indicate that these events are likely to contain information, thus sending signals about the features of the index additions to the market. To test this hypothesis, we compared the cumulative returns (CRs) for the added firms with the cumulative return for the market over the period (-10, 150)16.

0

5

10

15

20

25

30

35

401/

1/06

9/1/

06

5/1/

07

1/1/

08

9/1/

08

5/1/

09

1/1/

10

9/1/

10

5/1/

11

DJIMSIDJGSI

0

5

10

15

20

25

30

35

40

45

1/1/

06

11/1

/06

9/1/

07

7/1/

08

5/1/

09

3/1/

10

1/1/

11

DJIMI DJWSI

Figure 1. The Performance of Dow Jones Islamic Market Sustainability Index (DJIMSI), Dow Jones Islamic Market Index (DJIMI), Dow Jones Sustainability Index (DJGSI), and Dow Jones World Stock Index (DJWSI) during 1/1/2006 to 1/5/2011.

Table1. Cumulative abnormal returns and relevant statistics for stock additions to the DJIM index in Egypt.

This table presents the cumulative abnormal returns (CARs) around the index addition for the 25 Egyptian firms in our sample. Results are presented for the windows (-10, 0), (-5, 0), (0, 0), (0, +5), (0, 15), and (0, 30), where day 0 represents the addition date. The Generalized Sign Z-test is a test with the null hypothesis that the fraction of posi-tive cumulative returns is the same as in the estimation period. The Positive/Negative column reflects how many firms had positive cumulative abnormal returns in the window. The symbols $, *, **, and *** denote statistical significance at the 10%, 5%, 1% and 0.1% levels, respectively, using a 1-tail test. The symbols), >, etc., correspond to $,* and show the significance and direction of the Generalized Sign-Z test.

Scholes-Williams Market Model

Intervals MCARs t-StatisticsGeneralized Sign Z-test

Positive/ Negative

(-10, 0) 1.71% 1.12 1.15 15/10(-5, 0) 2.41% 1.75* 1.55$ 16/9)(0, 0) 2.85% 6.36*** 4.35*** 23/2 >>>(0, +5) 3.44% 2.63** 2.74** 19/6 >>(0, +15) 2.69% 1.72* 2.35** 18/7 >>(0, +25) 6.31% 2.69** 1.95* 17/8 >

Sadeghi


Figure 2 illustrates CRs for the portfolio of added stocks, compared with the market CRs during (-10, 150) for Egypt, showing the shares’ superior performance of 352% gain, compared with less than 48% for the market by Day 150. Figure 3 compares the performance of the same variables on a risk-adjusted basis, calculated using the Sharpe Ratio. According to this figure, the Sharpe Ratio for the shares shows a value of 34 compared with a ratio of 1.4 for the market.

Table 2 and Table 3 present mean cumulative abnormal returns (CARs) for the added and the deleted Australian firms, respectively. To test the robustness of our findings, we have used both the single-factor and Scholes-Williams market models as the benchmarks for estimating normal return. Our results show that the magnitudes of CARs and the level of their statistical significance from the application of the two methods are similar. Nevertheless, we report and discuss the results from the Scholes-Williams model to avoid non-synchronous trading bias, as a considerable proportion of shares included in this study are likely to trade less frequently.

Table 2 presents the estimated CARs for index additions in the pre- and post-event periods. The coefficient for CARs

accumulated during Day -10 to Day 0, is -1.22%, and is statistically significant at the 0.05 level. When the CARs coefficient is estimated over the shorter interval of Day -5 to Day 0, it increases slightly to -1.18% and remains statistically significant at the 0.05 level. CARs for Day 0 (the event day) and Day 0 to Day 5 increase to -0.32% and 0.21%, respectively. However, they are not significantly different from zero at the conventional levels.

CARs for the intervals Day 0 to Day 10 and Day 0 to Day 25 decline continuously, dropping to -4.04% and become highly significant at the 0.01 level. CARs for the entire window (Day -10 to Day 25) is -4.94% and significant at the 0.01 level. The temporary upward trend in CARs around the event day may have been caused by the positive reactions of some Muslim investors to the additions news. However, this reaction was perhaps not strong enough to fully offset a negative response from the market as a whole. The coefficients for generalised sign tests are consistent with the coefficients for t-statistics, although they are not as strongly significant as the later ones. It is mainly the coefficient for Day 0 to Day 25 and the entire event window (Day -10 to Day 25) that are statistically significant at the conventional level, indicating that the significance of our findings is robust to both parametric and non-parametric

–0.5

0

0.5

1

1.5

2

2.5

3

3.5

4

1 10 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145 154

Market IndexShares

Figure 2. Cumulative firm return and market return around day -10 to Day 150 egyptian stocks addition to DJIM index.

Figure 3. Risk adjusted cumulative firm return and market return around day -10 to day 150 Egyptian stocks addition to DJIM index

–5

0

5

10

15

20

25

30

35

40

1 10 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145 154

Market IndexShares



tests. Our findings are also consistent with the results of Clarke and Russell’s (2008) study on Socially Responsible Investing (SRI): they found significant negative CARs for DS400 additions that persisted at least 30 days after the events.

Table 3 presents the estimated CARs for index deletion in the pre- and post-event periods for Australian shares. The coefficient for CARs, accumulated during Day -10 to Day 0, is -1.57% and is marginally significant at the 0.10 level. The CARs coefficient estimated over the shorter interval of Day -5 to Day 0, increases to 2.11% and becomes statistically significant at the 0.05 level. This coefficient for Day 0 (the event day) is 0.47% and statistically significant at the 0.10 level. CARs for Day 0 to Day 5 is negative and statistically insignificant at the conventional levels. This coefficient quickly rises to 5.34% during the interval Day 0 to Day 10, and becomes statistically significant at the 0.05 level.

CARs increases further to 6.05% during the interval Day 0 to day 15, and to 7.45% during the interval Day 0 to Day 25, respectively. Both coefficients remain highly significant at the 0.01 level. CARs for the entire window (Day -10 to Day 25) is 8.55% and significant at the 0.01 level. The temporary downward trend in CARs after the event day may have been caused by the negative reactions of Muslim investors to the deletion news. However, this reaction did not seem to be strong enough to fully offset the positive response from the market as a whole.

Results in Table 2 and Table 3 show CARs of up to 25 days after additions and deletion, respectively. Some studies in the literature, such as one by Nesbitt (1994), suggest that the value of socially responsible investing may be more apparent in the long-run. To examine whether DJIM Index additions and deletions have any prolonged information effects on shares, we compared the cumulative returns

Table 2. Mean cumulative abnormal return and relevant statistics for stock additions to the DJIM index in Australia.

This table presents the mean cumulative abnormal returns (CARs) around the index addition for the 117 firms in our sample. Results are presented for the windows (-10, 0), (-5, 0), (0, 0), (0, +5), (0, +10), (0, 15), (0, 25), and (-10, 25), where day 0 represents the addition date. The third column is the precision-weighted cumulative mean abnormal return. The generalized sign Z is a test of the null hypothesis that the fraction of positive cumulative returns is the same as in the estimation period. The symbols $, * and ** denote statistical significance at the 10%, 5%, and 1% levels, respectively, using a 1-tail test.

Scholes-Williams Market Model

Intervals

Cumulative average abnormal

return (CAAR)

Precision- weighted

CAARt-

statisticsGeneralized sign Z-test

(-10, 0) -1.22% -1.06% -1.70* -0.65(-5, 0) -1.18% -0.78% -2.24* -1.21(0, 0) -0.32% -0.15% -1.16 -0.47(0, +10) -1.11% -0.36% -1.24 -1.21(0, +15) -1.62% -0.65% -1.55$ -1.39(0, +25) -4.04% -2.03% -2.71** -1.58$

(-10, +25) -4.94% -2.94% -3.02** -1.76*

Table 3. Mean cumulative abnormal return and relevant statistics for Australian stock deletions from the DJIM index.

This table presents the mean cumulative abnormal returns (CARs) around the index deletion for the 87 firms in our sample. Results are presented for the windows (-10, 0), (-5, 0), (0, 0), (0, +5), (0, +10), (0, 15), (0, 25), and (-10, 25), where day 0 represents the addition date. The third column is the precision-weighted cumulative mean abnormal return. The Generalized Sign Z is a test of the null hypothesis that the fraction of positive cumulative returns is the same as in the estimation period. The symbols $, * and ** denote statistical significance at the 10%, 5%, and 1% levels, respectively, using a 1-tail test.

Scholes-Williams market model

Intervals

Cumulative average abnormal

return (CAAR)

Precision- weighted

CAARt-

statisticsGeneralized sign Z-test

(-10, 0) 1.57% 1.36% 1.29$ 0.11(-5, 0) 2.11% 1.62% 2.51** 2.60*(0, 0) 0.47% 0.41% 1.35$ 1.19(0, +5) 0.04% -0.04% -0.05 -0.31(0, +10) 5.34% 2.24% 2.28* 1.40$

(0, +15) 6.05% 2.68% 2.33** 1.83*(0, +25) 7.45% 3.82% 2.82** 2.05*(-10, +25) 8.55% 4.77% 3.02** 2.06*

Sadeghi


–0.16

–0.14

–0.12

–0.1

–0.08

–0.06

–0.04

–0.02

0

1 10 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145 154

Market IndexShares

Figure 4. Cumulative return on a portfolio of added Australian shares, compared with cumulative return on the market for the 160-day period from Day -10 to Day 150 around the event day.

–90

–80

–70

–60

–50

–40

–30

–20

–10

01 15410 19 28 37 46 55 64 73 82 91 100 109 118 127 136 145

Market IndexShares

Figure 5. Cumulative risk adjusted return on a portfolio of added Australian shares, compared with risk-adjusted cumulative return on the market for the 160-day period from Day -10 to Day 150 around the event day. Risk-adjusted returns are estimated according to the Sharpe performance index.

Table 4. Measures of liquidity changes from pre- to post-stock additions to the DJIM index in Egypt.

This table presents the change of a variety of liquidity measures around the index addition day for an equally weighted portfolio of 25 Egyptian firms in our sample. Results are presented for the windows (1, 25), (1, 50), (1, 100), and (1, 150), compared with the control periods (-35, -10), (-60, -10), (-110, -10), and (-160, -10), respectively. The bid-ask mean difference represents the difference between average liquidity measures in each interval compared with the corresponding interval in the control period. The symbols $, *, **, and *** denote statistical significance at the 10%, 5%, 1%, and 0.1% levels, respectively, using a 1-tail test.

Intervals Liquidity measures

(1, 25) vs. (-35, -10)

(1, 50) vs. (-60, -10)

(1, 100) vs. (-110, -10)

(1, 150) vs. (-160, -10)

Standard Deviation (SD) 1.05% 0.85% 0.89% 0.85%SD (control period) 1.49% 2.56% 2.05% 1.80%SD change –0.44% –1.71%*** –1.16%*** –0.95%***Relative bid-ask spread 1.53% 1.36% 1.35% 1.36%Relative bid-ask spread

(control period)0.80% 0.71% 0.55% 0.45%

Bid-ask mean difference 0.73% 0.65% 0.80% 0.91%Average daily volume 40.24 53.97 130.73 202.37Average daily volume

(control period)25 50 100 150

Average daily volume change 60.95%* 7.94% 30.73%** 34.91%***Amivest liquidity measure 13.85 13.84 13.68 13.69Amivest liquidity measure

(Control period)13.29 13.32 13.81 13.88

Amivest liquidity measure change

0.56*** 0.52*** –0.13$ –0.19**



(CRs) for the added and deleted firms with cumulative return for the market over the period from Day -10 to Day 15017.

Figure 4 and Figure 5 provide long-term evidence of negative market reaction to the index addition. Figure 4 illustrates CRs for the portfolio of added stocks, compared with the market CRs during Day -10 to Day 150, showing the market’s superior performance of -7.8% compared with -13.1% for the shares by Day 150. Figure 5 compares the performance of the same variables on a risk-adjusted basis, calculated according to the Sharpe performance index (SPI). According to this figure, SPI for the market shows a figure of -42.9% compared with the SPI of -70.8% for the shares.

Table 4 provides evidence of changes in liquidity measures for Egypt. The results show a decline in the standard deviation of returns between 0.95% and 1.71%, accompanied by an increase in the volume of trade from 30.73% to 60.95%. Amivest liquidity measure changes also suggest an increase in the market liquidity over the short to medium term and a decline over the medium to long term. The coefficients for changes in the bid-ask spread is positive; however, they are not statistically significant. Overall, there is more evidence for improvement in the liquidity of the Egyptian stock market than for decline.

5. Concluding remarksCurrent paper reports the outcome investigating the sustainability and efficiency of Shariah –compliant investment from the global and cross-country perspectives. Our findings, thus far, suggest that Dow Jones Islamic Market Sustainability Index out performs Dow Jones Global Sustainability Index by less than 1% during 1/1/2006–1/5/2011. However, Dow Jones Islamic Market stock Index shows a much higher return of 22% compare to Dow Jones World Stock Market Index during the same period. The superior performances of Islamic Market indexes suggest that, relative to their counterparts within the family of Dow Jones Indexes, Shariah compliant investments are generally more resilient and sustainable in the long term.

In the cross country component of our study, we used data from eight countries (only one is reported in this paper) and an event study methodology to estimate cumulative abnormal returns in the days surrounding index additions and deletions for testing the price effects of market reaction. We also used several liquidity measures; including the bid–ask spread, the Amivest liquidity ratio, standard deviation of returns, and volume of trade to estimate changes in the liquidity of the added shares around these events. Our results show that stock prices respond positively to index additions for Muslim countries and negatively for non Muslim countries, both in the short and long terms. Further

Table 5. Measures of liquidity changes from pre- to post- Australian stock additions to the DJIM index.

This table presents the change of a variety of liquidity measures around the index addition for an equally weighted portfolio of 117 firms in our sample. Results are presented for the intervals in days (1–25), (1–50), (1–100), and (1–150), compared with the control periods (-35 to 10), (-60 to -10), (-110 to -10), and (-160 to -10), respectively. Day 0 represents the addition date. The mean difference represents the difference between average liquidity measures in each interval compared with the corresponding interval in the control period. The symbols $, *, **, and *** denote statistical significance at the 10%, 5%, and 1%, and 0.1% levels, respectively, using a 1-tail test.

Intervals Liquidity measures

Day 1 to 25 (-35 to -10)

Day 1 to 50 (-60 to -10)

Day 1 to 100 (-110 to -10)

Day 1 to 150 (-160 to -10)

Absolute bid-ask spread 6.02¢ 6.17¢ 6.24¢ 6.48¢Absolute bid ask (control period) 5.61¢ 5.63¢ 6.14¢ 5.84¢Absolute bid-ask mean difference 0.41¢* 0.54¢** 0.10¢ 0.64¢***Relative bid-ask spread 0.43% 0.45% 0.49% 0.51%Relative bid-ask (control period) 0.47% 0.46% 0.47% 0.46%Relative bid-ask mean difference –0.04% –0.01% 0.02%$ 0.05%**Average volatility (SD) 3.60% 3.51% 3.61% 3.66%Average volatility (SD) (control period)

3.45% 3.15% 2.87% 2.91%

Average volatility ratio 0.15% 0.36%** 0.74%*** 0.75%***Average daily volume 27.10 54.15 106.35 157.56Average daily volume (control period)

25.00 50.00 100.00 150.00

Average volume difference 8.40% 8.30%$ 6.35%* 5.04%*Zero-return 12.12% 11.33% 12.79% 13.15%Zero-return (control period) 8% 7.84% 12.77% 12.64%Zero-return mean difference 4.12%*** 3.49*** 0.02% 0.51%Amivest liquidity measure 12.44 12.38 12.40 12.33Amivest liquidity measure (Control period)

12.53 12.50 12.46 12.44

Amivest liquidity measure mean difference

–0.09*** –0.12*** –0.06*** –0.11***

Sadeghi


evidence from non Muslim countries suggests that stock market react positively to index deletions.

Observing negative abnormal return for index additions, and positive abnormal return for index deletions in Australia suggests that market in this country perceives these events as a value destroying, and value adding exercises, respectively. This view is in line with Friedman (1996) agency theory, perceiving any effort by companies to go beyond maximising their profit as a burden on their return. These opposing reactions can also be explained by differences in both fundamental and socio-cultural factors in Muslim vs. non Muslim countries. For instance, a company in the West world can become Shariah compliant by chance, or by force, not necessarily by choice. A low debt/equity ratio in the capital structure of companies in Western countries can make them Shariah-compliant. However, this may occur, perhaps, due to their inability to borrow money if they are relatively small. While a company in a Muslim world may intentionally borrow less to comply with Shariah- principles.

These findings have important implications for the development and growth of Islamic finance around the world. For example, if Western countries plan to promote themselves as a centre for Islamic finance, they need to overcome certain impediments to be successful. This includes reduction in psychological barriers, as well as revision in taxation laws and non-taxation regularities to ensure that they do not inhibit the development of Islamic finance. There is also need for a trained work force in financial sector (education in Islamic economics, finance, banking, insurance, accountancy, and business law), and ability to market Islamic financial products overseas once they are developed.

Notes1. HSBC Report, Islamic Banking and Finance Summit,

Reuters’ Offices, Dubai, 2009.2. S&P Press Release, 1st February 2010.3. Daniel Mahler, A.T. Kearney, Inc. Report, titled Green

Winners: The Performance of Sustainability-focused Companies in the Financial Crisis, 2009.

4. Transition Towards Sustainable Investing, World Economic Forum White paper, 2011.

5. DJIM Indices were introduced in 1999 as the benchmarks to represent Shariah-compliant portfolios.

6. Socially responsible fixed-income securities are found in conventional financial markets, while, at least in theory, they are banned by Shariah.

7. Refer to Sadeghi (2008) for more details.8. Companies from Islamic countries were added to Dow

Jones Islamic market Index in 2009 for the first time.9. Refer to Sadeghi (2010a) for more details.10. Refer to Sadeghi (2010b) for more details.11. This is a pragmatic compromise, rather than an ideal

situation from an Islamic perspective.12. The DJIM index constituents are screened from

around the globe and are mostly located in non-Muslim countries. For instance, from 2403 DJIM index constituents on 30th November 2009, 2204 originated in the non-Muslim world, especially in the West.

13. Refer to Beneish and Whaley (1996), Lynch and Mendenhall (1997), Kaul et al. (2000), and Wurgler and Zhuravskaya (2002) for more details.

14. The frequency of trading declines with the reduction in the sampling interval.

15. This allowed for slow responses from overseas Muslims that might cause delays in the market reaction to the index revision.

16. We believe that if index inclusion contains information, this information must have been reflected in share prices earlier than the event day and should extend for some time afterwards. As a result, we have used a sample of data that extends from 10 days before to 150 days after the event.

17. We believe that if index inclusion and exclusion contain information, this information must have been reflected in share prices earlier than the event day and should extend for some time afterwards. As a result, we have used a sample of data that extends from 10 days before to 150 days after the event.

References Amihud Y, Mendelson M. 1986. Asset pricing and the bid-

ask spread. Journal of Financial Economics. 17:223–49.

Amihud Y. 2002. Illiquidity and stock returns: Cross-section and time-series effects. Journal of Financial Markets. 5:31–56.

Asyraf Wajdi Dusuki AW, Abdullah NI. 2007. Maqasid al-Shari`ah, Maslahah, and Corporate Social Responsibility. The American Journal of Islamic Social Sciences. 24(1):25–45.

Becchetti LL, Ciciretti I, Hasan. 2009. Corporate social responsibility and shareholder’s value: An empirical analysis. Bank of Finland Research Discussion Papers. 1.

Beneish M, Whaley R. 1996. An anatomy of the “S&P 500 Game: The effects of changing the rules. Journal of Finance. 51:1909–30.

Boehmer E, Poulsen A. 1991. Event-study methodology under conditions of event-induced variance. Journal of Financial Economics. 30:253–72.

Chen HG, Noronha VS. 2004. The price response to S&P 500 Index additions and deletions: Evidence of asymmetry and a new explanation. Journal of Finance. 59:1901–29.

Clarke S, Russell W. 2008. Stock price effects associated with socially responsible index constituent changes. Paper presented at Midwest Finance Association Conference. San Antonio, Texas, USA.

Dhillon U, Johnson H. 1991. Changes in the Standard and Poors 500 list. Journal of Business. 64:75–85.

Friedman M. 1996. The social responsibility of business is to increase profits. In: Schwartz MS (Ed.). Beyond Integrity: A Judeo-Christian Approach. Zondervan Publishing House. Grand Rapids.

Harris L, Gurel E. 1986. Price and volume effects associated with changes in the S&P 500: New evidence for the existence of price pressures. Journal of Finance. 41:815–30.



Hasan M, Dridi J. 2010. Islamic Banking: Put to the Test. Finance & Development. 47(4):45–47.

Hegde S, McDermott J. 2003. The liquidity effects of revisions to the S&P 500 Index: An empirical analysis. Journal of Financial Markets. 6:413–59.

HSBC Report. 2009. Islamic Banking and Finance Summit, Reuters’ Offices, Dubai.

Jain P. 1987. The effect on stock price of inclusion in or exclusion from the S&P 500. Financial Analysts Journal. 43:58–65.

Karpoff J. 1987. The relation between price changes and trading volume: A survey. Journal of Financial and Quantitative Analysis. 22:109–26.

Kothari S, Warner J. 2004. Econometrics of event studies, Working Paper. (Tuck School of Business at Dartmouth, Hanover, USA).

Merton R. 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance. 42:483–510.

Nesbitt S. 1994. Long-term rewards from shareholder activism: a study of the CalPERS effect. Journal of Applied Corporate Finance. 6:75–80.

Price Waterhouse Coopers International Limited Report. 2009. Shariah-compliant funds: A whole new world of investment.

Sadeghi M. 2011a. Investment opportunities and stock liquidity: evidence from DJIM index additions in the Gulf States. Investment Management and Financial Innovations. 8(1):50–59.

Sadeghi M. 2011b. Shariah-compliant Investment and Shareholders’ Value: An Empirical Investigation. Global Economy and Finance Journal. 4(1):44–61.

Sadeghi M. 2012. Are Faithful Investors Rewarded by the Market Place? Evidence from Australian Shariah-compliant Equities. Qualitative Research in Financial Markets, forthcoming.

Sadeghi M. 2008. Financial Performance of Shariah-Compliant Investment: Evidence from Malaysian Stock Market. International Research Journal of Finance and Economics. 20:15–26.


Cite this chapter as: Sufian F, Zulkhibri M, Majid A (2015). The nexus between economic freedom and Islamic bank performance: empirical evidence from the MENA banking sectors. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

The nexus between economic freedom and Islamic bank performance: Empirical evidence from the MENA banking sectorsFadzlan Sufian1, Muhamed Zulkhibri, Abdul Majid2

1Universiti Islam Antarabangsa Malaysia.2Economic Research and Policy Dept., Islamic Development Bank. Mailing address: 9th Floor, Economic Research and Policy Department, Islamic Development Bank, P.O Box 5925, Jeddah, 21432, Saudi Arabia, e-mails:[email protected]; [email protected], Tel: +966-2-646-6533; Fax: +966-2-506047132

All findings, interpretations, and conclusions are solely of the authors’ opinion and do not necessarily represent the views of the institutions.

Abstract - The present study provides new empirical evidence on the impact of economic freedom on Islamic banks’ performance. The empirical analysis focuses on Islamic banks operating in the MENA banking sectors during the period 2000–2008. We find that the larger, more diversified, and better capitalized Islamic banks tend to be relatively more profitable, while credit risk and expense preference behaviour seem to exert negative impact. The findings suggest that greater financial freedom positively influence the profitability of Islamic banks operating in the MENA banking sectors. Interestingly, the impact of monetary freedom is negative implying that higher (lower) monetary policy independence reduces (increases) Islamic banks’ profitability, providing support to the benefits of government interventions.

Keywords: economic freedom, Islamic banks, profitability, panel regression analysis, MENA

JEL Classification: G21; G28

1. IntroductionIslamic banking is a relatively recent addition to the global financial markets. Its conventional brick and mortar root can be traced back to the early 1960s when Myt Ghamar Bank was formed in Egypt in 1963. Between 1963 and 1971 the bank provided Muslims with a place to deposit their savings in accordance to the Syari’a principles1. Despite its humble beginning, Islamic banks have blossomed throughout the world and are looked upon as a viable alternative system which has many things to offer.

Although it was initially developed to fulfill the needs of Muslims, Islamic banking has now gained universal acceptance. According to El-Qorchi (2005), the number of Islamic financial institutions increased from a single institution in 1975 to approximately 486 financial institutions operating in more than 75 countries worldwide2. Total assets of Islamic financial institutions are estimated at US$250 billion and are tipped to be growing at 15% per year, three times the rate of conventional banks. The rapid growth rate confirms the growing importance of Islamic banking and finance in the global financial markets.

The Islamic banks operate in markets characterized by competition-inhibiting government regulation and in a protected banking environment. Islamic banking, being a participatory type of banking system, has entered on the global banking market in full force. In recent years, market conditions in Islamic banking have undergone extensive changes from both the demand and supply sides. On the demand side, customers have become more sophisticated, value-oriented, and price sensitive, while on the supply side, the globalization of financial markets has been accompanied by governmental deregulation, financial innovation, and automation.

These two factors have resulted in an increase in the number of competitors, cost reductions, and profit declines. The revolution in information technology, mainly in internet banking has enabled the larger financial institutions to penetrate markets and to increase their market share within both national and overseas markets by providing competitive products at lower prices. Furthermore, Islamic equity-type financial instruments are competing with conventional banking products and now face strong

Sufian et al.


competition from both banks and non-bank financial institutions. This also accentuates competition within the financial services industry.

It is reasonable to assume that these developments posed great challenges to Islamic banks as the environment in which they operates in has changed rapidly. This could sensibly have an impact on the determinants of their performance. Despite considerable development of the Islamic banking sector, empirical works on Islamic banks’ performance is still in its infancy. The knowledge of the underlying factors which influences the Islamic banking sector’s performance is essential given the growing importance of Islamic banking and finance in the global financial markets. It is therefore essential not only for the managers of the Islamic banks, but for numerous stakeholders such as the central banks, bankers associations, governments, and other financial authorities to help them identify and formulate policies to improve the performance of the Islamic banking sector, particularly in the MENA region3.

On the perspective of economic freedom, economic theories suggest that economic freedom tend to affect incentives, productive effort, and the effectiveness of resource use.4 Economists and economic historians have argued that since the time of Adam Smith, central ingredients for economic progress are the freedom to choose and supply resources, competition in business, trade with others, and secure property rights (North and Thomas, 1973). Within the context of the MENA region, it can be observed from Table 1 that the region has achieved modest improvement in economic freedom during the year 20105. It can be seen from Table 1 that Bahrain retained the top ranking within the region and managed to be ranked in the world Top 10, while Qatar ranks in the world top 30.

The ongoing transformations of innovative and reform-oriented states such as Bahrain, Qatar, Kuwait, and Oman may pave the way for a more robust and dynamic regional economic growth in the region. On different scale, Jordan and Oman registered the highest gains in economic freedom and Qatar’s improvement to 70.5, moved it from the category of “moderately free” to “mostly free”, while Syria’s improvement lifted its designation from “repressed” economy to “mostly unfree”. However, no other MENA countries are rated as having “mostly free” economies. Nearly half of the region falls into the “mostly unfree” category and two countries, namely Libya and Iran, ranked among the world’s most repressed economies. The institutional problems, such as lack of investment and financial freedom and weak systems for protecting property rights and preventing corruptions continue to degrade the region’s overall economic freedom and economic potential.

The purpose of the present paper is to extend the earlier works on the performance of the Islamic banking sector in the MENA region and to establish empirical evidence on the impact of economic freedom. The paper also investigates to what extent the performance of Islamic banks is influenced by internal factors (i.e. bank-specific characteristics) and to what extent by external factors (i.e. macroeconomic conditions and economic freedom). Although studies on economic freedom is vast in the literature (e.g. Heckelman and Knack, 2009; Altman, 2008; Powell, 2003; Adkins et al. 2002; De Haan and Sturm, 2000; Heckelman and

Stroup, 2000; Heckelman, 2000; De Haan and Siermann, 1998), these studies have mainly examined the impact of economic freedom on economic growth. On the other hand, virtually nothing has been published to examine the impact of economic freedom on the performance of the conventional or Islamic banking sectors. This limitation is somewhat surprising given the importance of bank lending in promoting economic growth and development (e.g. Ben Naceur and Ghazouani, 2007; Beck and Levine, 2004; Rajan and Zingales, 1998) and given the impact that economic freedom is likely to have on the banking sector.

The paper is divided into five sections. The following section presents the literature review. Section 3 describes the data, sources, and empirical settings. In section 4 we present the results and finally, section 5 concludes.

2. Review of the literatureThe empirical evidence on the performance of the conventional banking sectors is extensive. To date, the numerous studies have mainly focused on the U.S. banking sector (e.g. DeYoung and Rice, 2004; Stiroh and Rumble, 2006; Hirtle and Stiroh, 2007; Tregenna, 2009) and the banking sectors of the western and developed countries (e.g. Williams, 2003; Pasiouras and Kosmidou, 2007; Kosmidou et al. 2007; Hawtrey and Liang, 2008; Kosmidou, 2008; Kosmidou and Zopounidis, 2008; Athanasoglou et al. 2008; Albertazzi and Gambacorta, 2008; Kasman et al. 2010). On the other hand, empirical works on the Islamic banking sector is still in its infancy. Typically, studies on Islamic bank performance have focused on theoretical issues and the empirical works have relied mainly on the analysis of descriptive statistics rather than rigorous statistical estimation (El-Gamal and Inanoglu, 2005).

Hussein (2003) provides an analysis of the cost efficiency features of Islamic banks in Sudan. By using the stochastic cost frontier approach, he estimates cost efficiency for a sample of 17 banks over the period 1990 and 2000. The results show large variations in the cost efficiency of Sudanese banks with the foreign owned banks being the most efficient, while the state owned banks being the most cost inefficient. The empirical findings suggest that the small banks are relatively more efficient compared to their large bank counterparts. In addition, banks with a higher proportion of musharakah and mudharabah finance relative to total assets tend to exhibit efficiency advantages.

In another study on the Sudanese Islamic banking sector, Hassan and Hussein (2003) examine the efficiency of the Sudanese banking system during the period of 1992 and 2000. They employed a variety of parametric (cost and profit efficiencies) and non-parametric Data Envelopment Analysis (DEA) methods to a panel of 17 Sudanese banks. They found that the average cost and profit efficiencies under the parametric method were 55% and 50% respectively, while it was 23% under the non-parametric method. During the period under study, they suggest that the Sudanese banking system has exhibited 37% allocative efficiency and 60% technical efficiency, suggesting that the overall cost inefficiency of the Sudanese Islamic banks were mainly due to technical (managerially related) rather than allocative (regulatory).


The nexus between economic freedom and Islamic bank performance

Tabl

e 1.

Eco

nom

ic fr

eedo

m in

dex

for M

ENA

.

Cou

ntry

Wor

ld

Ran

kR

egio

nal

Ran

king

2011

O

vera

ll

Scor

e

Cha

nge

fr

om

2010

Bus

ines

s

Free

dom

Trad

e

Free

dom

Fisc

al

Free

dom

Gov

ernm

ent

Spen

ding

Mon

etar

y

Free

dom

Inve

stm

ent

Free

dom

Fina

ncia

l Fr

eedo

mPr

oper

ty

Rig

hts

Free

dom

fr

om

Cor

rupt

ion

Labo

ur

Free

dom

Bahr

ain

10

177

.71.

477

.482

.899

.880

.274

.075

.080

.060

.051

.097

.0Q

atar

27

270

.51.

570

.382

.499

.878

.171

.945

.050

.070

.070

.067

.0O

man

34

369

.82.

169

.483

.698

.568

.169

.555

.060

.050

.055

.089

.1Jo

rdan

38

468

.92.

865

.878

.892

.760

.981

.470

.060

.055

.050

.074

.2Is

rael

43

568

.50.

866

.187

.862

.344

.878

.480

.070

.070

.061

.064

.3U

AE

47

667

.80.

567

.382

.699

.979

.176

.535

.050

.050

.065

.072

.4Sa

udi

Ara

bia

54

766

.22.

086

.182

.299

.474

.664

.340

.050

.045

.043

.077

.0

Kuw

ait

61

864

.9-2

.864

.481

.699

.969

.769

.355

.050

.050

.041

.067

.9Le

bano

n 8

9 9

60.1

0.6

57.5

80.5

91.0

64.9

77.7

60.0

60.0

25.0

25.0

59.0

Mor

occo

93

1059

.60.

475

.775

.867

.874

.676

.565

.060

.040

.033

.027

.2Eg

ypt

96

1159

.10.

164

.574

.089

.665

.360

.865

.050

.040

.028

.053

.6Tu

nisi

a10

012

58.5

-0.5

80.2

53.5

73.7

77.6

77.3

35.0

30.0

50.0

42.0

65.7

Yem

en12

713

54.2

-0.2

73.7

81.6

83.2

44.5

82.3

45.0

30.0

30.0

21.0

50.9

Alg

eria

132

1452

.4-4

.569

.472

.883

.562

.475

.420

.030

.030

.028

.052

.9Sy

ria

140

1551

.31.

955

.965

.484

.685

.369

.720

.020

.030

.026

.055

.8Ir

an17

116

42.1

-1.3

69.4

44.8

81.1

76.0

60.7

0.0

10.0

10.0

18.0

50.7

Liby

a17

317

38.6

-1.6

20.0

85.0

80.3

44.5

71.0

10.0

20.0

10.0

25.0

20.0

Not

e: E

ach

one

of th

e 10

free

dom

is g

rade

d us

ing

a 0

to 1

00 s

cale

, whe

re 1

00 re

pres

ents

the

max

imum

free

dom

. A s

core

of 1

00 s

igni

fies

an e

cono

mic

env

iron

men

t or

set o

f pol

icie

s th

at is

mos

t con

duci

ve to

eco

nom

ic fr

eedo

m. M

any

of th

e 10

free

dom

s ar

e ba

sed

on q

uant

itat

ive

data

that

are

con

vert

ed d

irec

tly in

to a

sco

re.

Sour

ce: T

he H

erit

age

Foun

dati

on.

Sufian et al.


El-Gamal and Inanoglu (2004) employ the stochastic frontier approach to estimate the cost efficiency of Turkish banks over the period 1990–2000. The study compared the cost efficiencies of 49 conventional banks with four Islamic special finance houses (SFHs). The Islamic firms comprised around 3% of the Turkish banking market. Overall, they suggest that the Islamic financial institutions to be the most efficient. This could be explained by their emphasis on Islamic asset-based financing which led to low non-performing loans ratios.

The study by Hassan (2006) is among the few performed to examine the efficiency of Islamic banks in a cross-country setting. He employs both the parametric (Stochastic Frontier Approach) and non-parametric (Data Envelopment Analysis) methods to examine the efficiency of banks in the sample. The findings indicate that during the period 1993–2001, Islamic banks have exhibited a relatively higher profit efficiency compared to cost efficiency. He suggests that the main source of inefficiency is allocative rather than technical. The results indicate that the overall inefficiency was output related. The results indicate that on average the Islamic banking industry is relatively less efficient compared to their conventional counterparts.

While the above outlines the literature that employs advanced modelling techniques to evaluate Islamic banks’ performance, one should also note that there is a growing body of literature that covers the general performance features of Islamic banks. Such studies include those by Hassan and Bashir (2003) who look at the determinants of Islamic banks’ performance and show that Islamic banks to be just as efficient as their conventional bank peers if one uses standard accounting measures such as the cost-to-income ratio. Other studies that followed similar approach are those by Sarker (1999) who examines the performance and operational efficiency of Bangladeshi Islamic banks, while Bashir (1999) investigates the risk and profitability of two Sudanese banks.

Bashir (1999) and Bashir (2001) performed regression analyses to examine the underlying determinants of Islamic banks’ performance. By employing bank level data from the Middle East, the results indicate that the performance of banks, in terms of profits, is mostly generated from overhead, customer short-term funding, and non-interest earning assets. Furthermore, Bashir (2001) claimed that since deposits in Islamic banks are treated as shares, reserves held by banks propagate negative impacts such as reducing the amount of funds available for investment. In essence, the findings from this literature are that Islamic banks are at least as efficient as their conventional bank counterparts and in most cases are relatively more efficient.

The above literature reveals the following research gaps. First, the majority of these studies have concentrated on the conventional banking sectors and the banking sectors of the western and developed countries. Second, empirical evidence on the developing countries banking sectors, particularly the Islamic banking sectors are relatively scarce. Finally, virtually nothing has been published to examine the impact of economic freedom on the Islamic banking sector. In light of these knowledge gaps, the present paper provides new empirical evidence on the impact of economic freedom on the performance of Islamic banks operating in the MENA countries banking sectors.

3. Data and methodologyThe present study employs an unbalanced annual bank level data of all Islamic banks operating in the MENA countries covering the period 2000–2008. The financial statements of Islamic banks operating in the MENA banking sectors are collected from the Bankscope database of Bureau van Dijk’s company. The macroeconomic variables are retrieved from the IMF Financial Statistics (IFS) and the World Bank World Development Indicator (WDI) databases while economic freedom variables are extracted from The Heritage Foundation.

Measure of performanceFollowing Ben Naceur and Goaied (2008), Kosmidou (2008), and Abbasoglu et al. (2007) among others, the dependent variable used in this study is Return on Assets (ROA). ROA shows the profit earned per dollar of assets and most importantly, reflects management ability to utilize banks financial and real investment resources to generate profits (Hassan and Bashir, 2003). For any bank, ROA depends on the bank’s policy decisions as well as other uncontrollable factors relating to the economy and government regulations. Rivard and Thomas (1997) suggest that bank profitability is best measured by ROA, since it is not distorted by high equity multipliers and represents a better measure of the ability of firms to generate returns on its portfolio of assets.

Internal determinants The bank specific variables included in the regression models are LLP/TL (loans loss provisions divided by total loans), EQASS (book value of stockholders’ equity as a fraction of total assets), NIE/TA (total overhead expenses divided by total assets), LOANS/TA (total loans divided by total assets), and LNTA (log of total assets).

The ratio of loan loss provisions to total loans (LLP/TL) is incorporated as an independent variable in the regression analysis as a proxy of credit risk. The coefficient of the LLP/TL variable is expected to enter the regression models with a negative sign. In this vein, Miller and Noulas (1997) point out that the greater the exposure of banks to high risk loans, the higher would be the accumulation of unpaid loans and profitability would be lower. Miller and Noulas (1997) suggest that decline in loan loss provisions are in many instances the primary catalyst for increases in profit margins. Furthermore, Thakor (1987) also suggests that the level of loan loss provisions is an indication of the bank’s asset quality and signals changes in future performance.

The EQASS variable is included in the regression models to examine the relationship between profitability and bank capitalization. Strong capital structure is essential for banks in developing economies, since it provides additional strength to withstand financial crises and increased safety for depositors during unstable macroeconomic conditions. Furthermore, lower capital ratios in banking imply higher leverage and risk and therefore greater borrowing costs. Thus, the profitability level should be higher for the better capitalized bank.

The ratio of non-interest expenses over total assets, NIE/TA, is used to provide information on the variations of bank



operating costs. The variable represents total amount of wages and salaries, as well as the costs of running branch office facilities. The relationship between the NIE/TA variable and profitability levels is expected to be negative, because the more productive and efficient banks should be able to keep their operating costs low. Furthermore, the usage of new electronic technology, like ATMs and other automated means of delivering services, may have caused expenses on wages to fall (as capital is substituted for labor).

An important decision that the managers of Islamic banks must take refers to the liquidity management and specifically to the measurement of their needs related to the process of deposits and loans. For that reason, the ratio of total loans to total assets (LOANS/TA) is used as a measure of liquidity. Higher figures denote lower liquidity. Without the required liquidity and funding to meet obligations, a bank may fail. Thus, in order to avoid insolvency problems, banks often hold liquid assets, which can be easily converted to cash. However, liquid assets are usually associated with lower rates of return. It would therefore reasonable to expect higher liquidity to be associated with lower bank profitability.

The LNTA variable is included in the regression models as a proxy of size to capture for the possible cost advantages associated with size (economies of scale). In the literature, mixed relationships are found between size and profitability, while in some cases a U-shaped relationship is observed. LNTA is also used to control for cost differences related to bank size and for the greater ability of the large bank to diversify. In essence, LNTA may lead to positive effects on bank profitability if there are significant economies of scale. On the other hand, if increased diversification leads to higher risks, the variable may exhibit negative effects.

External determinants If analysis is done in a static setting, they may fail to capture developments in the regulatory environment and in the marketplace, which may have changed the underlying production technology and the associated production functions. Furthermore, different banking forms could demonstrate different reactions to environmental changes. Hence, the change in the financial landscape and structure, etc., may vary across banking groups (Saunders et al. 1990; Button and Weyman-Jones, 1992; Berger, 1995). To measure the relationship between economic and market conditions and Islamic banks’ performance, LNGDP, INFL, CR3, and Z-SCORE variables are used.

Gross domestic product (GDP) is among the most commonly used macroeconomic indicator to measure total economic activity within an economy. The GDP is expected to influence numerous factors relating to the supply and demand for loans and deposits. Favourable economic conditions will affect positively on the demand and supply of banking services, but will have either positive or negative influence on bank profitability levels.

Another important macroeconomic condition which may affect both the costs and revenues of banks is the inflation rate (INFL). Staikouras and Wood (2003) points out that inflation may have direct effects i.e. increase in the price

of labour and indirect effects i.e. changes in interest rates and asset prices on the profitability of banks. Perry (1992) suggests that the effects of inflation on bank performance depend on whether the inflation is anticipated or unanticipated. In the anticipated case, the profit rates are adjusted accordingly resulting in revenues to increase faster than costs subsequently positive impact on bank profitability. On the other hand, in the unanticipated case, banks may be slow to adjust their interest rates resulting in a faster increase of bank costs compared to bank revenues and consequently negative effects on bank profitability6.

To examine the impact of concentration on Islamic banks’ performance, the CR3 variable is introduced in the regression models. The CR3 ratio is calculated as the total assets held by the three largest banks in the country. The variable is used to examine the impact of asset concentration in the national banking sector on the profitability of Islamic banks. The Structure-Conduct-Performance (SCP) theory posits that banks in a highly concentrated market tend to collude and therefore earn monopoly profits (Molyneux et al. 1996). Berger (1995) points out that the relationship between bank concentration and performance in the U.S. depends critically on what other factors are held constant. According to the industrial organization literature, a positive impact is expected under both collusion and efficiency views (Goddard et al. 2001).

The Z-Score (Z-SCORE) variable is used as a proxy of bank soundness. The index measures how many standard deviations a bank is away from exhausting its capital base (a distance-to-default measure). The Z-Score is a popular measure of soundness because it combines banks’ buffers (capital and profits) with the risks they face in a way that is grounded in theory (Cihak et al. 2009). A higher Z-Score implies a lower probability of insolvency, providing a more direct measure of soundness than, for example, simple leverage measures (Cihak et al. 2009). This index combines in a single indicator: (i) profitability, given by a period average return on assets (ROA); leverage measure, given by the period average equity-to-asset ratio (K) (equity here is defined as total equity from the balance sheet of a bank); and return volatility, given by the period standard deviation of ROA (Vol. (ROA)) i.e. Z ROA K

Vol ROA.( )= + where ROA (profitability) is a period average of ROA, K (leverage measure) is the period average equity-to-asset ratio, and Vol. (ROA) is the return volatility given by the period standard deviation of ROA. A higher (lower) Z-SCORE indicates lower (higher) risk (De Nicolo et al. 2003).

Economic freedom measurements In simple terms, economic freedom is a conceptual measure of the private ownership and market allocation of resources, in lieu of government ownership and control. Expressing the sentiment of many, including the originators of the economic freedom index, Berggren (2003) defines economic freedom as “the degree to which an economy is a market economy—that is, the degree to which it entails the possibility of entering into voluntary contracts within the framework of a stable and predictable rule of law that upholds contracts and protects private property, with a limited degree of interventionism in the form of government ownership, regulations, and taxes”.

Sufian et al.


In regression model 2, OVER_FREE is introduced to examine the impact of overall economic freedom on the performance of the Islamic banks operating in the MENA banking sectors. OVER_FREE is the overall economic freedom index and is defined by multiple rights and liberties. The index uses 10 specific freedoms, namely Business freedom, Trade freedom, Fiscal freedom, Government size, Monetary freedom, Investment freedom, Financial freedom, Property rights, Labor freedom, and Freedom from corruption.

Besides the overall economic freedom index, we have selected three other indices which are closely related to the financial sector. These include BUSI_FREE, MONE_FREE, and FINA_FREE indices. BUSI_FREE is the business freedom index. The index measures how free entrepreneurs are to start businesses, how easy it is to obtain licenses, and the ease of closing a business. Impediments to any of these three activities are deterrents to businesses and therefore to job creations. MONE_FREE is the monetary freedom index. The index combines a measure of price stability with an assessment of price controls. Both inflation and price control distorts market activity. Price stability without microeconomic intervention is an ideal state of a free market. FINA_FREE is the financial freedom index. The index is a measure of banking security as well as independence from government’s control. State ownership of banks and other financial institutions such as insurer and capital markets is an inefficient burden and political favoritism has no place in a free capital market. All these indices have 0 to 100 scales, where 100 represents maximum freedom. A score of 100 signifies an economic environment, or set of policies that is most conducive to economic freedom.

Finally, CORR_FREE is introduced in regression model 6 to assess the impact corruption on the performance of Islamic banks. CORR_FREE is the freedom from corruption index. The index is based on quantitative data that assess the perception of corruption in the business environment, including levels of governmental, legal, judicial, and administrative corruption. Similar to the BUSI_FREE, MONE_FREE, and FINA_FREE indices, the CORR_FREE index also takes a value of between 0 and 100, where 100 represent the maximum freedom. Table 1 contains the summary statistics of the variables used to proxy profitability and its determinants.

Econometric specificationSince the panel data cover many heterogenous banks and time periods, the possible correlation between the regressors and bank-specific effects, the endogeneity of regressors with respect to idiosyncratic shock and the heteroscedasticity of the disturbance term (idioscyncratic shock) would result in a biased and inconsistent estimation with Ordinary Least Square (OLS) estimation technique. The OLS estimator would result in an upward estimate of the coefficient while the within-group estimator would be downward biased (Blundell et al. 1992). A natural technique for dealing with variable that are correlated with the error term is to instrument them.

Berger et al. (2000) points out that bank profitability tend to persist over time reflecting impediments to market competition, informational opacity, and sensitivity to macroeconomic shocks. Furthermore, Garcia-Herrero

et al. (2009) suggest that potential endogeneity could be a problem when assessing bank profitability determinants. For instance, the more profitable banks may have sufficient resources to provision for their non-performing loans. The more profitable banks may also find it easier to increase their customer base via successful advertising campaigns and could hire the most skilled personnel, and therefore enhances their profitability levels (Garcia-Herrero et al. 2009).

Arellano and Bond (1991) proposed an efficient Generalized Methods of Moment (GMM) estimator that uses instruments of which the validity is based on the orthogonality between the lagged values of the dependent variable and the errors. The technique eliminates the unobserved bank heterogeneity by estimating the equation in first-differences and to control for possible endogeneity problem by using the model’s variables lagged by one or more periods as instruments. We employ the GMM estimator as proposed by Arellano and Bond (1991) to ensure efficiency and consistency of the estimations. Therefore, a dynamic GMM model is adopted via the inclusion of a lagged dependent variable among the regressors to capture the persistence of bank profitability over time reflecting impediments to market competition, informational opacity, and/or sensitivity to regional/macroeconomic shocks (Berger et al. 2000).

The baseline model is formulated as follows:

X M Eit i t it t t t i it, 1 ∑∑∑π a λπ b γ δ µ υ e= + + + + + + +- (1)

where i = 1, 2, …, N (number of firms) and t = 1, 2, …, T (time period). In the specification, pit

denotes the profitability of bank i at time t; pt-1 indicates a one period lagged profitability; Xi,t is vector exogenous bank-specific regressors: Mt is a vector of country specific variables; Ei is a vector of country specific economic freedom variables; mt is a time fixed effect; vi

is an unobserved banks’ fixed effect; ei,t is a serially uncorrelated error term.

We use several tests proposed by Arellano and Bond (1991) to check whether the instruments are properly chosen and the assumptions underlying the model holds. Our estimations rely on the fact that the disturbances follow an MA(1) process and there is no second order autocorrelation (m2) together with Sargan/Hansen tests of over-identifying restriction (J-test) to examine the validity of the instruments used in the regression models.

Extending Eq. (1) to reflect the variables as described in Table 2, the baseline model is formulated as follows:

ROA LOANS TA LNTALLP TL NIE TA EQASS

LNGDP INFL CR Z SCOREOVER FREE BUSI FREE

MONE FREE FINA FREECORR FREE

// /

3_ _

_ __

jt jt jt

jt jt jt

t t t t

t t

t t

t jt

0 1 2

3 4 5

6 7 8 9

9 10

11 12

13

b b bb b bb b b bb bb bb e

= + ++ + ++ + + + -+ ++ ++ + (2)

Table 3 provides information on the degree of correlation between the explanatory variables used in the panel regression analysis. In general, the matrix shows that the



Tabl

e 2.

Des

crip

tive

of t

he v

aria

bles

use

d in

the

regr

essi

on m

odel

s.

Vari

able

Des

crip

tion

Mea

nSt

d. D

ev.

Sour

ces/

Dat

abas

e

RO

AA

pro

xy m

easu

re o

f ban

k pr

ofita

bilit

y m

easu

red

as th

e re

turn

on

aver

age

to

tal a

sset

s of

the

bank

in y

ear t

. 2

.577

3.7

98Ba

nkSc

ope

Inde

pend

ent

Inte

rnal

Fac

tors

LLP/

TLLo

an lo

ss p

rovi

sion

s/ to

tal l

oans

. An

indi

cato

r of c

redi

t ris

k, w

hich

sho

ws

ho

w m

uch

a ba

nk is

pro

visi

onin

g in

yea

r t re

lati

ve to

its

tota

l loa

ns.

8.3

2114

.13

Bank

Scop

e

EQA

SSA

mea

sure

of b

ank’

s ca

pita

l str

engt

h in

yea

r t, c

alcu

late

d as

equ

ity/

tota

l as

sets

. Hig

h ca

pita

l ass

et ra

tio

is a

ssum

ed to

be

indi

cato

r of l

ow le

vera

ge

and

ther

efor

e lo

wer

ris

k.

21.2

2723

.458

Bank

Scop

e

NIE

/TA

Cal

cula

ted

as n

on-in

tere

st e

xpen

se/

tota

l ass

ets

and

prov

ides

info

rmat

ion

on

the

effici

ency

of t

he m

anag

emen

t reg

ardi

ng e

xpen

ses

rela

tive

to th

e

asse

ts in

yea

r t. H

ighe

r rat

ios

impl

y a

less

effi

cien

t man

agem

ent.

3.7

55 3

.247

Bank

Scop

e

LOA

NS/

TAA

mea

sure

of l

iqui

dity

, cal

cula

ted

as to

tal l

oans

/ to

tal a

sset

s. T

he ra

tio

in

dica

tes

wha

t per

cent

age

of th

e as

sets

of t

he b

ank

is ti

ed u

p in

loan

s in

ye

ar t.

48.5

8323

.706

Bank

Scop

e

LNTA

The

natu

ral l

ogar

ithm

of t

he a

ccou

ntin

g va

lue

of th

e to

tal a

sset

s of

the

ba

nk in

yea

r t.

8.1

17 2

.666

Bank

Scop

e

Exte

rnal

Fac

tors

LNG

DP

Nat

ural

loga

rith

m o

f gro

ss d

omes

tic

prod

ucts

. 6

.057

4.2

37IM

F

Inte

rnat

iona

l Fi

nanc

ial S

tati

stic

sIN

FLTh

e ra

te o

f infl

atio

n. 2

.246

2.1

54IM

F

Inte

rnat

iona

l Fi

nanc

ial S

tati

stic

sC

R3

The

thre

e la

rges

t ban

ks a

sset

con

cent

rati

on ra

tio.

0.7

30 0

.151

IMF

In

tern

atio

nal

Fina

ncia

l Sta

tist

ics

Z-SC

OR

ETh

e Z-

Scor

e in

dex.

Is u

sed

as a

pro

xy m

easu

re o

f the

ban

king

sec

tor’

s ri

sk to

de

faul

t.11

.027

6.8

32IM

F

Inte

rnat

iona

l Fi

nanc

ial S

tati

stic

s

Econ

omic

Fre

edom

OV

ER_F

REE

Ove

rall

econ

omic

free

dom

is d

efine

d by

mul

tipl

e ri

ghts

and

libe

rtie

s ca

n

be q

uant

ified

as

an in

dex

of le

ss a

bstr

act c

ompo

nent

s. T

he in

dex

uses

10

sp

ecifi

c fr

eedo

ms,

som

e as

com

posi

tes

of e

ven

furt

her d

etai

led

and

qu

anti

fiabl

e co

mpo

nent

s.

60.4

6412

.575

Her

itag

e Fo

unda

tion

(w

ww

.her

itag

e.or

g/in

dex)

(Con

tinu

ed)

Sufian et al.


Tabl

e 2.

(Co

ntin

ued)

BUSI

_FR

EEBu

sine

ss fr

eedo

m m

easu

res

how

free

ent

repr

eneu

rs a

re to

sta

rt b

usin

esse

s,

how

eas

y it

is to

obt

ain

licen

ses,

and

the

ease

of c

losi

ng a

bus

ines

s.

Impe

dim

ents

to a

ny o

f the

se th

ree

acti

viti

es a

re d

eter

rent

s to

bus

ines

s

and

ther

efor

e to

job

crea

tion

.

64.8

9014

.763

Her

itag

e Fo

unda

tion

(w

ww

.her

itag

e.or

g/in

dex)

MO

NE_

FREE

Mon

etar

y fr

eedo

m c

ombi

nes

a m

easu

re o

f pri

ce s

tabi

lity

wit

h an

ass

essm

ent

of p

rice

con

trol

s. B

oth

infla

tion

and

pri

ce c

ontr

ols

dist

ort m

arke

t act

ivit

y.

Pric

e st

abili

ty w

itho

ut m

icro

econ

omic

inte

rven

tion

is th

e id

eal s

tate

for t

he

free

mar

ket.

75.8

2912

.318

Her

itag

e Fo

unda

tion

(w

ww

.her

itag

e.or

g/in

dex)

FIN

A_F

REE

Fina

ncia

l fre

edom

is a

mea

sure

of b

anki

ng s

ecur

ity

as w

ell a

s in

depe

nden

ce

from

gov

ernm

ent c

ontr

ol. S

tate

ow

ners

hip

of b

anks

and

oth

er fi

nanc

ial

inst

itut

ions

suc

h as

insu

rer a

nd c

apit

al m

arke

ts is

an

ineffi

cien

t bur

den,

an

d po

litic

al fa

vori

tism

has

no

plac

e in

a fr

ee c

apit

al m

arke

t.

47.3

5426

.329

Her

itag

e Fo

unda

tion

(w

ww

.her

itag

e.or

g/in

dex)

CO

RR

_FR

EEFr

eedo

m fr

om c

orru

ptio

n is

bas

ed o

n qu

anti

tati

ve d

ata

that

ass

ess

the

pe

rcep

tion

of c

orru

ptio

n in

the

busi

ness

env

iron

men

t, in

clud

ing

leve

ls o

f go

vern

men

tal l

egal

, jud

icia

l, an

d ad

min

istr

ativ

e co

rrup

tion

.

46.1

3424

.140

Her

itag

e Fo

unda

tion

(w

ww

.her

itag

e.or

g/in

dex) correlation between the bank specific variables are not

strong, implying that multicollinearity problems are not severe. Kennedy (2008) points out that multicollinearity is a problem when the correlation is above 0.80 which is not the case here. However, it is worth noting that the LNGDP variable is highly correlated to most of the economic freedom variables. To address this concern, we have also estimated all regression models by excluding the macroeconomic variables. Furthermore, due to the high correlation between the economic freedom variables, the regression models are estimated by including the each economic freedom indicator at a time, rather than estimating all economic freedom variables concurrently.

4. Empirical findingsThe regression results focusing on the relationship between bank profitability and the explanatory variables are presented in Table 4. The reliability of our econometric methodology depends critically on the validity of the instruments, which can be evaluated with Sargan’s test of overidentifying restrictions, asymptotically distributed as c2 in the number of restrictions. A rejection of the null hypothesis that instruments are orthogonal to the errors would indicate that the estimates are not consistent (Baum et al. 2010)7. We also present test statistics for the first and second order serial correlations in the error process. In a dynamic panel data context, second order serial correlation should not be present if the instruments are appropriately uncorrelated with the errors (Baum et al. 2010).

It can be observed from Table 4 that for all the estimated models, the Sargan statistics for overidentifying restrictions and the Arrelano–Bond AR(2) tests shows that our instruments are appropriately orthogonal to the error and no second order serial correlation is detected respectively. Furthermore, the highly significant of the lagged dependent variable’s coefficient confirms the dynamic character of the model specification, thus justifying the use of dynamic panel data model estimation.

Concerning the liquidity results, the empirical findings suggest a negative sign of the coefficient of the LOANS/TA in the baseline regression model. As higher (lower) figures of the ratio denote lower (higher) liquidity levels, the results imply that the relatively less (more) liquid banks tend to exhibit higher (lower) profitability levels. On the other hand, the empirical findings also suggest that the coefficient of the variable is positive when we control for overall economic freedom, business freedom, financial freedom, and freedom from corruption. However, the results should be interpreted with caution since the coefficient of the variable is not significant at any conventional levels in any of the regression models estimated.

It can be observed from Table 4 that the coefficient of the LNTA variable entered the baseline regression model with a negative sign and is statistically significant when we control for economic and financial market conditions lending support to Spathis et al. (2002), Dogan and Fausten (2003), and Kosmidou (2008). Moreover, the earlier studies have concluded that marginal cost savings could be achieved by increasing the size of the banking firm, especially as markets develop (Berger et al. 1987; Boyd and Runkle, 1993; Miller and Noulas, 1997; Athanasoglou et al. 2008). In this vein,



Tabl

e 3.

Cor

rela

tion

mat

rix

for t

he e

xpla

nato

ry v

aria

bles

.

The

nota

tion

use

d in

the

tabl

e be

low

is d

efine

d as

follo

ws:

LO

AN

S/TA

is u

sed

as a

pro

xy m

easu

re o

f loa

ns in

tens

ity,

cal

cula

ted

as to

tal l

oans

div

ided

by

tota

l ass

ets;

LN

TA is

a

prox

y m

easu

re o

f siz

e, c

alcu

late

d as

a n

atur

al lo

gari

thm

of t

otal

ban

k as

sets

; LLP

/TL

is a

mea

sure

of b

ank

risk

cal

cula

ted

as th

e ra

tio

of to

tal l

oan

loss

pro

visi

ons d

ivid

ed

by to

tal l

oans

; NII

/TA

is a

mea

sure

of b

ank

dive

rsifi

cati

on to

war

ds n

on in

tere

st in

com

e, c

alcu

late

d as

tota

l non

-inte

rest

inco

me

divi

ded

by to

tal a

sset

s; N

IE/T

A is

a p

roxy

m

easu

re fo

r co

sts,

cal

cula

ted

as n

on-in

tere

st e

xpen

ses

divi

ded

by to

tal a

sset

s; E

QA

SS is

a m

easu

re o

f cap

ital

izat

ion,

cal

cula

ted

as b

ook

valu

e of

sha

reho

lder

s eq

uity

as

a fr

acti

on o

f tot

al a

sset

s; L

NG

DP

is n

atur

al lo

g of

gro

ss d

omes

tic

prod

ucts

; IN

FL is

the

rate

of i

nflat

ion;

OV

ER_F

REE

is th

e ov

eral

l eco

nom

ic fr

eedo

m in

dex;

BU

SI_F

REE

is

the

busi

ness

free

dom

inde

x; M

ON

E_FR

EE is

the

mon

etar

y fr

eedo

m in

dex;

FIN

A_F

REE

is th

e fin

anci

al fr

eedo

m in

dex;

CO

RR

_FR

EE is

the

free

dom

from

cor

rupt

ion

inde

x.

LLP/

TL

EQA

SSN

IE/T

ALO

AN

S/

TALN

TALN

GD

PIN

FLC

R3

Z-SC

OR

EO

VER

_ FR

EEB

USI

_ FR

EEM

ON

E_

FREE

FIN

A_

FREE

CO

RR

_ FR

EE

LLP/

TL1.

000

0.49

3**

0.04

8-0

.369

**-0

.293

**-0

.169

*-0

.237

**-0

.097

-0.0

520.

044

0.09

50.

160*

0.03

10.

059

EQA

SS1.

000

0.40

7**

-0.0

26-0

.555

**-0

.479

**-0

.432

**0.

007

-0.1

41*

0.48

4**

0.51

4**

0.41

9**

0.49

6**

0.38

7**

NIE

/TA

1.00

0-0

.204

**-0

.265

**-0

.068

-0.1

290.

077

-0.3

26**

0.14

8*0.

238*

*0.

058

0.20

2**

0.05

2LO

AN

S/TA

1.00

00.

193

0.02

50.

036

-0.0

15 0

.109

-0.0

32-0

.094

-0.0

27-0

.117

0.06

7LN

TA1.

000

0.67

0**

0.61

9**

-0.0

920.

184*

*-0

.624

**-0

.636

**-0

.622

**-0

.713

**-0

.476

**LN

GD

P1.

000

0.75

4**

-0.1

30*

0.07

6-0

.841

**-0

.759

**-0

.791

-0.8

47**

-0.6

57**

INFL

1.00

00.

142*

0.04

6-0

.324

**-0

.094

-0.5

69**

-0.1

35*

-0.1

40*

CR

31.

000

-0.1

61**

-0.0

840.

002

-0.1

000.

181*

*-0

.177

**Z-

SCO

RE

1.00

00.

000

-0.1

420.

045

-0.2

63**

0.22

0**

OV

ER_F

REE

1.00

00.

831*

*0.

815*

*0.

846*

*0.

869*

*BU

SI_F

REE

1.00

00.

639*

*0.

794*

*0.

732*

*M

ON

E_FR

EE1.

000

0.68

1**

0.61

8**

FIN

A_F

REE

1.00

00.

608*

*C

OR

R_F

REE

1.00

0

Not

e: T

he ta

ble

pres

ents

the

resu

lts

from

Pea

rson

cor

rela

tion

coe

ffici

ents

. **

and

* in

dica

tes

sign

ifica

nce

at 1

% a

nd 5

% le

vels

, res

pect

ivel

y.

Sufian et al.


Eichengreen and Gibson (2001) suggest that the effect of a growing bank’s size on performance may be positive up to a certain limit. Beyond this point the effect of size could be negative due to bureaucratic and other reasons.

As expected, the impact of credit risk (LLP/TL) is negative (statistically significant at the 10% level) suggesting that Islamic banks with higher credit risk tend to exhibit lower profitability levels. The results imply that Islamic banks should focus more on credit risk management, which has been proven to be problematic in the recent past. Serious banking problems have arisen from the failure of financial institutions to recognize impaired assets and create reserves to write off these assets. An immense help towards smoothing these anomalies would be provided by improving the transparency of the banking sector, which in turn will assist banks to evaluate credit risk more effectively and avoid problems associated with hazardous exposure.

Similarly, the empirical findings seem to suggest that expense preference behaviour measured by NIE/TA has consistently exhibit a negative relationship. The finding is in consonance with the bad management hypothesis of Berger and DeYoung (1997). Low measure of efficiency is a signal of poor senior management practices, which apply to input-usage and day-to-day operations. Clearly, efficient cost management is a prerequisite to improve the profitability of Islamic banks operating in the MENA banking sectors. Furthermore, most of the MENA countries banking sectors have not reached the maturity level required to link quality effects from increased spending to higher earnings.

Referring to the impact of capitalization, it can be observed from Table 4 that EQASS exhibits a positive relationship. The result is consistent with the previous studies by among others Isik and Hassan (2003), Goddard et al. (2004), and Kosmidou (2008) providing support to the argument that the well capitalized banks face lower costs of going bankrupt, thus lowers their cost of funding or that they have lower needs for external funding resulting in a higher profitability level. Nevertheless, strong capital structure is essential for banks in emerging economies since it provides additional strength to withstand financial crises and increased safety for depositors during unstable macroeconomic conditions (Sufian, 2009). However, it should be noted that the coefficient of the variable is not significant at any conventional levels in any of the regression models estimated.

The empirical findings seem to suggest that LNGDP has positive and significant impact on the profitability of Islamic banks operating in the MENA countries, lending support to the association between economic growth and banking sector’s performance. The high economic growth could have encouraged Islamic banks to lend more and improve the quality of their assets. The demand for financial services tends to grow as economies expand and societies become wealthier. Likewise, it can be observed from Table 4 that the coefficient of the INFL variable exhibits a positive sign in the baseline regression model, implying that during the period under study the levels of inflation have been anticipated by Islamic banks operating in the MENA banking sectors. This allows bank managements the opportunity to adjust the profit rates accordingly and consequently earn higher profitability.

Turning to the impact of banking sector’s concentration, it can be observed from Table 4 that the coefficient of the three banks concentration ratio (CR_3) has consistently exhibit a positive sign and becomes statistically significant when we control for freedom from corruption (CORR_FREE). Within the context of the MENA Islamic banking sector, the empirical findings clearly lend support to the SCP hypothesis. The SCP hypothesis states that banks in a highly concentrated market tend to collude and therefore earn monopoly profits (Short, 1979; Gilbert, 1984; Molyneux et al. 1996). It can be observed from Table 4 that the impact of banking sector risk (Z-SCORE) is positive and highly significant. The result is in consonance with the findings of among others Boyd and De Nicolo (2006) lending support to the stringent capital requirements of Basel II. From the policymaking point of view, the empirical findings calls for a more effective policymaker’s role in reducing excessive bank risk exposures and at the same time to induce more efficient risk management practices by Islamic banks operating in the MENA banking sectors.

Does greater economic freedoms foster bank performance?To address the issue whether economic freedom matters in determining the performance of Islamic banks operating in the MENA banking sectors, we re-estimate Eq. (2) to include the economic freedom indices variables discussed in Section 3. The results are presented in columns 3 to 7 of Table 4. As observed, the empirical findings presented in column 3 of Table 4 suggest that the coefficient of the overall economic freedom (OVER_FREE) variable is negative, but is not statistically significant at any conventional levels.

Concerning the impact of business freedom (BUSI_FREE) on the profitability of Islamic banks, the empirical findings presented in column 4 of Table 4 indicate that the coefficient of the BUSI_FREE variable is positive. The results imply that the greater ability to start, operate, and close businesses fosters the performance of Islamic banks. Clearly, the greater ability to set up new businesses in the MENA countries is a prerequisite to improve the performance of the Islamic banking sector.

Referring to the impact of monetary freedom (MONE_FREE), it is interesting to note that the coefficient of the variable is negative. If anything could be delved, the empirical findings indicate that higher (lower) government intervention in the market increases (reduces) the profitability of Islamic banks operating in the MENA banking sectors. A stable and reliable monetary policy is crucial to business environment, as it may help firms and societies to make investment, savings, and other long-term plans. High inflation rates not only confiscate wealth, but also distort pricing, misallocate resources, and raise the cost of doing business. Furthermore, the value of a country’s currency largely depends on the monetary policy of its government. A monetary policy that endeavors price stability and puts inflation at bay, enables firms to rely on the market prices for their future investments plans.

As expected, the coefficient of the financial freedom (FINA_FREE) variable entered the regression model with



Tabl

e 4.

Pan

el G

ener

aliz

ed M

etho

ds o

f Mom

ents

(G

MM

) re

gres

sion

resu

lts.

CRZ

SCO

RE

OVE

RFR

EE3

_RO

ALO

ANS

TALN

TALL

PTL

NIE

TAEQ

ASS

LNG

DP

INFL

BUSI

FREE

MO

NE

FREE

FIN

AFR

EECO

RRFR

EE/

//

__

__

tt

t8

99

jtjt

jtjt

jtjt

tt

tt

tt

jt

0

12

34

5

67

1011

1213

bb

b+

+-

+b

bb

bb

bb

bb

bb

be

=+

++

++

++

++

++

+

The

nota

tion

use

d in

the

tabl

e be

low

is d

efine

d as

follo

ws:

LO

AN

S/TA

is u

sed

as a

pro

xy m

easu

re o

f loa

ns in

tens

ity,

cal

cula

ted

as to

tal l

oans

div

ided

by

tota

l ass

ets;

LN

TA is

a p

roxy

mea

sure

of

siz

e, c

alcu

late

d as

a n

atur

al lo

gari

thm

of t

otal

ban

k as

sets

; LLP

/TL

is a

mea

sure

of b

ank

risk

cal

cula

ted

as th

e ra

tio

of to

tal l

oan

loss

pro

visi

ons

divi

ded

by to

tal l

oans

; NII

/TA

is a

mea

sure

of

ban

k di

vers

ifica

tion

tow

ards

non

inte

rest

inco

me,

cal

cula

ted

as to

tal n

on-in

tere

st in

com

e di

vide

d by

tota

l ass

ets;

NIE

/TA

is a

pro

xy m

easu

re fo

r co

sts,

cal

cula

ted

as n

on-in

tere

st e

xpen

ses

divi

ded

by to

tal a

sset

s; E

QA

SS is

a m

easu

re o

f cap

ital

izat

ion,

cal

cula

ted

as b

ook

valu

e of

shar

ehol

ders

equ

ity

as a

frac

tion

of t

otal

ass

ets;

LN

GD

P is

nat

ural

log

of g

ross

dom

esti

c pr

oduc

ts; I

NFL

is

the

infla

tion

rat

e; O

VER

_FR

EE is

the

over

all e

cono

mic

free

dom

inde

x; B

USI

_FR

EE is

the

busi

ness

free

dom

inde

x; M

ON

E_FR

EE is

the

mon

etar

y fr

eedo

m in

dex;

FIN

A_F

REE

is th

e fin

anci

al

free

dom

inde

x; C

OR

R_F

REE

is th

e fr

eedo

m fr

om c

orru

ptio

n in

dex.

ON

E ST

EP S

YS-G

MM

(1)

(2)

(3)

(4)

(5)

(6)

(7)

CO

NST

AN

T5.

073*

(1

.68)

-1.2

91

(-0.

20)

6.87

9 (0

.42)

-12.

792

(-0.

98)

5.57

5 (0

.41)

-9.4

84

(-1.

28)

-4.2

59

(-0.

47)

Bank

Cha

ract

eris

tics

RO

A t-

10.

126

(1.1

9)0.

155*

(1

.79)

0.11

9*

(1.6

5)0.

152*

(1

.75)

0.14

3**

(2.3

2)0.

107*

* (1

.96)

0.11

5 (1

.43)

LOA

NS/

TA-0

.014

(-

0.84

)-0

.016

(-

0.72

)0.

004

(0.1

9)0.

004

(0.1

5)-0

.000

(-

0.01

)0.

002

(0.0

9)0.

025

(0.8

3)LN

TA-0

.130

(-

0.68

)-0

.871

***

(-2.

60)

0.25

6 (0

.47)

-0.2

48

(-0.

42)

-0.3

37

(-0.

55)

-0.2

68

(-0.

45)

-0.0

59

(-0.

09)

LLP/

TL-0

.056

* (-

1.67

)-0

.026

(-

0.97

)-0

.009

(-

0.34

)0.

004

(0.1

3)-0

.002

(-

0.09

)0.

001

(0.0

4)0.

006

(0.1

5)N

IE/T

A-0

.599

**

(-2.

22)

-0.8

07**

* (-

3.31

)-0

.529

* (-

1.62

)-0

.665

* (-

1.74

)-0

.736

**

(-2.

38)

-0.5

70*

(-1.

80)

-0.6

41*

(-1.

86)

EQA

SS0.

025

(0.9

0)0.

009

(0.2

7)0.

040

(1.0

8)0.

012

(0.2

9)0.

012

(0.2

8)0.

009

(0.2

3)0.

063

(1.4

9)

Econ

omic

Con

diti

ons

LNG

DP

0.59

1**

(2.0

1)-0

.771

(-

0.91

)0.

428

(0.8

1)-0

.015

(-

0.04

)1.

100*

(1

.71)

-0.5

93

(-1.

23)

INFL

0.02

6 (0

.29)

-0.0

19

(-0.

19)

-0.0

06

(-0.

07)

-0.0

28

(-0.

32)

-0.1

12

(-0.

86)

-0.0

19

(-0.

26)

CR

39.

487

(1.4

0)11

.918

(1

.37)

9.47

8 (1

.14)

8.98

9 (1

.12)

0.01

4 (0

.00)

16.2

62*

(1.8

7)Z-

SCO

RE

0.18

1**

(2.2

0)0.

280*

* (2

.47)

0.25

0**

(2.5

0)0.

251*

* (2

.38)

0.27

5***

(2

.60)

0.28

0**

(2.5

3)

(Con

tinu

ed)

Sufian et al.


Tabl

e 4.

(Co

ntin

ued)

Econ

omic

Fre

edom

OV

ER_F

REE

-0.2

64

(-1.

16)

BUSI

_FR

EE0.

078

(0.7

1)M

ON

E_FR

EE-0

.124

(-

1.20

)FI

NA

_FR

EE0.

127*

* (1

.97)

CO

RR

_FR

EE-0

.173

***

(-2.

58)

Wal

d c2

38.4

1***

137.

76**

*99

9.69

***

1169

.36*

**54

3.96

***

183.

61**

*23

7.74

***

AR

(1)

p-va

lue

0.89

30.

192

0.18

20.

171

0.24

10.

334

0.02

7A

R(2

) p-

valu

e0.

457

0.77

30.

530

0.96

60.

985

0.69

30.

243

Sarg

an p

-val

ue0.

231

0.87

80.

987

0.75

90.

858

0.59

80.

999

No.

of O

bser

vati

ons t-

114

811

194

9494

9494

Valu

es in

par

enth

eses

are

z-s

tati

stic

s.

***,

**, a

nd *

indi

cate

s si

gnifi

canc

e at

1, 5

, and

10%

leve

ls, r

espe

ctiv

ely.

a statistically significant positive sign, suggesting that banking security as well as independence from government control exerts positive impact on Islamic banks’ profitability. The more banks are controlled by the government, the less free they are to engage in essential financial activities that facilitate private sector led economic growth.

Finally, it is observed from column 6 of Table 4 that the coefficient of the freedom from corruption (CORR_FREE) variable exhibits a negative sign (statistically significant at the 1% level). The empirical findings from this study clearly suggest that corruption (e.g. corruption in the business environment, including levels of governmental, legal, judicial, and administrative) has significant negative impact on the profitability of Islamic banks operating in the MENA banking sectors.

Robustness checksIn order to check for the robustness of the results, we carry out several sensitivity analyses. First, in light of Holmes et al. (2008) arguments, we remove all the macroeconomic and market conditions variables from the regression models and repeat Eq. (2). The regression results are presented in Table 5. All in all, it can be observed that the coefficients of the baseline regression models stay mostly the same: the sign and the order of magnitude remained similar and significant as in the baseline regression models. As observed, the empirical findings suggest that the coefficient of the OVER_FREE, BUSI_FREE, and MONE_FREE entered the regression models with a negative sign, but are insignificant. From column 4 of Table 5 it can be observed that the coefficient of the financial freedom (FINA_FREE) retains its positive sign and is significant. On a similar vein, the empirical findings suggest that CORR_FREE exhibits the same negative sign.

Second, it is also interesting to examine the persistence of the explanatory variables over time. To do so, we lag all the explanatory variables by one period and repeat Eq. (2). The results are presented in Table 6. As can be seen, the coefficients of the baseline variables remain stable as in the baseline regression model. It is also worth noting that the coefficient of the LNTA variable is now significant in four out of the six models estimated, while LLP/TL has consistently exhibit negative and significant impact on Islamic banks’ profitability levels. The empirical findings suggest that the impact of capitalization (EQASS) is positive in all of the regression models estimated.

It is also interesting to note that the impact of overall economic freedom (OVER_FREE) is now positive and significant. The empirical findings seem to support the notion that economic freedom is a key to the creation of an environment that allows a virtuous cycle of entrepreneurship, innovation, and sustained economic growth and development to flourish. Furthermore, economies with higher levels of economic freedom are likely to enjoy higher living standards (Holmes et al. 2008). Holmes et al. (2008) points out that a higher level of economic freedom is associated with a higher level of per capita GDP. They also suggest that countries which increase their levels of freedom tend to experience faster growth rates and the freest economies also have lower rates of unemployment and inflation. However, the coefficient



Table 5. Panel Generalized Methods of Moments (GMM) regression results.

ROA LOANS TA LNTA LLP TL NIE TAjt jt jt jt jt= + + + + β β β β β0 1 2 3 4/ / /_

+ + + ++ +

β β β ββ β

5 6 7 8

9 9

3EQASS LNGDP INFL CRZ SCORE OVER FR

jt t t t

t- EEE BUSI FREE MONE FREE FINA FREE CORR FREt t t t+ + + +β β β β10 11 12 13_ _ _ _ EEt jt+ e

The notation used in the table below is defined as follows: LOANS/TA is used as a proxy measure of loans intensity, calculated as total loans divided by total assets; LNTA is a proxy measure of size, calculated as a natural logarithm of total bank assets; LLP/TL is a measure of bank risk calculated as the ratio of total loan loss provisions divided by total loans; NII/TA is a measure of bank diversification towards non interest income, calculated as total non-interest income divided by total assets; NIE/TA is a proxy measure for costs, calculated as non-interest expenses divided by total assets; EQASS is a measure of capitalization, calculated as book value of shareholders equity as a fraction of total assets; LNGDP is natural log of gross domestic products; INFL is the inflation rate; OVER_FREE is the overall economic freedom index; BUSI_FREE is the business freedom index; MONE_FREE is the monetary freedom index; FINA_FREE is the financial freedom index; CORR_FREE is the freedom from corruption index.

ONE STEP SYS-GMM

(1) (2) (3) (4) (5)

CONSTANT 4.095(0.45)

4.410(0.92)

10.278(1.45)

-5.241(-1.23)

6.055(1.44)

Bank Characteristics

ROA t-1 0.151***(2.98)

0.140**(2.44)

0.163***(3.04)

0.111** (2.21)

0.194*** (3.07)

LOANS/TA -0.001(-0.05)

-0.002(-0.13)

-0.001(-0.05)

0.012 (0.85)

0.009 (0.48)

LNTA 0.048(0.16)

0.022(0.11)

-0.132(-0.48)

0.544** (2.07)

-0.064(-0.27)

LLP/TL -0.041(-1.26)

-0.042(-1.33)

-0.038(-1.37)

-0.017(-0.55)

-0.040(-1.42)

NIE/TA -0.827*(-1.71)

-0.823**(-2.06)

-0.911**(-2.27)

-0.725*(-1.76)

-0.975**(-2.07)

EQASS 0.050**(2.50)

0.050**(2.13)

0.043*(1.77)

0.057*** (2.60)

0.068*** (3.44)

Economic Freedom

OVER_FREE -0.024(-0.26)

BUSI_FREE -0.023(-0.73)

MONE_FREE -0.075(-1.39)

FINA_FREE 0.052**(2.40)

CORR_FREE -0.058*(-1.62)

Wald c2 55.82 59.30 66.63 50.43 93.00AR(1) p-value 0.556 0.584 0.654 0.934 0.505AR(2) p-value 0.470 0.558 0.334 0.859 0.422Sargan p-value 0.126 0.128 0.148 0.151 0.681No. of Observationst-1 129 129 129 129 129

Values in parentheses are z-statistics. ***, **, and * indicates significance at 1, 5, and 10% levels, respectively.

of the variable is insignificant. Similarly, the empirical findings seem to suggest that the coefficient of the BUSI_FREE is significantly related to the profitability of Islamic banks operating in the MENA banking sectors. However, it can also be observed from columns 5 and 6 of Table 6 that financial freedom and freedom from corruption loses their explanatory power.

Third, we restrict our sample to banks with more than three years of observations. All in all, the results remain qualitatively similar in terms of directions and significance levels. Finally, we address the effects of outliers in the sample by excluding the top and bottom 1% of the sample. The results continued to remain robust in terms of directions and significance levels.8

Sufian et al.


Table 6. Panel Generalized Methods of Moments (GMM) regression results.

ROA LOANS TA LNTA LLP TL NIE TAjt jt jt jt jt= + + + + β β β β β0 1 2 3 4/ / /_

+ ++ + + +

β ββ β β β

5 6

7 8 9 93EQASS LNGDP

INFL CR Z SCORE OVER FRjt t

t t t- EEE BUSI FREE MONE FREEFINA FREE CORR FRE

t t t

t

+ ++ +

β ββ β

10 11

12 13

_ __ _ EEt jt+ e

The notation used in the table below is defined as follows: LOANS/TA is used as a proxy measure of loans intensity, calculated as total loans divided by total assets; LNTA is a proxy measure of size, calculated as a natural logarithm of total bank assets; LLP/TL is a measure of bank risk calculated as the ratio of total loan loss provisions divided by total loans; NII/TA is a measure of bank diversification towards non interest income, calculated as total non-interest income divided by total assets; NIE/TA is a proxy measure for costs, calculated as non-interest expenses divided by total assets; EQASS is a measure of capitalization, calculated as book value of shareholders equity as a fraction of total assets; LNGDP is natural log of gross domestic products; INFL is the inflation rate; OVER_FREE is the overall economic freedom index; BUSI_FREE is the business freedom index; MONE_FREE is the monetary freedom index; FINA_FREE is the financial freedom index; CORR_FREE is the freedom from corruption index.

ONE STEP SYS-GMM

(1) (2) (3) (4) (5) (6)

CONSTANT 1.290(0.58)

-10.061*(-1.73)

-6.400**(-2.14)

1.873(0.21)

-4.750(-1.25)

0.559(0.21)

Bank Characteristics

ROA t-1 0.530***(3.34)

0.405***(2.59)

0.492**(2.55)

0.456***(2.75)

0.384**(2.14)

0.484***(3.11)

ROA t-2 -0.156**(-1.93)

-0.182***(-3.45)

-0.199***(-3.46)

-0.167***(-2.67)

-0.149**(-2.23)

-0.160**(-2.31)

LOANS/TA 0.017(0.57)

0.006(0.22)

0.024(0.70)

0.012(0.45)

0.012(0.51)

0.015(0.52)

LOANS/TAt-1 -0.027(-0.89)

-0.003(-0.12)

-0.019(-0.64)

-0.011(-0.43)

-0.003(-0.11)

-0.011(-0.40)

LNTA -2.492*(-1.64)

-1.465(-0.98)

-2.223*(-1.84)

-2.686*(-1.84)

-1.361(-0.62)

-3.135*(-1.77)

LNTAt-1 2.582* (1.63)

1.910(1.26)

2.649** (2.05)

2.854**(1.98)

1.814(0.87)

3.334* (1.85)

LLP/TL -0.078***(-2.84)

-0.066**(-2.35)

-0.061**(-2.31)

-0.076***(-3.53)

-0.069***(-3.87)

-0.082*** (-4.38)

LLP/TL t-1 -0.019(-0.67)

-0.013(-0.73)

-0.026(-1.35)

-0.014 (-0.61)

-0.001(-0.04)

-0.014(-0.55)

NIE/TA -0.375(-1.09)

-0.044 (-0.16)

-0.178(-0.56)

-0.096(-0.30)

-0.090(-0.31)

-0.124(-0.37)

NIE/TA t-1 0.269(0.82)

-0.020 (-0.05)

0.172(0.39)

0.032(0.07)

-0.071(-0.15)

0.101(0.24)

EQASS 0.077**(2.42)

0.076***(2.89)

0.074**(2.54)

0.087***(2.98)

0.085***(3.29)

0.087***(2.97)

EQASS t-1 0.009(0.24)

0.027 (0.77)

0.030(0.81)

0.013(0.38)

0.022(0.68)

0.022(0.56)

Economic Freedom

OVER_FREE 0.113*(1.61)

BUSI_FREE 0.057**(2.03)

MONE_FREE -0.026(-0.34)

FINA_FREE 0.031(1.06)

CORR_FREE -0.023(-0.80)

(Continued)



5. Concluding remarksBy using an unbalanced bank level panel data, the study attempts to examine the impact of economic freedom on the performance of Islamic banks operating in the MENA banking sectors during the period 2000–2008. We find that the larger, more diversified, and better capitalized banks are relatively more profitable. The empirical findings seem to suggest that efficient cost management is a prerequisite to improve the profitability of Islamic banks operating in the MENA banking sectors. Similarly, we find that higher credit risk has negative and significant influence on the profitability of Islamic banks operating in the MENA banking sectors. The results suggest economic conditions exert negative impact on Islamic banks’ profitability levels when we control for overall economic freedom, monetary freedom, and freedom from corruption. We also find that the level of inflation has positive impact when we control for monetary and financial freedom.

The findings from this study seem to suggest that greater financial and business freedom exerts positive impacts on the profitability of Islamic banks operating in the MENA banking sectors. The positive coefficient of the financial freedom variable indicate that higher (lower) freedom on the activities that Islamic banks could undertake increases (reduces) their profitability, which is consistent with the view that less regulatory control allows banks to engage in various activities enabling banks to exploit economies of scale and scope and generate income from non-traditional sources. Furthermore, higher freedom on entrepreneurs to start businesses is conducive to job creation and consequently increases Islamic banks’ profitability. Interestingly, the impact of monetary freedom is negative implying that higher (lower) monetary policy independence reduces (increases) Islamic banks’ profitability, providing support to the benefits of government interventions. In essence, although price stability without intervention is an ideal state for a free market, the empirical findings from this study clearly lend support to the benefits of government interventions in the markets.

The findings of this study present considerable policy relevance. In view of the increasing competition attributed to the more liberalized banking sector, bank managements as well as the policymakers will be more inclined to find ways to obtain the optimal utilization of capacities as well as making the best use of their resources, so that these resources are not wasted during the production of banking products and services. The findings pointed to the need for bankers to choose flexible operating environment and economic system favouring the rapid development of a vibrant banking sector to maximize their performance. From the regulatory perspective, the performance of the Islamic banking

sector will be based on their operating performance. Therefore, policy direction is expected to point towards enhancing the resilience and performance of the banking institutions with the aim of intensifying the robustness and stability of the Islamic banking sector.

Within MENA countries, enhancing economic freedom is of an importance policy if the region is to attract more financial investments, improve weak financial infrastructure and enhance the banking system performance. The current state of the financial sector in MENA which mainly controlled by state owned banks and dominate banking activities (up to 95% of assets in several countries in the MENA region) resulting in poor services, high costs, and weak financing of new investments and trade. As the MENA region competes for economic benefits for its citizens in the new global economy, it is important that the policy makers in these countries to improve their quality of governance and transparency; to promote a legal system that protects shareholders and creditors rights; and to enhance their economic freedom.

Notes1. The basic tenets and principles of Islamic banking

are built upon the avoidance of usury (riba’) and the prohibition of impermissible activities as clearly mentioned in the Quran, the Islam’s holy book and the traditions of Prophet Muhammad (sunnah): “Believers! Do not consume riba’, doubling and redoubling…” (3.130); “God has made buying and selling lawful and riba’ unlawful…” (2:274).

2. The estimates of the number of Islamic financial institutions vary considerably between institutions. For instance, the International Monetary Fund (IMF) estimate that the number of Islamic financial institutions has increased to more than 300, while the Association of Islamic Banking Institutions Malaysia (AIBIM) suggests that there are around 486 Islamic financial institutions around the world.

3. MENA stands for Middle East and North Africa region comprises of Algeria, Bahrain, Egypt, Iran, Iraq, Israel, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Saudi Arabia, Syria, Sudan, Tunisia, UAE, and Yemen.

4. Islam has laid down some principles and prescribed certain limits for the economic activity of man so that the entire pattern of production, exchange, and distribution of wealth may conform to the Islamic standard of justice and equity.

5. The Economic Freedom Index is released by The Heritage Foundation and The Wall Street Journal.

6. Islamic banks income must not be uncontaminated by usury (riba’). Thus, in the case of the Islamic banking sector, it is reasonable to assume that the interest rate to be the profit rate.

Table 6. (Continued)

Wald c2 567.29*** 1685.09*** 2902.11*** 3870.54*** 1471.55*** 1070.45***AR(1) p-value 0.049 0.057 0.047 0.035 0.099 0.038AR(2) p-value 0.465 0.152 0.202 0.097 0.378 0.181Sargan p-value 0.219 0.229 0.306 0.061 0.049 0.114No. of Observationst-2 111 98 98 98 98 98

Values in parentheses are z-statistics.***, **, and * indicates significance at 1, 5, and 10% levels, respectively.

Sufian et al.


7. Following Garcia-Herrero et al. (2009) among others, we instrument for all regressors. The macroeconomic characteristics are treated as exogenous (see among others Baum et al. 2010).

8. To conserve space, we do not report the regression results in the paper but are available upon request.

ReferencesAbbasoglu, O.F., Aysan, A.F. and Gunes, A. (2007)

“Concentration, Competition, Efficiency, and Profitability of the Turkish Banking Sector in the Post-Crisis Period,” Banks and Bank Systems, 2(3), 106–115.

Adkins, L., Moomaw, R. and Savvides, A. (2002) “Institutions, Freedom and Technical Efficiency,” Southern Economic Journal, 69(1), 92–108.

Albertazzi, U. and Gambacorta, L. (2009) “Bank Profitability and the Business Cycle,” Journal of Financial Stability, 5(4), 393–309.

Altman, M. (2008) “How Much Economic Freedom is Necessary for Economic Growth? Theory and Evidence,” Economics Bulletin, 15(2), 1–20.

Athanasoglou, P.P., Brissimis, S.N. and Delis, M.D. (2008) “Bank Specific, Industry Specific and Macroeconomic Determinants of Bank Profitability,” Journal of International Financial Markets, Institutions and Money, 18(2), 121–136.

Bashir, A.H.M. (1999) “Risk and Profitability Measures in Islamic Banks: The Case of Two Sudanese Banks,” Islamic Economic Studies, 6(2), 1–24.

Bashir, A.H.M. (2001) “Assessing the Performance of Islamic Banks: Some Evidence from the Middle East,” paper presented at the American Economic Association Annual Meeting, New Orleans, Louisiana.

Baum, C.F., Caglayan, M. and Talavera, O. (2010) “Parlia-mentary Election Cycles and the Turkish Banking Sector,” Journal of Banking and Finance, 34(11), 2709–2719.

Ben Naceur, S. and Goaied, M. (2008) “The Determinants of Commercial Bank Interest Margin and Profitability: Evidence from Tunisia,” Frontiers in Finance and Economics, 5(1), 106–130.

Berger, A.N. (1995) “The Relationship Between Capital and Earnings in Banking,” Journal of Money, Credit and Banking, 27(2), 432–456.

Berger, A.N. and DeYoung, R. (1997) “Problem Loans and Cost Efficiency in Commercial Banks,” Journal of Banking and Finance, 21(6), 849–870.

Berger, A.N. and Humphrey, D.B. (1997) “Efficiency of Financial Institutions: International Survey and Directions for Future Research,” European Journal of Operational Research, 98(2), 175–212.

Berger, A.N., Bonime, S.D., Covitz, D.M. and Hancock, D. (2000) “Why Are Bank Profits So Persistent? The Roles of Product Market Competition, Information Opacity and Regional Macroeconomic Shocks,” Journal of Banking and Finance, 24(7), 1203–1235.

Berger, A.N., Hanweck, G.A. and Humphrey, D.B. (1987) “Competitive Viability in Banking: Scale Scope and Product Mix Economies,” Journal of Monetary Economics, 20(3), 501–520.

Berggren, N. (2003) “The Benefits of Economic Freedom: A Survey,” Independent Review, 8(2), 193–211.

Blundell, R. and Bond, S. (1998) “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models,” Journal of Econometrics, 87(1), 115–143.

Boyd, J. and Runkle, D. (1993) “Size and Performance of Banking Firms: Testing the Predictions Theory,” Journal of Monetary Economics, 31(1), 47–67.

Boyd, J.H. and De Nicolo, G. (2006) “The Theory of Bank Risk Taking and Competition Revisited,” The Journal of Finance, 60(3), 1329–1343.

Button, K.J. and Weyman-Jones, T.G. (1992) “Ownership Structure, Institutional Organization and Measured X-Efficiency,” The American Economic Review, 82(2), 439–445.

ihák, M., Maechler, A., Schaeck, K. and Stolz, S. (2009) “Who Disciplines Bank Managers?” Working Paper, International Monetary Fund.

De Haan, J. and Siermann, C.L.J. (1998) “Further Evidence on the Relationship between Economic Freedom and Economic Growth,” Public Choice, 95(3–4), 363–380.

De Haan, J. and Sturm, J.E. (2000) “On the Relationship between Economic Freedom and Economic Growth,” European Journal of Political Economy, 16(2), 215–241.

DeNicolo, G., Bartholomew, P., Zaman, J. and Zephirin, M. (2003) “Bank Consolidation, Internationalization, and Conglomeration: Trends and Implications for Financial Risk,” Financial Markets, Institutions & Instruments, 13(4), 173–217.

DeYoung, R. and Rice, T. (2004) “Non-Interest Income and Financial Performance at US Commercial Banks,” Financial Review, 39(1), 101–127.

Dogan, E. and Fausten, D.F. (2003) “Productivity and Technical Change in Malaysian Banking: 1989–1998,” Asia-Pacific Financial Markets, 10(2–3), 205–237.

Eichengreen, B. and Gibson, H.D. (2001) “Greek Banking at the Dawn of the New Millennium,” Discussion Paper, Center of Economic and Policy Research.

El Qorchi, M. (2005) “Islamic Finance Gears Up”, Finance and Development, 42(4), 46–49.

El-Gamal and Inanoglu (2004) “Islamic Banking in Turkey: Boon or Bane for the Financial Sector,” Proceedings of the Fifth Harvard University Forum on Islamic Finance, Cambridge: Center for Middle Eastern Studies, Harvard University.

El-Gamal, M.A. and Inanoglu, H. (2005) “Efficiency and Unobserved Heterogeneity in Turkish Banking” Journal of Applied Econometrics, 20(5), 641–664.

Garcia-Herrero, A., Gavila, S. and Santabarbara, D. (2009) “What Explains the Low Profitability of Chinese Banks?” Journal of Banking and Finance, 33(11), 2080–2092.

Gilbert, R. (1984), “Bank Market Structure and Competition – A Survey,” Journal of Money, Credit and Banking, 16(4), 617–645.

Goddard, J., Molyneux, P. and J. Wilson (2004) “Dynamic of Growth and Profitability in Banking,” Journal of Money, Credit and Banking, 36(6), 1069–1090



Goddard, J., Molyneux, P. and Wilson, J.O.S. (2001) European Banking: Efficiency, Technology and Growth. New York: John Wiley and Sons.

Hassan, M.K and Hussein, K.A. (2003) “Static and Dynamic Efficiency in the Sudanese Banking System,” Review of Islamic Economics, 14, 5–48.

Hassan, M.K. (2006), “The X-Efficiency of Islamic Banks,” Islamic Economic Studies, 13(1–2), 49–77.

Hassan, M.K. and Bashir, A.H.M. (2003) “Determinants of Islamic Banking Profitability,” Paper Presented at the 10th ERF Annual Conference, 16th–18th December, Morocco.

Hawtrey, K. and Liang, H. (2008) “Bank Interest Margins in OECD Countries,” The North American Journal of Economics and Finance, 19(3), 249–260.

Heckelman, J.C. (2000) “Economic Freedom and Economic Growth: A Short-Run Causal Investigation,” Journal of Applied Economics, 3(1), 71–91.

Heckelman, J.C. and Knack, S. (2009) “Aid, Economic Freedom and Growth,” Contemporary Economic Policy, 27(1), 46–53.

Heckelman, J.C. and Stroup, M.D. (2000) “Which Economic Freedoms Contribute to Growth?” KYKLOS, 53(4), 527-544.

Hirtle, B.J. and Stiroh, K.J. (2007) “The Return to Retail and the Performance of US Banks,” Journal of Banking and Finance, 31(4), 1101–1133.

Holmes, K.R., Feulner, E.J., and O’Grady, M.A. (2008) 2008 Index of Economic Freedom. The Heritage Foundation: Washington, D.C.

Hussein, K.A. (2003) “Operational Efficiency in Islamic Banking: The Sudanese Experience,” Working Paper, Islamic Research and Training Institute (IRTI), Islamic Development Bank.

Isik, I. and Hassan, M.K. (2003) “Efficiency, Ownership and Market Structure, Corporate Control and Governance in the Turkish Banking Industry,” Journal of Business Finance and Accounting, 30(9–10), 1363–1421.

Kasman, A., Tunc, G., Vardar, G. and Okan, B. (2010) “Consolidation and Commercial Bank Net Interest Margins: Evidence from the Old and New European Union Members and Candidate Countries,” Economic Modelling, 27(3), 648–655.

Kennedy, P. (2008) A Guide to Econometrics. Blackwell Publishing: Malden, Massachusetts.

Kosmidou, K. (2008) “The Determinants of Banks’ Profits in Greece during the Period of EU Financial Integration,” Managerial Finance, 34(3), 146–159.

Kosmidou, K. and Zopounidis, C. (2008) “Measurement of Bank Performance in Greece,” South Eastern Europe Journal of Economics, 6(1), 79–95.

Kosmidou, K., Pasiouras, F. and Tsaklanganos, A. (2007) “Domestic and Multinational Determinants of Foreign Bank Profits: The Case of Greek Banks Operating Abroad,” Journal of Multinational Financial Management, 17(2–3), 1–15.

Miller, S.M. and Noulas, A. (1997) “Portfolio Mix and Large Bank Profitability in the USA,” Applied Economics, 29(4), 505–512

Molyneux, P., Altunbas, Y. and Gardener, E.P.M. (1996) Efficiency in European Banking. John Wiley & Sons, Chichester.

North, D., Thomas, R.P., (1973) The Rise of the Western World: A New Economic History. Cambridge: Cambridge University Press.

Pasiouras, F. and Kosmidou, K. (2007) “Factors Influencing the Profitability of Domestic and Foreign Commercial Banks in the European Union,” Research in International Business and Finance, 21(2), 222–237.

Perry, P. (1992) “Do Banks Gain or Lose from Inflation,” Journal of Retail Banking, 14(2), 25–40.

Powell, B. (2003) “Economic Freedom and Growth: The Case of the Celtic Tiger,” Cato Journal, 22(3), 431–448.

Rivard, R.J. and Thomas, C.R. (1997) “The Effect of Interstate Banking on Large Bank Holding Company Profitability and Risk,” Journal of Economics and Business, 49(1), 61–76.

Sarker, M.A.A. (1999) “Islamic Banking in Bangladesh: Performance, Problems, and Prospects,” International Journal of Islamic Financial Services, 1(3), 15–36.

Saunders, A., Strock, E., and Travlos, N.G. (1990) “Ownership Structure, Deregulation, and Bank Risk Taking,” Journal of Finance, 45(2), 643–654.

Short, B.K. (1979) “The Relation Between Commercial Bank Profit Rate and Banking Concentration in Canada, Western Europe and Japan,” Journal of Banking and Finance, 3(3), 209–219.

Spathis, C., Kosmidou, K. and Doumpos, M. (2002) “Assessing Profitability Factors in the Greek Banking System: A Multicriteria Methodology,” International Transactions in Operational Research, 9(5), 517–530.

Staikouras, C. and Wood, G. (2003) “The Determinants of Bank Profitability in Europe,” Paper presented at the European Applied Business Research Conference, Venice, 9–13 June.

Stiroh, K.J. and Rumble, A. (2006) “The Dark Side of Diversification: The Case of US Financial Holding Companies,” Journal of Banking and Finance, 30(8), 2131–2161.

Sufian, F. (2009) “Determinants of Bank Efficiency During Unstable Macroeconomic Environment: Empirical Evidence from Malaysia,” Research in International Business and Finance, 23(1), 54–77.

Thakor, A. (1987) “Discussion”, Journal of Finance, 42(3), 661–663.

Tregenna, F. (2009) “The Fat Years: The Structure and Profitability of the US Banking Sector in the Pre-Crisis Period,” Cambridge Journal of Economics, 33(4), 609–632.

Williams, B. (2003) “Domestic and International Determinants of Bank Profits: Foreign Banks in Australia,” Journal of Banking and Finance, 27(6), 1185–1210.


Cite this chapter as: Anouze A L (2015). Efficiency of performance of banks in the Gulf region before, during and after crises (financial and political). In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Efficiency of performance of banks in the Gulf region before, during and after crises (financial and political)Abdel Latef AnouzeDepartment of Management and Marketing, College of Business and Economics, Qatar University, E: [email protected]

Abstract - Only a few cross-country empirical studies have been conducted to measure the perfor-mance of commercial banks especially before, during, and after crises (financial or political). This study makes an attempt to fill the gap in the literature by investigating the impacts of crises on Gulf Corporate Council (GCC) commercial banks’ performance over the period 1997-2007. The rationale behind this selection is that the GCC countries within this period witnessed two major crises: a political crisis (the second Gulf war) and a financial crisis (the current global crisis). Clearly, it is important that a manager recognizes the best bank policy in the face of each crisis that could help both bankers and regulators in managing these crises. Also, the banking system within GCC countries comprises two different operating banking systems, Islamic and conventional. As both are operating in similar environments, it is of interest to examine whether one can make judgments concerning the success of their competitive strategies, and other management-determined factors by using performance measures.

Two different evaluation methods are computed to measure bank performance: data envelopment analysis (DEA), and classification and regression tree (CART). The overall results show that conventional banks perform well during a political crisis, whereas Islamic banks performed better during the financial crisis. However, this difference is not statistically significant, which means that GCC commercial banks can be equally competitive when it comes to technical efficiency. Also, there is no statistically significant relationship between bank geographical location and its efficiency score. Moreover, the results confirm that large and small size GCC commercial banks are more efficient than medium-sized banks. Out of the 24 environmental factors included in the study to investigate the relationship between environmental factors (internal and external) and bank performance, only 15 factors are considered to be important in predicting fully-efficient banks.

Keywords: data envelopment analysis, classification and regression tree, bank performance, Islamic bank, GCC countries

1. IntroductionIn the light of the on-going international financial crisis, and the large costs generated for national and international financial systems, it is essential to assess the performance of the financial sector in order to avoid the financial disaster becoming more complicated. Assessing banks’ performance would help managers examine the success of managerial decisions that they have taken before, during and after the crisis; to better understand their management effectiveness, and it would provide them with valuable reference for improving their performance. Also, such assessment would help managers to measure the success of these decisions

compared with those made by their counterparts during same period. On the other hand, it also helps policy makers to develop a strong and healthy environment for the banking sector by examining the impact of economic and financial reforms that have taken place. Meanwhile, investors want to see how well a bank is performing before potentially investing in it. A high stock price alone is not a sufficient measure to use; they have to see how well a bank is performing too. Therefore, if a bank is to survive and succeed, it should learn the status of its efficiency and how it compares with counterparts in same country or other countries. Hence, to identify appropriate financial

Anouze


decisions that will achieve better allocation of financial resources in a more efficient and effective manner, it is important to assess bank performance at the country and/or international level. A number of international empirical studies have been conducted to measure the performance of the banking sector before, during, and after crises (Mercan et al. 2003; Jeon and Miller 2004 and 2005). However, all of these studies among others were carried out prior to the current global financial crisis. Therefore, this study attempts to fill the gap in the literature by assessing GCC commercial banks’ performance before, during and after the crises to guide bank managers and other stakeholders, such as policy makers and investors, in their decisions.

There is a substantial body of literature discussing different methods applied to evaluate the performance of banks (e.g., Anouze 2010; Fethi and Pasiouras 2010; Berger and Humphrey 1997). Reviewing 130 studies of the efficiency of financial institutions, Berger and Humphrey (1997) classified these methods according to the technical approach employed into parametric, such as the stochastic frontier approach (SFA), and nonparametric, such as data envelopment analysis (DEA). Application of these methods alone to evaluate banks’ performance determines efficiency scores but gives no details of factors related to inefficiency, especially if these factors are in the form of non-numeric variables such as the operating style of the banking sector (Emrouznejad and Anouze 2010). This study proposes a comprehensive performance evaluation framework based on managerial, financial, and macroeconomic indicators to measure and predict banks’ performance. It allows exploration and discovery of meaningful, previously hidden information from given data. It integrates Data Envelopment Analysis (DEA) with the Classification and Regression Tree (CART) technique. DEA is a nonparametric method for measuring the performance of Decision Making Units (DMUs) such as banks, hospitals, universities, or services. It groups data into inputs and outputs to produce a productive efficiency frontier against which an individual bank or the banks of an entire country can be benchmarked. Input variables within the DEA context are resources to be minimized, while output variables are product or services to be maximized in order to achieve a high efficiency score. The DEA efficiency score is a relative measure, which is derived for each bank from the DEA based on the quality of transforming the inputs into outputs. CART, on other hand, is a nonparametric data-mining technique which allows meaningful information to be explored and discovered from a given data set. Unlike the DEA model, in which each case needs to be compared, CART produces results that can easily be applied to determine the efficiency of a bank. A unique feature of CART is that it illustrates the data in the form of a decision tree so that the results can be presented in the form of diagrams that are easy to understand. Integration of the two techniques would help stakeholders to assess, predict and identify the banks that are most likely to be troubling or, on the other hand, outperforming. Hence, stakeholders would have an overall understanding of banks’ performance and, consequently, better improvement policies could be developed for unsuccessful banks.

2. Literature review: Banking performanceAlthough there is a huge volume of published research on banking efficiency, little effort has been made to conduct

studies of the impact of financial or political crises on banking performance. To our knowledge, this is the first study to explore the combined effect of financial and political crises on banking performance. However, our work contributes and relates closely to several branches of literature on bank performance, including studies of the impact of the financial crisis, bank health, and financial regulations on banking performance.

Few research studies have explored the impacts of the current financial crisis on of bank performance. Xiao (2009) used qualitative and quantitative tools to examine the performance of French banks during 2006–2008. The findings showed that French banks were not immune but proved relatively elastic to the global financial crisis. Beltratti and Stulz (2009) studied the bank stock return across the world during the period from the beginning of July 2007 to the end of December 2008; they found that large banks with more deposit financing at the end of 2006 exhibited significantly higher returns during the crisis. Cornett, McNutt and Tehranian (2010) analyzed the internal corporate governance mechanisms and the performance of US banks before and during the financial crisis; they found that the largest banks faced the largest losses during the crisis. Dietrich and Wanzenried (2011) examined how bank-specific characteristics, industry-specific, and macroeconomic factors affected the profitability of Swiss commercial banks over the period from 1999 to 2009; their results provide some evidence that the financial crisis did have a significant impact on banks’ profitability.

These studies, among others, used a regression analysis and limited bank performance to a single indicator—such as profit, capital, and deposit to assets ratio—to measure bank performance during or after the crisis. Although regression analysis is a useful tool, it tells nothing about how to improve the performance, nor which is the best practice during or after the crisis. Also, it only counts for a single indicator, whereas banks could aim to maximize more than one indicator during their financial transactions. Furthermore, none of these studies have investigated the performance of the GCC banking sector during the financial crisis.

Other researchers paid particular attention to the impact of financial regulation on bank performance. Policy makers introduce such regulations to develop a healthy environment that increases competition and improves banking sector efficiency. Although are numerous studies have examined the impact of financial regulations on banks’ performance, the overall impact of financial regulation is ambiguous. Huang, Hsiao, Cheng and Change (2008); Brissimis, Delis and Papankiolaou (2008); Koutsomanoli-Filippaki, Margaritis and Staikouras (2009); Hsiao, Chang, Cianci and Huang (2010); and Zhao, Casu and Ferrari (2010), among others, found that deregulation improves banking performance and stimulates competition in the financial market. In contrast, findings from other studies such as those of Fukuyama and Weber (2002); Halkos and Salamouris (2004); Park and Weber (2006): and Fu and Heffernan (2009) show a decline in bank efficiency during a period of financial reform.

Finally, others researchers studied the real effects on bank performance of deterioration in bank health or competition


Efficiency of performance of banks in the Gulf region before, during and after crises (financial and political)

during the financial crisis. Almeida, Campello, Laranjeira, and Weisbenner (2009) and Duchin, Ozbas and Sensoy (2010) studied the effect of the recent financial crisis on corporate investment. Their results show that corporate investment declined significantly following the onset of the crisis. Berger and Bouwman (2010) examined the effect of pre-crisis bank capital ratios on banks’ ability to survive financial crises, market shares, and profitability during the crises. Their findings show that capital helps banks of all sizes during banking crises; possession of higher capital helped banks to increase their probability of survival, market shares, and profitability. Gryglewicz (2011) studied the impact of both liquidity and solvency concerns on corporate finance, and showed how changes in solvency affect liquidity, and also how liquidity concerns affect solvency through capital structure choice.

These studies, among others, provide a comprehensive examination of the effects of a financial crisis on bank efficiency (see table A1 in appendix for summary of previous literature). Evaluation of this literature that presented in table helped us to select the appropriate variables to be included in our analysis. However, the impact of the crisis on bank performance has not yet been fully analyzed. Moreover, most of the reviewed research made use of statistical models to study the impact of the financial crisis on firms. Statistical models make some assumptions about statistical distribution or propensities of the data, but most financial data do not meet the statistical requirements of certain statistical models; also, statistical tests are sensitive to sample size. On the other hand, an operations research technique makes no assumptions about statistical distribution, and it is more accurate when testing complex or large samples (Demyank and Hasan 2010). Therefore, use of such a technique to study the impacts of the financial crisis on banking performance tends to be more appropriate in practical situations.

3. Methodology

Data Envelopment Analysis (DEA)DEA is a nonparametric relative performance evaluation method developed by Charnes, Cooper and Rhodes in (1978) considering constant return to scale (CRS). The CRS model compares banks’ performance based only on overall efficiency assuming constant returns to scale; however, it ignores the fact that different banks could be operating at different scales. To overcome this drawback Banker, Charnes and Cooper (1984) introduced the variable returns to scale (VRS) model that is similar to the CRS Model, but it ensures that an efficient bank is only benchmarked against banks of similar size, while in the CRS model a bank may be benchmarked against banks which are substantially larger or smaller than it. Subsequently, the original DEA models (CRS and VRS) have undergone many modifications and developments. Most of these developments occurred when the deficiencies of the original model were exposed during its application to solving real life problems (Thompson, Singleton, Thrall and Smith 1986).

To introduce the DEA-VRS model, assume there are n banks (j = 1,…, n) using m inputs (xij i = 1,…m) and producing s outputs (yrj, j = 1,…s). DEA measures the technical efficiency of bank j0 compared with n peer group of banks, as illustrated in model 1a and 1b.

In the input and output oriented DEA models (model 1a and 1b, respectively) bankj0 is assessed under variable returns to scale, where the efficiency of bankj0 is the optimal value of ? in Model 1a and 1/? in Model 1b (Thanassoulis 2001).

One of the key concerns when we have a variable that takes positive values for some banks and negative values for others is that its absolute value should rise or fall for the bank to improve its performance, depending on whether the bank concerned has a positive or negative value on that variable (Emrouznejad, Anouze, and Thanassoulis 2010a). For example, in the case of an output variable, if the bank has a positive value (profit) the output should rise to improve further but it should fall in absolute value as long as it continues to be negative (loss). To overcome this problem Emrouznejad and Anouze (2009) and Emrouznejad et al. (2010a and 2010b) treated each variable that has positive value for some banks and negative for others as consisting of the sum of two variables, and proposed a semi-oriented radial model (SORM).

Classification and Regression Tree (CART)CART is the commonly used decision tree in data mining that was developed by Breiman, Friedman, Olshen and Stone (1984) and further improved by Ripley (1996). In principle, CART is similar to regression analysis since both are used for prediction. However, CART has some advantages over the regression model:

• A model generated by a CART is easier to understand and relatively simple to interpret for non-statisticians (Breiman et al. 1984; Torgo 1997; Han and Kamber 2001)

• There are no assumptions to be made regarding the underlying distribution of values of the predictor variables as it is a nonparametric technique

• CART can handle numerical data, as well as categorical, with either ordinal or non-ordinal structure.

These are important features of CART as they will eliminate analyst time which would otherwise be spent determining whether variables are normally distributed and making transformations if they are not; specifically, it is important to use CART with DEA since DEA scores are skewed to one side.

Model 1a: Standard Input Oriented DEA – VRS

Model 1b: Standard Output Oriented DEA – VRS

Min θ subject to j

n

j ij ij

j

n

j rj rj

j

n

j

j

x x i

y y r

j

=

=

=

∑

∑

∑

≤ ∀

≥ ∀

=

≥ ∀

1

1

1

0

0

1

0

l θ

l

l

l ; ,

;

;

θ free

Max φ subject to j

n

j ij ij

j

n

j rj rj

j

n

j

j

x x i

y y r

j

=

=

=

∑

∑

∑

≤ ∀

≥ ∀

=

≥ ∀

1

1

1

0

0

1

0

l

l

l

l

;

; ,

;

φ

φ free

To reach to the CRS from model 1a and 1b one can remove the following constraint from the above model l j

j

n=

=∑ 11

Anouze


To validate the results generated by CART, the dataset is partitioned into two datasets, training and validation (Han and Kamber 2001). The data then go into two major phases of process: growth and pruning (Kim and Koehler 1995). In the growth phase, CART constructs a tree from the training dataset. In this phase, either each leaf node is associated with a single class, or further partitioning of the given leaf would result in the number of cases in one or both subsequent nodes being below some specified threshold. In the pruning phase the CART generated in the growth phase is improved in order to avoid over-fitting. Also in this phase, the CART result is evaluated against the validation dataset in order to generate a sub-tree with the lowest error rate.

There are several criteria for measuring CART results. The predictive accuracy of a CART is commonly measured by R-squared (average squared error); however, simplicity and stability are also important measures for a CART. Simplicity refers to the interpretability of the CART and is often based on the number of leaves in the CART. Stability of a CART refers to obtaining similar results for the training and validation datasets. One way to assess the stability of the CART can be by comparing the predicted mean value of the target variable (based on the training dataset) and the corresponding value for the validation dataset for each rule of the CART (Han and Kamber 2001).

Proposed methodology (DEA with CART)Figure 1 illustrates the proposed analysis, that is, DEA and CART. The DEA stage is to compute the efficiency score of each bank using DEA. Accordingly, the banks are categorized into two groups: efficient banks (target = 1) and inefficient banks (target = 0). In the CART stage the classified efficiency score (0 or 1) is used as the target of CART while the environmental (explanatory) variables is used as an input. However, an accurate CART requires a large dataset, whereas our sample was limited to 60 banks. Therefore, a new stage was introduced before the CART stage to increase the original dataset using the bootstrapping technique. Thus, we randomly selected 60 banks (by replacement) and repeated this sampling 61 times to achieve 3660 banks, so ensuring better accuracy

on the predicted CART results. The 3660 banks were divided into the two datasets, training and validation, by the ratio of 7:3 (Zhou and Jiang 2003; Emrouznejad and Anouze 2010).

Data description and analysisBanking industries in Gulf state countriesThe early banking sector in the GCC countries experienced much foreign ownership primarily by British banks with branches extending across all six GCC countries. Local banks were uncommon as there was insufficient experience. Subsequently, governments adopted central banking systems to strength local banks and to eliminate foreign involvement. Today there are 68 local banks operating in GCC countries. These banks can be grouped according to their operating style (mode of running financial transactions) into two groups: Islamic and conventional banks. Unlike conventional banks, Islamic banks run their financial transactions free of interest (i.e., no interest rate is taken or given against any financial transaction). Among the 68 local banks, 18 are Islamic banks and 50 are conventional banks.

Figure 2 illustrates the share of Islamic and conventional banking assets within each country. Saudi Arabia is the largest investor in the GCC, holding 32% of the total bank assets, with nine conventional banks and two Islamic banks, and had total assets of US$ 239,095 million in 2007. The UAE, with fifteen conventional and five Islamic banks, and total assets of US$ 224,542 million is the second largest investor in the area. Bahrain follows, with nine conventional and six Islamic banks and total assets of US$ 108,307 million, along with Kuwait, which has seven conventional and three Islamic banks and total assets of US$ 108,174 million. Qatar is in fourth position, with four conventional and two Islamic banks and total assets of US$ 56,429 million, which represents only 7% of the total GCC assets. Finally, Oman has only six conventional banks and total assets of US$22,259 million, this representing only 3% of the total assets. Although our study aimed to include all GCC commercial banks, eight banks are excluded on the basis of lack of availability of data, the remainder comprising 48 conventional and 12 Islamic banks.

Figure 1. Integrated data envelopment analysis and classification and regression tree.



Data DescriptionSelection of proper input and output variables to define and measure bank performance is always an extremely important decision (Mercan et al. 2003). It is especially so when using DEA, as different results may obtain from different sets of variables. Traditional bank behaviour theories described banks as accepting deposits from households, and making loans to investors (Diamond and Dybvig 1983; Diamond 1984; Gorton and Winton 2003). Yet, the change in bank involvement in markets and bank behaviour during crises requires a new theory of financial intermediation. In this study, the input variables include fixed assets, non-earning assets, and deposit, while the output variables are investments, loans, off-balance sheet, and net profit.

The selected input and output variables varied over the study period. It can be seen from table 1 that the minimum value of fixed assets—which is one of the inputs—is US$ 0.03 million whereas the maximum value is US$ 413.34 million, with average US$ 7.28 million and standard deviation of US$ 24.16 million. Similarly for other variables, for example, the net profit, the minimum net profit (loss) is US$ -289.01 million and the maximum value is US$ 195.97 million, with average of US$ 8.70 million and standard deviation of US$ 21.52 million. Given the long time period analysed, such variation would be expected; nonetheless, since DEA models are sensitive to observations it is likely that significant levels of variation would also be found in banks’ performance.

Data Envelopment Analysis (DEA)

Overall performance of GCC commercial banksTo study bank performance before, during and after the crises one grand-frontier (common-frontier) is computed for all banks in all countries. The grand-frontier provides a trend in the efficiency of banks, which would not be available if we computed the efficiency of banks using a separate frontier for each year. The approach employed, therefore, provides variations in the efficiency of banks over both time and space. This comparison across time and countries is on the same principle as the global frontier used by Portela and Thanassoulis (2010). A VRS output-oriented model is used to measure banks’ efficiency, since the CRS model is not possible in technologies where negative data can exist (Portela, Thanassoulis and Simpson 2004). The efficiency score obtained for all GCC commercial banks at the individual bank level is aggregated to obtain the annual average efficiency scores of all banks, and this is thenaggregated at country level and at operating style level.

For better capturing of bank performance during the crises, the study period (1998–2007) is divided into four periods:

1. Before the political crisis (second Gulf war, 1998–2002).2. Political crisis (2003).3. After political crisis (2004–2006).4. During the financial crisis (2007).

Figure 2. GCC commercial banks: Share of assets, (2007 data).

Table 1. Descriptive analysis of input and output variables (2007).

Inputs/Outputs Variables Mean Std Dev Min Max

Inputs US$M Fixed Assets 7.28 24.16 0.03 413.34Non-earning assets 21.86 55.08 0.00 609.61Deposits 424.11 940.28 0.00 11,161.00

Outputs US$M Investments 226.01 525.43 0.00 5,766Loans 256.26 531.37 1.27 7,528.63Off-balance sheet 166.87 423.91 0.00 4,619.70Net profit 8.70 21.52 -289.01 195.97

Anouze


It can be seen from able 2 that the overall average efficiency score is 85.6% for all banks (60 banks); this suggests that with the same level of inputs and by adopting best practices, GCC commercial banks can, on average, increase their outputs by 14.4% (i.e., 100–85.6%). However, the potential increment in outputs from adopting best practices varies from bank to bank. In general, GCC commercial banks have the scope of producing 1.17 times (i.e., 1/0.856) as much outputs from the same level of inputs.

The literature on technical efficiency provides no consensus on how efficiency in banking varies through time in response to market forces (Berger 1993). However, since the study period covers a long and turbulent time (including the second Gulf war in 2003 and the 2007 financial crisis), it is expected that the political and financial crises will dominate the market forces.

It can also be seen from table 2 that, of the 60 commercial banks covered in this study, there are ten banks which are fully-efficient over the entire study period. The overall results show relatively low average efficiency scores; nevertheless, it is possible to detect a slight improvement in the efficiency levels over the study period (+2.2% between 1998 and 2004). In general, the table shows

that the technical efficiency remains relatively stable over the period 1998–2003, then improved a little to reach its highest level (92%) during 2004, while the period 2005–2007 witnessed a volatility of the efficiency score, reach 79% at the end of the period. The year 2005, that is, two years after the Gulf war (political crisis) exhibits a decreased technical efficiency (77%) across all banks studied. It seems that, over time, banks were wasting more resources on average, relative to best practice technical frontiers for the industry.

To find out whether the efficiency scores show a particular trend during the period 1998–2007, the question is whether the mean efficiency score increased since 1998. In fact, Figure 3 shows that the trend of mean efficiency scores decreased over time. It moved in the same direction over the period 1998–2002 (before the political crisis), then declined a little to reach 86% during the second Gulf war. It fluctuated over the study period 2004–2006, reaching its highest level in 2004, and deteriorating to the lowest efficiency level in 2005–2006. The mean efficiency score further declined in 2007 (the year of the financial crisis) to reach 79%. Although 2004seemed to be an atypical year, it is important to note that the performance of GCC commercial banks varied over the study period.

Table 2. Summary of banks’ technical efficiency.

Bank Code Efficiency score

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 Average

Average 89 88 89 88 87 86 92 77 81 79 85.6No of efficient banks 27 30 26 26 27 27 29 24 26 26 10

Figure 3. Technical efficiency of GCC banks during study period (1998–2007).



Another appropriate way to study the trend is by looking at the mean and the standard deviation of technical efficiency. If the banking markets of the GCC became more alike during the ten-year period under consideration, an increase in mean technical efficiency and a decrease in the spread of technical efficiency would be expected. Table 3 shows that the exact mean technical efficiency was relatively stable for the period 1998–2003, and then reached its highest level in 2004. The lowest efficiency score was exhibited during the 2005, which is two years after the second Gulf crisis, then fluctuated below the average for the last two years. The standard deviation was relatively stable for the period 1998–2003, and then reached its lowest level in 2004. The standard deviation tends to be low when average technical efficiency is high, and vice versa. These results strongly support the view that traditional efficiency techniques based on pooled frontier efficiency scores tend to estimate the actual efficiency levels of each bank.

Islamic and conventional banks performance before and during the crisesTo compare commercial bank performance based on their operating style, whether Islamic or conventional, the

efficiency score of all banks at the individual bank level is aggregated at the operating style level to obtain the annual average efficiency scores of Islamic and conventional banks, as illustrated in the figure below.

Figure 4 shows that the Islamic banks outperformed the conventional banks for the first four years (1998–2001), and thereafter their performance declined. It reached its lowest level of the study period (78.6%) by 2003 (second Gulf war). The performance improved to reach 88% by 2004; however, it was still below the performance of the conventional banks. Subsequently, the Islamic banks appeared to be ahead of the GCC commercial banks, with an average efficiency score of around 89.3%.

For further analysis and comparison between the performance of Islamic and conventional banks over the study period, a Mann-Whitney rank sum test was applied. The Mann-Whitney test, which is an alternative to the independent group t-test, is a nonparametric (distribution-free) test for testing whether the number of times scores from one sample are ranked significantly higher than scores from another, unrelated, sample. Similar to many non-parametric tests, it uses the ranks of the data rather than their raw values to calculate the statistic. For this test, the efficiency score is considered as the group variable and the bank operating style as the test variable.

The results of the Mann-Whitney test reveal that there is no significance difference in bank efficiency performance due to differences in operating style. Hence, the null hypothesis that the two efficiency scores have the same value of median is rejected at the 5% level of significance.

Table 3. Statistical descriptive of average overall technical efficiency.

1998– 2002

2003 2004– 2006

2007

Mean 88.02 86.20 83.37 79.30Std Dev 14.04 16.50 20.03 24.60

Figure 4. Performance of Islamic and conventional banks before and during the financial crisis.

Anouze


Efficiency of commercial banks across GCC countriesTo measure commercial bank technical efficiency across countries the efficiency score for all banks is aggregated at country level to get the annual average efficiency scores for each country. Figure 5 shows that the Kuwaiti banks outperform other countries banks before and during the second Gulf crisis. Thereafter, Kuwaiti banks decline, becoming the worst performers of all GCC countries, and then become even worse during the financial crisis. Although there is tough competition between Saudi and UAE commercial banks as they appear to be following the same pattern before and during the second Gulf war, UAE banks’ performance deteriorated during and after the financial crisis.

Qatari banks performed badly before and during the second Gulf war; however, performance increased rapidly after the crisis, but declined again during the financial crisis. The performance of Bahraini and Omani banks followed the same behaviour before, during and after the second Gulf war, whereas the Omani banks outperformed all other GCC commercial banks during the financial crisis.

For further investigation of the efficiency score across GCC countries, we adopted the Kruskal-Wallis rank test (Sueyoshi and Aoki 2001) to examine whether or not scores vary among countries. The Kruskal Wallis X^2 statistics are 6.952 (p = 0.224), meaning that there is no statistically significant relationship between the geographical location of a bank and its efficiency scores.

Previous analyses have been directed mainly at bank managers; however, regulators may require different

information in order to assist them in developing a strong and healthy environment. Similarly, investors want to know where to invest their money in a way that will maximize their return.

Classification and Regression Tree (CART) analysisThe first stage results show the differences in inefficiency among banks in the six countries. In this stage an efficient score is treated as a target variable, while the internal and external environmental factors are considered as predictors for the CART algorithm. These factors were identified from the related literature and include economical, financial and political factors. Data of 24 factors was collected and tested to determine the appropriate factors to include in the CART analysis. Correlation tests showed a high correlation between numbers of factors. For example, number of branches and number of employees were highly correlated so we included only a number of branches to reflect the size of banks. Also, the price/book value and price earnings ratio was highly correlated so we included only a price/book value factor to reflect the size of the stock market price for each bank. Therefore, fifteen factors were considered as input factors for the CART algorithm (see table 6).

All factors as an input of CART algorithmWe built different CART models with a different selection of input factors for CART with the efficiency score as target. First, we included all factors as inputs and efficiency classification as output. Figure 6 shows the importance of variables. The fifteen environmental factors were considered to be important in predicting the fully-efficient banks; only seven of these factors are considered as primary splitters for the decision tree. Assets structure is the most important factor (100%), followed by financial strength (92%) and ROA (91%), whereas operating style, population density, size, and support rating have low importance. This suggests that banks should give more importance to their assets structure as it is one of the most important factors for the efficiency of banks.

Figure 7 shows the predicated accuracy of the generated tree:

Out of 3,660 cases, 1586 cases are actually efficient and predicted to be efficient; and 2074 cases are inefficient

Table 5. Results of Kruskal-Wallis test.

Bank Location

N Mean Rank

x2 d.f. Asymp. Sig.

Bahrain 11 34.32 6.952 5 0.224Kuwait 9 19.94Oman 5 39.20Qatar 6 32.00Saudi Arabia 9 36.00UAE 20 28.05

Figure 5. Bank performance across GCC countries before and during the financial crisis.

Table 4. Mann-Whitney test for 2007 results.

Bank Type Sample Size

Mean Rank

Mann-Whitney U

Z-value

Islamic 12 29.6 245.5 -0.82Conventional 48 34.04



and predicted to be so. This means that the accuracy in predicting the efficient and inefficient banks is 100%, which represents a high level of confidence. Certain of the rules extracted for efficient and inefficient banks are as follows:

4. Rules for efficient banksBanks are efficient (total of 1586 cases) if:

1. Financial strength is greater than or equal 4.0, ROA is greater than or equal to 2.59, and country is less than 4 (122 cases).

2. Financial strength is greater than or equal 4.0, ROA is greater than or equal to 2.59, country is greater than or equal to 4, and internal growth is greater than or equal to 4 (61 cases).

3. Financial strength is greater than or equal to 4.0, ROA is less than 2.59, internal growth is greater than

Table 6. Statistical description of environmental factors.

Variable Descriptive Statistics

Variable type Minimum Maximum Mean Std. Deviation

Establish Date Categorical 1 5Country Categorical 1.00 6.00Inflation Numerical 3.60 14.00 8.61 4.71Population Density Categorical 0.70 23.60Operating Style Categorical 1.00 2.00Internal Growth Numerical 0.27 45.15 14.93 8.74GDP Growth Numerical 1.90 8.40 6.34 1.98Bank Size Categorical 1 3Return on Assets (ROA) Numerical -2.53 8.28 2.76 1.53Return on Equity (ROE) Numerical -34.18 33.37 17.79 8.86Financial Strength Numerical 1.00 13.00 7.90 4.36Support Rating Categorical 1.00 4.00Loan to Deposit Ratio Numerical 28.50 1,904.35 138.76 263.59Market Share Numerical 0.00 8.44 1.67 1.80Asset Structure Numerical 0.02 3,534.00 209.70 518.82

Established date: Banks are grouped according to their established date into 5 groups to capture the age affect: group 5 banks established before 1960; group 4 (1960–1970); group 3 (1970–1980); group 2 (1980–1990) and group 1 (1990–2000). It is expected to have strong positive relationship between bank performance and the established date; the older are the more efficient. Country: Although, GCC countries mostly have the same regime, it is expected to have a variation in efficiency score according to the bank geographical location due to differences in each country regulations. Inflation: is an indicator of macroeconomic stability, and is directly related to the interest rate levels and, thus, interest expense and revenue. Population density: is measured as a ratio of country population to the GCC countries total populations. It is believed that banks in heavily populated countries are more likely to operate closer to their optimal size than banks in less populated country. Hence it is easier for bank management to sustain higher efficiency levels in heavily populated areas than in less populated. Operating style: to capture the efficiency of Islamic rule and regulations. Internal growth rate: is calculated as the percentage of retained profits of the year on the equity at the beginning of the year. Bank size: is measured by the bank total assets, which classified into three groups hence, the larger banks (with total assets more than US $15,000 Million), medium size (with total assets between US $5,000 – 15,000 Million) and small size (total assets less than US $5,000 Million). Profitability ratios: we measure this variable using return on assets (ROA) and return on equity (ROE). Financial strength rating: it provides an opinion of a bank’s intrinsic safety, soundness and risk profile (Arab banking and finance, 2007). It takes a scale from AAA (extremely strong finance and highly attractive operating environment) to D (extremely weak financial condition and untenable position). Support rating: it assesses the possibility that the bank will receive enough financial assistance from the government or private owners in the event of difficulties to enable them to meet their financial obligations. It takes a scale from 1 (very likely) to 5 (very unlikely) (Arab banking & finance, 2007). Loan/Deposit: loan-to-deposit ratio is a measure of the extent to which banks are able to transform deposits into loans. It is mainly used to measure the loan and deposit fund utilization of banks. Market Share: is the ratio of total deposit of each bank to total deposit of all banks. Asset structure: is the ratio of tangible assets to the total assets.

Anouze


or equal to 5.66, and established date is greater than or equal 4 (100 cases).

5. Rules for inefficient banksBanks are inefficient if:

1. Financial strength is greater than or equal 4.0, ROA is greater than or equal to 2.59, country is greater than or equal to 4, and the internal growth is less than 4.44.

2. Financial strength is greater than or equal 4.0, ROA is less than 2.59, and internal growth is less than 5.66 (854 cases).

External Factors as a Single Input of CART AlgorithmTo investigate the impact of the economic and political factors (external) on bank performance, CART is drawn by including only the external factors. All the external environmental factors are considered to be important in setting the rules for fully-efficient banks. Operating style and established date are the most important factor, followed by inflation (89.14%). Support rating and GDP growth seems to have medium importance whereas country and total population density have low. The predictive accuracy of the generated tree is 92%, which represents a high level of confidence. The rules of efficient and inefficient banks that extracted as follow:

6. Rules for efficient banksBanks are efficient if:

1. Established date is greater than or equal to 5, GDP growth is less than 7.95%, inflation is less than 5.72, country less than or equal to 4, support rating is greater than or equal to 2.5 but less than or equal to 3.5, and operating style is 1 (61 cases).

2. Established date is greater than or equal to 5, GDP growth is less than 7.95%, inflation is less than 5.72, country less than or equal to 4, support rating is greater than or equal to 2.5 but less than or equal to 3.5, and operating style is 2 (61 cases that represent 16.7%).

3. Established date is greater than or equal to 5, GDP growth is less than 7.95%, inflation is less than 5.72, country less than or equal to 4, support rating is greater than or equal to 2.5 but less than or equal to 3.5, and operating style is 2 (183 cases).

7. Results and DiscussionThe overall technical efficiency for all GCC commercial banks over the study period is 85.6%. It reaches its highest level in 2004, which is one year after the second Gulf crisis. The reason behind this unexpected improvement in performance could be due to the injection of more money into the market through policy makers and regulators deciding to produce more oil in order to avoid failure of the banking sector or bankruptcy after the Gulf crisis. Therefore, the banking sector performed well, until the regulators stopped the injection of funds, when performance declined to reach its lowest level over the study period. It is worth noting that the performance of the banking sector in countries like Saudi Arabia (the largest oil producer), Qatar (the largest gas producer), and Oman all improved after the second Gulf war. The performance of banks in all GCC countries deteriorated during the financial crisis, except for Omani commercial banks, which reached their highest performance level during the crisis.

Figure 6. Factors importance in predicting fully-efficient banks.

Figure 7. Predicated accuracy of tree.



The highest average efficiency score is for the Saudi banks, at around 89.8%, followed by banks of UAE, which have an efficiency score of 86.3%. There seems to be tight competition between Omani and Bahraini commercial banks, which have average efficiency scores of 85.7% and 85.1%, respectively. Banks operating in Qatar are the least efficient banks, with a score of around 81.3%.

Although, not really comparable as they differ in terms of frontier, inputs and output variables, and the study period, these results are in the line with the research of Al Shammari (2003), who found that the banks of Saudi Arabia and UAE are ahead of those in the other GCC countries, while Qatar and Bahrain have the poorest performing banks.

When the GCC commercial banks efficiency scores are compared with those of their counterparts in other countries (e.g., Singapore banks -95%; Japan -87%; Germany -92%; Peru -98%), the results show that, on average, those of the GCC banks are lower. Nevertheless, the results are relatively similar to the average efficiency for banks in industrial countries like France (84.3%), US (83%), UK (83.9%), Spain (82–84%), or other developed countries such as

Lebanon (84%) and China (85%) (Ariss 2008; Avkiran 2009; Burki and Niazi 2009; Emrouznejad and Anouze 2010; Emrouznejad and Anouze 2009; Hermes and Nhung 2008; Huang et al. 2010; Ismail, Davidson and Frank 2009; Koetter 2008)

The results suggest that, even though it is possible to detect a slight improvement in the overall efficiency scores, there are marked insignificant differences in bank efficiency levels across GCC countries. Islamic banks seemed to be more affected by the Gulf war than were conventional banks, whereas, during the international financial crisis, Islamic banks seemed to be the more resistant. This could be due to the level of involvement of the banks in the international financial institutes: Islamic banks might have less involved than conventional ones and, hence, they were less affected. Also, it could be due to the differences in the relation between bank and clients, which is based on profit-loss sharing in Islamic banks and based on fixed rate (interest rate) in conventional banks, so the latter made less profit compared with Islamic banks

Assets structure, followed by financial strength and ROA were most important, whereas, operating style, population density, size, and support rating were found to be of low importance. Considering only the external factors, the set of efficiency rules that allow the prediction that a bank is fully-efficient indicate that it should be old, and operate in a country with high GDP growth and a lower level of inflation. Such rules benefit regulators or policy makers in their quest to establish a healthy environment that will help their banking sector to achieve a high level of efficiency as well as be a regional financial hub. Managers could also benefit from this analysis in working to improve their bank performance. DEA produces information to guide an improvement policy for inefficient banks, and any such improvement may result in them being considered fully-efficient banks. Furthermore, investors will find such results in their interest as they will want to invest their money in such a way as to maximize returns. Therefore, managers, policy makers, investors and researchers are encouraged to use the proposed methodology to gain more information about the performance of the

banking sector and to establish a set of rules for the efficient operation of banks.

8. ConclusionThis paper investigates the performance of banks in the Gulf states before, during and after crises (political and financial). The study period (1998–2007) includes two crises: the second Gulf crisis (2003) and the global financial crisis (2007). This period allowed us to take look deeply into each bank’s performance under two different situations. The results show that the overall technical efficiency of all GCC commercial banks is relatively stable over time. The commercial banks of Saudi Arabia appear to be ahead of the GCC countries, followed by banks of UAE, whereas Qatar has the least efficient banks. However, there is no reason to believe that bank performance differs from a statistical perspective according to location. Also, different regulations (if any) that have been put in place within GCC countries during the crises have had more or less the same impact on banks’ performance. Furthermore, conventional banks performed better during the second Gulf crisis, whereas it was the Islamic banks that performed better during the global financial crisis. Nevertheless, from a statistical perspective,Islamic and conventional banks rank more or less are same.

Out of the 24 environmental factors, fifteen were tested and considered to be important, and only seven of them are viewed as primary splitters for the decision tree. Assets structure is the most important factor, followed by financial strength and ROA. The operating style, population density, size and support rating all have low importance. Testing only for the external environmental factors, operating style and established date are the most important factors, whereas country and total population density have low importance.

Finally, this study contributes to the theory in developing a comprehensive framework for measuring bank performance, and in identifying the most important factors that improve bank performance. The study also makes a practical contribution as it is the first to assess the impact of financial and political crises on banks of the Gulf states, and it provides useful information for banks managers, investors and policy makers for tracking banks’ efficiencies in order to maintain a sustainable growing sector, and in providing early warning signals of a bank that is potentially at risk.

The results of this study are limited to the selected banks and study period. Researchers are therefore encouraged to study the performance of the GCC banking sector after the current global financial crisis; also, to compare the performance of Islamic and conventional banks as the different financial tools used by each of them may lead to differences in performance.

ReferencesAl Shammari, S., Structure-conduct-performance and effici-

ency in Gulf Cooperation Council (GCC) banking markets, Unpublished PhD thesis, Wales University, 2003.

Almeida, H., Campello, M., Laranjeira, B., and Weisbenner, S., Corporate debt maturity and the real effects of the 2007 credit crisis, Unpublished working paper, University of Illinois, 2009.

Anouze


Anouze, AL., Evaluating productive efficiency: comparative study of commercial banks in Gulf countries, Unpublished PhD thesis, Aston University, 2010.

Ariss, R., Financial liberalization and bank efficiency: evidence from post-war Lebanon, Applied Financial Economics, 18, 2008. 931–946.

Avkiran, N., Removing the impact of environment with units-invariant efficient frontier analysis: An illustrative case study with intertemporal panel data, Omega, 37, 2009. 535–544.

Banker, R., Chang, H., and Lee, S-Y., Differential impact of Korean banking system reforms on bank productivity, Journal of Banking and Finance, 34, 2010. 1450–1460.

Banker, R., Charnes, A., and Cooper, W., Some models for estimating technical and scale inefficiencies in data envelopment analysis, Management Science, 30, 1984. 1078–92.

Beltratti, A., Stulz, R., Why did some banks perform better during the credit crisis? a cross-country study of the impact of governance and regulation, ECGI’s Finance Working Paper No. 254/2009, 2009.

Berger, A., and Humphrey, D., Efficiency of financial institutions: International survey and directions for future research, European Journal of Operational Research, 98, 1997. 175–212.

Berger, A., Distribution-free estimates of efficiency in the U.S. banking system and tests of the standard distributional assumptions, Journal of Productivity Analysis, 4, 1993. 261–292.

Breiman, L., Friedman, J., Olshen, R. and Stone, C., Classification and regression trees, Pacific Grove, Wadsworth-Monterey, USA, 1984.

Brissimis, S. Delis, M. and Papankiolaou, N., Exploring the nexus between banking sector reform and performance: Evidence from newly acceded EU countries, Journal of Banking & Finance, 32, 2008. 2674–2683.

Burki, Abid and Niazi, G.S., Impact of financial reforms on efficiency of state-owned, private and foreign banks in Pakistan, Applied Economics, 1–14 URL: http://dx.doi.org/10.1080/00036840802112315, 2009.

Casu, B., and Girardone, C., integration and efficiency convergence in EU banking markets, Omega, 38, 2010. 260–267.

Charnes, A., Cooper, W., and Rhodes, E., Measuring the efficiency of decision making units, EJOR, 2, 1978. 429–444.

Chiou, C-C., Effects of financial holding company act on bank efficiency and productivity, Neurocomputing, 72, 2009. 3490–3506.

Chiu, Y-H., and Chen, Y-C., The analysis of Taiwanese bank efficiency: incorporating both external environment, Economic Modelling, 26, 2009. 456–463.

Cornett, M., McNutt, J. and Tehranian, H., The financial crisis, internal corporate governance, and the performance of publicly-traded U.S. bank holding companies, unpublished working paper, Boston College, 2010.

Das, A., and Ghosh, S., Financial Deregulation and Profit Efficiency: A Nonparametric Analysis of Indian Banks. Journal of Economics and Business, 61, 2009. 509–528.

Demyank, Y., and Hasan, I., Financial crisis and bank failures: a review of prediction methods, Omega, 38, 2010.315–324.

Diamond, D., and Dybvig, P., Bank runs, deposit insurance and liquidity, Journal of Political Economy, 91, 1983. 401–419.

Diamond, D., Financial intermediation and delegated monitoring, Review of Economic Studies, 51, 1984. 393–414.

Dietrich, A., Wanzenried, G., Determinants of bank profitability before and during the crisis: Evidence from Switzerland, J. Int. Financ. Markets Inst. Money, doi:10.1016/j.intfin.2010.11.002, 2011.

Duchin, R., Ozbas, O., and Sensoy, B., Costly external finance, corporate investment, and the subprime mortgage credit crisis, Journal of Financial Economics, 97, 2010. 418–435.

Emrouznejad, A, Anouze, AL, and Thanassoulis, E., A semi-oriented radial measure for measuring the efficiency of decision making units with negative data, using DEA, EJOR, 200, 2010a. 297–304.

Emrouznejad, A. and Anouze, AL., A note on the modeling the efficiency of top Arab banks, Expert Systems with Applications, 36, 2009. 5741–5744.

Emrouznejad, A. and Anouze, AL., Data envelopment analysis with classification and regression tree: a case of banking efficiency, Expert Systems, 27, 2010. 231–246.

Emrouznejad, A., Amin, G., Thanassoulis, E. and Anouze, AL., On the boundedness of the SORM DEA models with negative data, EJOR, 206, 2010b. 265–268.

Fethi, M. and Pasiouras, F., Assessing bank performance with operational research and Artificial Intelligence techniques: A survey, EJOR, 204, 2010. 189–198.

Fu, X. and Heffernan, S., The effects of reform on China’s bank structure and performance, Journal of Banking and Finance, 33, 2009. 39–52.

Fukuyama, H. and Weber, W., Estimating output allocative efficiency and productivity change: application to Japanese banks, European Journal of Operational Research, 137, 2002. 177–190.

Fukuyama, H., and Weber, W., A Slacks-based inefficiency measure for a two-stage system with bad outputs, Omega, 38, 2010. 398–409.

Gorton, G., and Winton, A. Financial intermediation, In G. Constantinides, M. Harris, and R. (Stulz, Handbook of the Economics of Finance North Holland-Amsterdam, 2003. 431–552.

Grifell-Tatjé, E., Profit, Productivity and distribution: differences across organizational forms—The case of Spanish banks, Socio-Economic Planning Sciences, 38, 2010. 1–12.

Gryglewicz, S., A theory of corporate financial decisions with liquidity and solvency concerns, Journal of Financial Economics, 99, 2011. 365–384.



Halkos, G. and Salamouris, D., Efficiency measurement of the Greek commercial banks with the use of financial ratios: a data envelopment analysis approach, Management Accounting Research, 15, 2004. 201–24.

Han, J and Kamber, M., Data Mining Concepts and Techniques, San Francisco: Morgan Kaufmann Publishers, 2001.

Hermes, N. and Nhung, T., The impact of financial liberalization on bank efficiency: evidence from Latin America and Asia, Applied Economics, URL: http://dx.doi.org/10.1080/00036840802112448, 2008.

Hsiao, H-C, Chang, H, Cianci, A, and Huang, L-H., First financial restructuring and operating efficiency: Evidence from Taiwanese commercial banks, Journal of Banking & Finance, 34 (2010), pp. 1461–1471.

Hsiao, H-C., Chang, H., Cianci, A., and Huang, L.-H., First financial restructuring and operating efficiency: evidence from Taiwanese commercial banks, Journal of Banking and Finance, 27 (2010), pp. 1461–1471

Huang, L-H., Hsiao, H-C., Cheng, M-A., and Chang, S-J., Effects of financial reform on productivity change, Industrial Management & Data Systems, 108, 2008. 867–886.

Huang, T-H, Liao, Y-T and Chiang, L-C., An examination on the cost efficiency of the banking industry under multiple output prices’ uncertainty, Applied Economics, 42, 2010. 1169–1182.

Ismail A, Davidson, I, and Frank, R., Operating performance of European bank mergers, The Service Industries Journal, 29, 2009. 345–366.

Jeon, Y. and Miller, M., Performance of domestic and foreign banks: The case of Korea and the Asian financial crisis, Global Economic Review, 34, 2005. 145–165.

Jeon, Y. and Miller, M., The effect of the Asian financial crisis on the performance of Korean nationwide banks, Applied Financial Economics, 14, 2004. 351–360.

Kim, H. and Koehler, G., Theory and Practice of Decision Tree Induction, Omega, 23, 1995. 637–652.

Koetter, M., The stability of bank efficiency rankings when risk preferences and objectives are different, The European Journal of Finance, 14, 2008. 115–135.

Koutsomanoli-Filippaki, A., Margaritis, D., and Staikouras, C., Efficiency and productivity growth in the banking industry of Central and Eastern Europe, Journal of Banking & Finance, 33, 2009. 557–567.

Lozano-Vivas, A., and Pastor, T., Do performance and environmental conditions act as barriers for cross-border banking in Europe? Omega, 38, 2010. 275–282.

Mercan, M., Reisman, A., Yolalan, R. and Emel, A., The effect of scale and mode of ownership on the financial performance of the Turkish banking sector: results of a DEA-based analysis, Socio-Economic Planning Sciences, 37, 2003. 185–202.

Park, K. and Weber, W., A note of efficiency and productivity growth in the Korean banking industry, 1992–2002, Journal of Banking & Finance, 30, 2006. 2371–86.

Portela S, Thanassoulis, E, and Simpson, G., A Directional distance approach to deal with negative data in DEA: an application to bank branches, JORS, 55, 2004. 1111–1121.

Portela, S. and Thanassoulis, E., Malmquist-type indices in the presence of negative data: An application to bank branches, Journal of Banking & Finance, 34, 2010. 1472–1483.

Ray, S., and Das, A., Distribution of Cost and Profit Efficiency: Evidence from Indian Banking, European Journal of Operational Research, 201, 2010. 297–307.

Ripley, B., Pattern recognition and neural networks, Cambridge University Press, Cambridge, 1996.

Shleifer, A. and Vishny, R., Unstable banking, Journal of Financial Economics, 97, 2010. 306–318.

Siriopoulos, C., and Tziogkidis, P., How do Greek banking institutions react after significant events? – A DEA Approach, Omega, 38, 2010. 294–308.

Staub, R., Souza, G., and Tabak, B., Evolution of bank efficiency in Brazil: A DEA approach, European Journal of Operational Research, 202, 2010. 204–213.

Sueyoshi T. and Aoki, S., A use of a nonparametric statistic for DEA frontier shift; the Kruskal and Wallis rank test, Omega, 29, 2001. 1–18.

Thanassoulis, E., Introduction to the Theory and application of data envelopment analysis: a foundation text with integrated software, Kluwer Academic Publisher, Massachusetts, USA, 2001.

Thompson, R. Singleton, F. Thrall, R. and Smith, B., Comparative site evaluations for locating a high-energy physics lab in Texas, Interfaces, 16, 1986. 35–49.

Torgo, L. (1997), Functional Models for regression tree leaves, Proceedings of the Fourteenth International Conference on Machine Learning (ICML 1997), Nashville, Tennessee, USA, July 8–12, 1997.

Xiao, Y., French Banks Amid the Global Financial Crisis, IMF Working Paper, WP/09/201, 2009).

Zhao, T., Casu, B., and Ferrari, A., The impact of regulatory reforms on cost structure, ownership and competition in Indian banking, Journal of Banking & Finance, 34, 2010. 246–254.

Zhou, Z. and Jiang, Y., Medical diagnosis with C4.5 Rule preceded by artificial Neural network ensemble, IEEE Transactions on information Technology in Biomedicine, 7, 200337–42.

Anouze


Appendix

Table A-1. Summary of selected studies.

Study CountryStudy Period Approach Inputs Outputs

Banker, Chang, & Lee (2010)

Korea 1995–2005 Intermediation (i) interest expense and (ii) other operating expense

(i) interest revenue, & (ii) other operating revenue

Casu & Girardone, (2010)

European Countries

1997–2003 Intermediation (i) Personnel expenses, (ii) other administrative expenses, (iii) interest paid, (iv) non-interest expenses.

(i) total loans and (ii) other earning assets.

Chiou (2009)

Taiwan 1999–2004 Intermediation (i) staff, (ii) fix asset, (iii) bank deposits (including current deposits, savings deposits, time deposits, check deposits, & other deposits), & (iv) salary expense.

(i) Provision of loan services (business & individual loans), (ii) investments (iii) interest revenue and (iv) non-investment revenue.

Chiu & Chen (2009)

Taiwan 2002–2004 Intermediation (i) Number of employees, (ii) total deposits, (iii) fixed assets

(i) Total amount of loans, (ii) total investment, (iii) non-interest revenue.

Das & Ghosh (2009)

India 1992–2004 Intermediation (i) deposits, (ii) labor, (iii) capital/fixed assets (iv) equity

(i) Loans & advances, (ii) investments, (iii) other income.

Fukuyama & Weber (2010)

Japan 2000–2006 Production and intermediation

1st stage: (i) labor, (ii) physical capital, (iii) financial capital 2nd stage: (i) deposits

1st stage: (i) deposits 2nd stage: (i) loans, and (ii) securities investments, and (iii) other business activities.

Grifell-Tatjé (2010)

Spain 1994–2004 Intermediation (i) Real operating profit from intermediation activities, (ii) real gross loan and financial income, and (iii) average value of loans & financial investments.

(i) financial expense (interest on deposits, loans, labor expense)

Hsiao et al. (2010)

Taiwan 2000–2005 Intermediation (i) interest expenses, (ii) non-interest expenses, and (iii) total deposits.

(i) interest revenue, (ii) non-interest revenue, and (iii) total loans.

Lozano- Vivas & Pastor (2010)

European Countries

2004 Production (i) labour, (ii) funds and (iii) physical capital

(i) loans, and (ii) other earning assets

Ray & Das (2010)

India 1996–2006 Intermediation (i) deposits, (ii) labor, (iii) capital/fixed assets (iv) equity & reserves

(i) investments, (ii) earning advances, and (iii) other income

Siriopoulos & Tziogkidis (2010)

Greece 1995–2003 Intermediation (i) Personnel expenses, (ii) provisions, (iii) operational expenses.

(i) Financial claims, (ii) operational income, (iii) net income before taxes.

Staub, Souza, & Tabak (2010)

Brazil 2000–2007 Intermediation (i) Labor, (ii) capital, & (iii) purchased funds.

(i) outputs, (ii) loans and (iii) investments,


Cite this chapter as: Srairi S, Kouki I, Harrathi N (2015). The relationship between Islamic bank efficiency and stock market performance: Evidence from GCC countries. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

The relationship between Islamic bank efficiency and stock market performance: Evidence from GCC countriesSamir Srairi1, Imen Kouki2, Nizar Harrathi3

1Associate Professor, Finance, Faculty of Law, Economics and Management of Jendouba, (LAREQUAD), Tunisia, E-mail: [email protected] Professor, Finance, Higher Institute of Management, (LAREQUAD), Tunisia, E-mail: [email protected] Professor of Quantitative Methods, Faculty of Economic Sciences and Management of Nabeul, (LAREQUAD), Tunisia

Abstract - Using data envelopment analysis (DEA), this paper estimates the efficiency of 25 Islamic banks operating in Gulf Cooperation Council (GCC) countries during the period 2003–2009. It also examines the relationship between the efficiency of Islamic banks and the performance of their stock. The results suggest that efficiency measures, particularly technical and pure technical efficiency, have increased over the period of study but remain low as compared to conventional banks. The inefficiency of Islamic banks can be attributed to pure technical inefficiency rather than to scale inefficiency. We also find that large and small banks are more efficient than medium banks in terms of overall technical efficiency. Furthermore, the empirical findings show that both technical and pure technical efficiency changes are positively related to share returns, while changes in scale efficiency have no impact on stock performance. Finally, the regression also indicates a significant and positive association between market return and the book-to-market equity ratio with share prices.

Keywords: banking, technical efficiency, stock performance, Islamic banks, data envelopment analysis, GCC countries

1. IntroductionAn Islamic bank is an institution that mobilizes and invests financial resources according to Shariah. Islamic banking transactions are based on six basic principles: prohibition of interest, risk sharing, money as potential capital, prohibition of speculative behaviour, sanctity of contracts, and Shariah approved activities (Iqbal 1997).

Islamic banking, which started to operate from the 1960s, exists today in all regions of the world, particularly in the Middle East and Southeast Asia. According to the report of the Blominginvest bank, which was established in February 2009, more than 390 Islamic financial institutions are spread across 75 countries with total assets estimated to be close to $1 trillion by 2010. The rating agency, Moody’s Investors’ Service, forecast that Islamic bank assets worldwide will reach $4 trillion within five years. The Islamic financial system is considered to be one of the fastest growing financial and economic sectors in the

world. During the last decade, the Islamic banking industry has grown at a remarkable pace, at 20–30% per year being three times the rate for conventional banks. According to many reports, the rapid and continued growth of Islamic banking is driven by multiple factors such as: increasing demand from a large number of Muslims; increasing oil wealth of Muslim countries; low banking penetration in Muslim majority nations; increasing demand from non-Muslim customers and countries; and the support of government and regulatory bodies for the development and promotion of Islamic banking.

Furthermore, the Islamic financial system has been less affected than the traditional system by the latest economic and financial crisis (2008), due mainly to its profit-loss sharing principle, and also because of its strict prohibition of investments in risky instruments, such as toxic assets and derivatives. In addition, according to an IMF survey (2010) and Chapra (2009), Islamic banks have contributed to

Srairi et al.


financial and economic stability during the global financial crisis. The strong performance of Islamic banks over recent years has encouraged several universal banks in developed countries to add Islamic products to their conventional banking industry, through Islamic banks windows or Islamic banking subsidiaries.

In view of the rapid growth of Islamic banks, several issues are revealed about the performance of these financial institutions. In addition, as Islamic banking was introduced as a parallel system of conventional banks in the majority of countries, the performance of the new form of banking may have an impact on the soundness and stability of the banking system as a whole (Mariani 2010). Moreover, the last economic and financial crisis has turned the focus towards Islamic financial institutions which, according to many sources, have showed stronger resilience than conventional banks (e.g., Moody’s; IMF working paper 2010). Despite the strong position of Islamic banks, several studies (Iqbal 2007; Iqbal and Van Greuning 2007) have identified weaknesses and vulnerabilities among Islamic banks in the areas of risk management (operational risk; weak internal control processes) and human resource issues (quality of management; technical expertise; professionalism). Therefore, it will be interesting to analyse the performance of Islamic banks during the last decade in order to provide some guidelines for managers, investors and policy makers to improve the efficiency of these banks and to formulate managerial strategies and public policies. Therefore, the aim of this study is to investigate the efficiency of Islamic banks operating in Gulf Council Cooperation (GCC) countries during the period 2003–2009, and to examine the relationship between the efficiency of Islamic banks and the performance of their stock. To our knowledge, this is the first study which analyses the relationship between efficiency and share performance in the context of Islamic banks in GCC countries.

To gain a better understanding of the Islamic banking sector in GCC countries, our analysis is conducted in two steps. First, by employing Data Envelopment Analysis (DEA) as a non-parametric approach, we estimate the technical efficiency of 25 GCC Islamic banks under the profit-oriented method which defines cost variables as inputs, and revenue variables as outputs. In addition, to analyse the sources of inefficiency of these banks, we calculated pure technical efficiency and scale efficiency as two components of technical efficiency. We chose a period of six years between 2003 and 2009 in order to investigate the evolution of the efficiency of Islamic banks over time. Moreover, in this study we attempt to compare the efficiency measures of Islamic financial institutions according to their size in terms of total assets. Following several studies concerning the conventional banking industry (e.g., Haddad et al. 2010; Pasiouras 2008; Beccali et al. 2006), in the second stage of this paper we investigat the potential association between Islamic banks’ efficiency and their share prices. To meet this objective, we regress annual stock returns calculated as the sum of daily share returns on efficiency scores obtained in the first step, adding some control variables.

This paper presents some interesting points compared with some other studies on Islamic banking efficiency in GCC countries. First, our sample comprises more than 90% of GCC Islamic banks assets, which makes it the most

comprehensive database on the GCC Islamic banking industry. Also, to the best of our knowledge, this is the first study that relates the efficiency of Islamic banks in GCC countries to their stock prices. Finally, our paper also attempts to study the impact of the recent economic and financial crisis on the performance of GCC Islamic banks, and compares the efficiency of large, medium and small banks.

2. Literature reviewTwo streams of literature are discussed in this study, the first concerning the efficiency of Islamic banks, the second being relevant to the relationship between bank efficiency and share performance.

Studies on Islamic bank efficiencyWhile there is wide discussion in the literature on bank efficiency within the conventional bank sector, particularly for the developing countries and, to a smaller degree, the transition economies, the work on Islamic banks remains limited. Even with the development of the Islamic banking sector in several regions of the world, few studies have evaluated the efficiency of the new form of banking, and noneconcern the relationship between bank efficiency and share performance.

According to Bashir (2007) and Sufian et al. (2008), the majority of studies on Islamic banks have focused on the concept issues describing the underlying principles (Al-Omar and Iqbal 2000; Zahar and Hassan 2001; Lewis, 2008) and performance measures using the traditional financial ratios of these type of banks (Bashir 2001; Olson and Zoubi 2008; Srairi 2009). A few studies have utilized frontier analysis techniques rather than traditional methods to estimate the efficiency of Islamic banks. Using both the stochastic frontier approach (SFA) and the DEA models, Hassan (2007) estimated a variety of parametric techniques (cost, profit efficiency, and productivity)to a panel of 43 Islamic banks operating in 22 countries during the period 1993–2001. He found that Islamic banks are relatively more efficient in generating profits compared with control costs. In fact, the score of profit efficiency was found to be about 84%, while for cost the efficiency was only 74%. The results also indicated that the major source of inefficiency was allocative inefficiency rather than technical inefficiency.

Mokhtar et al. (2008) used a non-parametric DEA technique and an intermediation approach to estimate the technical and cost efficiency of the fully-fledged Islamic banks as well as Islamic windows in Malaysia from 1997 to 2003. The main results of the study revealed that, although the fully-fledged Islamic banks were more efficient than the Islamic windows, the two types of Islamic banks were still less efficient than the conventional banks. This finding also showed that the average efficiency of the overall Islamic banking sector increased over the survey period.

Employing the DEA model, Sufian et al. (2008) examined the technical efficiency and its components (pure technical efficiency and scale efficiency) of 37 Islamic banks operating in 16 MENA and Asian countries during the period 2001–2006. The results suggest that pure technical


The relationship between Islamic bank efficiency and stock market performance: Evidence from GCC countries

inefficiency dominated scale inefficiency of Islamic banks during all years except for the year 2006. On the other hand, the authors found that the MENA Islamic banks exhibited higher technical efficiency compared to their Asian Islamic bank counterparts.

A more recent study concerning GCC countries was conducted by Srairi (2010), who employed a SFA model with country-specific environment variables and estimated the cost and profit efficiency of 71 commercial banks during the period 1999–2007. The empirical results indicated that, on average, the conventional banks are more efficient in terms of cost and profit than the Islamic banks. This study also revealed that both conventional and Islamic banks in Arab Gulf countries are relatively more efficient in generating profits than in controlling costs.

Bank efficiency and share performanceWhile there is an extensive literature examining several issues on bank efficiency, such as the impact of liberalization on the efficiency of banks (e.g., Chen et al. 2005; Das and Ghosh 2006; Paul and Kourouche 2008), the sources of bank inefficiency (e.g., Grigorian and Manole 2006; Pasiouras 2008; Sufian 2009), the comparison of the efficiency of banks according to country (e.g., Fries and Taci 2005; Kasman and Yildirim 2006; Inui et al. 2008), ownership structure (e.g. Isik and Hassan 2003; Bonin et al. 2005; Kyj and Isik 2008), and comparison of type of bank (foreign and domestic: Havrylchyk 2006; new and old: Canhoto and Dermine 2003; conventional and Islamic: Srairi 2010), only a limited number of papers have investigated the impact of the efficiency of banks on stock performance, and none of these papers have concerned Islamic banks. The relationship between the efficiency of banks and stock performance within the conventional banking sector has been studied both on the basis of an individual country and for a cross-section of countries.

Haddad et al. (2010) estimated the monthly efficiency and productivity of 24 listed Indonesian banks and their market performance using the non-parametricSlack-Based Model (SBM) approach over the period January 2006 to July 2007. They found that the stock market values of the banks were in accordance with their performance. The results also indicated a positive correlation between the index of the Indonesian stock exchange (JCI) and bank efficiency. On the other hand, the findings suggest that Indonesian banks with foreign ownership tend to be less efficient than their domestic counterparts.

Using both DEA and SFA methods, Xiang and Shamsudding (2009) calculated the technical, cost and profit efficiency of nine publicly-listed Australian banks over the period 1997–2007, and analyzed the potential link between these efficiency scores and stock returns. They observed that an improvement in cost and profit efficiency, calculated using the SFA model, increased bank stock performance. However, the DEA efficiency scores were uncorrelated with stock returns.

Pasiouras et al. (2008) examined the association between the efficiency of ten Greek banks and their share performance between 2000 and 2005. The authors used the DEA technique (profit-oriented approach) and computed

three efficiency levels: technical efficiency under constant returns to scale (CRS); technical efficiency under variable returns to scale (VRS); and scale efficiency. The results indicated that annual changes in technical efficiency (under CRS or VRS) were positively related to stock returns, while changes in scale efficiency had an insignificant impact on share performance. Erdem and Erdem (2008) used a DEA with intermediation approach, and found no association between stock price returns and change in economic efficiency for Turkish banks.

Across international financial markets, Beccali et al. (2006) used both SFA and DEA approaches to estimate cost efficiency for a sample of banks operating in five European countries (France, Germany, Italy, Spain and United Kingdom) in the year 2000. The results suggested that the change in the prices of bank shares reflects percentage changes in cost efficiency, particularly those derived from DEA. More recently, Liadaki and Gaganis (2010), who employed a larger sample (15 EU countries and 171 banks) and a longer time period (2002–2006) than Becalli et al. (2006), estimated the cost and the profit efficiency by using the SFA model and taking into account the macroeconomic and other country-specific characteristics. The main result of this study showed higher profit inefficiency (21%) than cost inefficiency (10%). This means that European banks are more efficient in controlling costs than in generating profits. However, Srairi (2010) found that profit efficiency scores are more informative to shareholders and investors in Gulf Arab countries. In fact, changes in profit efficiency have a positive and significant effect on stock returns, while there is no association between changes in cost efficiency and stock returns.

3. Methodology and dataIn this study, we employ a three-stage procedure to analyse the efficiency of Islamic banks and the relation to share price performance:

1. A non-parametric approach (DEA technique) is used to estimate efficiency scores with an input-oriented model.

2. Annual stock returns are calculated on the basis of daily share returns in order to measure the share performance for each bank.

3. The relationship between bank efficiency and stock performance is examined by regressing the annual return on stock against the yearly change of efficiency levels.

DEA modelFrom the literature, it is apparent that two models are used to examine the efficiency of banks. Parametric techniques, such as Stochastic Frontier Analysis (SFA), Thick Frontier Approach (TFA), and Distribution Free Approach (DFA), use econometric tools and specify the function form for the cost or profit function. On the contrary, non-parametric approaches, such as Data Envelopment Analysis (DEA) and Free Disposable Hull Analysis (FDHA), do not make an assumption concerning the functional form of the frontier, and use a linear program to calculate the efficiency level. The small size of our sample pushed us to adopt the DEA technique, which was first introduced by

Srairi et al.


Charnes et al. (1978). According to Avkiran (1999), DEA is thought to work well with fewer data, fewer assumptions, and limited sample sizes. Furthermore, DEA does not require any specification of the functional form on the data to construct the production frontier, and the distribution forms of errors (Bauer et al. 1998). However, DEA has some limitations. This technique is very sensitive to outlying observations, and all deviations from the frontier indicate inefficiency (Havrylchyk 2006). Moreover, the DEA approach does not allow for any error in the data and, in consequence, it may overstate the true levels of relative inefficiency for some entities (Drake and Hall 2003; Berger and Mester 1997). Despite its limitations, we propose that DEA is a robust tool for examining the efficiency of Islamic banks in GCC countries.

DEA is a deterministic model that can be used to examine the relative efficiency of a number of entities (decision-making units: DMUs) in the sample having the same multiple inputs and multiple outputs. To calculate the efficiency scores, a linear programming model is solved for each bank. The DEA model measures the efficiency of each DMU relative to all other DMUs, with the simple restriction that all DMUs lay on, or below, the efficiency frontier (Das and Ghosh 2006). If a DMU lies on the frontier, it is referred to as an “efficient unit”. Otherwise, it is DEA-inefficient. The value of the efficiency score for each DMU is ranged between zero and one. To define the best practice frontier, DEA can run under either constant returns to scale (CRS), or variable returns to scale (VRS). The main difference between these two models is the treatment of returns to scale. The VRS model, which was defined by Banker et al. (1984), compares each bank only with other banks operating in the same region of return to scale (banks of similar size). However, the CRS assumption is only justifiable when all banks are operating at an optimal scale. It means that a rise in inputs results in a proportionate rise in outputs. On the other hand, a DEA model can be constructed using an input-orientation (minimizing inputs) or output-orientation (maximizing outputs) approach.

The input-orientation approach is defined as the ability of the bank to obtain a given level of outputs by utilizing a minimum combination of inputs; the opposite approach analyzes the ability of banks to produce the maximum level of outputs, given the current level of inputs (Cooper et al. 2000). In this study, we adopt an input-oriented DEA technique because of the expressed interest of the Islamic banking sector in more control costs. Many studies (e.g., Archer and Abdel-Karim 2002; Kamaruddin et al. 2008) conclude that the cost of funds and labour in Islamic banks is higher compared with those in conventional banks.

The DEA approach permits calculation for each bank of the overall technical efficiency (TE) and its two components, pure technical efficiency (PTE) and scale efficiency (SE). PTE, also called “managerial efficiency”, represents the failure of the bank to extract the maximum output from its adopted input level and, hence, it relates to the ability of the manager to utilize the firm’s given resources (Drake and Hall 2003; Pasiouras 2008). SE, another indicator of efficiency, measures the proportional reduction in input usage if the bank can operate at a point where the production exhibits CRS (Kyj and Isik 2008). It can be computed by dividing TE under the assumption of CRS to the TE under the VRS

assumption (TE = PTE*SE). To calculate these efficiency scores, we employed the software DEAP version 2.1 developed by Coelli (1996).

Specification of inputs and outputsTo estimate the efficiency frontier using the DEA techni-que, we needed measures of inputs and outputs. In the literature, there has been little consensus over which inputs and outputs should be used with the DEA model and how they could be measured (Berger and Humphrey 1992). Consequently, several approaches are used in bank efficiency studies: the production approach, the intermediation approach, the operating approach, and the profit approach.

Following recent studies on bank efficiency (e.g., Drake et al. 2006; Pasiouras 2008; Sturm and Williams 2004), in this study we adopt the profit-oriented approach. This method focuses on revenues as well as costs. It also has the advantage of allowing a better understanding of the strategies used by banks to respond to the changes in environment. Accordingly, three inputs and two outputs are selected to estimate efficiency levels. Hence, the vector of inputs comprises: employee expenses (x1), other operating expenses (x2) and loan loss provisions (x3). The vector of outputs includes two variables: net interest income (y1 = interest income- interest expense) and other operating income (y2).

Bank efficiency and share performanceOnce the efficiency scores (TE, PTE, SE) and the annual share returns are computed,in the third stage of this study we examine the impact of the efficiency of Islamic banks on performance (e.g., Liadaki and Gaganis 2010; Sufian and Abdul-Majid 2009; Erdem and Erdem 2008). The relationship is checked using the following linear model:

RSit = a + β1CEit + β2MRjt + β3BSFit + eit (1)

where RSit is the annual return on bank i’s stock in year t. CEit represents the annual percentage change in bank efficiency and includes the technical (TE, model 1) or pure technical (PTE, model 2) or scale efficiency (SE, model 3) for bank i in year t. MRjt is the market return for the banking sector j in year t, and BSFit concerns some specific factors and includes two variables, LTAit, which is the size of bank i in year t measured as the natural logarithm of total assets, and BMit, which is the book-to-market equity ratio calculated as the ratio of the book value of a bank’s equity to its market value. The a intercept represents the constant of the model, bi is the parameters to be estimated and eit is the disturbance term calculated as follows:

eit = uit + vi

Since we have a panel regression combining cross section and time series data, we estimate this model by using a fixed effect model (ni which represents bank specific effect is fixed over time) and a random effect model (in the case ni is considered as an error term). The fixed effect model is tested by the Fisher (F) test, while the random effect model is examined by the Lagrange Multiplier (LM) test. If the null hypothesis of heteroscedasticity residual variance is rejected, the ordinary least square (OLS) regression



is favored. To choose between these two models, we calculated the Hausman test (H).

DataOur sample comprises 25 Islamic banks operating in five Gulf Arab countries (GCC) with six banks in Bahrain, eight banks in Kuwait, two banks in Qatar, two banks in Saudi Arabia, and seven banks in the United Arab Emirates, over the period 2003–2009. The choice of region is justified for many reasons: first, the GCC countries, which comprise six states (Bahrain, Kuwait, Qatar, Saudi Arabia, the United Arab Emirates, and Oman) hold the largest share (about 61.6%) of Islamic bank’ assets in the world ($263 billion in 2008). Saudi Islamic banks occupy the first place in terms of GCC Shariah-compliant assets (35%), followed by Kuwait (24%), the United Arab Emirates (19%), Bahrain (14%), and Qatar (8%). During the last decade, the Islamic banking sector in GCC countries had achieved strong growth in term of total assets (over 35%). Also, since 2002, the GCC region has been in a relatively strong position (7% growth between 2002 and 2008) and is expected to continue at the same pace and to launch huge projects of more than $1 trillion during the next decade. Finally, while the GCC states provide opportunities in many sectors and offer ample liquidity in the banking sector, Islamic banks are expected to further diversify their products and services and so attract a wider clientele. In addition, the Islamic financial system will continue to spread to investment banking, project finance, capital markets, insurance, wealth management and micro-finance (Iqbal 2007).

The annual data of Islamic banks (financial statements) used to calculate the efficiency scores are collected from Bankscope Database of Bureau Van Dijk’s Company. The daily stock prices and market index are obtained from Datastream. Since Gulf countries have different currencies, all the annual financial values are converted into US dollars using appropriate average exchange rates for each year. Also, to ensure comparability of data across countries, all values are deflated to the year 2003 using each country’s consumer price index (CPI).

Table 1 summarises the mean of inputs and outputs employed in the DEA model and also presents the average value of stock returns and control variables used in the

regression over 2003 to 2009. The table shows a great increase of all inputs and outputs during the period of study. In fact, we note that employee expenses, the other operating expenses, the net interest income, and other operating income have risen about 200%, 211%, 153%, and 243%, respectively. The loan loss provision was constant during 2003–2007 and grew rapidly during the two last years of the study period (2008 and 2009). It is interesting to note that the crisis did not have the same effect on Islamic banks as is reported for conventional banks (Blominvest bank report 2009), the income of Islamic banks exhibiting only a small decrease of 4%. Finally, we note an increase of more than 25% of the average rate of assets.

4. Empirical resultsThe analysis of the empirical findings on the efficiency of Islamic banks in GCC countries is structured in two main parts. First, we estimate the overall technical efficiency and its components, measured by DEA method, and evaluate its evolution over time. Further, we attempt to examine the efficiency of Islamic banks according to their size. In the second part, we extend the analysis by examining the relationship between efficiency scores of Islamic banks and their share performance

DEA efficiency measuresIn this section, we examine the efficiency scores of Islamic banks calculated under the profit-oriented approach and obtained by the DEA technique. In order to analyse the evolution of the efficiency of Islamic banks between 2003 and 2009, we chose to construct a common frontier for all banks in the sample; the implicit assumption was of an absence of technical change during the period of study. In this approach, the efficiency of each bank observed in different years is estimated in relation to a common benchmark technology (Canhoto and Dermine 2003).

Table 2 provides a summary of annual means of efficiency indexes over 2003–2009 classified by year (panel A) and by size (panel B). As can be seen from this table, overall technical efficiency scores exhibit an upward trend from 2003 to 2009. The mean of TE varies from 61.2% (2003) to 68.5% (2009) with an average equal to 65.5%. This result appears to show an improvement of the efficiency of

Table 1. Summary statistics of dataset used in the study (average values).

2003 2004 2005 2006 2007 2008 2009

Panel A: inputs and outputsa

– Employee expenses (x1) 23.47 25.21 32.7 48.35 67.63 73.89 70.27– Other operating expenses (x2) 19.66 21.26 36.47 43.77 52.78 58.83 61.22– Loan loss provision (x3) 19.95 19.05 17.71 13.16 18.92 61.73 120.47– Net interest income (y1) 103.92 116.50 168.35 200.29 272.66 275.72 262.78– Other operating income (y2) 27.86 38.62 77.86 113.12 157.13 132.67 95.82

Panel B: control variables and stock returnb

– Total assets (US$ Millions) 2707 3065 3835 5200 7124 9062 9739– Book-to-market equity – 2.15 0.90 1.35 1.45 2.23 1.53– Annual stock return – 44.82 67.84 -27.33 16.56 -81.25 -15.33

a = variables in US$ million; b = all variables are in percentages, except where indicated.

Srairi et al.


Islamic banks during the period of study. Indeed, efficiency scores, particularly TE and PTE, increased by 12% and 13% on average, respectively, while scale efficiency remained constant. However, during 2008 and 2009, these measures are constant but slightly changed and increased by 1.5% and 1.7%, respectively, compared to 2007. It is apparent that the last financial and economic crisis has affected the performance of Islamic banks, but to a lesser extent than for conventional banks. According to Hasan and Dridi (2010), “the initial impact of the crisis on Islamic Banks’ profitability in 2008 was limited. However, with the impact of the crisis moving to the real economy, Islamic Banks in some countries faced larger losses compared to their conventional peers”.

Despite the increase in efficiency of Islamic banks between 2003 and 2009, the average of the input waste is large and equal to 34.5%. Therefore, there is still room for improvement in the performance of these banks through more efficient use of resources. Indeed, the efficiency scores of Islamic banks in GCC countries are low compared not only to conventional banks (Srairi, 2010; Rosly and Abu Baker 2003) but also to Islamic banks in other countries. For instance, Kamaruddin et al. (2008) found that the average of technical efficiency of the Malaysian Islamic banks is 93% for the period 1998–2004. In a recent study of Islamic banks in MENA and Asian countries, Sufian et al. (2008) found that Islamic banks in Indonesia during the period 2001–2006 are the most efficient from the Asian region, exhibiting a mean technical efficiency of 92.3%. However, several studies (e.g., Mohammed et al. 2008; Hassan et al. 2009) suggested that there are no significant differences between the overall efficiency results of conventional compared with Islamic banks.

The decomposition of overall technical efficiency into PTE and SE components provides information on the source of technical inefficiency. Table 2 reveals that the pooled means for PTE and SE during the period analyzed are of 77.3% and 85.5%, respectively. The result shows that

the inefficiency in Islamic banks could be attributed to pure technical inefficiency (29.3%) rather than to scale inefficiency (17%). It means that Islamic banks in GCC countries are managerially inefficient in controlling costs but manage their inputs efficiently. This finding of the dominant impact of managerial inefficiency over scale inefficiency is also reported in other studies, for example, Sufian et al. (2008) for Islamic banks in MENA and Asian countries; Kyj and Isik (2008) for the Ukrainian banking industry; and Zaim (1995) for Turkish banks. According to several studies (e.g., Bashir 2007; Iqbal 2007), the inefficiencies in Islamic banks can also be attributed to many other causes such as: limited number of short-term instruments; shortage of products for medium and long term maturities; portfolios of Islamic banks being concentrated on equity and non-interest based financing, and especially focused on trade financing; small size of banks; weak management; and lack of proper risk-monitoring systems.

Furthermore, we attempt in this study to identify the nature of scale inefficiency, which can be due to increasing returns to scale (IRS) or decreasing returns to scale (DRS). Table 3 displays statistics for the number of banks in the different categories of scale economies, and also presents the returns to scale of banks classified by size. According to the figures in this table, only 19% of Islamic banks operate at their optimal scale (CRS) and the majority of banks are scale-inefficient (58% at DRS and 23% at IRS). It is also interesting to note that the share of the banks experiencing economies of scale (IRS) and diseconomies of scale (DRS) are relatively constant during the sample period. The results confirm those shown in Table 2, relative to the stability of scale efficiency of Islamic banks over the period of study. Panel B of Table 3 also indicates that the majority of Islamic small banks (83%) exhibited IRS (53%) or CRS (30%), while the medium and large banks operated at DRS (80%). It means that increasing the activities and size of Islamic small banks may bring significant cost savings and, in consequence, improve the technical efficiency of these

Table 2. Efficiency scores by year and size (average values).

TE PTE SE

Mean SD Mean SD Mean SD

Panel A: by year2003 0.612 0.147 0.718 0.136 0.855 0.1162004 0.643 0.195 0.738 0.149 0.864 0.1382005 0.650 0.141 0.751 0.143 0.883 0.2002006 0.642 0.112 0.778 0.141 0.839 0.1502007 0.671 0.115 0.799 0.106 0.847 0.1312008 0.681 0.162 0.813 0.128 0.838 0.1272009 0.685 0.085 0.817 0.138 0.852 0.128Panel B: by sizeSmall banks 0.669 0.156 0.676 0.153 0.990 0.157Medium banks 0.653 0.118 0.762 0.123 0.840 0.140Large banks 0.686 0.159 0.779 0.147 0.885 0.142

Overall 0.655 0.140 0.773 0.137 0.855 0.144

TE = technical efficiency; PTE = pure technical efficiency; SE = scale efficiency.



banks, in contrast to the case of expansion by the medium and large banks. A similar finding has been made for other countries such as Singapore (Rezvanian and Mehdian 2002), Turkey (Isik and Hassan 2002) and India (Rezvanian et al. 2008).

In order to compare the efficiency scores of banks according to their size, we categorized the sample banks into three groups based on their total assets, with an approximate number of banks in each category. The first group comprises nine small banks with an asset size of less than $3 billion. The second group includes medium banks (eight banks) whose assets are between $3 and $5 billion, while, the last group comprises large banks (eight banks) whose assets exceed $5 billion.

In terms of overall technical efficiency, panel B of Table 2 shows that large (68.6%) and small (66.9%) banks are the most efficient, while the medium banks presented the lowest mean TE of 65.3%. This is consistent with several studies which reported a significant positive association between size and efficiency (e.g., Drake and Hall 2003; Chen et al. 2005; Pasiouras 2008; Srairi, 2010). Large banks present some advantages over small and medium banks. According to Kyj and Isik (2008), “large banks may be able to hire a better management team, utilize better technology, be located in larger, more competitive markets, and have more diversified loan portfolio. Large banks, thus, may have lower default risk, and lower borrowing costs”. However, other studies found a negative (e.g., Christopoulis et al. 2002; Bonin et al. 2005) or no significant (e.g., Berger and Hannam 1998; Girardone et al. 2004) relationship between size and efficiency. On the other hand, the result indicates that large (77.9%) and medium (76.2%) banks are more pure technically efficient than small banks (67.6%). However the latter display a superior measure for scale efficiency, this being 10.5% and 15% higher than for medium and large banks, respectively. Consequently, it seems that Islamic small banks need more improvement in terms of managerial

practice, while Islamic medium banks need to increase their scale efficiency.

Efficiency and share performanceTo assess the relationship between the efficiency of Islamic banks and their share prices, we regress annual stock price returns on annual percentage change of efficiency scores, derived from DEA analysis, with other explanatory variables. Models 1, 2 and 3 in Table 4 present the regression results estimated by the fixed-effect model for technical, pure technical, and scale efficiency changes respectively. The results indicate that both technical and pure technical efficiency changes have a positive and statistically significant (1% for TE and 5% for PTE) effect on stock returns. Indeed, the share prices of Islamic banks respond positively towards improvement in managerial efficiency. Hence, it seems that information regarding the efficiency of banks is reflected in the stock prices of banks. In fact, in an efficient market, share prices incorporate all publicly available information (Fama 1970). Thus, according to Beccalli et al. (2006) and others, efficient banks can better improve their share price performance than inefficient banks. So, our results are in line with several studies in other countries which found a positive association between technical efficiency change and share performance (e.g., Pasiouras et al. 2008 for Greek banks; Xiang and Shamsudding 2009 for Australian banks; Sufian and Abdul Majid 2009 for China banks). However, other researches (e.g., Liadaki and Gaganis 2010 for European banks; Ioannidis et al. 2008 for Asian and Latin American banks; Chu and Lim 1998 for Singapore banks) show that changes in stock returns reflect changes only in profit efficiency rather than in cost efficiency. According to Liadaki and Gaganis (2010), these results can be explained by the fact that rational shareholders and investors are more interested by the profit of banks as an indicator of the future dividends. Moreover, cost efficiency reflects the

Table 3. Return to scale in Islamic banks by year and size.

Years

DRS IRS CRS Total of

banksNb. of banks % share Nb. of banks % share Nb. of banks % share

Panel A: by year2003 15 62 6 25 3 13 242004 15 62 5 21 4 17 242005 14 58 6 25 4 17 242006 13 52 6 24 6 24 252007 12 48 7 28 6 24 252008 15 60 5 20 5 20 252009 15 60 5 20 5 20 25Total 99 58 40 23 33 19 172Panel B: by sizeSmall banks 10 17 32 53 18 30 60Medium banks 44 79 3 5 9 16 56Large banks 45 80 5 9 6 11 56

DRS: decreasing returns to scale; IRS: increasing returns to scale; CRS: constant returns to scale.

Srairi et al.


capability of managers but it is not directly observed in the stock market.

From Table 4 (model 3), it is also noted that the estimated coefficient of scale efficiency change is positive but it is not statistically significant. It means that scale efficiency does not have any impact on a bank’s share returns. This finding is also confirmed by the coefficient of bank size which is insignificant in all of the regression models. A similar result was also found by Pasiouras et al. (2008) and Sufian and Abdul Majid (2009).

With regard to the control variables and their influence on stock returns, Table 4 indicates that market return in all models has the expected sign and a significant power to explain the variation in stock prices. This result, which is consistent with previous studies (e.g., Xiang and Shamsuddin 2009; Erdem and S.Erdem 2008), shows that stock price returns of Islamic banks are positively related to the overall performance of the market. On the other hand, the association between the ratio of book-to-market value (BM) and share performance is positive and significant at the 5% level for all models. However, our results are different from the study of Xiang and Shamsuddin (2009) concerning Australian banks which found a negative sign

of BM, implying a possibility of market expectation of systematic risk.

5. DiscussionMany policy implications and recommendations can be derived from the results of this paper. First, since Islamic banks in GCC countries exhibited a lower level of efficiency compared to conventional banks, it is necessary for these institutions to promote and enhance their functioning in several areas (Bashir 2007; Iqbal 2007). Islamic banks are still operating with a limited number of instruments for the short-term, and there is a shortage of products for medium- to long-term maturities. In this regard, Islamic banks have to offer new products and modes of finance that enhance risk management and portfolio diversification. Due to limited size and resources, Islamic banks are unable to reap the benefits of economies of scale and are also unable to afford high cost management information systems to assess and monitor risks. Accordingly, Islamic banks have to perform strategic alliances with other Islamic financial institutions, and collaborate with conventional banks which are more sophisticated in terms of financial engineering. Further, to better control and reduce their costs, Islamic banks need to invest more in technology, to develop innovative

Table 4. Regression results of equation (1).

RSit = a + β1CEit + β2MRit + β3BSFit + eit

RSit is the annual return on bank i’s stock in year t. CEit represents the annual percentage change in bank efficiency and includes the technical (TE, model N°1) or pure technical (PTE, model N°2) or scale efficiency (SE, model N°3) for bank i in year t. MRjt is the market return for banking sector j in year t and BSFit concerns some specific factors and includes two variables: LTAit is the size of bank i in year t measured as the natural logarithm of total assets and BMit is the book-to-market equity ratio calculated as the ratio of the book value of a bank’s equity to its market value.

Variables Model 1 Model 2 Model 3

Constant: a 2.15 2.03 2.11(2.75)* (2.62)* (2.70)*

Annual change in efficiency scoresTechnical Efficiency: TE 3.14 – –

(2.30)**Pure Technical Efficiency: PTE – 7.07 –

(2.61)*Scale Efficiency: SE – – 0.83

-0.83Control VariablesMarket return: MR 0.84 0.86 0.87

(9.05)* (8.88)* (8.96)*Size of bank: LTA 0.71 0.56 0.51

(1.61) (0.91) (1.60)Book to market equity ratio: BM 7.31 7.35 7.12

(2.55)** (2.49)** (2.41)**Adjusted R2 0.617 0.597 0.499F value 38.66* 35.29* 35.60*Nb. Observations 125 125 125F value 48.3 46.4 43.2LM 385.2 391.3 380.4Hausman test 36.1 33.8 31.2

T-statistics are between parentheses; *, and ** indicate statistical significance at 1%, and 5% respectively.



methods in terms of risk management, and to increase the efficiency of their staff by investing in training and development. The results also show that there has been an improvement in efficiency of Islamic banks over the period of liberalization in Gulf countries. Therefore, authorities in this region should continue to reinforce financial reforms, increase economic integration between countries, and undertake constructive policy actions to develop Islamic capital markets which help to integrate Islamic financial institutions into regional and international financial systems. Finally, while there is a positive association between the performance of Islamic banks and their stock price returns, it appears that efficiency measures contain important and helpful information which could be used by managers of banks, shareholders and investors.

6. ConclusionsIslamic banking is viewed as competitive and an alternative to the conventional banking system in many states of the world, particularly in GCC and some Asian countries. In addition, during the last decade, Islamic banking assets have been growing at a faster pace (an average annual growth of 20%) than the overall banking system, with the expectation that it will play an increasing important role in the coming years. Moreover, the Islamic financial system has proved to be the least affected by the last economic and financial crisis. In the light of these considerations, it is important to assess and analyse how Islamic banks have performed during the past few years.

In the present study, we estimate the efficiency of 25 GCC Islamic banks over the period 2003–2009. By using a non-parametric DEA technique, under the profit oriented approach, we calculate technical, pure technical and scale efficiencies to study the evolution of these efficiency measures across time and to analyse the size efficiency relationship. Additionally, this paper attempts to investigate the influence of the performance of Islamic banks in terms of efficiency on their stock prices. Several important findings emerge from this present study. The results indicate that the average technical efficiency was equal to 66% and that there was a rising trend for both TE and PTE, suggesting that Islamic banks in GCC countries improved their efficiency during the survey period. This was the period where the processes of liberalization of the GCC financial system were realised at an accelerated pace. Overall, we also find that inefficiency in Islamic banks is attributed mainly to pure technical inefficiency (29%) rather than scale inefficiency (17%). Thus, it seems that Islamic banks are managerially inefficient in controlling their costs and their inputs. It is interesting to note that the majority of Islamic banks are scale-inefficient and are either small- or medium-sized, which implies that these banks can achieve cost savings and improve their efficiency by increasing their size and scale of operations. Furthermore, our findings regarding the impact of size on the efficiency of Islamic banks suggest that, while large banks are more managerially and technically efficient than small banks, they are also less scale-efficient than the smaller banks. In terms of pure technical efficiency, large-sized Islamic banks seem also to be the most efficient ones, followed by the medium banks. In this regards, it appears that small banks need to improve their managerial practices, while medium banks have to increase their scale efficiency.

Using the efficiency scores of Islamic banks, we analysed the link between efficiency change and stock returns. The results derived from the fixed effect model show that percentage changes in the prices of bank stocks reflect percentage changes in both technical and pure technical efficiency. However, we do not find any significant relationship between scale efficiency and stock returns. Thus, our results seem to support the argument that stock returns respond positively towards improvement in managerial efficiency, but do not react towards changes in scale efficiency (Sufian and Abdul Majid 2009). Hence, the efficiency of a bank’s operation provides significant information about its share price performance, which is not explained by market movements.

One implication of the findings is that managerially-efficient banks should be more profitable and therefore generate greater shareholder returns. This is in line with the efficient market theory that, in an efficient market, a change in cost efficiency should be incorporated in the price formation process. Finally, the study also revealed that market return and ratio of book-to-market value have a positive impact on stock returns.

ReferencesAl-Omar, F.A., Iqbal, M. Some strategic suggestions for

Islamic banking in the 21st century, Review of Islamic Economics, 9, 2000. 37–56.

Archer S, Abdel-Karim R. Islamic Finance. Euromoney Books, UK. 2002.

Avkiran, N. K The evidence on efficiency gains: The role of mergers and the benefits to the public, Journal of Banking and Finance, 23, 1999. 991–1013.

Bashir, A.H.M. Assessing the performance of Islamic banks: Some evidence from the Middle East, in American Economic Association Annual Meeting, New Orleans, USA, 2001.

Bashir, A.H.M. Islamic banks participation, concentration and profitability: Evidence from MENA countries. Working Paper 0402, 2007.

Banker, R. Charnes, A. Cooper, W “Some models for estimating technical and scale inefficiencies in data envelopment analysis”. Management Science, 1984. 1078–1092.

Bauer, P. W. Berger, A. N. Ferrier, G. D. Humphrey, D. B. Consistency conditions for regulatory analysis of financial institutions: A comparison of frontier efficiency methods, Journal of Economics and Business, 50, (1998). 85–114.

Becalli, E. Casu, B. Giraradone, C. Efficiency and stock performance in European banking, Journal of Business Finance and Accounting, 33, 2006. 245–262.

Berger, A. N., Humphrey, D. B. Measurement and Efficiency issues in commercial banking, in Z. Griliches (Ed.) Measurement issues in the service sectors, National Bureau of Economic research, University of Chicago Press, 1992. 245–279.

Berger, A. N. Hannam, T. H. The efficiency cost of market power in the banking industry: A test of the quiet life and

Srairi et al.


related hypotheses, Review of Economics and Statistics, 80, 1998. 454–465.

Berger, A. N. Mester, L. J. Inside the black box: What explains differences in the efficiencies of financial institutions, Journal of Banking and Finance, 21, 1997. 895–947.

Blominvest Bank Islamic banking in the MENA region. 2009.

Bonin J. P., Hassan I., Wachtel P. Bank performance, efficiency and ownership in transition countries. Journal of Banking and Finance, 29, 2005.31–53.

Canhoto, A. Dermine, J. A note on banking efficiency in Portugal: New vs. old banks, Journal of Banking and Finance, 27, 2003. 2087–2098.

Chapra, M. U. The global financial crisis: Can Islamic finance help, NewHorizon, January-March, 170, 2009.

Charnes, A. Cooper, W. Lewin, A. Seiford, L. Data envelopment analysis: Theory, methodology, and application, Kluwer Academic Publishers, Boston, 1994.

Charnes, A. Cooper, W. Rhodes, E. Measuring the efficiency of decision making units, European Journal of Operational Research, 2, 1978. 429–444.

Chen, X. Skully, M. Brown, K. Banking efficiency in China: Application of DEA to pre- and post-deregulation era: 1993–2000, China Economic Review, 16, 2005. 229–245.

Christopoulos, D.K., Lolos, S.E.G. Tsionas, E.G. Efficiency of Greek banking system in view of the EMU: A heteroscedastic stochastic frontier approach, Journal of policy model, 24, 2002. 813–829.

Chu, S. Lim, G. Share performance and profit efficiency of banks in an oligopolistic market: Evidence from Singapore, Journal of Multinational Financial Management, 8, 1998. 155–168.

Coelli, T. A guide to DEAP version 2.1, A data envelopment analysis, (Computer Program), CEPA, Working Paper 96/08, University of New England, Armidale, 1996. Available at: www.une.edu.au/econometrics/cepa.htm.

Cooper, W. W. Seiford, L. M. Tone, K. Data envelopment analysis. Kluwer Academic publishers, Boston, 2000.

Das, A. Ghosh, M. S. Financial Deregulation and efficiency: An empirical analysis of Indian banks during the post reform period, Review of Financial Economics, 15, 2006. 193–221.

Drake, L. Hall, M. J. B. Efficiency in Japanese banking, Journal of Banking and Finance, 27, 2003. 891–917.

Drake, L. Hall, M. J. B. Simper, R. The impact of macroeconomic and regulatory factors on bank efficiency: A non-parametric analysis of Hong Kong’s banking system, Journal of Banking and Finance, 30, 2006. 1443–1466.

Erdem C, Erdem M.S. Turkish banking efficiency and its relation to stock performance. Applied Economics Letters, 15, 2008. 207–211.

Fama, E F. Efficient capital markets: A review of theory and empirical work. Journal of Finance, 25, 1970. 383–417.

Fare, R. Grosskopf, S. Lovell, C.A The measurement of efficiency of production, Kluwer Academic Publishers, Boston, 1985.

Fields, J. A. Murphy, N. B. Tirtiroglu, D. An International comparison of scale economies in banking: Evidence from Turkey, Journal of Financial Services Research, 4, 1993. 157–168.

Fries S, Taci A. Cost efficiency of banks in transition: Evidence from 289 banks in 15 post-communist countries. Journal of Banking and Finance, 29, 2005. 55–81.

Girardone C, Molyneux P, Gardener E P M. Analyzing the determinants of bank efficiency: the case of Italian banks, Applied Economics, 36, 2004. 215–227.

Grigorian, D. A. Manole, V. Determinants of commercial bank performance in transition: An application of data envelopment analysis, Comparative Economic Studies, 48, 2006. 497–522.

Hadad, M. Hall, M. Kenjegalieva, K. Santoso, W. Satria, R. Simper, R. Bank efficiency and stock Market performance: An analysis of listed Indonesian banks, Review Quantitative Finance and Accounting, DOI.10.1007/s11156–010–0192–1, 2010.

Hassan, K. M. The X-efficiency of Islamic banks, Islamic Economic Studies, 2, 2007. 49–77.

Hasan, M. Dridi, J The effects of the global crises on Islamic and conventional banks: A comparative study, International Monetary fund, Working Paper /10/201, 2010.

Hassan, T. Mohamed, S. Bader M. Efficiency of conventional versus Islamic banks: Evidence from the Middle East, International Journal of Islamic and Middle Eastern Finance and Management, 2, 2009. 46–65.

Havrylchyk, O. Efficiency of the Polish banking industry: Foreign versus domestic banks, Journal of Banking and Finance, 30, 2006. 1975–1996.

Inui, T. Park, J. Hyun-Han, S International comparison of Japanese and Korean banking efficiency: Comments and discussion, Seoul Journal of Economics, 21, 2008. 1–16.

International Monetary Fund Survey Islamic banks: More resilient to crisis, Washington, D.C., 2010.

Ioannidis, C. Molyneux, P. Pasiouras, F. The relationship between bank efficiency and stock returns: Evidence from Asia and Latin American, University of Bath, School of Management, Working paper 2008.10, 2008.

Iqbal, Z. Islamic financial systems, Finance and Development, 34, 1997. 42–45.

Iqbal, Z. Challenges facing Islamic financial industry, Journal of Islamic Economics, Banking and Finance, 3, 2007. 1–14.

Iqbal, Z. Van Greuning, H Analyzing banking risk for Islamic banks, World Bank Publications, Washinghton, USA, 2007.

Isik, I. Hassan, K. M. Efficiency, ownership and market structure, corporate control and governance in Turkish banking Industry, Journal of Business Finance and Accounting, 30, 2003. 1363–1421.

Kamaruddin B H, Safa M S, Mohd R. Assessing production efficiency of Islamic banks and conventional bank Islamic windows in Malaysia. International Business Management Research, 1, 2008. 31–48.



Kasman A, Yildirim C. Cost and profit efficiency in transition banking: the case of new EU members. Applied Economics, 38, 2006. 1079–1090.

Kyj, L. Isik, I. Bank x-efficiency in Ukraine: An analysis of service characteristics and ownership, Journal of Economics and Business, 60, 2008. 369–793.

Liadaki, A. Gaganis, C. Efficiency and stock performance of EU banks: Is there a relationship? Omega, 38, 2010. 254–259.

Lewis, M (2008). In what ways does Islamic banking differ from conventional finance, Journal of Islamic Economics Banking and Finance, 4, 9–22.

Mariani, A.M. The input requirements of conventional and Shariah-compliant banking, The International Journal of Banking and Finance, 7, 2010. 51–78.

Mohamad, S. Hassan, T. Bader M. Efficiency of conventional versus Islamic banks: international evidence using stochastic frontier approach, journal of Islamic Economics, Banking and Finance, 24, 2008. 107–130.

Mokhtar, H. S. Abdullah, N. Alhabshi, S. M. Efficiency and competition of Islamic banking in Malaysia, Humanomics, 24, 2008. 28–48.

Moody’s (2008). Islamic banks in the GCC: A comparative analysis.

Olson, D. Zoubi, T. Using accounting ratios to distinguish between Islamic and conventional banks in the GCC region, The Intermediation Journal of Accounting, 43, 2008. 45–65.

Pasiouras, F Estimating the technical and scale efficiency of Greek commercial banks: The impact of credit risk, off-balance sheet activities, and international operation, Research in International Business and Finance, 22, 2008. 301–318.

Pasiouras, F. Liadaki, A. Zopounidis, C. Bank efficiency and share performance: Evidence from Greece, Applied Financial economics, 18, 2008. 1121–1130.

Paul, S. Kourouche, K. Regulatory policy and the efficiency of the banking sector in Australia, The Australian Economic Review, 41, 2008. 260–271.

Rezvanian, R. Mehdian, S. (2002). An examination of cost structure and production performance of commercial banks in Singapore, Journal of Banking and Finance, 26, 79–98.

Rezvanian, R., Rao, N., Mehdian S. M. Efficiency change, technological progress and productivity growth of private, public and foreign banks in India: Evidence from the post-liberalization era, Applied Financial Economics, 18, 2008. 701–713.

Rosly, SA. Abu Baker, MA. Performance of Islamic and mainstream banks in Malysia, International Journal Social Economics, 30, 2003. 1249–1265.

Srairi, S. A comparison of the profitability of Islamic and conventional banks: The case of GCC countries, Bankers, Markets, Investors, 98, 2009. 16–27.

Srairi, S. Cost and profit efficiency of conventional and Islamic banks in GCC countries, Journal of Productivity Analysis, 34, 2010. 45–62.

Sufian, F. Determinants of bank efficiency during unstable macroeconomic environment: Empirical evidence from Malaysia, Research in International Business and Finance, 23, 2009. 54–77.

Sturm, J. E., Williams B. Foreign bank entry, deregulation and bank efficiency: Lessons from the Australian experience, Journal of Banking and Finance, 28, 2004. 1775–1799.

Sufian, F. Abdul Majid, M.Z. Bank efficiency and share prices in China: Empirical evidence from a three-stage banking model, International Journal of Computational Economics and Econometrics, 1, 2009. 23–47.

Sufian, F. Noor, M.A Abdul Majid, M.Z. The efficiency of Islamic banks: Empirical evidence from the MENA and Asian countries Islamic banking sectors, The Middle East Business and Economic Review, 20, 20081–19.

Xiang, D. Shamsuddin, A. Efficiency and stock market performance of Australian banks, paper presented at the proceedings of the Asian Finance Association 2009 Conference. Brisbane, June 30 to July 3, 2009.

Zahar, T.S. Hassan, K.H. A comparative literature survey of Islamic finance and banking, Finance Markets institutions and instruments, 10, 2001155–199.

Zaim, O. The effect of financial liberalization of the efficiency of Turkish commercial banks, Applied Financial Economics, 5, 1995. 257–264.


Cite this chapter as: Aytug H, Ozturk H (2015). Conventional banks versus Islamic banks: What makes the difference? In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Conventional banks versus Islamic banks: What makes the difference?Huseyin Aytug1, Huseyin Ozturk2

1Department of Economics, University of California at Santa [email protected]

Abstract - This paper investigates the determinants of banking profitability in the Turkish banking sector between 2003 and 2011. In addition, we calculate the effect of being an Islamic bank on banking profitability, which allows us to differentiate conventional and Islamic banks. We introduce the method of propensity score matching to the banking literature in order to estimate the average treatment effect (ATE) of being an Islamic bank in Turkey where there exists a dual banking system. The results show that in terms of return on asset (ROA) and return on equity (ROE), being an Islamic bank does not create any difference. However, being an Islamic bank turns out to have a significant and negative effect on net interest margin (NIM). These results have many policy implications in the Turkish banking industry where Islamic banks mimic others to be one of the leading examples.

Keywords: average treatment effect, propensity score matching, Islamic bank, profitability

JEL Classification: C51, G15

1. IntroductionIslamic finance is a source of funding that complies with Islamic jurisprudence. This source of funding has already been in practice in countries where the majority of the population is Muslim; however, the importance of Islamic finance has prevailed in the global financial system recently. Although the distinction between Islamic and conventional finance needs deeper understanding, the main difference within Islamic finance can be attributed to the Islamic idealism of creating a moral economy where profits come from commerce or real transactions not from money lending or speculative transactions. While conventional finance helps direct the flow of capital to investment opportunities that are supposed to provide the highest return in the market, Islamic finance allegedly seeks socio-economic optimality. Another principle of Islamic finance is that investment is expected to produce an optimal socio-economic outcome in line with Islamic norms.

The globalization has affected the scope and breadth of Islamic capital markets in a way that its pervasiveness has reached a global scale. Moreover, the financial crisis, which has recently erupted and hit the global finance and economy severely, has paved the way for other means of financial schemes. Islamic finance, which has stayed relatively resilient, became a focus of the global financial system since then.1 The bulk of Islamic funds staying in financial hubs of many Gulf States and in global financial

networks promises a lucrative and stable source of funding. Despite its growing popularity, Islamic banking remains a small part of the total financial sector and will likely remain so due to lack of penetration in the market and strong competition that is challenged from the conventional banking system.

The most explicit distinction of Islamic finance from conventional finance, where the return is based upon interest, is the prohibition of riba (interest based lending) and gharar (speculation or uncertainty). The underlying contracts of Islamic products differ from their conventional counterparts in structure as well. For example murabaha (a sale-based instrument), which is similar to a conventional loan, involves the purchase of an asset by a bank and its sale to a client at a cost plus a pre-determined profit. The structure of a sale has important legal implications according to the Islamic rules, which extensively dictate the terms of risk and mutual consent. Other sale-based products include the financing of commissioned manufacturing or construction (istinaa) and the forward-sale (salam), and these types of products require structural differences from their conventional equivalents. Lease-based instruments (ijara) are similar to the traditional leasing with certain distinctions and equity-based financial intermediation, which is known as mudaraba, takes place through profit and loss arrangements (Warde, 2010). Another product that recently became a regular budget financing instrument,

Aytug and Ozturk


sukuk, is an asset-based instrument of which tangible assets are specified according to the type of the issuance. In comparison with conventional debt instruments, these products do not pledge fixed income or an interest-based income stream. Yet, it collects returns based on the collection of lease or sale of certain assets that are specified beforehand. These types of products are predominantly used in the Middle East, but their prevalence is visible in East Asian countries as well as developed European markets.

Islamic finance in Turkey is yet a more recent issue. Changes in domestic financial systems and public sensibilities have allowed participation banking to gradually become more visible. They have also emerged resilient in the context of two periods of economic turmoil: the domestic financial crisis of 2001 and the global financial crisis of 2008. The severe banking crisis in 2001 did not have as much inverse effects on participation banks as the conventional banks. The contagious effect of the 2001 crisis had a limited effect on participation banks due to lack of interbank activities by participation banks. The growing financial capacity of the religiously conservative public has also been another factor that made participation banks attractive (Hardy, 2012). These types of banks became the sole option for those people who are resistant to conventional banking. The increasing level of associated client portfolios and deposits have enabled participation banks to reach over 4.5 percent of market share in total assets from nearly 1 percent in 2001. Recently a new legislation passed to facilitate Islamic banking in the private sector, and government officials have indicated interest in issuing sovereign sukuk (rent certificate), comparable to bonds, for funding central government budget requirements.

Table 1–2 gives an idea about the market share and the growth of conventional and Islamic banks between 2004 and 2011. It is striking that the market share of Islamic banks in assets, equity and loans has grown from 1% to 5% approximately and the growth rate of these fundamentals have been higher than the conventional banks. Despite the fact that the growth rate has declined after the crisis in 2008, it is still higher than its counterparts.

Islamic banking in Turkey followed a parallel path to the history of the Turkish economy. Turkey had a highly centralized economy, whereby state institutions owned and managed most important industries until the beginning of the 1980s (Ozturk et al., 2010). However, the 1980s saw a period of liberalization of the tightly controlled Turkish economy. As part of a plan to attract more foreign direct investment from the Arab Gulf states, a decree passed in 1983 legalized the operation of special finance houses to provide interest-free banking. These institutions were highly regulated but they did not gain the same status as conventional banks, e.g., they were not covered by the insurance scheme the other banks in the system utilized and could not invest in government securities by its nature.2

Following the Asian financial crises in the late 1990s, the Turkish economy experienced tremendous volatility, which caused some consolidation of the special finance houses. Poor regulation, the accumulation of public debt, and politically driven lending habits contributed to a severe financial crisis in Turkey in 2001. This affected all strands of the domestic banking sector, although conventional banks Ta

ble

1. M

arke

t sha

re a

nd g

row

th in

ass

ets,

equ

ity

and

loan

s of

Isla

mic

and

com

mer

cial

ban

ks.

Mar

ket

Shar

e

IB

Mar

ket

Shar

e

CB

Gro

wth

: A

sset

s

IB

Gro

wth

: A

sset

s

CB

Mar

ket

Shar

e:

Equi

ty

IB

Mar

ket

Shar

e:

Equi

ty

CB

Gro

wth

: Eq

uity

IB

Gro

wth

: Eq

uity

C

B

Mar

ket

Shar

e:

Loan

s

IB

Mar

ket

Shar

e:

Loan

s

CB

Gro

wth

: Lo

ans

IB

Gro

wth

Lo

ans

C

B

2011

4.39

95.6

131

.25

24.6

84.

2095

.98

17.7

912

.42

5.76

94.2

430

.15

35.5

120

104.

1895

.82

25.2

117

.05

4.02

95.9

826

.90

22.6

65.

9894

.02

28.4

022

.91

2009

3.92

96.0

831

.32

16.1

43.

8996

.11

25.3

725

.56

5.74

94.2

624

.33

7.30

2008

3.48

96.5

237

.18

23.9

83.

8996

.11

51.8

314

.91

4.99

95.0

138

.31

36.3

220

073.

1696

.84

38.6

815

.87

2.98

97.0

254

.38

22.2

44.

9395

.07

51.8

928

.00

2006

2.65

97.3

560

.47

28.4

22.

3797

.63

81.5

012

.96

4.18

95.8

267

.95

49.8

620

052.

1397

.87

43.6

524

.80

1.49

98.5

116

.62

2.8

33.

7596

.25

60.2

043

.62

2004

1.86

98.1

445

.21

24.2

61.

5598

.45

48.2

526

.97

3.38

96.6

267

.69

45.8

720

04–2

008

2.45

97.5

547

.00

23.3

42.

1097

.90

50.1

921

.00

4.06

95.9

461

.93

41.8

420

08–2

011

3.99

96.0

131

.24

20.4

64.

0096

.00

30.4

718

.89

5.62

94.3

830

.30

25.5

120

04–2

011

3.22

96.7

839

.12

21.9

03.

0596

.95

40.3

319

.94

4.84

95.1

646

.11

33.6

7


Conventional banks versus Islamic banks: What makes the difference?

suffered more than the special finance houses since they had a much larger role in the overall economy. Although the crisis had a major negative effect on the entire banking sector, the special finance houses also suffered. Turkish holding company Ulker purchased Faisal Finans in 2000, changing its name to family Finance House. Then, Ihlas Finans filed for bankruptcy in 2001 as the liquidity crisis in Turkey reached its peak. This showed that the participation banks were not immune to crises even tough they functioned with a different business model. The reasons for this were twofold; the participation banks were not decoupled from the whole financial system due to business connections with other banks and the economic crisis hit the overall economy. Yet, it is worth mentioning here that the prohibition of holding public debt protected the participation banks from a worse shock than what could have otherwise ensued. In the initial stages of the financial crisis, Ihlas and other participation banks, that could have held liquid government securities, were thus not as greatly affected as conventional banks. However, when the liquidity crisis hit, Ihlas exposure led it to a collapse, and it experienced a traditional bank run on its deposits.

The 2001 crisis led to a rehabilitation of Turkey’s financial system, and the parliament passed a new law in order to discipline the overall banking system (Law No. 4389). In addition to strengthening banking regulations and creating new oversight bodies for conventional banks, Law No. 4389 founded the Union of Private Finance Houses in order to address common issues among participants and provide a level of state control for the sector. All special finance institutions were required to become members of this association, but they still lacked many of the privileges that conventional banks had, such as the provision of deposit insurance. In 2006, banking law No. 5411 officially replaced the term “special finance institutions” with the name “participation banking.” Participation Banks Association of Turkey was established and employed with the unification of Private Finance Houses. The new law created a savings

deposit insurance fund for participation banks as well. In doing so, the insurance scheme began to cover the whole banking system.

These changes may represent a shifting paradigm in the level of acceptance for participation banking in Turkey. Participation banking emerged stronger after each of these periods of instability. Evidently, participation banking’s role in the economy will also likely grow as Turkey considers options for attracting investors from the Gulf region who are currently highly liquid in terms of capital and also religiously conservative. Therefore, the recent developments in Islamic banking bring old debates to discussion again whether they are really different from the conventional banks or not.

The current literature on Islamic banking addresses the issue of whether the distinction between conventional banks and Islamic banks is only their names or the distinction also appears in their business model. As a bold example, Kassim et al. (2009) questions the argument whether Islamic banks are not susceptible to the interest rate changes as compared to their counterparts where both of them operate in tandem. Their research question emanates from the basic proposition whether they differentiate in behavior to common macro-economic shocks. We aim to contribute to the current literature by putting Turkey into the center and investigating the profitability issue with a focus on Islamic and conventional bank differentiation. Although the determinants of profitability in conventional banks in Turkey have been a subject of some research, there have been a few studies conducted regarding profitability of the participation banks in the literature (Macit, 2012). Islamic banking in Turkey is under an interesting transformation that is reflected in their asset and equity growth. This transformation opens a new debate: what are the determinants of this change? Do Islamic banks really differ from their counterparts? To what extent do the dynamics associated with the performance of Islamic

Table 2. Banks’ size and market share (*share of assets/**share of equity/***share of loans).

Banks

2011-Q4* 2007-Q4* 2011-Q4** 2007-Q4** 2011-Q4*** 2007-Q4***

IB/CB Overall IB/CB Overall IB/CB Overall IB/CB Overall IB/CB Overall IB/CB Overall

Akbank 11.51 10.98 12.15 11.75 12.68 12.14 14.43 13.98 10.58 9.97 13.20 12.55Denizbank 3.10 2.96 2.66 2.57 2.85 2.73 1.98 1.92 3.38 3.18 3.71 3.53Finans Bank 3.98 3.80 3.72 3.60 4.11 3.94 3.57 3.46 4.56 4.29 5.05 4.81HSBC 2.08 1.98 2.39 2.31 2.02 1.93 2.75 2.66 2.08 1.96 3.33 3.17ING Bank 1.81 1.73 2.23 2.16 1.71 1.63 1.73 1.68 2.31 2.18 3.03 2.89Sekerbank 1.24 1.18 1.08 1.05 1.06 1.01 1.18 1.14 1.28 1.21 1.29 1.23TEB 3.28 3.13 2.10 2.03 3.04 2.91 1.24 1.20 3.86 3.64 2.45 2.33TR Ziraat 13.84 13.21 14.42 13.94 9.52 9.11 9.82 9.52 10.75 10.13 7.70 7.32TR Garanti 12.63 12.05 12.04 11.64 12.70 12.15 9.37 9.07 12.62 11.88 13.27 12.62TR Halk 7.85 7.49 7.17 6.93 6.24 5.97 5.96 5.78 8.46 7.97 6.46 6.14TR Is 13.93 13.39 14.29 13.81 12.94 12.39 14.43 13.98 13.79 12.99 12.12 11.52TR Vakiflar 7.68 7.33 7.56 7.30 6.72 6.43 7.11 6.89 8.63 8.13 8.37 7.96Yapi Kredi 9.31 8.88 8.97 8.67 8.45 8.09 6.67 6.47 10.20 9.61 10.17 9.66Albaraka Turk 18.65 0.86 18.99 0.64 16.22 0.69 22.58 0.70 17.71 1.03 18.56 0.91Bank Asya 30.65 1.41 32.21 1.08 34.51 1.48 36.12 1.13 32.06 1.86 29.74 1.46Kuveyt Turk 26.57 1.22 19.90 0.67 23.22 0.99 16.44 0.51 25.01 1.45 20.37 1.02Turkiye Finans 24.12 1.11 28.90 0.97 26.05 1.12 24.86 0.77 25.22 1.47 30.97 1.52

Aytug and Ozturk


banks differs from their counterparts? These questions will be addressed in this study.

We use a broad set of data to investigate the determinants of bank profitability. The bank specific variables that may have an impact on bank profitability are selected in accordance with the current literature (e.g., Athanasoglou et al. (2008); Garca-Herrero et al. (2009); Dietrich and Wanzenried (2011). The macroeconomic and industry specific variables that are considered to be influential on bank efficiency are also included. The recent financial crisis is also considered by separating the whole period as two sub-periods: pre crisis and post crisis. This approach is similar to Hasan and Dridi (2011); Dietrich and Wanzenried (2011), who separate the whole period as pre crisis and post crisis. One of the main contributions of this paper is that we introduce the method of propensity score matching to the banking literature in order to estimate the average treatment effect (ATE) of being an Islamic bank where there exists a dual banking system. The results are compared with the ordinary least squares (OLS) estimation results. We make use of a unique and up-to-date database by combining quarterly conventional bank and participation bank data. The results provide insightful policy making implications and will be discussed in the upcoming sections. All in all in our study seeks to examine a series of questions. First, the issue of being an Islamic bank on bank profitability will be mainly discussed. This issue deserves particular attention due to motivation discussed above. Beyond this main scope, the determinants of profitability of the Turkish banking industry will be explored. This is also crucial to study considering the limited number of studies specific to Turkish banking. Last but not the least, the effect of recent financial crisis on profitability of banks will be explored. The recent financial crisis that caused great havoc in the banking sector of many countries – e.g., many defaults or bail-outs have taken place very recently especially in developed country space – reasonably might affect the profitability.

The remainder of the study is organized as follows: the second section will explain the data. The data set we compile is the largest data set as to the best of our knowledge. The third section will discuss the methodology in some detail. The ATE methodology and the model specifications will be discussed in this section. The fourth section will discuss the results and policy implications. The policy implications and policy recommendation will be provided together specific to the Turkish case. The fifth section will conclude.

2. Literature reviewIn the studies, which investigate the Islamic banking and conventional banking dualism, the recent findings reveal that there is not a fundamental difference in terms of their routine activities. In other words, they show similar responses to basic impulses, e.g., their profitability measures respond to market interest rates in a similar way. Ergec and Arslan (2011) contend that Islamic banks, relying on interest-free banking, shall not be affected by the interest rates; however, in concurrence with the previous studies, the article finds that the Islamic banks in Turkey are visibly influenced by interest rates. This study differentiates from Kassim et al. (2009) in one respect. Kassim et al. (2009), claim that the primary reason why Islamic banking may become more stable compared to conventional banks is that they are not affected by the fluctuations on interest

rates. In other words, Islamic banks are expected to be more stable than the conventional banks where Islamic banking is not influenced by interest rates. Stability in demand for money holds some positive effects in terms of efficiency in monetary policies and the financial stability in the system. On the other hand, Kia and Darrat (2007) refer to two major reasons why interest-free Islamic banks contribute to the stability more than the others. A first reason is related with the demand for money whereas the second one is an assessment with the balance sheet perspective. Of these factors determining demand for money, interest rates appear to be the most effective component for speculation. Thus, interest free banking reduces stability in the banking system since Islamic banks shy away from interest rate. As for the balance sheet perspective, along with changes in the interest rates, the banks revalue their assets before liabilities. The loan interest rates respond to a change in the interest rate much earlier than the savings interest rates. In this case, the revaluation of balance sheet entries, a key component for profit maximization, makes the impact of interest rate changes more vulnerable. However, in the Islamic banking system, there is no need for revaluation of balance sheet entries because there is no risk of interest rate. All in all, for these two primary reasons, it is anticipated that the banking system, dominated by conventional banking, is more unstable. To test for the stability issue in the Islamic banking system, there are also plenty of studies investigating the country experiences from interest rate change perspective. All these studies present empirical findings suggesting that demand for money is more stable in banking systems where Islamic banks are the key players.3

Since the early works by Short (1979), Kwast and Rose (1982) and Bourke (1989), a considerable amount of recent studies have investigated some of the major determinants of bank profitability. The empirical studies have focused their analyses either on cross-country evidence or on the banking system of individual countries. The studies by Molyneux and Thornton (1992), Demirguc and Huizinga (1999), Goddard, et al. (2004), Micco, et al. (2005), and Pasiouras and Kosmidou (2007) investigate a panel data set. Studies by Berger, et al. (1987), Berger (1995), Michelle (1997), Bennaceur and Goaied (2008), Athanasoglou, et al. (2008) and Garca-Herrero, et al. (2009) center their analyses on single country cases.

Bank profitability is vastly measured as a function of return on asset (ROA) or return on equity (ROE).

Some works also add net interest margin (NIM) as a complementary measure that it related with profitability. The literature classifies determinants as being internal and external determinants. The internal determinants include bank-specific variables of which the intrinsic features of individual banks compose this bloc. The external variables reflect factors that are expected to affect the profitability of banking industry although banks are not capable of controlling them.

In most studies, variables such as bank size, risk, capital ratio and operational efficiency are used as internal determinants of banking profitability. Pasiouras and Kosmidou (2007) find a positive and significant relationship between the size and the profitability of a bank. This is due to the fact that larger banks are likely to have a higher degree of product



and loan diversification than smaller banks. Furthermore, large banks benefit from economies of scale. Berger, et al. (1987), provide evidence that costs are reduced only slightly by increasing the size of a bank and those very large banks often encounter scale inefficiencies. Micco, et al. (2005) find no correlation between the relative bank size and the return on assets for banks, i.e., the coefficient is always positive but never statistically significant. Therefore the impact of being a big bank in size on profitability is mix.

If it is the risk that is concerned in this literature, Abreu and Mendes (2002) who examined banks in some European countries, find that the loans-to-assets ratio, as a proxy for risk, has a positive impact on the profitability of a bank. Bourke (1989) and Molyneux and Thornton (1992), among others, find a negative and significant relationship between the level of risk and profitability. This result reflects the fact that banks that are exposed to high risk also have a higher accumulation of non-performing loans, and non-performing loans lower the returns of the affected banks.

Another bank feature that is suggested to be effective on the profitability of banks is the asset composition of the banks. Empirical evidence by Bourke (1989), Demirguc and Huizinga (1999), Abreu and Mendes (2002), Goddard, et al. (2004), Bennaceur and Goaied (2008), Pasiouras and Kosmidou (2007) and Garca-Herrero, et al. (2009) indicate that the best performing banks are those that maintain a high level of equity relative to their assets. The authors explain this relation with the observation that banks with higher capital ratios tend to face lower costs of funding due to lower potential bankruptcy costs. I.e., equity has the lowest order to be paid during liquidation.

One more bank-specific variable is the ownership of a bank. Private or state owned banks are the most visited separation in the literature. This separation has many insightful findings. For instance, the owner-ship structure plays an important role in explaining banking profitability. Micco, et al. (2007) found that the separation of banks as privately owned or state-owned provides insights in examining bank performance. According to their results, state-owned banks operating in developing countries tend to have a lower profitability, lower margins, and higher than comparable privately owned banks. In industrialized countries, however, this relationship is found to be much weaker. Iannotta, et al. (2007) point out that government owned banks exhibit a lower profitability than privately owned banks. Therefore, bank ownership structure is also important regardless of being developed or developing, yet the analysis suggests a different degree of significance. Another strand of bank ownership is being a foreign bank. The international connection of a bank may have a significant impact on profitability. This is plausible in a sense that foreign banks have easy access to a pool of funds abroad that domestic banks cannot easily reach. Yet in terms of profitability issues, being a foreign bank is found to have differing impact on profitability. Demirguc and Huizinga (1999) suggest a significant relationship, on the other hand Bourke (1989) and Molyneux and Thornton (1992) find this relationship insignificant.

Many studies in profitability literature take—such as central bank interest rate, inflation, GDP growth etc.— another bloc of variables that affect bank profitability.

Most studies have shown a positive relationship between inflation, central bank interest rates, GDP growth, and bank profitability (e.g., Bourke (1989), Molyneux and Thornton (1992), Demirguc and Huizinga (1999), Athanasoglou, et al. (2008), Albertazzi and Gambacorta (2009)). As per the effects of macroeconomic variables, the effect of inflation rate on bank profitability depends upon whether the inflation is anticipated or unanticipated. In the case of an anticipated inflation bank profits may improve as the banks may adjust the price of lending according to the inflation rate. However, an unanticipated inflation may have negative effects. Bourke (1989) and Molyneux and Thornton (1992) find that a higher inflation rate is associated with better profitability indicators. Recently, Macit (2012) has recorded similar findings.

To measure the effects of market structure on bank profitability, the structure conduct and performance (market-power) hypothesis states that increased market power yields monopoly profits. The inverse relation between the degree of market concentration and degree of competition has been the underlying assumption of the structure conduct performance hypothesis. The bank concentration is discussed in some studies. The Herfindahl Hirschman index was used to proxy the level of competition in the industry (see Dietrich and Wanzenried (2011)). According to the results of Bourke (1989) and Molyneux and Thornton (1992), the bank concentration ratio shows a positive and statistically significant relationship with the profitability of a bank and is, therefore, consistent with the structure conduct performance paradigm. In contrast, the results of Demirguc and Huizinga (1999) and Staikouras and Wood (2002) indicate a negative but statistically insignificant relationship between bank concentration and bank profits. Likewise, the estimations by Berger (1995) and E.C. and P.C. (2003) contradict the structure-conduct performance hypothesis. As briefly discussed above, the determinants of bank profitability can be defined with three blocks of variables briefly discussed above. The literature is more or less similar in terms of the data selection. The variation of data employed in the analysis is rather sparse. Therefore the existing literature provides a comprehensive examination of the effects of bank-specific, industry-specific, and macroeconomic determinants on bank profitability. In this study, we take being an Islamic bank as the centre and investigate the effect of being an Islamic bank on profitability. This contribution is vital in investigating the on-going transformation in the banking sector of Turkey. Considering the lack of studies on banking profitability in the case of Turkey, especially the special attention given to Islamic banks, this study bridges an important gap. Dietrich and Wanzenried (2011), underlines the relative scarcity of literature that discusses the effect of the recent financial crisis on bank profitability. In our study, we also address this issue by making pre-crisis and post-crisis analysis. The data employed in our analysis is in line with the current literature with small variations that will be discussed in the next section.

3. DataThere is a well-established set of determinants available to investigate the profitability of the banking system in Turkey. To examine the effect of being an Islamic bank on banking profitability, we rely on our data set in line

Aytug and Ozturk


with current literature. The data is gathered from the quarterly unconsolidated balance sheets of banks that operated between 1994Q1 and 2011Q4. The balance sheets are obtained from The Banks Association of Turkey and Banking Regulation and Supervision Agency. There are 29 conventional banks and 4 Islamic banks in 1994Q1–2011Q4. Macroeconomic variables are from Central Bank of Turkey and the Undersecretariat of Treasury who are responsible for economy management in Turkey. As per the effects of macroeconomic variables, the real interest rate on government bonds used to proxy interest rate. The daily interest rate of government bonds is not available. Therefore we use one-year T-bill rates at the data issue. These will proxy for the interest rates for each and every quarter. The level of foreign exchange rate is the USD/TRY rate and an increase in exchange rate implies depreciation in Turkish Lira.

Islamic banks constitute a small portion of the banking system in Turkey but potential to grow, may be in size rather than number. Of these four banks, two of them are open to public and are daily traded in the stock market. Three of these four banks are foreign and only one bank is domestic. In terms of bank specific determinants, we look at various different variables, namely the ratio of equity to total assets, the ratio of net loans to total assets, log of real assets, and the ratio of non-performing loans to total loans. To control for macroeconomic determinants of profitability, we use GDP

growth, level of foreign exchange rates, consumer inflation, and real interest rate. Turkey has experienced great growth during the period subject to analysis. Therefore, GDP growth is expected to have a positive impact on profitability regardless of being conventional or Islamic.

Industry-wise we do not include concentration measures in our analysis. The Turkish banking industry constitutes a competitive market without a dominant group of banks or single bank. Dietrich and Wanzenried (2011) take concentration into their analysis. Yet, their focus is on the Swiss banking industry, where there exists a huge concentration. As per industry specific variables we use growth measures. The reason why we employ growth measures is that the growth in the banking industry was visible in the Turkish case in the last decade. The recapitalisation was the major theme of the banking industry (see Yeldan, 2007).

We define state bank as the base and present three dummies for private bank, state banks and Islamic banks. In doing so, we aim at controlling for industry specific effects on bank profitability. The variables that are used in our analysis are detailed in Table 3.

Table 4 provides descriptive statistics for our sample showing the observations, means and standard deviations of all variables. Observations are divided into two groups

Table 3. Variables used in the Empirical Analysis

Variable Notation Measure

Bank-Specific Variables Return on asset ROA Net income/total assetsReturn on equity ROE Net income/total equityNet interest margin NIM Net interest income/total

assetsCapital adequacy ETA Equity/total assetsAsset quality LTA Loans/total assetsAsset size NPLTA—log (TA) Non-performing loans/total

assets (Natural logarithm of total assets)

Liquidity LQD Liquid assets/total assetsDeposits DPTA Deposits/total assetsCredit risk CR Loans loss provision/loansLiquidity risk LR Loans/deposits

Market-Specific Variables Total asset growth TAG Growth of total assetsTotal equity growth TEG Growth of total equityTotal loans growth TLG Growth of total loans

Macroeconomic Variables GDP growth GDPGR Growth of GDPInflation rate INF Inflation rateExchange rate FX Exchange rate between

TL and $Real interest rate IR Real interest rate

Dummy Variables Islamic bank dummy IB 1=Islamic banks, 0=Normal banks

Ownership dummy PD 1=Privately owned; 0=Publicly owned

FD 1=Foreign owner; 0=Domestic owner

Crisis dummy CD 1=After the crisis; 0=Before the crisis



Table 4. Descriptive Statistics

Conventional Banks Islamic Banks

Variable Obs Mean Std. Dev. Min Max Obs Mean Std. Dev. Min Max

ROA 1044 0.013 0.027 –0.176 0.322 144 0.007 0.003 0.001 0.024ROE 1044 0.074 0.128 –1.786 0.465 144 0.058 0.033 0.007 0.236NIM 1044 0.030 0.028 –0.0217 0.169 144 0.017 0.009 0.005 0.063ETA 1044 0.174 0.140 0.037 0.916 144 0.118 0.024 0.073 0.183LTA 1044 0.413 0.212 0.001 0.847 144 0.694 0.089 0.436 0.835NPLTA 1044 0.023 0.022 0.000 0.177 144 0.036 0.022 0.005 0.135LOGTA 1044 6.574 0.890 4.346 8.231 144 6.569 0.391 5.664 7.236LQD 1044 0.217 0.182 0.009 0.812 144 0.080 0.050 0.016 0.235DPTA 1044 0.513 0.235 0.000 0.903 144 0.786 0.068 0.322 0.880CR 1044 0.029 0.146 0.000 3.095 144 0.008 0.002 0.001 0.012LR 1044 1.138 1.883 0.000 34.532 144 0.897 0.186 0.518 2.012

as conventional banks with 1044 observations and Islamic banks with 144 observations. Simple inspection of the table shows that conventional banks are likely to perform better in profitability since the means of return on asset, return on equity and net interest margin are higher. For conventional banks, ROA is 1.3%, ROE is 7.4% and NIM is 3% on average. For Islamic banks, ROA is 0.7%, ROE is 5.8% and NIM is 1.7%. On the other hand, credit risk and liquidity risk are lower for Islamic banks as one may expect this result because of the risk-sharing principle. While credit risk and liquidity risk are 2.9% and 114% for conventional banks, 0.8% and 89% for Islamic banks. Another issue that is worth underlining is the total cumulative asset growth, which is higher for Islamic banks than conventional banks, being 31.28% and 25.64% respectively. Correlation matrix for independent variables is presented in Table 5. Correlations among most of the variables are quite low, signalling multi-collinearity does not create significant bias in any of our analyses.

4. MethodologyIn the banking literature, it is quite common to use ordin ary least square (OLS) methods. The recent studies also employ a generalized method of moments (GMM) since the profit persistence is the common feature of banking data. The main focus of our study is to observe the differing behavior of Islamic banking in profitability. To observe the differing behavior, the sample can be split into two sub-samples, or pooled estimation can be done through assigning a dummy variable to examine the “being an Islamic bank” effect. However, one of the main problems in such analysis is the selection bias or non-random selection that may have been produced by OLS or GMM estimators. Specifying some certain banks as Islamic banks and investigating the effect of being an Islamic bank is a self-selection choice, therefore it is possible to have self-selection bias.

In order to overcome this possibility, we introduce the meth od of propensity score matching developed by Rosenbaum and Rubin (1985) to the banking literature in order to estimate the “average treatment effect” (ATE) of being an Islamic bank in Turkey where there exists a dual banking system. The purpose of this method is to create a control group that is similar to a treatment group, and the similarity among banks will be assessed

by calculating the propensity score that is defined as the conditional probability of receiving a treatment despite the unavailability of experimental data. In the present context, while Islamic banks constitute the treatment group, conventional banks constitute the control group, and being an Islamic bank is defined as receiving the treatment.

IBjt is defined as a dummy variable whether bank j is an Islamic bank at time t. While, y jt

1 denotes the profitability of bank j that is an Islamic bank at time, y jt

0 denotes the profitability of bank j at time t that is a conventional bank. The average treatment effect of being an Islamic bank is defined as:

τ jt jt jty y= −1 0 , (1)

If both states of the world, y jt1( ) and y jt

0( ) , were observable, the average treatment effect would be estimated with no trouble. Nevertheless, due to unobservability, the average treatment effect (τ) would be equal to the difference of mean outcomes y y1 0−( ) . Since either of states of world, y jt

1( ) and y jt0( ) , are observable,

we use propensity score matching to calculate the average treatment effect. As argued in Rosenbaum and Rubin (1985), a vector of covariates, Z, can be used to compare Islamic and conventional banks. Z is defined as:

y y IB Z IB Zjt jt

1 0 1 0 1, Pr , ,⊥ | |=( ) ∈( ) (2)

where ⊥ denotes independence. However it is not possible to find observations with identical values for all covariates in Z. In order to eliminate this problem, they suggest using propensity score matching, which uses probability of bank pairs that receive the treatment (IB) on the characteristics of the pair. I estimate the probability of a bank-pair that receives the treatment (IB) using the following logit specifications.

p IB F BSF MSF MACjt =( ) = ( )1 , , , (3)

where BSF is a vector of bank-specific variables includes capital adequacy, asset quality, credit risk, etc. MSF is a vector of market specifics variables, including total asset growth, total loans growth and total equity growth, and

Aytug and Ozturk


Tabl

e 5.

Cor

rela

tion

mat

rix

RO

AR

OE

NIM

ETA

LTA

NPL

TALO

GTA

LQD

DPT

AC

RLR

IBPD

FDTA

GR

TEG

RTL

GR

GD

PGR

INF

FXIR

RO

A1.

000

RO

E0.

643

1.00

0N

IM0.

246

0.16

71.

000

ETA

0.42

00.

016

0.22

61.

000

LTA

-0.1

890.

046

0.03

70.

464

1.00

0N

PLTA

0.17

4-0

.003

0.24

80.

376

0.02

41.

000

LOG

TA0.

016

0.26

50.

000

0.42

10.

379

-0.0

331.

000

LQD

0.13

60.

078

0.11

30.

424

0.54

90.

019

-0.5

07 1

.000

DPT

A-0

.254

0.00

60.

081

-0.5

180.

581

0.00

00.

357

-0.4

091.

000

CR

0.13

10.

024

0.07

40.

424

0.25

80.

137

-0.1

13 0

.199

-0.1

971.

000

LR0.

066

0.13

60.

004

0.06

20.

020

-0.0

86-0

.081

0.0

36-0

.375

-0.0

661.

000

IB0.

084

0.04

30.

163

0.13

70.

415

0.19

2-0

.002

-0.2

530.

373

-0.0

51-0

.045

1.00

0PD

0.02

40.

200

0.00

00.

162

0.11

8-0

.044

-0.4

210.

161

0.05

50.

042

-0.0

820.

117

1.00

00FD

0.02

60.

098

0.03

30.

077

-0.1

64-0

.091

-0.4

900.

366

-0.1

85-0

.071

0.08

7-0

.011

0.32

61.

0000

TAG

R0.

029

0.03

50.

030

-0.0

630.

141

-0.0

060.

007

-0.0

900.

127

-0.0

26-0

.008

0.28

10.

033

-0.0

031.

0000

TEG

R0.

081

0.10

40.

115

0.00

60.

050

0.03

6-0

.036

-0.0

620.

081

-0.0

24-0

.018

0.20

70.

024

-0.0

020.

111

1.00

0TL

GR

-0.0

270.

034

0.05

9-0

.067

0.05

0-0

.042

-0.0

38-0

.094

0.05

8-0

.021

0.01

00.

142

0.01

7-0

.002

0.77

0-0

.182

1.00

0G

DPG

R-0

.076

0.07

70.

108

-0.0

43-0

.065

-0.0

19-0

.050

-0.0

51-0

.006

0.03

3-0

.004

0.00

00.

000

0.00

00.

184

-0.1

170.

501

1.00

0IN

F-0

.017

0.01

30.

017

0.00

6-0

.256

0.06

8-0

.199

0.00

6-0

.026

-0.0

11-0

.042

0.00

00.

000

0.00

0-0

.308

0.05

2-0

.288

0.01

51.

0000

FX-0

.050

0.03

40.

009

0.00

10.

022

0.06

60.

098

0.06

90.

014

0.05

1-0

.048

0.00

00.

000

0.00

0-0

.219

-0.1

03-0

.359

-0.1

240.

048

1.00

0IR

-0.0

280.

034

0.01

9-0

.027

-0.2

520.

005

-0.2

20-0

.014

-0.0

30-0

.013

-0.0

220.

000

0.00

00.

000

-0.2

950.

017

-0.2

42-0

.028

0.82

4-0

.105

1.00

0



MAC is a vector of macroeconomic variables, including GDP growth, inflation rate, foreign exchange rate between the Turkish lira and the US dollar and real interest rate. The logit model is used here to identify the effect of Islamic banking on banking profitability. By using the logit model we can compare banking profitability of conventional and Islamic banks since both types of banks are similar in terms of their propensity scores. Next step is to use a matching technique in order to estimate missing counterfactuals by using obtained propensity scores, y yjt tjt

1 1,( ) or y yjt tjt0 0,( ). There are

several methods used as matching technique but we use three of them in our analysis, Nearest-Neighbor Matching, Stratification Matching and Kernel Matching. Thanks to these techniques we will be able to observe whether different matching algorithms result in different treatment effects. The average treatment effect of Islamic banking is given by

τ TT E E y IB z E y IB z= =[ ]− =[ ]{ }1 01 0| |, , , (4)

5. ResultsTo test the relationship between bank profitability and the bank specific, market specific and macroeconomic determinants described earlier, we first estimate a linear regression model in the following form:

ROA BSF MSF MACit it it it= + + + +α β β β ε1 2 3 (5)

ROE BSF MSF MACit it it it= + + + +α β β β ε1 2 3 (6)

NIM BSF MSF MACit it it it= + + + +α β β β ε1 2 3 (7)

Tables 6,7 and 8 present regression results for three different dependent variables (ROA, ROE and NIM). In our analysis, the coefficients are estimated for the entire, pre-crisis and post-crisis periods. For each time period, we estimate three regression models while the first one only includes bank specific factors, the second one includes both bank and market specific factors, and the third one includes bank-specific, market-specific and macroeconomics factors. Table 6 depicts the regression results when ROA is the dependent variable. According to Table 6, the determinants of profitability vary over the periods, and being an Islamic bank does not have a statistically significant effect on ROA except in the third model for the full sample. Capital adequacy, asset size, credit risk, liquidity risk, total asset growth, total equity growth, foreign exchange rate have statistically significant effects on ROA. The results are in line with the recent literature (Athanasoglou et al., 2008).

Table 7 shows the results when banking profitability is measured as return on equity (ROE). It is obvious that the determinants of profitability do not vary over periods as much as they do for ROA. However for ROE, being an Islamic bank never plays a significant role, even though the coefficients are higher compared to the coefficients in Table 6. On the other hand, being a private bank is one of the determinants of profitability over periods. Interestingly, liquidity risk is an important factor for ROE while it is not for ROA.

Table 8 treats NIM as a dependent variable and presents regression results. The determinants of NIM and ROE are very similar as they are revealed. ETA, LTA and NPLTA are positively correlated with NIM and ROE. The most interesting result is being an Islamic bank is negatively

Table 6. Regression results: ROA is the measure of bank profitability

Dependent Variable ROA Full sample (1994–2011) Before the crisis (1994–2007) After the crisis (2008–2011)Independent Variables Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

IB -0.0053 -0.0081 -0.0091 0.0098 -0.0139 -0.0140 -0.0021 -0.0036 -0.0033ETA 0.0661 0.0631 0.068 0.0759 0.0709 0.0674 0.0898 0.0916 0.0934LTA 0.014 0.0143 0.0187 0.0316 0.0325 0.0312 -0.0076 -0.0022 0.0000NPLTA -0.0329 -0.0250 -0.0161 0.0025 0.0170 0.0128 -0.0673 -0.1036 -0.1071LOGTA 0.0028 0.0027 0.006 0.0053 0.0042 0.0026 0.0071 0.0079 0.0097LQD 0.0043 0.0044 0.0071 0.0164 0.0155 0.0122 0.0043 0.0066 0.0101DPTA -0.0016 -0.0028 -0.0035 0.0082 0.0063 0.0056 0.0075 0.0059 0.0051CR -0.0235 -0.0229 -0.0211 -0.0059 -0.0053 -0.0040 -0.0214 -0.0189 -0.0162LR 0.0003 -0.0003 -0.0004 -0.0004 -0.0005 -0.0005 0.0027 0.0028 0.0031PD -0.0045 -0.0042 -0.0024 -0.0069 -0.0073 -0.0082 0.0007 0.0011 0.0023FD 0.0003 0.0002 0.0025 0.0033 0.0026 0.0018 0.0001 0.0008 0.0018TAGR 0.0579 0.0533 0.1075 0.0826 0.0642 0.1019TEGR 0.0314 0.0318 0.0141 0.0118 0.0238 0.0098TIGR -0.0209 -0.0006 -0.0558 -0.0188 -0.0879 -0.1709GDPGR -0.0002 -0.0016 0.0002INF 0.0003 0.0003 0.0003FX -0.0086 0.0284 -0.0155IF 0.0001 -0.0004 -0.0001R-SQ 0.18 0.18 0.19 0.14 0.15 0.16 0.39 0.43 0.45PROB>CHI2 (F test) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Aytug and Ozturk


Table 7. Regression results: ROE is the measure of bank profitability

Dependent Variable ROA Full sample (1994–2011) Before the crisis (1994–2007) After the crisis (2008–2011)

Independent Variables Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

IB -0.0116 -0.0256 -0.0306 -0.0023 -0.0177 -0.0220 -0.0170 -0.0267 -0.0260ETA 0.1389 0.1204 0.1424 0.2284 0.2026 0.1702 0.1966 0.2042 0.2298LTA 0.062 0.0643 0.0877 0.11778 0.1244 0.1119 -0.1393 -0.1109 0.0999NPLTA -0.5497 -0.5206 -0.5001 0.486 -0.4218 -0.4526 -0.1330 -0.3459 -0.2325LOGTA 0.0233 0.0228 0.0400 0.0422 0.038 0.0298 0.0623 0.0694 0.0926LQD 0.0070 0.0068 0.0188 0.0130 0.0065 -0.0181 0.0248 0.0354 0.05655DPTA -0.0005 -0.0072 -0.0113 0.0137 -0.0238 -0.0301 0.1387 0.1318 0.1216CR -0.0509 -0.047 -0.0369 -0.0120 -0.0090 -0.0015 -0.0157 -0.0057 0.0169LR 0.0075 -0.0075 -0.0077 -0.0105 -0.0107 -0.0115 0.0452 0.0458 0.0465PD -0.0806 -0.0794 -0.0694 -0.0894 -0.0907 -0.0936 0.0004 0.0057 0.0206FD 0.0104 0.0099 0.0220 0.0354 0.0331 0.0297 0.0028 0.0083 0.0225TAGR 0.2695 0.2350 0.5297 0.3850 0.2357 0.3929TEGR 0.1898 0.1901 0.0979 0.1059 0.1601 0.0953TIGR -0.1402 -0.0217 -0.3472 -0.0486 -0.3472 -0.647GDPGR -0.0011 -0.0099 0.0003INF 0.0030 0.0013 0.0061FX -0.0452 0.0421 -0.0718IF 0.0001 -0.0008 -0.0010R-SQ 0.12 0.13 0.15 0.14 0.15 0.17 0.23 0.26 0.26PROB>CHI2 (F test) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

Table 8. Regression results: NIM is the measure of bank profitability

Dependent Variable ROA Full sample (1994–2011) Before the crisis (1994–2007) After the crisis (2008–2011)

Independent Variables Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

IB -0.019 -0.0238 -0.0244 -0.0110 -0.0176 -0.0185 -0.0229 -0.0272 -0.0272ETA 0.0263 0.0197 0.0229 0.0383 0.0293 0.0226 0.0203 0.0239 0.0239LTA 0.0165 0.0165 0.0182 0.0136 0.0139 0.0056 0.0233 0.0355 0.0385NPLTA 0.1572 0.1651 0.1754 0.0977 0.1258 0.1313 0.3258 0.2664 0.2905LOGTA -0.0011 -0.0017 0.0002 0.0033 0.0017 -0.0026 0.000 0.0012 0.0033LQD 0.0167 0.0157 0.0168 0.0211 0.0203 0.0134 0.0231 0.0259 0.0315DPTA 0.0084 -0.0062 0.0059 0.0026 -0.0007 -0.0014 0.0025 -0.0003 -0.0012CR -0.0163 -0.0153 -0.0142 -0.0077 -0.0065 -0.0059 -0.0076 -0.0025 0.0019LR -0.0004 -0.0005 0.0005 -0.0006 -0.0007 -0.0008 0.0032 0.0033 0.0038PD -0.0021 -0.0020 -0.0008 -0.0004 -0.0001 -0.0017 0.0003 0.0005 0.0018FD -0.0027 -0.0032 -0.0018 0.0045 0.0056 -0.0081 0.0024 0.0038 0.0050TAGR 0.1326 0.1123 0.1311 0.1098 0.1681 0.3036TEGR 0.0472 0.0520 0.0401 0.0453 0.0447 0.0249TIGR -0.0685 -0.0329 -0.0510 -0.0128 -0.185 -0.4159GDPGR -0.0003 -0.0025 0.0008INF 0.0001 0.0009 0.0023FX -0.0006 0.0289 -0.0190IF 0.0002 -0.0003 -0.0001R-SQ 0.11 0.14 0.17 0.13 0.15 0.14 0.22 0.30 0.35PROB>CHI2 (F test) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00



correlated with NIM and it is always significant at a 99% level for the full sample and after the crisis. Nevertheless, being an Islamic bank lowers the profitability when it is measured as net interest margin (NIM). The effect of Islamic banks was pretty limited or not significant before the crisis.

Capital adequacy (ETA) has a positive and significant impact on ROA and ROE in all periods with different models. The findings imply that banks that have higher capital adequacy are more profitable. This may be related to the fact that banks with higher capital adequacy tend to be more credible and operate with lower costs. The specific finding for the Turkish case constitutes a different dimension in explaining the solvency risk (capital adequacy) on profitability. Pervan, et al. (2012) find that capital adequacy ratio is negatively related with return in their analysis. They conclude that higher capital adequacy implies lower profitability. They point out that a higher level of bank capital provides safety and over-caution in the banking business and it reduces profitability in the Macedonian banking system. Likewise, our results show that capital adequacy doesn’t have a significant effect on NIM after the crisis. While it has a positive and significant effect on NIM for the full sample, the relationship between capital adequacy and NIM is blurry before and after the crisis with different models.

In terms of liquidity risk, the findings are quite specific to the Turkish case and are not in accordance with studies in the literature (Pervan et al., 2012). While liquidity risk has a positive and significant effect on ROA and ROE only after the crisis, it has a negative effect on ROE before the crisis and for the full sample. The banks with less liquidity risk are deemed to be more profitable, since the banks that have higher loan to deposit ratios are expected to produce more returns due to interest revenues. However, the Turkish case proves to be different than the theory, and Turkish banks are more tempted to invest in government assets that proposed higher yields (Aysan and Ceyhan, 2007). Therefore, the liquidity risk implies a positive relationship between liquidity and profitability as it is in some extent not in line with recent findings.

In the model where ROA is dependent, it is interesting to note that the coefficient of non-performing loans to asset reveals a negative relationship with bank profitability and is statistically significant only after the crisis. The estimated coefficients are negative for all the samples. When the ROE is a dependent variable, the coefficients are still negative, but they are statistically significant only for the full sample. The empirical finding is in contrast with the skimping hypothesis of Berger and DeYoung (1997). Berger and DeYoung (1997) suggest that under the skimping hypothesis, a bank maximizing the long run profits may rationally choose to have lower costs in the short run by skimping on the resources devoted to underwriting and monitoring loans but bear the consequences of greater loan performance problems. Therefore, cost minimization in the short run may not bring long-term profitability. Yet, the findings are plausible in Turkey’s case where the non-performing loans are low. Therefore, holding more funds kept for potential losses in the future increases banks cost, thus

diminishing bank profitability. However the effect on NIM reveals a different picture since the coefficients are always positive and significant.

The sector specific variables, the asset and equity growth positively affect profitability in Turkey. In aggregate terms, as the banking system as a whole scales up, the asset and equity size, individual banks tend to be more profitable. This also implies that asset and equity growth is distributed evenly among banks. Therefore, no specific bank or banking group dominated the others. In contrary to this, total loans growth has a negative and significant effect on profitability.

In both estimated regressions where ROA and ROE is the dependent variable, asset size is found to have a significant and positive impact on profitability. Hauner and Peiris (2005) suggest two potential explanations for this impact. First, if it relates to market power, large banks should pay less for their inputs, i.e., lower cost as mentioned in capital adequacy case. Second, there may be increasing returns to scale gains through the allocation of fixed costs.

The deposit to total assets ratio has a negative but insignificant impact on bank profitability. This might be an indication of the fact that the link between deposit and lending is not efficiently operated. Turkish banking systems have suffered from the short term character of its deposit base. The maturity of deposit is mainly short and cannot be efficiently converted into lending, i.e., higher income earnings.

Our results regarding the impact of ownership on profitability support the findings of Micco, et al. (2007) and Iannotta, et al. (2007), who point out that state banks exhibit a lower profitability than privately owned banks. The case also holds for foreign banks. Foreign banks are also more profitable than state banks. The findings shed light on the inefficiency of state banks in the past where state banks were mandated as the lender of unprofitable and politically driven projects. For instance, the duty losses that were one of the main causes of the 2001 banking crisis were an indication of how state banks operated with political bias. One of the main purposes of this paper is to find the average treatment effect (ATE) of being an Islamic bank. Regression results give us some insights about how the ATE might look for different measures of profitability. According to the regression results Islamic banks play an insignificant role when the measure is ROA and ROE. In addition to this, Islamic banks lower the profitability when the measure is NIM.

Estimates of treatment effects are presented in Table 9. As we mentioned above we use the different algorithms to calculate the treatment effect for the sake of robustness. First, the treatment effect proves that there is no relationship between Islamic banks and ROA. The coefficient is always zero. Second, Islamic banks are positively correlated with ROE and the magnitude is higher. However, OLS coefficients are insignificant. Third, the ATE on NIM varies over periods. It is always positive before the crisis and negative after crisis. For the entire period, different algorithms prove the negative relationship between Islamic banks and NIM.

Aytug and Ozturk


6. ConclusionThis paper provided evidence on the implications of Islamic banking in Turkey. As a baseline analysis and in parallel with the literature, the determinants of profitability in Turkish banking have been found to be familiar with the previous studies. However, our main difference comes with the role of being an Islamic bank. In our study we investigated the effect of being an Islamic bank on bank performance in terms of various bank profitability measures. Specifying some certain banks as Islamic bank and investigating the effect of being an Islamic bank is a conscious “self-selection” choice, therefore have a potential to have “self-selection bias”. In order to overcome this possibility, we introduced the method of propensity score matching to the banking literature in order to estimate the average treatment effect (ATE) of being an Islamic bank in Turkey where there exists a dual banking system. In order to estimate the average treatment effect propensity score matching has been conducted. The results are compared with the OLS results. According to the OLS results, Islamic banks have a negative relationship with NIM and a positive relationship with ROA. The effect of being an Islamic bank on ROE is insignificant on the other hand. Propensity score matching technique confirms some of OLS results. The ATE of being an Islamic bank, on ROA and ROE, are positive, yet it is blurry but negative on NIM. The results provided evidence of increasing performance of Islamic banks in the Turkish banking system. Therefore, although not in total asset size – these banks signal the efficacy of Islamic banks in the near future.

Several potential research themes emerge as well. Firstly, the blurry results of being an Islamic bank on NIM have to be identified in some detail. This discrepancy could provide further insights about Islamic banks operations if deeply

investigated. Second, the results of the Turkish case have to be compared with other country cases by identifying more proper econometric methodologies as introduced in this paper. The issue of “self-selection bias” needs to be tested in other country cases as well. As a continuation of the concurrent problem of “self-selection bias,” the profitability of Islamic banks should be compared with the conventional banks in a synthetic control group framework introduced by Abadie and Gardeazabal (2003).

Notes1. See Hasan and Dridi (2010) for differing performance

of conventional and Islamic banks during and after the financial crisis in countries where conventional banks and Islamic banks jointly operate.

2. After the decree, initial institutions were Bahrain-based Al Baraka Turk and Saud-based Faisal Finans Kurumu, which each opened subsidiaries in 1985. The Kuwaiti-based Kuveyt Turk Kurumu began operations in 1989. Eventually, special finance houses began lending with mainly domestic capital, including: Anadolu Finans (1991), Ihlas Finans (1995) and Asya Finans (1996).

3. See Darrat (1988), Zuberi (1992), Darrat and Salaaming (1990), Kia and Darrat (2007).

ReferencesAbadie A, Gardeazabal J. (2003) “The Economic Costs of

Conflict: A Case Study of the Basque Country. American Economic Review. 93(1):113–132.

Abreu M, Mendes, V. (2002). Commercial bank interest margins and profitability: evidence from EU countries. Working Paper, Porto.

Table 9. Estimates of treatment effect

1994–2011 1994–2007 2008–2011

RE NM S OLS NM S OLS NM SROA Islamic Banking -0.005 0 0 -0.014 0 0 0.005 0.001 0.001

No. Obs 1188 160 1080 660 84 600 528 68 478Treated 144 144 144 80 80 4 64 64 62Controls 1044 16 936 580 4 596 464 4 416

1994–2011 1994–2007 2008–2011

RE NM S OLS NM S OLS NM SROE Islamic Banking -0.010 0.004 0.003 -0.006 -0.007 0.012 0.017 0.015 0.013


1994–2011 1994–2007 2008–2011

RE NM S OLS NM S OLS NM SNIM Islamic Banking -0.023 -0.003 -0.004 -0.018 0.001 0.011 -0.023 -0.005 -0.005




Albertazzi U, Gambacorta L. (2009) Bank Profitability and the Business Cycle. Journal of Financial Stability. 5(4):393–409.

Athanasoglou PP, Brissimis SN, Delis MD. (2008) Bank-Specific, Industry-Specific and Macroeconomic Determinants of Bank Profitability. Journal of International Financial Markets, Institutions and Money. 18(2):121–136.

Aysan AF, Ceyhan SP. (2007) Market Disciplining Role of Crisis on the Restructuring of the Turkish Banking Sector. Technical Report.

Bennaceur S, Goaied M. (2008) The Determinant of Commercial Bank Interest Margin and Profitability: Evidence from Tunisia. Frontiers in Finance and Economics. 5(1):106–130.

Berger AN. (1995) The Profit-Structure Relationship in Banking-Tests of Market-Power and Efficient-Structure Hypotheses. Journal of Money, Credit and Banking. 27(2):404–31.

Berger AN, DeYoung R. (1997) Problem Loans and Cost Efficiency in Commercial Banks. Journal of Banking and Finance. 21(6):849–870.

Berger AN, Hanweck GA, Humphrey DB. (1987) Competitive Viability in Banking: Scale, Scope and Product Mix Economies. Journal of Monetary Economics. 20(3):501–520.

Bourke P. (1989) Concentration and Other Determinants of Bank Profitability in Europe, North America and Australia. Journal of Banking and Finance. 13(1):65–79.

Darrat A, Salaaming M. (1990) Islamic Banking: An Outline of Some Conceptual and Empirical Aspects. Savings and Development. 19:185–192.

Darrat AF. (1988) The Islamic Interest-Free Banking System: Some Empirical Evidence. Applied Economics. 20(3):417–425.

Demirguc A, Huizinga H. (1999) Determinants of Commercial Bank Interest Margins and Profitability: Some International Evidence. World Bank Economic Review. 13(2):379–408.

Dietrich A, Wanzenried G. (2011) Determinants of Bank Profitability Before and During the Crisis: Evidence from Switzerland. Journal of International Financial Markets, Institutions and Money. 21(3):307–327.

EC M, PC R. (2003) Determinants of Greek Commercial Banks Profitability, 1989–2000. Spoudai. 53(1):84–94.

Ergec EH, Arslan BG. (2011) Impact of Interest Rates on Islamic and Conventional Banks: The Case of Turkey.

Garca-Herrero A, Gavil S, Santabrbara D. (2009) What Explains the Low Profitability of Chinese Banks? Journal of Banking and Finance. 33(11):2080–2092.

Goddard J, Molyneux P, and Wilson JOS. (2004) Dynamics of Growth and Profitability in Banking. Journal of Money, Credit and Banking. 36(6):1069–90.

Hardy L. (2012) The Evolution of Participation Banking in Turkey. Online Journal on Southwest Asia and Islamic Civilization. Winter 2012.

Hasan M, Dridi J. (2011) The Effects of the Global Crisis on Islamic and Conventional Banks: A Comparative Study. Journal of International Commerce, Economics and Policy (JICEP). 2(02):163–200.

Hauner D, Peiris SJ. (2005) Bank Efficiency and Com petition in Low-Income Countries: The Case of Uganda. IMF Working Papers 05/240. International Monetary Fund.

Iannotta G, Nocera G, Sironi A. (2007) Ownership Structure, Risk and Performance in the European Banking Industry. Journal of Banking and Finance. 31(7):2127–2149.

Kassim S, Majid M, Yusof R. (2009) Impact of Monetary Policy Shocks on the Conventional and Islamic Banks in Dual Banking System: Evidence from Malaysia. Journal of Economic Cooperation and Development. 30(1):41–58.

Kia A, Darrat AF. (2007) Modelling Money Demand Under the Profit-Sharing Banking Scheme: Some Evidence on Policy Invariance and Long-Run Stability. Global Finance Journal. 18(1):104–123.

Kwast ML, Rose JT. (1982) Pricing, Operating Efficiency, and Profitability Among Large Commercial Banks. Journal of Banking and Finance. 6(2):233–254.

Macit F. (2012) Bank Specific and Macroeco no mic Determinants of Profitability: Evidence From Partici-pation Banks in Turkey. Economics Bulletin. 32(1): 586–595.

Micco A, Panizza U, Yaez M. (2005) Bank Ownership and Performance Does Politics Matter? Working Papers Central Bank of Chile 356. Central Bank of Chile.

Micco A, Panizza U, Yanez M. (2007) Bank Ownership and Performance – Does Politics Matter? Journal of Banking & Finance. 31(1):219–241.

Michelle NC, Wheelock D. (1997) Why Does Bank Performance Vary Across States? Review. (Mar):27–40.

Molyneux P, Thornton J. (1992) Determinants of European Bank Profitability: A Note. Journal of Banking and Finance. 16(6):1173–1178.

Ozturk H, Gultekin-Karakas D, Hisarciklilar M. (2010) The Role of Development Banking in Promoting Industrialization in Turkey. Region et Developpement. 32:153–178.

Pasiouras F, Kosmidou K. (2007) Factors Influencing the Profitability of Domestic and Foreign Commercial Banks in the European Union. Research in International Business and Finance. 21(2):222–237.

Pervan M, Curak M, Poposki K. (2012) Industrial Con-

centration and Bank Performance in an Emerging Market: Evidence from Croatia. In: Advances in Finance and Accounting. Tomas Bata University, Czech Republic.

Rosenbaum PR, Rubin DB. (1985) Constructing a Control Group Using Multivariate Matched Sampling Methods that Incorporate the Propensity Score. The American Statistician. 39(1):33–38.

Short BK. (1979) The Relation Between Commercial Bank Profit Rates and Banking Concentration in Canada, Western Europe, and Japan. Journal of Banking and Finance. 3(3):209–219.

Aytug and Ozturk


Staikouras C, Wood G. (2002) The Determinants of European Bank Profitability. International Business and Economics Research Journal. 3:57–68.

Warde I. (2010) Islamic Finance in the Global Economy. 2nd Edition. Edinburgh University Press. Edinburgh.

Yeldan E. (2007) Patterns of Adjustment Under the Age of Finance: The Case of Turkey as a Peripheral Agent of Neoliberal Globalization. Technical report.

Zuberi HA. (1992). Interest-Free Banking and Economic Stability. The Pakistan Development Review. 31:1077–1087.


Cite this chapter as: Calkan Z T (2015). Estimating expected returns on Mudaraba time deposits of Islamic banks. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Estimating expected returns on Mudaraba time deposits of Islamic banksZeynep Topaloglu CalkanAdjunct Assistant Professor, Georgetown University – Qatar, Phone: +974 5568 9211, [email protected]

Abstract - On the deposit side, Islamic banks work on a mudaraba (partnership) contract, where the depositor and the bank are business partners. While in conventional banks the depositor is provided with a fixed interest rate, in Islamic banks the depositor can only discover his return when the investment period is over. This fundamental distinction brings forth a disadvantage for Islamic banks while competing with their conventional counterparts in the market.

On the other hand, most of the credits extended by Islamic banks follow a murabaha (cost-plus sale with deferred repayment) contract, and the banks specify profit rate on the credits from the beginning. Using this information we have developed a forecast model to quote the depositors their expected returns on mudaraba time deposits within a 95% confidence interval at the beginning of the investment term. Besides increasing competitive advantage, estimating expected returns will assist Islamic banks in their risk management and asset-liability management.

Keywords: Islamic banking, estimating returns on Islamic time deposits

1. IntroductionThe financial crisis of 2007–2008 triggered an extended global recession, which still distresses the economic activities. This crisis also evoked questions about morality of the current global financial system. The research after the crisis shows that financial institutions of the West, with the help regulators and the rating agencies, deceived their clients and the public.1 Islamic finance and its ethical principles emerged as an alternative model because most of the elements that caused and extended the crisis are not permissible by Shariah rule.2 Although the virtue of the Islamic financial system was invigorated after the crisis, the scholars and practitioners of Islamic finance have been emphasizing its value long before. Considering its merit, the insignificant size of Islamic financial assets compared to global financial assets makes us question what can be and should be improved by practitioners of the area of Islamic finance to reach its deserved levels. Here, in this paper, we propose a tool to improve competitiveness of Islamic banks. This is an attempt to facilitate the process of reflecting intrinsic value of Islamic finance to statistical figures.

Except a few countries3 where the entire internal financial system has been transformed, Islamic banks have to compete with their conventional counterparts in attracting customers. Factors like lower penetration4 and fewer

financial instruments place Islamic banks in an unfavorable position compared to their competitors.

One of the structural factors that cause difficulties for Islamic bankers is the inability to provide a specific rate of return to depositors. Depositors are used to knowing the interest rate and are most of the time comparing different investment alternatives based on their rate of return. This becomes challenging especially when banks are interested in attracting customers to long-term savings accounts. Depositors of Islamic banks need to wait up until the end of their investment period in order to observe their return, whereas in conventional banks depositors know how much they will receive at the beginning of their investment.

The reason behind this lack of information is the underlying Shariah contract in Islamic savings accounts. Typically two types of contracts, wadia (safe-keeping) and mudaraba (profit-loss sharing), are used in these accounts. Briefly, wadia is a contract where the bank acts as a safe keeper to depositor’s funds with a permission to utilize them at its own risk, and mudaraba is a type of profit-loss sharing contract where a depositor invests money into a partnership with the bank, where the bank provides its labor/services to the partnership in an agreement to share proceedings – either profit or loss – later on. In both of these contracts it is forbidden to specify a rate of return in the beginning as

Calkan


it nullifies the essence of these contracts. Besides Shariah prohibition, in reality it is not possible to know the value of ingenerated profits/losses.

Currently, the banks that use these two contracts in their profit sharing investment accounts (PSIA) do not share a profit rate with their deposit clients. They only share historical information, mainly previous term’s profit distribution. In this paper, we argue it is possible for Islamic banks to provide an anterior rate to their clients without breaking Shariah rules by sharing already available information in their accounts.5

Although it is not permitted and possible to specify the profits in advance, it is still possible to make precise estimations. The reason behind it lies in the way PSIA funds are being utilized in the Islamic banks. Scholars argue that Islamic banks are initially established in order to promote sharing risk not only on the depositor side but also on the financing side through profit/loss sharing system. However due to reasons like mismatch between asset and liability duration, moral hazard and adverse selection issues, today the main mode of financing in Islamic finance is not one of the profit/loss sharing contracts (musharaka or mudaraba) but one of the trade contracts namely murabaha. In murabaha contract the Islamic financial institution (IFI) acts as a tradesman. It purchases the items from the vendor and sells them to its client with a mark-up in a deferred payment plan. There are several requirements of murabaha contract that have to be satisfied by the bank and the client; however, for our research, the significance of this contract is its feature to provide ex-ante profit rate. Both parties know and agree on how much profit will be charged by the bank on the item well before the transaction takes place. Most of the time Islamic banks follow conventional loan rates when quoting their mark-up price since they are in competition with them.

Therefore even if IFIs cannot specify the profit rate that they will distribute to PSIA holders, on the financing side they know how much money they will make when they sell the murabaha contract. It is possible to use this knowledge to generate a reliable estimation about their profit distribution.

In the following sections of the paper the research proceeds as follows: In the second section, the main discussions about profit loss sharing accounts in the literature have been presented. In the third section the methodology to estimate PSIA return has been developed, and in the last section the results and conclusion have been shared with the readers.

2. Literature reviewProfit sharing investment accounts (PSIA) have been a topic of interest among scholars in the field of Islamic finance. There are mainly two issues, one of which is accounting treatment and risk management implications of PSIAs, while the other is the profit smoothing applications of Islamic Financial Institutions (IFI) against the risk of losing PSIA investors, in other words displacement risk.

Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI) in its Financial Accounting Standard

No. 66 defines PSIAs as a new category between liability and owners’ equity since PSIA investors are not regular depositors receiving a fixed return but they are profit/loss-sharing partners. On the other hand PSIA investors do not have managerial and voting rights as owners of the bank. Akacem and Gilliam (2002), Kahf (2005), Sultan (2006), Ayub (2007), Ibrahim (2007), Shubber and Alzafiri (2008), in agreement with the definition provided by AAOIFI, states PSIAs are an equity-like instrument compared to conventional deposits. This attribute of the PSIAs brings challenges in terms of accounting treatment of them. Atmeh and Ramadan (2012) critically evaluate accounting treatment of mudaraba returns by AAOIFI, and they explain how some of its implications deteriorate the reliability and fairness of the financial statements and how they compare AAOIFI standards with IFRS in that respect. In application we observe not all banks stick to the AAOIFI standards. Rating Agency Malaysia (RAM) in its Research Report (October–December 2007) comparing Malaysian and Middle Eastern IFIs, points out reporting PSIAs as a liability on the balance sheet as one of the differences of Malaysia.

Besides accounting treatment, PSIAs equity-like structure brings out different risk implications compared to conventional risk management. Since PSIA investors are supposed to bear the associated loss in these accounts, capital adequacy calculation, and the approach towards interest rate risk and credit risk differs in IFIs. Khan and Ahmed (2001) provide a detailed analysis of the risk management in the Islamic financial industry. Archer and Abdel Karim (2009) especially call attention to the regulatory challenges faced in Europe and North America, where there is no special regulation for IFIs. Ariffin and Kassim (2011) through a survey on selected banks uncover the areas for improvement in the risk management practices within IFIs.

On the other hand, Sundararajan (2005) with his cross-country study argues PSIAs are subject to a considerable amount of smoothing on their return, which in turn implies the investment risks of the banks that are not fully shared by the PSIA investors. He challenges whether IFIs really need a distinct risk treatment compared to conventional banks.

Second major issue about PSIAs is the return equalization activities of IFIs. As pointed out by Sundararajan (2005), and Archer et al. (2010), in order to compete in the market and avoid withdrawal risk by providing market-related returns to PSIAs, IFIs employ a variety of return smoothing activities. Additionally, in Jordan, Malaysia and Qatar, the central bank requires IFIs to manage PSIAs in a way not to reflect losses to the investors and to “smooth” returns. Therefore due to regulatory requirements or market pressure, IFIs are driven to use a combination of methods like conservative investment strategies, profit equalization reserves (PER), (a reserve account formed out of profits of PSIA to smooth profits), investment risk reserves (IRR), (another similar account to cover periodic losses) or a donation to PSIA holders from the share of the owners. Besides Sundararajan (2005), many scholars like Zuobi and Al-Khazali (2007), Taktak et al. (2010), Farook et al. (2012) confirm that IFIs pursue income-smoothing activities to compete with market-based deposit interest


Estimating expected returns on Mudaraba time deposits of Islamic banks

rates. Taktak et al. (2010) demonstrate unlike conventional banks, IFIs do not use Loan Loss Provisions (LLP) but use PER and IRR to provide steady returns. On the other hand Zuobi and Al-Khazali (2007) report GCC banks that use LLP to smooth their returns.

The evidence for profit distribution management revealed recently by Farook et al. (2012) represents the strongest support in the literature. They have utilized an extended dataset covering 37 banks in 17 countries. They have shown most IFIs do really manage profit distributions, with IFIs in Brunei, Malaysia and UAE showing lower average profit distribution management, while in Bahrain, Indonesia, Pakistan, Saudi Arabia, IFIs presenting a higher average profit distribution management. Percentage of Muslim population, financial development, market concentration, depositor reliance and age of the Islamic bank are the most significant factors to answer the question why IFIs engage in income smoothing.

The existence of income smoothing activities by itself proves the competitive pressure on Islamic banks. Literature shows competing with the conventional banks is one of the most important challenges of IFIs and they utilize various methods to overcome this challenge. In this study we will provide a model to forecast PSIA returns in the short-run to provide a marketing tool for IFIs.

3. MethodologyIn order to define PSIA returns, we will use a simplified version of the framework mandated by the Turkish Banking Authority to the IFIs in Turkey. The reasons behind choosing the Turkish framework follow:

1. PSIA funds are not mixed with bank equity capital2. Profit/loss distribution is done on a daily basis3. Instead of PER and IRR, LLP is used to smooth returns

which makes it comparable to conventional banks

Before stating the formula for PSIA returns, we need to define some of the terminology that will be used in our specification.

Unit Value (ut): It is assumed to be equal to 100 on the first day IFI accepted deposits to its PSIAs. It changes daily based on the distribution of profit and loss to the account. We can consider unit value as an index to tell us how our funds are performing.7 There is one unit value for each day of operation. The change from one day to another reflects the percentage of profits or losses made by the bank during one day.

Unit Account Value (uat): It defines the current total monetary value of PSIAs for the bank, which can be calculated by multiplying unit value and account value. For every individual account holder, unit account value is equal to the amount of money they have deposited in the first day. Unit account value will grow each day based on bank’s performance.

uat = ut × at (1)

Account Value (at): It is the value calculated by dividing the amount of money deposited to PSIAs by that day’s unit

value. It represents the participation share of PSIA holder to the total funds. There is an account value for each individual account holder. The total of all depositors’ account values make the account value for the total funds.

After defining the basic terminology, we will now define the formula to calculate unit value. As stated, unit value is basically an index to track the performance of PSIAs. It can be said that by converting PSIAs to an index value, the Turkish Banking Authority requires Islamic banks to remove the size effect and concentrate on the performance. In our study we use the following simplified formula for unit value:

u

ua R c yat

t t t t

t+ =

+ - +1

( ) (2)

where ut+1 represents unit value for time t + 1, the next day, uat represents unit account value for time t, Rt represents revenues to PSIA for time t, Ct represents costs to PSIA for time t, yt represents reserves for time t and at represents account value for time t.8

Rt, revenues to PSIA, is defined as a combination of profits accrued to the PSIA account within current business day plus any annulment of former provisions and/or reserves. Ct, costs to PSIA, is defined as a combination of provisions imposed by regulations, plus payments to deposit insurance fund and loan loss provisions. Revenues and costs are reflected on the calculation starting with the first day of the loan. Therefore even if the bank does not receive any installments, the PSIA account will start recording a profit when the loan is extended. For the purposes of simplicity we will assume all loans follow a murabaha contract.9

After calculating unit value, we define daily returns to our PSIA funds with the following formula:

r

u uut

t t

t+

+=−

×11 100 (3)

In this calculation for unit value, PSIA costs and reserves are constant percentages of PSIA revenues given by the regulatory authority. Therefore the only variants in this formula are unit account value, which represents the amounts deposited to the bank and PSIA revenues, representing the returns from the loans extended by the bank. Unit account value fluctuates as the depositors withdraw their money or deposit new funds to their accounts. PSIA revenues increase as new murabaha loans are extended to the clients. Therefore in our analysis we will simulate these two variables and the unit value and daily returns to PSIA funds will adjust.

4. Results and conclusion

DataIn our analysis, we will use the formula generated in the methodology section to simulate the daily average PSIA returns of Turkish participation banks.10 Currently there are four participation banks operating in Turkey. They publish daily returns to their PSIA funds for the

Calkan


trailing one-month, three months, six months and one year for different currency classes in their websites. We have taken four banks’ average data for trailing one-month (31 days) returns to Turkish Lira PSIA funds for the period between January 15, 2012, and February 15, 2012. In our simulation we try to estimate returns during this one-month period. We have chosen one-month maturity, since these accounts hold the highest amount of funds compared to other maturities. The date selection is arbitrary. Turkish lira has been chosen as the selected currency since the analysis is being done with Turkish data. Figure 1 presents the path followed by PSIA returns for our selection.

First simulationIn our simulation we have assumed the following:

At t = 0 At t >0Unit account

valueua0 : TL6 billion Endogenous

Unit value u0 : 100 EndogenousAccount value a0 : 60 million Assumed constantPSIA revenues R0 : TL 2 million Will be estimatedPSIA costs C0 : 15% of revenues ExogenousReserves y0 : 5% of revenues Exogenous

Unit account value represents the deposits collected by our average participation bank. Since we are using the data from the beginning of 2012 for TL accounts, we have referred to TL deposits collected at the end of 2011. Total value of TL deposits collected has been TL 24.04 billion.11 In our simulation we have taken the average of this value and used TL 6 billion12 as unit account value.

Unit value is assumed to be 100. Since it is an index to follow the performance of the funds, instead of its absolute value, the changes from day to day are significant for our analysis.

Assuming a unit account value of TL 6 billion and unit value of 100, account value is calculated to be TL 60 million.13 With the separation of unit value and account value, it is now possible to track changes in deposits in two ways. Unit value captures the changes due to the investment returns (profits and losses), and account value captures the fluctuation in the funds deposited to the bank (deposits and withdrawals).

The percentages 15% and 5% for PSIA costs and reserves assumed in our simulation are an approximation based on the rules imposed by the Turkish regulatory authority. Turkish regulatory authority requires banks to set aside reserves under the names of special provision, general provision, federal deposit insurance premium and precautionary provision. Based on Turkish monetary and macroeconomic policy, the rates on these provision expenses can be altered; a total of 20% is a fair approximation to the reality. The ratios for PSIA costs and reserves are kept constant in all simulations since they represent the best legal approximation.

For PSIA revenues our selection of TL 2 million daily return is derived from 6.75%, with a starting monthly PSIA return given in Figure 1. TL 2 million daily revenue is equal to 0.03% daily return on TL 6 billion pool of funds. Using 70/30 profit sharing ratio between the bank and the depositor, it refers to a 6.72%14 monthly PSIA return.

In our first simulation, we have assumed the account values to be constant at t = 0 value, meaning that there are no deposits or withdrawals in PSIA funds. Therefore we only need to simulate the trend in PSIA revenues. In 2011, participation banks have extended a total of TL 41.14 billion15 in loans. This will approximately refer to daily TL 40 million16 of loans. The revenue on these loans depends on the profit rate charged by the bank to the borrower. There is going to be a negative correlation between the rate and the loan amount. As the rate on the loan goes

Figure 1. Average PSIA returns for 4 Turkish participation banks.



down there will be more demand for loans and vice versa. Also the profit rate has to be set in accordance with market conditions. Islamic banks using their internal data can make analysis on this negative correlation and find out the relationship between the profit rate and the amount of loans they can extend in detail. However in our analysis we will use market interest rate as a reference point since participation banks are competing with conventional banks in attracting customers. Average monthly commercial loan rate for the period January 15th –February 15th 2012

in Turkey was 1.27%17. Thus, if the bank extends TL 40 million of loans at a rate of 1.27% everyday, PSIA revenues are increasing around TL 17,000 per day. Of course this is a rough estimation. Therefore in a way to provide the 95% confidence interval to PSIA returns curve given in Figure 1, we have constructed a range for the trend of PSIA returns. If mean and standard deviation of PSIA revenue change stays in the area given in Figure 2 Panel A, the bank can provide a profit rate quote that is within 95% confidence interval of the actual realized value. In order to demonstrate this, we

Figure 2. Estimating the trend in PSIA revenues: Panel A gives the area for mean-standard deviation combination, Panel B uses 4 sample points from Panel A to construct confidence interval for PSIA returns.

Calkan


have plotted confidence interval for the actual PSIA return for four different points in the area given.

The result for the first simulation represents the possibility of providing a profit rate quotation to the deposit holders in Islamic banks. As our analysis suggests the banks can do this with a level of flexibility in their revenue estimations. Here in our analysis we have used market interest rate and four banks’ average for other variables. Islamic banks on the other hand have access to historical data about individual loans with profit rates, loan values, payment information and so on. Using this extra information it is even possible for them to provide a better estimate for PSIA returns.

Second simulationIn the second simulation, we release our assumption of constant account value, in other words now there are deposits to and withdrawals from PSIA funds. In the first simulation account value was assumed constant at TL 60 million along the estimation period. Islamic banks in applying this framework can use their internal data about the fluctuations in the PSIA funds. However for our analysis we will use market growth rate as the trend. We have extracted the trend in TL deposits for participation banks for a period two months prior to our estimation period. Weekly deposits data is available at Central Bank’s website18. Deposits data is equivalent of unit account value in our framework. However we calculate unit account value endogenously using equation 1. Therefore we will use the trend in deposits as an approximation to trend in account value. Figure 3 displays the path followed by TL deposits of Participation Banks. As we have been doing all along, here again we have divided total TL deposits value by four to reach an average value for one of the four banks in the market.

Therefore our assumptions in this simulation are as follows:

At t = 0 At t >0Unit account

valueua0 : TL6 billion Endogenous

Unit value u0 : 100 EndogenousAccount value a0 : 60 million Linked to deposit

growthPSIA revenues R0 : TL 2 million Will be estimatedPSIA costs C0 : 15% of revenues ExogenousReserves y0 : 5% of revenues Exogenous

With the introduction of PSIA funds variation, mean and standard deviation combinations for revenues have been updated. Figure 4 Panel A displays the new range for revenues that makes it possible for the Islamic bank to provide a profit rate estimate within 95% confidence interval of the actual value. In Panel B, four different mean and standard deviation combinations from Panel A have been used to plot confidence intervals for the return estimate. Actual value lies within 95% confidence interval in each of them.

This second simulation demonstrates a more accurate estimate for realized values. We have incorporated variations in PSIA funds due to new deposits and withdrawals. Although we have used market growth rate as an approximation to PSIA variations, recognizing it as an exogenous variable, in reality banks have some level of control over their deposits. They are affected from market developments, but by charging a higher rate they can attract more deposits and they are always free to refuse new deposits. We have assumed in our analysis that they

Figure 3. Average TL deposits of participation banks.



do not have any control. Figure 5 demonstrates the effect of change in deposit assumption from one simulation to another. It plots Panel A of Figure 2 and 4 on top of each other. As can be seen, when we assume growing deposits, the bank needs to receive more credit return meaning they will provide more murabaha credits to the clients. This figure also shows that Islamic bank can use deposit growth as another policy tool. The changes in deposit growth will move the policy region up or down. Also, orange colored

area provides a more conservative estimate, which suits both growth assumptions.

In both of the analyses we have considered PSIA revenues as a policy tool. The bank can change the profit rate they distribute to deposit holders by increasing or decreasing the mark-up rate they charge on murabaha credits. Although the demand for credit depends on the rate they charge, there is still room for policy making. Therefore, using this

Figure 4. Estimating the trend in PSIA revenues: Panel A gives the area for mean-standard deviation combination, Panel B uses 4 sample points from Panel A to construct confidence interval for PSIA returns.

Calkan


tool, Islamic banks can also compete with each other and with conventional banks in attracting deposits, besides providing a profit rate quotation.

6. ConclusionIn this study we have studied the possibility of providing an ex-ante return rate quotations to deposit holders at Islamic banks. Simulation results suggest, it is possible to offer a reliable forecast within a 95% confidence interval of upcoming actual PSIA returns to the clients. The benefits of this study are three-fold. The most important benefit of providing this information is increasing Islamic banks’ competitive advantage compared to their conventional counterparties. Secondly, using these simulation techniques

IFIs can improve their fund management by controlling the amount of funds deposited to PSIAs and setting their mark-up rates according to their PSIA return targets. Furthermore these simulation methods can be used to better manage risks associated with asset liability management and rate of return.

Appendix: Detailed unit value calculationThe information below is extracted from Annex 1 regulation on the principles and procedures for accepting, withdrawal of deposits and participation funds as well as the prescribed deposits, participation funds custody and receivables in Turkish banking legislation.

Figure 5. Changes in revenues.

New UUA + R - (C + Y)

ADetails of Numerator

UA Unit Account Value

-

C PSIA Costs (a+b+c+d)R PSIA Revenues (a+b+c+d+e) a Special Provision Expenses

a Participation Share of Dividend Incomes b General Provision Expenses

a.1 Dividend Incomes Procured from LoansExtended Arising from PSIA c

Deposit Insurance Fund (DIF) PremiumExpenses

a.2 Profit Equivalent to Extended Fund Surplus d Precautionary Provision Expensesb Collections Made from Loans Cancelled

c Cancellations of Special Provisions YAmounts/ Reserves Allocated from Profitto be Distributed to PSIA

d Cancellations of General Provisions

eProvision Cancellations Set Aside from Profitsto be Distributed to PSIA

DenominatorA 0 Account Value



PSIA Revenues (a + b + c + d):

a) Participation Share of Dividend Incomes: Dividend amount equivalent to extended fund surplus is deducted from dividend incomes procured from extended loans arising from participation account. The amount found is separated on a currency type basis according to its weight in total participation accounts. The amount found by multiplying the separated amount by the ratio of account owner’s participation in profit defines the amount in dividend income falling to the share of participation accounts.1. Dividend Incomes Procured from Extended Loans

Arising from PSIA: This is the dividend income procured from funds extended arising from PSIA on currency basis. Whether or not delay funds collected for those not paid in their maturity among those funds or dividends deprived of as well as income from required reserves shall be taken into consideration as dividend income in the unit value calculation of PSIA shares are determined in PSIA contracts.

2. Profit equivalent to extended fund surplus: the amount found by multiplying by the ratio calculated by dividing the dividend income procured from loans extended arising from PSIA on a currency basis to the sum of funds extended, with extended fund surplus.

b) Collections Made from Loans Cancelled: The amount falling to the share of participation accounts, from the collections made concerning cancelled loans from loans extended arising from PSIA.

c) Cancellation of Special Provisions: The amount relating to PSIA, among cancelled amounts of special provisions set aside for loans arising from PSIA classified as non-performing loans pursuant to the regulation on principles and procedures for determination of qualifications of loans and other receivables by banks and provisions to be set aside.

d) Provision Cancellations Set Aside from Profits to be Distributed to PSIAs: It is the amount cancelled for meeting SDIF premium and special and general provisions of provisions monitored in amounts set aside from profits to be distributed to PSIAs.

PSIA Costs (a + b + c + d +e):

a) Special Provision Expenses: It is the part fall to the share of PSIA of general provisions set aside for PSIA emanated loans classified as NPL pursuant to the regulation on principles and procedures relating to for determination of qualifications of loans and other receivables by banks and provisions to be set aside.

b) General Provision Expenses: It is the part that falls to the share of participation accounts of general provisions set aside for PSIA emanated loans pursuant to the regulation on principles and procedures relating to determination of qualifications of loans and other receivables by banks and provisions to be set aside.

c) DIF (Deposit Insurance Fund) Premium Expenses: It is the part that falls to the share of DIF premium participation accounts.

d) Precautionary Provision Expenses: It is the amount of precautionary provision to be used in meeting the part fall to the share of DIF premium PSIA and special and

general provisions from the total amount of income items stated in (b), (c) and (d) sub paragraph of the PSIA revenues explanation. These provisions set aside are recorded to the account of amounts set aside from profit to be distributed to PSIAs included in communiqué on uniform chart of account and its explanation to be implemented by participation banks.

Amounts Allocated from Profit to be Distributed to PSIAs:

It is the provision amount allocated within the scope of the provision of the article 14(3) of the regulation on principles and procedures relating to determination of qualifications of loans and other receivables by banks and provisions to be set aside from profit amounts to be distributed to participation accounts by calculation date of unit values.

Notes 1. Wall Street and Financial Crisis: Anatomy of

Financial Collapse, report by US Senate Permanent Subcommittee on Investigations, 2011, Available at: http://www.hsgac.senate.gov//imo/media/doc/Financial_Cr is is/FinancialCr is isRepor t .pdf?attempt = 2.

2. Kayed and Hassan (2011), Al Mamun and Mia (2012). 3. Iran, Sudan and Pakistan. 4. Average market share of Islamic banks in MENA region

is 14%. 5. Some of the banks might be using this information in

their internal policy making at the moment. 6. AAOIFI Statement of Financial Accounting No. 2:

Concepts of Financial Accounting for Islamic Banking and Financial Institutions.

7. Turkish Banking Authority’s approach to PSIA return calculation is similar to return calculation in a fund.

8. See appendix for detailed calculation of unit value. 9. As majority of the loans in the Islamic banking

are provided with murabaha system, we will assume all loans given follows murabaha contract for the simplicity.

10. In Turkey, Islamic Financial Institutions are called participation banks.

11. Data from the website of Participation Banks Association of Turkey: www.tkbb.org.tr.

12. 24 billion/4 banks = 6 billion.13. 60 million = 6 billion/100.14. 0.03% × 360days = 9.60% total return; 70% × 9.60% =

6.72% depositors share.15. Data from the website of Participation Banks

Association of Turkey: www.tkbb.org.tr16. 41.14 billion/4 = 10.28 billion per bank; 10.28/250

(business days in a year)~40 million per day.17. From Turkish Central Bank Electronic Data Distri-

bution System.18. From Turkish Central Bank Electronic Data Distri-

bution System.

ReferencesAkacem M, Gilliam L. (2002) Principles of Islamic Banking:

Debt Versus Equity Financing. Middle East Policy. 9(1):124–138.

Calkan


Al-Mamun Md, Mia MAH. (2012) Origin of & Solution to Global Financial Meltdown: An Islamic View. International Journal of Business and Management. (7):12.

Archer S, Abdel Karim RA. (2009) Profit-Sharing Investment Accounts in Islamic Banks: Regulatory Problems and Possible Solutions. Journal of Banking Regulation. 10:300–306.

Archer S, Abdel Karim RA, Sundararajan V. (2010) Supervisory, Regulatory, and Capital Adequacy Implications of Profit-Sharing Investment accounts in Islamic Finance. Journal of Islamic Accounting and Business Research. 1(1):10–31.

Ariffin NM, Kassim SH. (2011) Risk Management Practices and Financial Performance of Islamic Banks: Malaysian Evidence. Paper presented at The 8th International Conference on Islamic Economics and Finance, Doha, December 19–21.

Atmeh MA, Ramadan AH. (2012) A Critique on Accounting for the Mudarabah Contract. Journal of Islamic Accounting and Business Research. 3(1):7–19.

Ayub M. (2007) Understanding Islamic Finance. John Wiley & Sons, Ltd. England.

Farook S, Hassan MK, Clinch G. (2012) Profit Distribution Management by Islamic Banks: An Empirical Investigation. The Quarterly Review of Economics and Finance. 52(3):333–347.

Ibrahim SH. (2007) IFRS vs. AAOIFI: The Clash of Standard? Munich Personal RePEc Archive (MPRA). No. 12539.

Kahf M. (2005) Basel II: Implications for Islamic Banking. Paper presented at the 6th International Conference on Islamic Economics and Banking, Jakarta, November 22–24.

Kayed RN, Hassan MK. (2011) The Global Financial Crisis and Islamic Finance. Thunderbird Int’l Bus Rev. 53:551–564.

Khan F. (2010) How ‘Islamic’ is Islamic banking? Journal of Economic Behavior & Organization. 76(3):805–20.

Khan T, Ahmed H. (2001) Risk Management – An Analysis of Issues in Islamic Financial Industry. Islamic Development Bank-Islamic Research and Training Institute. Occasional Paper No. 5. Jeddah.

Shubber K, Alzafiri E. (2008) Cost of Capital of Islamic Banking Institutions: An Empirical Study of a Special Case. International Journal of Islamic and Middle Eastern Finance and Management. 1(1):10–19.

Sultan SAM. (2006) An Overview of Accounting Standards for Islamic Financial Institutions. Finance Bulletin. Jan-Mac.

Sundararajan V. (2005) Risk Measurement and Disclosure in Islamic Finance and the Implication of Profit Sharing Investment Accounts. Paper presented at the 6th International Conference on Islamic Economics, Banking and Finance. Jakarta, November 22–24.

Taktak NB, Zouari SBS, Boudriga A. (2010) Do Islamic Banks Use Loan Loss Provisions to Smooth Their Results? Journal of Islamic Accounting and Business Research. 1(2):114–12.

Turkish Banking Legislation: Annex 1 to Regulation on the Principles and Procedures for Accepting, Withdrawal of Deposits and Participation Funds as Well as the Prescribed Deposits, Participation Funds Custody and Receivables. Available at: <http://www.bddk.org.tr/WebSitesi/english/Legislation/9656engdepositandparticipation_fundannex_08_06_2010.pdf>. Accessed: 10 May 2013.

Turkish Banking Legislation: Regulation on Procedures and Principles for Determination of Qualifications of Loans and other Receivables by Banks and Provisions to be Set Aside. Available at: <http://www.bddk.org.tr/WebSitesi/english/Legislation/8836eng_provisions_to_be_set_aside_13_06_2011.pdf>. Accessed: 10 May 2013.

Turkish Banking Legislation: Communiqué on Uniform Chart of Account and its Explanation to be Implem-ented by Participation Banks. Available at: <http://www.bddk.org.tr/WebSitesi/english/Legisla tion /8802eng_participationbanks_uniform_chart_ofaccounts_08_06_2011.pdf>. Accessed:10 May 2013.

Zoubi TA, Al-Khazali O. (2007) Empirical Testing of the Loss Provisions of Banks in the GCC region. Managerial Finance. 33(7):500–511.


Cite this chapter as: Nienhaus V (2015). Self-adjusting profit sharing ratios for Musharakah financing. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Self-adjusting profit sharing ratios for Musharakah financingVolker NienhausHonorary Professor, University of Bochum, ICMA Centre, University of Reading (United Kingdom), Tel. +49 201 8695750, [email protected]

Abstract - Banks avoid participatory financing due to serious information asymmetries, adverse selection and moral hazard problems resulting in negative impacts for the return on capital provided. Even financing instruments with a participatory legal form such as musharakah sukuk have been stripped of their risk sharing substance and become functional equivalents of interest-bearing bonds. Several authors have addressed these issues, but some proposals are applicable only for (listed) joint stock companies, while others imply Shariah compliance issues. To overcome these limitations, a “self-adjusting profit sharing ratio” is proposed, based on building blocks found in AAOIFI Shariah standards for musharakah financing and musharakah sukuk. These building blocks allow a (surprisingly) wide range of discretionary adjustments of participatory contracts, provided the contracting parties come to an agreement in re-negotiations of the contractual terms. This requires an agreement on a fair distribution of profits. What the parties consider a fair distribution is already known when the contract is initially concluded: It determines the parties’ profit shares based on their profit expectations at this point in time. The AAOIFI building blocks allow the structuring of a formula for the profit sharing ratio, which automatically adjusts to changes in the expected or actual profit. It thus ensures continuously a profit distribution in line with the initially agreed-upon principles of fairness. The formula can be calibrated such that the financing party gets under “normal” circumstances a return in line with a predetermined benchmark (e.g., the market rate of fixed term financings plus a risk mark-up) while the financed party has the advantages of an “insurance” against losses and unrestricted upside gains. Thus, financing instruments or sukuk with new risk/return profiles and some participatory elements could be structured so as to overcome the problems caused by information asymmetries in “pure” PLS financings.

Keywords: Islamic finance, musharakah, profit and loss sharing, information asymmetries

1. Discrepancies between theory and practice of Islamic financeIslamic economists consider finance based on profit and loss sharing (PLS) (participatory finance) as the genuine Islamic mode of finance and the major factor distinguishing Islamic from conventional finance. Indeed, an economic system where PLS is the dominant mode of financing would have different qualities with regard to efficiency, stability and distribution compared to a conventional interest- and debt-based system. However, the practice of Islamic finance (which was factually Islamic banking until the early 2000s) was and is very different from this ideal model.

• PLS financing hardly ever takes place in Islamic banks. Bad experiences of pioneering banks and theoretical explanations of adverse selection and

moral hazard issues in PLS financing (where the ratio of profit sharing is fixed in advance and not changed afterwards) can explain the abstinence from participatory modes of financing.

• Banks apply the PLS principle only in the deposit business, i.e., in contracts of mudarabah-based investment accounts, but even there the practice was quite different from the model: Fluctuations of investment returns were not passed on to the account holders but rather smoothed out by recourse to reserves (investment risk reserves and profit equalization reserves) and, in worst cases, by interest-free loans or even “voluntary” gifts of the shareholders to the account holders. This was done to avoid massive withdrawals by disappointed investment account holders which would have

Nienhaus


created serious problems for the bank. With ex ante announced “anticipated” returns which were typically congruent with the realized ex post returns, and with the coverage of investment accounts by capital protecting deposit insurance schemes in some jurisdictions, mudarabah-based investment accounts got the “look and feel” of conventional interest-bearing deposits.

• With the growing popularity of sukuk in the 2000s, a new option for PLS financing emerged in the Islamic capital market: musharakah sukuk. But again, the practice converted this participatory instrument into an Islamic bond with factually fixed returns for the sukuk holders. The applied techniques are explained in detail later in this paper.

Unfortunately, Islamic bankers and Shariah scholars never shared the enthusiasm of Islamic economists for PLS finance in practice. It is not that Shariah scholars did not allow PLS arrangements: Mudarabah and musharakah contracts (and modern derivatives thereof) are explicitly approved as Shariah compliant. However, the approval was done in a way which opened the door widely for Islamic bankers to convert the participatory concepts into close functional equivalents of conventional interest-based and factually risk-free modes of financing. In practice, Shariah scholars have approved the conversion of “equity-based” sukuk (participatory instruments with a variable return) into “Islamic bonds” (debt instruments with fixed costs).

The growth of Islamic finance over the last decade was driven by the inroad of Western financial institutions into this new and seemingly lucrative segment of the global financial industry. Many Islamic financial institutions are run by CEOs and management teams who have been socialized in conventional finance before they converted to Islamic finance. In addition, many executives and staff of Islamic financial institutions were hired from the conventional sector. Individuals and teams who were successful in conventional finance before they joined the Islamic finance industry were familiar with strategies and instruments for a good or even outstanding performance of their financial institution. This was probably the reason why they were lured away for their previous employment, and shareholders expect a continuation of such a management performance in the new Islamic environment. Thus, it should not come as a surprise that Islamic bankers who were socialized in the conventional system tried to replicate those strategies and instruments which they had applied successfully in their previous position in conventional finance.

The Islamic economics literature was often too theoretical, abstract, macro-oriented or prescriptive to be of much use for practitioners who were looking for more effective instruments in commercial, for-profit financial institutions. The Islamic economists have not been able to convey their enthusiasm for PLS instruments to Islamic bankers – and probably also not to Shariah scholars. While participatory finance is often seen by Islamic economists as the core of Islamic finance—as sale- and rent-based modes of finance are mentioned only on the sidelines in their models—it is exactly the opposite from the perspective of Shariah scholars: sale- and rent-based contracts have been elaborated over centuries in extensive detail, while financing based on profit and loss sharing were sidelined.

Islamic economists have recognized that the practice of Islamic finance deviates substantially from the ideal, but they did not simply criticize practitioners for this unfortunate development. They analyzed the reasons for the observable discrepancy (mainly agency problems in anonymous markets), and they came forward with a number of proposals suggesting how to solve the identified problems and promote participatory finance.

• The second chapter summarizes and comments on a number of such contributions. Looking at the rapid growth of musharakah sukuk in the 2000s, it seemed that at least the Islamic capital market had overcome the agency problems and moved toward the PLS ideal. Unfortunately, this was not the case.

• The third chapter takes a closer look at the structuring of these sukuk and explains how equity-based instruments could be converted into functional equivalents of debt instruments with predetermined costs and capital guarantees. After a critique of prevailing practices by a prominent Shariah scholar in 2007, and a resolution of the Accounting and Auditing Organisation of Islamic Financial Institutions (AAOIFI) in 2008, the issuance of musharakah sukuk dropped sharply from then until today. Despite the recovery of the sukuk market in recent years, the once dominant form of sukuk has become marginal by today. This is a deplorable signal because mudarabah and musharakah sukuk were the only financing instruments of significant quantitative weight in the Islamic finance markets which upheld at least the PLS form.

• The fourth chapter outlines the mechanics of a musharakah sukuk concept, which is based on the PLS principle but uses building blocks found in the Shariah standards of AAOIFI to overcome agency problems. The concept allows the structuring of a financial instrument that brings differing commercial interests of contracting parties in balance and keeps this balance by a self-adjusting profit sharing ratio. The automated ratio adjustment takes place whenever new information on the expected or actual performance (=profit) of the financed venture becomes available.

2. The agency problem of participatory finance in anonymous marketsParticipatory finance is a generic term for different forms of financing on the basis of profit and loss sharing, for example mudarabah and musharakah bank financing or mudarabah and musharakah sukuk.

Bacha (1997) sees mudarabah financing as a hybrid of elements of debt and equity financing. Somehow the agency problems of both are combined in this hybrid:

• The equity agency problem is that the mudarib has a strong incentive to “produce” costs which accrue to him as benefits. This goes on as long as the marginal utility from fringe benefits or perks exceeds the corresponding reduction in the mudarib’s share of the profit. In addition, since mudarabah financing is for a specified project of an existing firm, the firm may have various possibilities to shift overhead and


Self-adjusting profit sharing ratios for Musharakah financing

other costs to the mudarabah project, thus reducing the profits which have to be shared with the bank (without reducing the overall profit of the firm).

• The debt agency problem is one of moral hazard: Debt (or more general: external capital at costs lower than the return on total capital) can leverage the return of equity substantially (even if a certain percentage of the profit goes to the provider of the external funds). External mudarabah funds have not only a factual but even a contractual loss absorbing quality. Consequently, a high leverage by mudarabah funding would create incentives for a firm to embark on projects with high profit potentials and high risks because the upside chances surpass the downside risk for the equity holders.

Bacha concludes that mudarabah financing has more agency problems than a pure debt or a pure equity financing and therefore is an inferior option for the capital providing party. To overcome the agency problems of mudarabah in its genuine form, Bacha proposed the introduction of “equity kickers:” specified events or outcomes to trigger an equity-related provision in a financing contract. The equity kicker in a mudarabah contract would be a clause “whereby in the event of losses in the Mudarabah financed project, the Rab-Ul-Mal absorbs the losses but is ‘reimbursed’ for the amount of losses thru issuance of new equity by the Mudarib to him.” (Bacha 1997, 18). By this clause the mudarib transfers benefits to the rab al-mal so that in the end the rab al-mal’s will be shielded against “avoidable” losses (caused by profit reducing management practices or excessive risk taking), and his risk will decrease. For the mudarib the opportunity costs of profit compression or high risk ventures would increase, and this should solve or at least mitigate the agency problems.

A major drawback of the equity kickers approach is that it can be applied only for the mudarabah financing of joint stock companies. Start-ups and small and medium sized enterprises (SMEs) usually do not have this legal form. But by far the largest number of enterprises are SMEs, which play a prominent role for employment, income generation and poverty reduction in many Muslim (and non-Muslim) countries.

Another problem is the Shariah compliance of the equity kickers. Shariah principles prohibit a protection of the rab al-mal’s capital through a guarantee by the mudarib. Bacha claims that this is not the case because the equity kickers are no guarantee against losses. The rab al-mal “will make losses if the project makes losses – although it will be much less than under existing Mudarabah.” Ignoring the problem of a gharar-free determination of the value of the transferred equity in cases of loss, the fact remains that the equity has some value and thus is a partial compensation for the loss. This could be seen as equivalent to a partial guarantee of mudarabah capital by the mudarib – which would be a violation of a Shariah stipulation. But Bacha himself sees another violation of a Shariah requirement: “The one Shariah requirement that would not be met by the proposed arrangement is the requirement that in Mudarabah, the financier should absorb all the losses. Any proposal that seeks to overcome the problems of existing Mudarabah would invariably come up against this injunction.” This injunction is repeated over and again

in all AAOIFI standards dealing with mudarabah and musharakah, and later the IFSB took the same position. A protection against capital losses can only be given by an independent third party, not by the mudarib.

Ahmed (2002) also addresses agency problems. He proposes a contractual arrangement that deals with the moral hazard problem arising from an underreporting of profits. It shall be overcome by an incentive-compatible contractual arrangement. The financing bank will have the right to undertake a costly audit if the profit reported by the financed firm falls short of the expected profits (on which the parties have to agree). The auditing expenses will be shared ex ante by both parties and become a co-determining factor for the calculation of the profit sharing ratio. But they materialize only if the bank has reason to undertake the audit. If the audit shows that the firm has reported correctly, the bank will refund the firm’s share of the audit costs (as a reward for honesty). If the audit uncovers false reporting, i.e., if it becomes apparent that the actual profit is higher than the reported profit, the firm has to bear the full auditing costs, calculate the bank’s profits share on the basis of the actual profit and pay a significant fine as a (additional) penalty for the false reporting. This contractual arrangement shall reduce (if not eliminate) the moral hazard incentives for the financed firm.

By resolving the firm’s incentive to swindle, the model removes only one of several obstacles that deter Islamic financial institutions to engage more in “true” PLS financings. If the bank has fixed the profit sharing ratio, and the correctly reported profits fall short of the expectations, the bank may not receive the benchmark of, for example, murabaha or ijarah financings with low risk and predetermined returns (especially not if the bank insisted on an audit and has to bear the full auditing costs). This, however, may be to the advantage of the firm because the PLS financing costs would be less than the costs of sale/rent-based modes of financing with low risk and predetermined returns for the bank. This, by the way, creates another problem – a kind of an ex ante replacement of the discouraged ex post moral hazard: the higher the profit expectations of the bank, the lower the profit sharing ratio necessary to meet a given benchmark. If the firm is able to present its project in such a way that the bank becomes more optimistic about the future profit than the firm itself, and if the firm’s expectations are correct, then the results for the firm are similar to those of false reporting – but without any risk of penalties.

Hasan (2002) presented a model for the determination of the equilibrium profit share for a bank (as provider of mudarabah capital to a firm) in a mixed system (= where the firm has the option of an interest-based loan financing) under alternative constellations of the externally-set market rates of interest (which is a kind of benchmark), a risk premium (to be paid in case of loan financing), the so-called leverage ratio (= share of externally provided capital on interest or mudarabah base in the total capital of the firm), and the total return on capital of the firm. The equilibrium profit sharing ratio of the bank will be less than its loss sharing ratio, and it will vary inversely with the return on total capital and directly with the leverage ratio. The model clarified some interdependencies between variables, but it did not explain (without

Nienhaus


additional exogenous hypotheses) why the use of mud-arabah financing is so low or what could be done to increase its use. Therefore Hasan supplemented his model with contemplations on a change of Muslims’ attitudes towards very high-risk aversion in business matters (risk understood as uncertainty, which cannot be measured and insured). Since banks are no exceptions to this behavioral pattern, they are very reluctant to finance risky ventures in which both the earnings and the repayment of the provided capital are uncertain. To improve the situation, Hasan makes two proposals:

• The contracting parties should agree on a clause in the mudarabah contract “to treat the bank among the preferential creditors of the borrowing firm in case of insolvency …

• … alternatively, the bank can be allotted redeemable preference shares for the money advanced.” (Hasan 2002, 49–50).

It is very questionable whether Shariah could accommodate these provisions. For example, AAOIFI had dealt with the Shariah rules and requirements in the Financial Accounting Standard No. 3 on mudarabah financing and No. 4 on musharakah financing, both adopted in 1996. Neither the treatment of the bank as a preferential creditor nor the allotment of redeemable preference shares, as a kind of security, are compatible with the restrictive Shariah principles spelled out by AAOIFI. AAOIFI explicated the principles in more detail again in the Shariah Standards No. 12 on musharakah and No. 13 on mudarabah, both adopted in 2002.

Diaw, Bacha and Lahsasna (2012) address agency problems of participatory finance not in banking but for musharakah sukuk. They diagnose a fundamental incongruence between (formal) requirements of Shariah contracts and they aim “to reproduce the substance of a financial instrument that is repugnant to their nature and to the Islamic paradigm in finance” (p. 45). The proposed solution takes inspirations from convertible bonds and develops the idea of equity kickers (Bacha 1997) further. Basic agency problems arise in mudarabah or musharakah arrangements (bank financings or sukuk) because the financing contract covers only a limited period of time. It is in the interest of the mudarib or the managing party of a musharakah and the owners of the financed firm to keep as much of the realised profits within the firm during the period of the participatory financing. A wide range of possible measures to compress the reported profit that has to be shared with the capital provider – from cost allocation to false reporting – has already been indicated. These techniques reduce the value (payouts) of sukuk and increase the value of the firm. This is a problem for the capital provider if he cannot participate in the increase of the value of the firm. A solution would be a conversion clause (which is like an embedded option in the financing contract): If the sukuk performance falls short of a benchmark (for example, if the return on the sukuk is less than a stipulated minimum), the sukuk holders have the right (but not the obligation) to convert their sukuk into shares of the financed firm. The exercise of the option would not only prolong the initially temporary partnership to an indefinite period but would also grant ownership rights to the previously silent partner.

A weakness of this concept that it is applicable only for joint stock companies and it requires the readiness of the actual shareholders to accept a certain dilution of their ownership rights by the issuance of new or the transfer of existing stocks to new shareholders. This cannot be taken for granted, for example, for family-run businesses.

Further, one may challenge the fairness or balance of a conversion clause for cases where the underperformance is not due to profit skimming of the financed firm but to market forces which lead to an unexpectedly poor performance of the financed project. In such a situation the conversion option may be felt as an undue and very severe penalty for something that was beyond the control of the management and not caused by its “misbehavior.” The penalty is severe because the conversion gives the financing party a perpetual claim on parts of all future profits, which may exceed the amount of the initially provided capital by a multiple. While this may be very attractive for the capital provider, it is far from obvious that it would also be acceptable for the financed firm, especially if the murabahah or musharakah was only for a short to medium term.

Finally, a very similar result for overcoming the moral hazard issues, but without the requirement of a particular legal form of the financed firm (joint stock company) and the “penalty character” of the conversion clause (under adverse market conditions), could be achieved by longer term or even a perpetual sukuk with a redemption clause. The redemption clause would be a promise (wa’d) of the issuer to redeem sukuk certificates at their fair value on the demand of the sukuk holders. An increase of the value of the firm would be reflected in the fair value of the certificates.

It has to be mentioned here that Diaw, Bacha and Lahsasna (2012) offer another approach for the analysis of equity-based sukuk which is independent from the equity kickers proposal. It is based on the idea of variable profit sharing ratios. A few remarks on differences between their analysis and the model outlined below can be found at the end of this paper.

3. The lack of participatory financing in banking and capital marketWhile the agency problems of participatory modes of finance can hardly be ignored, it seems that none of the solutions developed in the academic literature have been applied in practice. While mudarabah financing is virtually non-existent in banking, musharakah sukuk expanded at an extraordinary rate over the first half of the 2000s and became the most widely used form of sukuk by 2007 (see chart below).

Profit and loss sharing in bankingThe virtual non-existence of participatory (= mudarabah or musharakah) bank financing may, on the one hand, be explained by the aforementioned agency problems and an adverse selection problem that is summarized in the box below (The “lemon problem” in mudarabah financing). These issues in their totality have not been solved in theoretical models.



On the other hand, Hasan’s reference to an exaggerated risk aversion in Muslim countries points to another issue in Islamic banking: The most widely used Shariah compliant alternative for interest-bearing savings and term accounts are mudarabah-based unrestricted investment accounts. Conceptually, losses from the investment of the investment account holders’ funds should be passed on to the investment accounts. The investment account holders are most probably risk averse, although they have signed a profit sharing and risk bearing contract: Muslims who were looking for an alternative for conventional savings and term deposits which could give them a Shariah compliant return have hardly an alternative to a mudarabah-based contract, which exposes their funds to a risk of loss. But to accept such a contract because no Shariah compliant alternative is available, does not imply that the account holders would ever want this risk to materialize. Most probably they expect from the bank that all conceivable forms of risk mitigation and risk avoidance are applied. Some Islamic bankers articulate such expectations very forcefully. They defend their banks’ policies of virtually risk free mark-up techniques only and the abandonment of mudarabah or musharakah in the financing business with their fiduciary duties towards the risk-averse investment account holders.

Mudarabah and Musharakah SukukIn view of the serious agency problems of participatory financing, the boom of musharakah sukuk is much more surprising than the lack of mudarabah or musharakah in bank financing. However, a closer look at the practice of

the issuance of mudarabah and musharakah sukuk offers a simple explanation: In spite of their participatory form and their classification as “equity-based” sukuk, most of these sukuk have never had a participatory substance. Hence, they did not suffer from the agency problems discussed in the academic literature. Instead, most of the musharakah sukuk were intentionally structured as functional equivalents of conventional bonds, i.e., as debt instruments with predetermined returns. On the other hand, “equity-based” sukuk are very flexible and allow (in contrast to most other types of sukuk such as murabahah or ijarah sukuk) the issuing of a security, which is not tied to the true or beneficial ownership of an existing specific tangible asset. Instead, mudarabah and musharakah sukuk create joint ventures for the investment of the sukuk capital in profit-generating Shariah approved assets, but these assets must not yet exist when the joint venture is formed. The assets can be created by the employment of the sukuk resources (i.e., the money paid by the sukuk subscribers), and the composition of the assets held by the joint venture can change over the life of the sukuk.

This flexibility regarding the underlying asset was the main attraction for practitioners and can explain the rapid growth of musharakah sukuk. Their popularity was not due to their equity structure which, in theory, brought them closer to the Islamic economists’ ideal of participatory finance. On the contrary, the equity elements, in particular the possible volatility of returns and the downside risk of a capital loss, were somewhat disturbing and have effectively been removed by contractual engineering.

The “lemon problem” in mudarabah financing

An entrepreneur can always compare the costs of Shariah-compliant funding based on the expected profit and a negotiated profit-sharing ratio on the one hand, and the costs of a fixed mark-up sale/rent financing for his project on the other hand. Less religious-minded entrepreneurs could also add riba-based financing alternatives offered by conventional banks. The mark-up is determined by competition (within the Islamic banking sector and/or between Islamic and conventional banks) and becomes the benchmark rate to which the profit sharing ratio has to be adjusted for a given expected profit. Roughly-speaking, both for the bank and the entrepreneur the mark-up multiplied by the amount of financing should equal the expected profit multiplied by the respective profit-sharing ratio. If the entrepreneur were able to convince the bank to become optimistic about his project and to expect a profit that is higher than the profit he himself expects realistically (without communicating this to the bank), then the bank would agree on a profit sharing ratio, which is too low (compared with the benchmark or the ratio based on a “realistic” profit expectation). In this setting, the entrepreneur would benefit from profit sharing financing. The bank has to take into account that in principle all customers who ask for a mudarabah financing have a strong incentive to present overly optimistic profit projections. The bank could protect itself to some degree against wrong profit projections by a thorough evaluation of business plans. But that requires human resources with a profound knowledge of the markets of their customers, and the experts of the bank should, on average, be better than the entrepreneurs themselves in predicting financial outcomes of business plans. Expert staff with such qualifications is hard to find, very expensive and probably even harder to retain (because these employees have all the qualities to become entrepreneurs by themselves). Therefore, the bank may take recourse to a less expensive protective mechanism, namely a simple “safety margin” on all profit-sharing ratios. But this will be anticipated by the entrepreneurs. If an entrepreneur presents a realistic profit projection, the “safety margin” on the bank’s profit sharing ratio will make the mudarabah financing more expensive for him than the mark-up financing. Entrepreneurs with good projects may not like to enter into the troubles of debating with the bank the credibility of their profit projections (in order to eliminate the safety margin). Instead, they prefer fixed-cost financing from the outset. In contrast, for entrepreneurs with weak projects, the mark-up financing may be too expensive, and they have a strong incentive to present an acceptable profit projection to the bank. In the end, the bank will have more weak than strong projects in its mudarabah portfolio, and it is highly probable that a number of weak projects will go bust so that the realised profit will fall short of the expectations. To avoid this, it is probably the best option not to enter into participatory financing at all. The market for mudarabah and musharakah financing will collapse (or never emerge).

Nienhaus


Debt character of equity-based Sukuk: Capital guaranteesThe debt character of equity-based sukuk was achieved by a (binding) promise of the obligor to repurchase the sukuk certificates at maturity (or in the event of a default) at their issuing price, respectively their face value. This eliminates contractually the risk of a capital loss of the rab al-mal, and it is a functional equivalent to the guarantee of the capital of one party by the other. Effectively, losses are not borne in proportion to the capital contributed to the venture, or even not borne at all by the capital providers. An arrangement with such a consequence can hardly meet the Shariah principle that only risk justifies return. The principle of loss sharing or loss bearing has been stated over and again—from classic legal manuals to contemporary AAOIFI standards (for example, Financial Accounting Standard No. 4 on musharakah financing, adopted in 1996, Shariah Standard No. 5 on guarantees, adopted in 2001, No. 12 on sharika (musharakah), adopted in 2002, and No. 17 on investment sukuk, adopted in 2003).

There is but one escape from the rule of loss bearing by the capital provider, namely a voluntary guarantee by an independent third party. The case study of the 2005 IDB sukuk by Mokhtar (2011, pp. 34–35) reveals that the criteria for the definition of an “independent third party” can be very formalistic and limp in economic substance. IDB had set up for its sukuk issuances the IDB Trust Services Limited in Jersey, a SPV with an authorized share capital of £10,000 and an issued share capital of £2. All the assets underlying the sukuk issuances were transferred from IDB to its SPV, and the prospectus of the 2005 sukuk advertised the fact that IDB was the unconditional and irrevocable guarantor of the sukuk issuance. The Shariah Board of IDB declared: “As it [IDB] is not the issuer of the Trust Certificates and is not a manager or participant, IDB can enter into contractual obligations which have the effect of guaranteeing the Aggregate Nominal Amount of the Trust Certificates and any Periodic Distribution Amounts in respect of the Trust Certificates.” Unfortunately, the Shariah resolution does not disclose whether the Shariah Board considered his decision in accordance with AAOIFI standards, and if not, how it would justify its different position.

Predetermined returns for equity-based SukukFor achieving predetermined returns, sukuk engineers used at least three different techniques:

• When actual profits fall short of the expected profits, sukuk managers often provided interest-free loans (which should be recovered later) in order to meet the expectations of the sukuk holders and to beef-up the payouts to them. AAOIFI made it clear that this practice is not Shariah compliant: “It is not permissible for the Manager of Sukuk, whether the manager acts as Mudarib (investment manager), or Sharik (partner), or Wakil (agent) for investment, to undertake to offer loans to Sukuk holders, when actual earnings fall short of expected earnings. It is permissible, however, to establish a reserve account for the purpose of covering such shortfalls to the extent possible, provided the same is mentioned in the prospectus.”

• Another technique that violates Shariah principles was applied in some musharakah sukuk with very

special purchase undertakings (PUs), which did not only guarantee the face value of the certificate at maturity, but factually also the expected profit. These PUs were not only triggered by the maturity of the sukuk but also when the venture was not performing well and the obligor failed to pay the expected profit. The exercise of the PU obliged the obligor to pay the outstanding principal plus any so far accrued but unpaid profit. Mokhtar (2011, p. 33) points out that “accrued profit is not necessarily actual profit earned. Profit accrued is the expected profit that is earned by the investors as time passes by.” Given that the prospectus indicated an expected profit accrued over the life of the sukuk, then the early redemption clause for the PU was effectively a guarantee of a (minimum) predetermined return. This converts the substance of an equity certificate into the equivalent of a conventional bond. It is obvious that this very special form of a “face value plus accrued profit PU” violates Shariah principles even more than a “plain face-value PU” at maturity.

• A more widely used technique for the conversion of a participatory instrument into a close equivalent of a bond with (nearly) predetermined returns for the sukuk holders in “good” years (where the mudarabah or musharakah generates a return that meets or surpasses an articulated profit expectation (=benchmark) were “incentive fees”: The profit share of the sukuk holders is set, for example, at 99%. Should the actual profit fall short of the expectation (benchmark), (nearly) all of the profit goes to the sukuk holders. But as soon as the actual profit exceeds the expected profit (which should be the normal situation), the amount of the actual profit that exceeds the benchmark is given to the sukuk manager as an incentive fee for “good management.” Suppose that the agreed upon expected profit is calculated on the basis of LIBOR plus a risk factor as the benchmark. Then this arrangement implies that sukuk holders will receive under “normal circumstances” the equivalent of the risk-adjusted market rate of interest. It may come as a surprise, but such a technique is in harmony with AAOIFI standards and thus should be considered Shariah compliant.

In November 2007 the chairman of AAOIFI’s Shariah Board criticized the prevailing practices that changed the substance of musharakah sukuk from an equity-based instrument (as it was conceived) into a functional equivalent of an interest-bearing bond, and in 2008 AAOIFI issued a resolution on sukuk. This resolution is a pointed summary of what was already contained in the AAOIFI standards.

Building blocks of “AAOIFI compliant” SukukHowever, AAOIFI did not summarize in this resolution a number of remarkable provisions in AAOIFI standards, which could be used as “building blocks” for the structuring of more “AAOIFI compliant” sukuk:

• The profit sharing formula has to be fixed at the beginning of a mudarabah or a musharakah: “It is a requirement that the mechanism for distributing profit must be clearly known in a manner that eliminates uncertainty and any possibility of dispute.”



• The profit distribution does not have to be based on a profit sharing ratio which is a simple percentage. The ratio can also be based on a more complex formula: “It is permissible for the partners to agree on the adoption of any method of allocation of profit, either permanent or variable, for example, by agreeing that the percentages of profit shares in the first period are one set of percentages and in the second period are another set of percentages, depending on the disparity of the two periods or the magnitude of the realised profit. This is allowed provided that using such a method does not lead to the likelihood of a partner being precluded from participation in profit.”

• Irrespective of the complexity of the initially accepted formula, it is “permissible for the parties to change the ratio of distribution of profit at any time and to define the duration for which the agreement will remain valid.”

• The changing of the distribution scheme can even be made when the actual profit of the venture is known at the end of the life of the sukuk: “The parties may bilaterally agree to amend the percentages of profit-sharing on the date of distribution. Also, a partner may relinquish, on the date of distribution, a part of the profit that is due to him in favor of another party.”

• “It is permissible for the issuer or the certificate holders to adopt permissible methods of managing risk, of mitigating fluctuation of distributable profits (profit equalisation reserve), such as establishing an Islamic insurance fund with contributions of certificate holders, or by participating in Insurance (Takaful) by payment of premiums from the income of the shares of Sukuk holders or through donations (tabarru’at) made by the Sukuk holders.”

The changing Sukuk landscapeInstead of using these building blocks to modify sukuk in such a way that they meet the 2008 AAOIFI Shariah

pronouncement on musharakah sukuk, the practice moved away from equity-based sukuk and switched to ijarah and murabaha sukuk since 2008.

The trends captured in the graph until 2009 continued: In 2012, the share of mudarabah and musharakah sukuk in global sukuk issuances (in US$) had declined to 15% while murabaha and similar sale-based sukuk accounted for 65% and ijarah sukuk for 16%.

The growing popularity of sale- and rent-based sukuk could be explained as follows:

• Murabaha sukuk are based on short term debt creating sale transactions, and a sufficiently large portfolio of such transactions should provide effectively a built-in protection (albeit not a formal guarantee) for the capital invested by the sukuk holders and accrued profit shares.

• Ijarah sukuk have a longer maturity and a significantly higher market risk that could lead to losses. However, AAOIFI allows for this type of sukuk a straightforward guarantee of the capital by a purchase undertaking of the issuer at face value.

Another reason for the departure from musharakah sukuk may be that the aforementioned options for more flexible profit distribution rules and re-negotiations were not practicable in the business environment in which sukuk flourished, namely the market of institutional investors and the interbank market. The market participants are mostly financial institutions which are hardly interested in participatory components in their contractual arrangements because that would imply an additional rate of return risk. Further, if predetermined returns are the objective, it may be less complicated and less risky to structure sale- or lease-based deals than to insert more complex profit distribution formulas into musharakah contracts. Finally, a readjustment

Figure 1. Trend in types of Sukuk issued.Source: Mokhtar 2011, p. 5

Nienhaus


of the profit sharing ratios at the day of distribution, i.e., when the distributable profit is known, is commercially equivalent to the fixing of the distribution of the profits in absolute amounts.

More important, a profit sharing ratio agreed upon at the beginning of the contract will generate one allocation of amounts (in absolute figures) on the distribution day, and a revised ratio will generate a different allocation of amounts. To accept the revised ratio implies a definite gain for one party and a definite loss for the other. It is extremely unlikely that banks or institutional investors would voluntarily give up profits which are legally due to them in favour of another bank or institutional investor. Against this background it is not surprising that the additional building blocks for mudarabah and musharakah sukuk were not utilized in the typical market environment for this kind of sukuk.

But the situation could be different under a different market setting and for institutions with somewhat different objectives. For example, “Modarabah Companies” were established in the 1980s in Pakistan. They can have a similar function as an SPV (set up, for example, by a financial holding company or a bank): As non-banks, they could collect funds through the issuance of mudarabah sukuk and invest these funds in profitable projects, for example in the leasing of equipment to manufacturing firms on the basis of ijarah contracts or in seed or growth financing on a musharakah basis. “Initially, there was a desire for the mudarabah sector to concentrate on funding SMEs in Pakistan, which were at times neglected by the banks. Unfortunately, due in part to the profit motive, it has steered more toward medium-sized enterprises and big-ticket funding, an area where banks are already involved. Rather than competing with banks, it may prove a competitive advantage to focus on smaller enterprises in a microfinance manner, as was originally envisioned.” (Khwaja 2009, 245).

Non-bank finance companies could utilize for their own funding sukuk structures. Their sukuk issuances will have similarities but also marked differences to mudarabah and musharakah sukuk of investment banks and actors in the interbank market.

• Finance companies which do not take deposits but issue sukuk would not need a full banking license in most jurisdictions, and they would not be subject to the strict Basel III capital adequacy and liquidity rules for banks.

• Their sukuk could be subscribed by institutional investors, but it is also possible to envisage a retail-oriented marketing. Malaysia has recently taken steps towards the creation of a retail sukuk market.

• Institutional investors have already shown their strong preference for predetermined returns. On the other hand, it seems reasonable to assume that also retail investors – similar to investment account holders – would be risk averse even if they purchase equity-type securities.

• The profit generating projects of sukuk-funded finance companies will be smaller than the big ticket transactions in the actual sukuk market. Small and medium sized enterprises (SME) could become a particular target group for the finance companies. On the one hand, SMEs may be still underserved

by commercial banks in many Muslim countries, in particular in those countries where the transformation of the economic system is on the political agenda. On the other hand, SMEs are usually individual or family run enterprises which do not have the legal form of a joint stock company and cannot get funding from the floating of shares.

• Two phases in the life of a SME can be of particular interest to sukuk-funded finance companies: the start-up phase and the expansion phase (after a SME has become established and grows in its market) with different risk/return profiles for funds provided. Both could be quite attractive for finance companies compared with financings in areas where stiff competition (by banks) has compressed margins. However, sufficient expertise in the industries of the financed SMEs is required for the assessment of the business plans for start-up or expansion.

• The finance companies could provide sale/rent-based financings (murabaha, istisna’, ijarah) to their target group, but they could also apply equity-based types of financing, for example mudarabah for start-ups, musharakah for growth financing. Participatory modes of finance could generate higher returns, but they are also associated with higher risks: general business risks, but also the risks resulting from the various agency problems outlined above.

Where SMEs are not joint stock companies, the agency problems in financing cannot be solved by those proposals which are based on a conversion of temporary external participatory capital to permanent equity by providing stocks to mudarabah or musharakah investors. Some SMEs may have the legal form of joint stock companies, but if they are owned by individuals or families who do not want to dilute the ownership structures through the issuance of new (or transfer of existing) stocks to other parties, they will hardly opt for finance contracts with equity kickers or similar conversion clauses.

4. A Musharakah Sukuk with a self-adjusting profit sharing ratioSince AAOIOFI has provided an interesting but underutilized toolbox for flexible mudarabah and musharakah sukuk structures, it is possible to reconcile the main interests of sukuk holders and financed entrepreneurs, namely:

• the interest of risk-averse sukuk holders in risk-mitigated predictable and stable returns,

• the interest of entrepreneurs• to get some financial relief in “bad times,” i.e.,

to have some risk-sharing elements in financing arrangements in a year of unexpectedly poor performance

• to have shared but uncapped upside potentials in “goods times,” i.e., to avoid a total skimming-off of profits in years of above-average performance

The basic idea of the musharakah sukuk with a self-adjusting profit sharing ratio is quite simple: If it is permissible to adjust the profit-sharing ratio at any time during the life of a mudarabah or musharakah sukuk by consent of the contracting parties, this must not be done in a discretionary manner (by separate negotiations after new



performance information become available). Instead, this can be automated by a simple formula which links the profit sharing ratio to the actual performance (in the simplest case to the actual profit) of the financed project or venture.

To calculate performance-dependent profit sharing ratios for a musharakah structure between an SPV and an enterprise, it must be known how much capital is provided externally by the SPV (respectively the sukuk holders) and how much internally by the enterprise (originator). At the beginning of the joint venture, a certain total profit is expected which has to be distributed to the external and the internal capital provider. The profit sharing ratio denotes the share of profit allocated to the external capital provider. This profit sharing ratio will not be fixed numerically directly but computed by a formula on which the parties have agreed. This formula is based on a distribution pattern for the profits about which the contracting parties have achieved a consensus. It could be, for example, a fixed relation between the rate of return for the external and for the internal capital, or it could tie one of the rates of return to an external benchmark. The distribution pattern, i.e., a particular relation between the rates of return or the link of one rate to a benchmark, has to be determined at the beginning of the joint venture, and this pattern (= distribution formula) should not be modified afterwards.

A model in which only ratios are agreed upon does not fix ex ante the absolute value (in currency units) of the profit for the contracting partners that would violate Shariah principles. It only determines the relative positions of the parties. The absolute volumes depend on the realized profits. To keep the relative positions of the parties in the agreed-upon proportion, it is necessary that the profit sharing ratio is adjusted automatically whenever new information on the expected or actual profit of the joint venture become available. This is automatically achieved by a simple formula. For the computation of the adjustable profit sharing ratios, the following variables are used:

Ke = external capital (provided by the SPV)Ki = internal capital (provided by the entrepreneur)K = Ke + Ki = total capitalP = total distributable profitPe = s ∙ P = profit allocated to the external capitalPi = (1 - s) P = profit allocated to the internal capitals = Pe/P = profit sharing ratio = share of total profits

allocated to the external capitalb = benchmark rate of returnre = Pe/Ke = rate of return for external capitalri = Pi/Ki = rate of return for internal capitalr = p/K = rate of return on total capital

Next an “objective function” has to be defined. The contracting parties may discuss various alternatives. For example, they may consider it as a just distribution that both parties achieve the same rate on return on their invested capital. A variant of this approach could be to give one party a predetermined bonus over the share of the other partner (for example, as a compensation for management efforts). Another plausible scenario could be that the provider of the external capital prefers a profit distribution which reduces the volatility of his own profit share and gives him a more stable revenue stream that meets a certain benchmark; for example, the return from an investment in fixed-income securities (such as ijarah sukuk); exceeding

profits will remain with the other party in good years, the value of its profit share would be reduced in bad years. The simple model outlined below indicates, inter alia, that the distribution parameters could also be calibrated with respect to leverage effects of external capital (provided the parties agree on such a pattern). These are only a few examples for a wide range of conceivable arrangements for the reconciliation of the revealed preferences and interests of the contracting parties in a participatory finance setting such as a musharakah (sukuk or bank financing). The preferences and interests are contractually recorded and translated into a structure that protects the distributional preferences of the contracting parties by automatic alignments of the distribution parameters (in particular the profit sharing ratio) whenever new performance information of the project or joint venture become available. The automatic adjustment obviates or replaces ex post re-negotiations on redistributions which are permissible according to AAOIFI, but very difficult to realise when the gains of one party are the losses of the other party.

The following is an illustration of the basic mechanism underlying the idea of an adjustable profit sharing ratio. Assume that the contracting parties had agreed on one of the following distribution rules (“objective functions” in the model):

The profit sharing ratio should be such that either

• the return on external capital is equal to the return on total capital, or

• the return on external capital is the same as the return on internal capital, or

• the return on external capital equates the benchmark b, or

• the return on internal capital is a multiple or fraction a of the return on external capital

The appropriate profit sharing ratios are determined as follows:

Cases (1) and (2), profit sharing ratio for re = ri = r:Case (3), profit sharing ratio for re = b:Case (4), profit sharing ratio for ri = a ∙ re:

A numerical example is given in the following table. It illustrates the influence of different benchmarks, of a surcharge on the relative return for one party, and of different (actual or expected) profits. The table does not change the relation between external and internal capital, but the relevance of the capital structure is clearly visible from the profit sharing formulas above.

The above sample did explain the basic mechanism of the model only for three simple objective functions. In practice, contracting parties might agree on more complex formulas (for example, with caps or equivalents of “incentive fees”). Admittedly, it would also be possible to define an objective in such a way that one party receives a fixed amount as profit share (provided the volume of the total profit is at least as large as the fixed amount) – which would not be permitted under Shariah. But discretionary re-negotiations can achieve the same result. Insofar the automated system is not better or worse than discretionary practices regarding a possible “misuse.”

Nienhaus


It is clear that the profit sharing mechanism will only work as long as profits are generated. It does not provide an effective protection against the need of “loss sharing” in an individual musharakah setting. In the interest of risk mitigation and loss avoidance, the management of a musharakah sukuk should, for example, diversify investments by financing not only one entrepreneur but a number of firms in different markets with uncorrelated market trends, and it may also invest some funds in risk-minimized fixed-return instruments. But risk management in Islamic finance in general is a topic which goes beyond the scope of this paper.

The idea of a variable profit sharing ratio can also be found in the paper of Diaw, Bacha and Lahsasna (2012). But their perspective is a theoretical analysis and not the outline of an implementable recommendation for contracting parties

in a musharakah structure. They present a general equation that indicates the directions in which different parameters influence the profit sharing ratio. They do not transform their general equation in such a way that it would become an objective function and could be used (after calibration) in negotiations on a distribution pattern. Instead, Diaw, Bacha and Lahsasna present a Monte Carlo simulation, which gives a feeling of implied dependencies and possible dynamics. In addition, they offer a backtesting, which shows relations between the average return on capital, an average benchmark (indicated opportunity costs or benefits), the average return to sukuk, return to equity and the profit sharing ratio. However, their input data for this exercise were taken from mudarabah sukuk, i.e., sukuk which do not incorporate a profit sharing ratio. Insofar it is not clear what the results of the backtesting can show.

Profit Sharing Ratios for Different Objective Functions

Objective function: re = r ( = ri) re = b re = bri = are, a = 1.1

ri = are, a = 0.8

Ke Assumption 400 400 400 400 400 400Ki Assumption 100 100 100 100 100 100K = Ke+Ki Definition 500 500 500 500 500 500P Assumption 50 50 50 50 50 50s = Pe/P DefinitionPe = s∙P Def. & result 40 48 32 39 42Pi = (1-s)P Def. & result 10 2 18 11 8b and a Assumption 0,120 0,080 1,100 0,800re = Pe/Ke Def. & result 0,100 0,120 0,080 0,098 0,104ri = Pi/Ki Def. & result 0,100 0,020 0,180 0,108 0,083r = P/K Definition 0,100 0,100 0,100 0,100 0,100 0,100

Solutions:

s = Ke/K re = r (= ri) 0,800

s = b(Ke/P) re = b 0,960 0,640

s = 1/(1+aKi/Ke) ri = are 0,784 0,833

Objective function: re = r ( = ri) re = b re = bri = are, a = 1.1

ri = are, a = 0.8

Ke Assumption 400 400 400 400 400 400Ki Assumption 100 100 100 100 100 100K = Ke+Ki Definition 500 500 500 500 500 500P Assumption 75 75 75 75 75 75s = Pe/P DefinitionPe = s⋅P Def. & result 60 48 32 59 63Pi = (1-s)P Def. & result 15 27 43 16 13b and a Assumption 0,120 0,080 1,100 0,800re = Pe/Ke Def. & result 0,150 0,120 0,080 0,147 0,156ri = Pi/Ki Def. & result 0,150 0,270 0,430 0,162 0,125r = P/K Definition 0,150 0,150 0,150 0,150 0,150 0,150

Solutions:

s = Ke/K re = r (= ri) 0,800

s = b(Ke/P) re = b 0,640 0,427

s = 1/(1+aKi/Ke) ri = are 0,784 0,833



It may be read such that it shows (on the basis of observed returns on equity and returns on sukuk capital) what profit sharing ratios would have been necessary in each period to achieve a certain benchmark return.

Diaw, Bacha and Lahsasna do not explicitly integrate changes of the expected (or actual) profit into their analysis. Changes of profits and those adjustments of the profit sharing ratio during the life of a participatory finance are the main focus here, which are required to sustain the initially stipulated profit distribution pattern.

Islamic economists have made many attempts to overcome agency problems in participatory finance in order to make them more acceptable to market participants, in particular to providers of funds. This is because many see participatory finance as the ideal form of Islamic finance and as the core element of a genuine Islamic financial system. Even if one would not go that far, participatory modes of finance such as mudarabah or musharakah could fill a gap, for example, in the start-up and growth financing of SMEs. In general, participatory finance gives entrepreneurs some financial relief in times of an unexpectedly poor performance of their business. This risk-reduction can enhance their willingness and ability to ramp up innovations (new products, processes or technologies) or to enter into new markets. This should not only generate private profits but also social benefits in terms of employment and income opportunities. The model presented here outlines a technique which could make participatory modes of financing more attractive to financiers as well as entrepreneurs seeking finance.

5. ConclusionAAOIFI has allowed a remarkable wide range of “corrective measures” in mudarabah and musharakah contracts which are deemed Shariah compliant: The contracting parties can re-negotiate factually all commercially relevant aspects of their contracts at any time. A main reason for such re-negotiations will be new information about the performance (profit) of the financed project or venture. The crux of re-negotiations is that they will be successful only if one party is willing to give up advantages (financial gains) of the contractual status quo. This voluntary redistribution is not very likely.

The proposal of a self-adjusting profit sharing ratio obviates the need for discretionary re-negotiations. It uses elements of the AAOIFI toolbox to structure a contractual arrangement that maintains the distribution pattern, which was initially agreed upon by the contracting parties. It does so by an automatic adjustment of the profit-sharing ratio whenever new information on the expected profit becomes available. This technique facilitates a solution of fundamental agency problems in participatory finance. It does not require a particular legal form of the financed firm (such as a joint stock company) and can be calibrated such that risk aversion of the financier can be factored in. A musharakah contract with a self-adjusting profit sharing ratio is incentive compatible insofar as it does not provide the incentives for profit compression by underreporting, by allocation of fixed cost, by fringe benefits, or by shifting profits into periods after the termination of the musharakah contract. All this is achieved without the need for a sophisticated accounting system of the financed enterprise.

For the financed entrepreneur it is beneficial that a participatory financing implies a kind of “embedded” insurance against unexpected downside risks (loss sharing). However, the loss absorbing qualities of a mudarabah or musharakah contract will not come for free but will be reflected in the costs of funds (e.g., by a risk premium added to a benchmark rate by the financier). On the other hand, the contract can be calibrated such that both parties can enjoy upside gains. This is in contrast to the widespread practice of a complete skimming off of gains by one party in musharakah sukuk. The financier can trade in upside opportunities for predictable and stable returns based on a risk-adjusted benchmark rate.

It is obvious that arrangements based on self-adjusting profit sharing ratios are no “real” profit and loss sharing contracts or “full” risk sharing partnerships. Insofar the approach is not the first best solution in an ideal world. However, under real-world circumstances first best models do not work, and the technical and Shariah merits or shortcomings of the approach should be discussed in relation to other real-world proposals to overcome the inherent agency problems of PLS or risk sharing arrangements.

ReferencesAAOIFI. (2008) Guidance Statement on Accounting for

Investments and Amendment in FAS 17. Available at: http://www.aaoifi.com/aaoifi_sb_sukuk_Feb2008_Eng.pdf.

AAOIFI. (2010a) Accounting, Auditing and Governance Standards for Islamic Financial Institutions. 1432 H – 2010. Manama: AAOIFI.

AAOIFI. (2010b) Shari’a Standards for Islamic Financial Institutions. 1432 H – 2010. Manama: AAOIFI.

Abdel-Khaleq AH, Crosby T. (2009) Musharakah Sukuk: Structure, Legal Framework and Opportunities. In: Abdulkader T (Ed.). Sukuk. Sweet & Maxwell Asia. Petaling Jaya. 187–222.

Ahmed H. (2002) Incentive-Compatible Profit-Sharing Contracts: A Theoretical Treatment. In: Munawar I, David T. Llewellyn DT (Eds.). Islamic Banking and Finance – New Perspectives on Profit-Sharing and Risk. Edward Elgar. Cheltenham. 40–54.

Bacha OI. (1997) Adapting Mudarabah Financing to Contemporary Realities: A Proposed Financing Structure. The Journal of Accounting, Commerce and Finance. 1(1). http://mpra.ub.uni-muenchen.de/12732/.

Casey P. (2012) Regulatory Lessons on Sukuk Financial Products, an Opinion. In: Ariff M, Munawar I, Shamsher M (Eds.). The Islamic Debt Market for Sukuk Securities: The Theory and Practice of Profit Sharing Investment. Edward Elgar. Northampton, MA. 99–118.

Diaw A, Obiyathulla IB, Ahcene L. (2012) Incentive-Compatible Sukuk Musharakah for Private Sector Funding. ISRA International Journal of Islamic Finance. 4(1):39–80.

Hasan, Z. (2002) Mudarabah as a Mode of Finance in Islamic Banking: Theory, Practice and Problems. Middle East Business and Economic Review. 14(2):41–53. http://mpra.ub.uni-muenchen.de/2951/.

Nienhaus


Ibn R. (1996) The Distinguished Jurist’s Primer, Vol. 2 (Bidayat al-Mujtahid). Translated by Imran Ahsan Khan Nyazee. Garnet. Reading.

Kapetanovic H, Muhamed B. (2009) Mudharabah Sukuk: Essential Islamic Contract, Applications and Way Forward. In Abdulkader T (Ed.). Sukuk. Sweet & Maxwell Asia. Petaling Jaya. 223–247.

Khwaja M. (2009) Mudharabah Sector Report – Pakistan. In Abdulkader T (Ed.). Sukuk. Sweet & Maxwell Asia. Petaling Jaya. 245–246 (= Annex to Kapetanovic and Becic 2009).

Mokhtar S. (2011) Application of Wa’ad in Equity Based Sukuk: Empirical Evidence. ISRA Research Paper 20. ISRA. Kuala Lumpur.

Demetriadies DG, Tyser CR, Effendi IH (Translators). (2001) The Mejelle – Being an English Translation of Majallah El-Ahkam-i-Adliya and a Complete Code on Islamic Civil Law. The Other Press. Kuala Lumpur.

Saeed A, Salah O. (2012) History of Sukuk: Pragmatic and Idealist Approaches to Structuring Sukuk. In: Ariff M, Iqbal M, Shamsher M (Eds.) The Islamic Debt Market for Sukuk Securities: The Theory and Practice of Profit Sharing Investment. Edward Elgar. Northampton, MA. 42–66.


Cite this chapter as: Rizvi S A R, Arshad S (2015). Indexing government debt to GDP: A risk sharing mechanism for government financing in Muslim countries. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Indexing government debt to GDP: A risk sharing mechanism for government financing in Muslim countriesSyed Aun R. Rizvi1, Shaista Arshad2

1INCEIF, Lorong Universiti A, Kuala Lumpur, Malaysia, Phone: +60136145752, Email: [email protected], Institute of Islamic Banking and Finance, Kuala Lumpur, Malaysia, Phone: +60133771370, Email: [email protected]

Abstract - Over the past decades much effort and research has gone into establishing a viable set of Islamic financial institutions. An area of utmost importance, which still has gaping holes, is the development of instruments for government financing on a global level. Most Muslim countries, with the exception of a few Gulf countries, are heavily indebted with high reliance on multilateral financing primarily based on high interest rates. This vicious cycle of interest rates and debt have stunted the growth of these nations and worsened the conditions of the masses. This paper is an attempt at introducing the concept of risk sharing instrumentation in the form of GDP-linked Papers for Muslim governments for their financing through multilateral and supranational bodies. Without dwelling much on the Shariah technicalities, we attempt to discuss the potential benefits of transferring government debt in these forms of risk sharing instruments. The authors have attempted to represent the economic benefits of these risk-sharing mechanisms for a set of four indebted Muslim countries with empirical proof. Through this paper, we endeavour to initiate a thought provoking and practical discussion for further development of these instruments for the betterment of Islamic countries.

Keywords: risk sharing instruments, sovereign debt, GDP-linked paper, Muslim countries

JEL Classification Codes: F34, H63

1. IntroductionIslamic finance has progressed phenomenally over the past few decades, moving from a small market in the Arab world to a much larger global presence. Marked with impressive growth, the presence of Islamic finance is estimated to be US$1.6 trillion and is expected to grow at the rate of 15% over the next few years.

The viability of Islamic finance as an alternative was put to test in the wake of the recent financial crisis, where in the midst of bank failures and stock market crashes, institutions practicing Islamic finance managed to stay afloat and rather provided a buffering effect to the crisis. Several economists argue that due to the innate laws of Islamic finance it was able to withstand the financial crisis much better than its conventional counterpart was.

Yet despite this impressive record, Islamic finance only accounts for 1% of the global financial system. Some scholars are of the opinion that this is due to the substantial indebtedness of Muslim nations. Narrowing in on the Organization of Islamic Centre (OIC) member countries, a study revealed 19 from 39 Heavily Indebted Poor Countries (HIPC) are part of the OIC (IMF, 2010). Furthermore, a mere 6 out of 57 OIC member countries, in the Gulf Cooperation Council (GCC) are prepared to meet financial shocks of sorts. The remainder are heavily reliant on borrowing from the international market to pay off their debts.

Additionally, these countries are dependent on Western financial intermediaries with high interest rates to meet their debt requirements, transacting mainly in dollars. During times of economic instability, liability dollarization (see Calvo et al., 2003; Calvo and Reinhart, 1999; Durdu

Rizvi and Arshad


and Mendoza, 2005) and relying heavily on high interest borrowings can backfire for these countries, as they are more likely to have long-term shocks on the economy leading to higher cost of borrowings. Alternatively, they may need to restructure or renegotiate the terms of borrowing or lastly, be forced to employ pro-cyclical policies that would lead to disastrous economic and social outcomes.

One way to avoid any of the above possibilities is to index their debt against real economic variables to reduce financial distress. Several economists advocate the value of indexing government debt reasoning that it would allow for an improved risk sharing among debtor countries and international creditors.

Amongst the key proponents of this instrument is Bailey (1983) who had initially suggested transferring the sovereign debt into claims on the exports of the country. Delving further on this, Krugman (1988) and Froot et al. (1989) studied the virtues of indexing sovereign debts to variables that are controlled by the country, partially or completely.

Shiller (1993) first proposed the concept of a true GDP-linked bonds, which fundamentally were perpetual claims on a fraction of a country’s GDP. Similarly, Borensztein and Mauro (2004) also sought to review the case for GDP growth-linked bonds, and Griffith-Jones and Sharma (2006) in their paper summarize the benefits of introducing GDP-linked bonds. Working on Shiller’s concept, Obstfeld and Peri (1998) propositioned members of the European Union to issue GDP-linked euro denominated securities. Likewise, Haldane and Quay (1999) argue in favour of indexing sovereign debt to commodity prices. Dreze (2002) advocated using GDP-indexed bonds as an integral part of restarting the debt of the poorest countries strongly.

While the idea of GDP-linked debt has been implemented to a limited extent, it received renewed impetus after a wave of financial and debt crisis that flooded emerging markets in the 1990s. Caballero (2002) has suggested the issuance of copper indexed bonds during Chile’s financial upheaval. He further recommends making the returns of debt instruments dependent on external shock variables that trigger crises (such as exports, regional economic growth, etc.) arguing that most emerging markets are affected by external shocks that they are not well-equipped to withstand.

Reverting back to our issue at hand, as most of Muslim nations are severely debt ridden, a fundamental question arises: Does Islamic finance have the aptitude to provide for the borrowing and financial needs of Muslim governments? Much has been written about the development and innovation of Islamic finance over the years. Mirakhor et al. (2011 and 2012) believes that risk sharing is the foundation of Islamic finance and is instrumental in the development of any financial products within the ambits of the Shariah.

A dearth in the present literature found concerns GDP-linked papers using the principles of Islamic finance. The potential benefit of using GDP-linked Sukuk was discussed by Diaw et al. (2012) where he proposed structuring the instrument based on forward ijarah. Mirakhor (2011) highlights the benefits of using macro-markets and similar GDP-linked papers in his works. Similarly, the prospective benefits from using an equity financed budget for Muslim

countries was discussed by Hasan and Siddiqui (1992) who highlighted that an equity financed deficit is not necessarily unstable and that the output of the economy can be enhanced by the government expenditure policies.

Keeping in mind the evident lack of research on this pertinent topic, this paper aims to assess the benefits of GDP-linked Sukuk, and tries to empirically show some of the key benefits these instruments can offer. This paper also attempts to address some of the concerns from the investor angle, to provide a case for these instruments as alternative to sovereign debt for Islamic countries.

The present paper is structured as follows: Succeeding the introduction is the motivation and justification for the study, which is followed by an insight into the need for GDP-linked Sukuks, benefits and fiqhi aspect in Section 3. Section 4 provides some simple empirical results on how the economies will benefit with these instruments, which is followed by discussion on the viability for investors in Section 5. Subsequently, the authors provide concluding remarks.

2. Motivation and justification of studyIn this study, the primary focus of the authors is to emphasis on the benefits of employing GDP-linked papers for Muslim countries that are emerging but debt ridden. With the substantial growth of Islamic finance over the years, it becomes imperative to introduce instruments that would benefit Muslim economies at large. In particular, the reliance of Muslim nations on high-interest rate borrowing has largely been neglected in Islamic finance, with little to no research available on alternatives.

The motivation for this paper arises from the need of evidence to support the practicality of linking real economic variables to sovereign debt for OIC countries. This will further allow us to examine the efficacy of such instruments on the economies of Muslim countries.

This paper is amongst the few that have discussed the use of GDP-linked instruments, primarily discussed by Diaw and Bacha (2011) who proposed Sukuk structures indexed with real economic variables. The use of real economic variable linked instruments, as proposed in this study, aims to quell instability in Muslim countries and help stabilize government spending while reducing the impact of interest borrowings.

Islamic finance theoretically has the capacity to provide an alternative to these Muslim countries that would allow them to restructure their debt and continue growing as an emerging economy. Islamic finance should aim to be the principal financial system for Muslim countries, hence raising the need for the development of new instruments and products to meet the needs of Muslim economies at both a micro- and macro-level.

This study is novel in this area as it is a humble attempt to initiate an empirical research-based argument to provide policy makers and economists an alternative Shariah compliant instrument as a replacement of the conventional sovereign debt.

In the current literature in conventional finance this area has not been studied in detail. Regarding Islamic finance,


Indexing government debt to GDP: A risk sharing mechanism for government financing in Muslim countries

it is very rarely studied area. But considering the structure of Muslim economies, which are primarily indebted economies, the issue of sovereign debt is of immense importance. The objective of this study is to provide empirically evidence on the viability and economic benefits of GDP-linked Sukuks.

3. GDP-linked Sukuk

Need for GDP-linked SukukGDP-linked Sukuk aim to provide a breathing room to economies when they go into recession and on the other side provide an automatic slowing-in mechanism when the economy goes into high growth zones and into overheat. Analysing the advantages of this instrument, we can find two sets of clear advantages from an economic point of view for the Muslim countries.

Firstly, since the payment of profit is linked to the actual growth of the economy, it can reduce the likelihood of crises. In comparison to the normal borrowing mechanism that Muslim economies are subjected to via plain vanilla sovereign bonds or multilateral agency debts, the payments are fixed, and any slowdown in economy results in a ballooning of debt.

In existing literature, there is an immense amount of evidence on the sovereign debt’s position of dependence on the economic growth. A glance at the history shows that slow growth has primarily underlined debt crises, like the Latin American crisis of 1980’s and the near complete default of the Heavily Indebted Poor Countries of 1990’s. In economic theory and its applications, the ratio of external debt to GDP is a significant predictor of impending crises. Detragiache and Spilimbergo (2001) in their study empirically show that an increase of 10% in the debt/GDP ratio has an impact of nearly 20% increment in the probability of the crisis. In another study by Easterly (2001), he argues with proof that a hundred basis point decline in GDP growth rate is associated with 1.5 times more debt rescheduling in the fifteen years that follow.

In recent times, Argentina in the global economic system has flirted with the concept of GDP-linked borrowings. Although the conventional GDP-linked bonds differ from the conceptual Sukuk proposed in this paper, the mechanics of economics can be argued for a benchmark and as an example. There appears to be opposing views on how much the Argentinian government benefited from GDP-linked borrowing, but economic numbers show a stable growth and recovery for the economy.

Secondly, GDP-linked Sukuk, as an alternative to the conventional debt, reduce the need of procyclical policies, which have stunted the growth of most Muslim countries post crises. It acts as automatic stabilizers in the case of an economic slowdown and in boom periods. Considering the example of Pakistan, post 2008stock market crash, and civil unrest owing to terrorism, the economy slowed down. Having immense amounts of external debt, which had a fixed interest payment due, the country had to revert to an IMF bailout plan, thus incurring another 10 billion dollar loan to meet its prior commitments. The increase in debt and the bailout plan came with specific conditions of fiscal tightening and tax reforms. These policies of fiscal

tightening and tax reforms and removal of subsidies provide a further negative impact on growth. A fiscal tightening by the government would reduce the expenditure in the country influencing the industrial sector’s growth.

Similarly, an increase in tax and removal of subsidies would reduce the disposable income of the people, thus reducing the consumption in the country. Both of these results end in amplifying the economic downturn. Calvo (2003) highlights some of these issues in his literature on sudden stops, where he argues that sovereign debt fails to serve as a device for inter-temporally smoothing the impact of such slowdowns.

Key benefits of GDP-linked SukukIn this section we try to highlight the key benefits GDP-linked Sukuk offer the issuing country from an economic perspective:

• In case the economy of the issuing country experiences a persistent low growth phase, the GDP-linked papers provide a cushion against mounting debt as in conventional borrowing. Since the payment due on these papers would vary with the growth rate, the Debt/ GDP ratio would only increase via a deficit borrowing of the country. This would lead to a less probability of default, and provide breathing room for policy makers to expand fiscally or through tax cuts to boost the economy.

• Usually emerging countries and HIPC countries (where most Muslim economies fall) face extreme difficulty for borrowing in times of crises due to weak economic numbers. With GDP-linked papers, since their liabilities would reduce, that would provide room for counter cyclical policies to boost the economy.

• As Barro (1995) highlights, a mechanism of linking GDP with the liability of the country can provide ease of maintaining smoother tax rates and provide basic infrastructure to the population in times of crises without increasing tax rates.

• It has been observed, in the case of many Islamic countries, as well as emerging economies, the tendency to go into frivolous expenditure in the times of high economic growth, which may lead to overheat of the economy followed by a recessionary phase to bring the economy back to equilibrium level. Amongst the key Islamic countries, the case of Indonesia and Malaysia, in the early 1990s is an example. The economic crash of 1997 affected the region bringing the growth number to negative in one year after a decade of double-digit growth. If the government is using GDP-linked paper, in times of high growth the payment on Sukuk will be higher as well, which will keep a check on excessive fiscal expansion.

Any instrument is not viable for financial markets until it provides benefits to both sides involved in the transaction. Although in this paper we aim to provide a concept of a possibility of using GDP-linked papers in Muslim economies, why it may work is provided in the following passage by listing the potential benefits to the buyer of this instrument.

• Diversification of investments: The economic cycles of different countries are far from being correlated perfectly. Although there exists regional correlation

Rizvi and Arshad


to some extent, but global perfectly correlated economic cycles is a utopian scenario. The investors in these papers can diversify their portfolio while keeping their investments in the real sector of the economy instead of financial papers only.

• An application of GDP-linked papers will provide a smoothing of the economic growth of the Muslim and emerging economies, reducing the probability and frequency of defaults. From an investor perspective, a default is a high cost of litigation and restructuring.

• While in developed countries, the stock markets provide an investment opportunity into the future prospects of the real sector, in Muslim economies, the stock markets are still at nascent stage. Issuance of GDP-linked Sukuk will provide an opportunity to investors to invest in the potential of the real economy.

Insight into structure and Fiqhi issuesThe proposal for a GDP-linked Sukuk has been raised in the corridors of Islamic finance previously in a paper by Diaw et al. (2011), where they have proposed an Ijarah based GDP-linked Sukuk for public sector funding and debt management. The proposal of this Sukuk takes into account this earlier study and is an attempt to further the idea. Whereas the previous study proposes an ijarah-linked paper, we propose a pure musharakah linked paper.

The idea of authors is to issue a musharakah paper by the government as the wakil of the country, where the underlying business is the economy as a whole. Keeping in mind the musharakah nature, where profit is paid on actual profit is the business, which in this case is the growth of the economy. This idea is subject to Shariah debate and approval, but in the understanding of the authors if we consider the economy as one business, there should not be a problem in sharing of the risk of the business (economy).

A key insight of this model is that in case the economy goes to zero growth or negative growth in any year, the investors would share in the losses, which would mean a reduction in their principle amount invested in the business (economy). This sharing in risk of the economy via a musharakah paper is different from indexation of the debt, since a debt is not created when a musharakah Sukuk is shared, but is similar to issuance of equity shares in the business (economy) of the country.

The issuance of indexation is a much-debated topic under the Shariah law, but is primarily focused on “debt” and not much has been discussed and debated on the issue of Equity shares linked to GDP. Bacha and Mirakhor (2012), provide a strong argument in favor of development of similar GDP-linked securities, linking it to welfare of Ummah through the concept of macro markets.

4. Empirical analysis

Empirical evidence for some Muslim economiesThe empirical section shows some primary basic economic numbers to analyze what benefit GDP-linked Sukuk may provide to the Muslim economies. To make our analysis, we take into account four countries, Malaysia, Indonesia, Iran and Turkey. These four countries throughout the

Islamic world comprise nearly 60% of the GDP of Islamic ex GCC countries. These four countries have experienced economic crashes and have been under external debt for the past few decades.

For a simple review of the benefits of a GDP-linked sovereign can be illustrated by taking the example of our sample countries. Suppose a simple musharakah based GDP-linked Sukuk was used for all financing of these four countries from 1985 onwards, including conversion of the existing debt to this instrument. The country would have to pay a profit rate of whatever the growth of the underlying business is – in this case the economy of the country. In any year where the business (economy) does not grow, there would be a zero payment of profit.

For the purpose of this example, we make the strong assumption that the composition of the debt and the changes in the debt levels does not have any impact on the behavior of any of the other variables in the economy during 1985–2010. What repayments the countries would have had to make and the interest cost savings is illustrated in the Figure 1. The graphs show, what was the actual interest payment and debt forgiveness in each year as a representative of the cost the country actually had to bear due to conventional mode of financing. In addition, the payments for GDP-linked Sukuk are shown for the same period and the savings as a percentage of the national income each year.

Looking at the Malaysian case, the Malaysian economy after a short decline in the mid-1980s bounced back and grew at a robust speed for the second half of the eighties and first part of nineties. This was followed by the infamous Asian financial crisis of 1997, which brought the ASEAN economies down to their knees in a matter of a year. With a couple of years of negative economic growth, the economy was propped back up with fiscal support policies and capital controls. Indonesia experienced a similar crisis but decided to take the multilateral agency bailout plan, and both economies have experienced a revival in economies in the last decade.

With liberalization policies in Indonesia, a massive influx of foreign investment supported by steady infrastructure development has resulted in high economic growth figures. Within these two economies, in our example for using GDP-linked Sukuks, the graphs show the crisis would have resulted in zero payments during the crisis years, providing room to the government for fiscal expenditure, to propel the economy further. Governments using GDP-linked Sukuk would have obtained a large reduction in the interest bill, leaving more room to avoid procyclical fiscal measures. Large interest savings would also have applied during the sudden slowdown of 2008–2009. In the case of Malaysia, the average profit payment over the 25-year period amounts to 5.99% as compared to 6.54% average interest payments that were paid by the Malaysian government. In addition to the lower average payment, over the period, Malaysia would have saved nearly 14% of its national income. In the case of Indonesia the difference between profit payment on GDP-linked Sukuk and interest payment is a mere 30 basis points annually, but the savings as a percentage of national income would have been 5%.

On the other hand, Turkey has had a unique economic growth case, where the real GDP has experienced high



Figure 1. Interest savings over the economic cycle.

–2%

0%

2%

4%

6%

8%

10%

12%

–2%

0%

2%

4%

6%

8%

10%

12%

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Savings in National IncomeGDP LinkedInterest Paid

Malaysia 1985-2010


Turkey 1985-2010

–2%

0%

2%

4%

6%

8%

10%

12%

–2%

0%

2%

4%

6%

8%

10%

12%

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

–2%

0%

2%

4%

6%

8%

10%

12%

–2%

0%

2%

4%

6%

8%

10%

12%

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010


Indonesia 1985-2010

–2%

0%

2%

4%

6%

8%

10%

12%

14%

–2%

0%

2%

4%

6%

8%

10%

12%

14%

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010


Iran 1985-2010

Rizvi and Arshad


volatility as compared to other countries. With its strategic location as the bridging country between Europe and Asia, Turkey has experienced internal economic crises during our sample period.

The 1980–88 Iraq-Iran war benefitted the Turkish economy. At that time, the Turkish economy increased trade with both countries and became the supply route for Iraqi oil exports. The end of war in 1988saw a sudden slowdown in the economy, which stabilized with the economic policy of import substitution of the then government. The economy again received a major blow during the Persian Gulf War, with a nearly $3 billion loss due to reduced trade. Saudi Arabia, Kuwait, and the United Arab Emirates (UAE) moved to compensate Turkey for these losses, and by 1992, the economy again began to grow rapidly, before it was plunged again into crisis in 1994, owing to government borrowings. Post crisis, the economy stabilized but experienced sharp downturns due to military assault in Iraq, the economic slowdown of Europe and the financial crisis of 2007. With this volatile nature of the economy, GDP-linked Sukuk, would have provided savings as a percentage of national income to the tune of 24% during this period, while the average payment per annum for GDP-linked paper is 4.77% compared to 6.83% average annual interest paid by Turkey.

The fourth country in the sample, Iran, with its decade-long war with Iraq in the 1980s and subsequent international sanctions owing to its nuclear research has experienced an economic situation marked by dogged growth, where it has been impacted by sanctions and restrictions on its oil export. Iran has the largest Islamic financial assets in the world, and using the Musharakah based GDP-linked Sukuk would have benefitted in aggregate savings of 6% over national income while the average profit payments on GDP-linked papers would have been 4.07% as compared to actual interest payments of 5.30% annually that Iran paid to external lenders.

Debt/GDP stabilization effectTo illustrate the benefits of the GDP-linked Sukuk, in stabilizing the Debt/GDP ratio of a country we perform a simple exercise. For our sample countries using the well-established identity in fiscal economics:

DY

r gDY

St

tt

t

tt= + -

--

-

( )1 1

1

where Dt is government debt, Yt is output, st is the primary surplus as a share of GDP, gt is the growth rate, and r is the interest/profit rate on the debt.

Using actual data for the time period under consideration, holding all other values of growth as constant (a very strong assumption), we simulate the new Debt/GDP levels for these four economies. Table 1 reports the standard deviation as a volatility measure of the Debt/GDP ratio during that period. It is evident that using the GDP-linked Sukuk would reduce the volatility in the Debt/GDP ratio for all the countries minus turkey, where it stays similar. Although the difference in the standard deviation doesn’t seem huge, but if it is translated in absolute dollar terms, it goes in excess of a US$1 billion for these countries.

Tables 2 and 3 represent the maximum and minimum value of the Debt/GDP ratio for a comparative analysis of actual and simulated using GDP-linked Sukuk. For our sample countries, we observe that during the 25 year period under study, the maximum levels of Debt/GDP would have been substantially lower for Indonesia and Malaysia by nearly 300 basis and 400 basis respectively. In the case of Turkey and Iran, the maximum level of Debt/GDP would have been considerably lower as well. Same trend holds for minimum level of Debt/GDP ratio, where the minimum levels would have been much less than the actual Debt/GDP ratios these countries experienced.

Table 1. Standard deviation of Debt/GDP.

Iran Indonesia Malaysia Turkey

Without GDP-linked papers

0.087 0.267 0.122 0.061

With GDP-linked papers

0.084 0.256 0.110 0.061

Table 2. Maximum value of Debt/GDP.



39.11% 158.69% 77.48% 57.62%


38.63% 155.50% 73.43% 55.78%

Table 3. Minimum value of Debt/GDP.



2.93% 27.57% 29.30% 32.80%


2.67% 28.34% 29.04% 31.35%

The earlier results highlight the potential benefits of having GDP-linked Sukuk as a replacement of conventional debt. Nevertheless, the illustrations in earlier tables have been provided under strong assumption of keeping all factors constant. In reality, the potential benefits may be large owing to interacting nature of economic variables. We have assumed that the primary surplus remained the same as actual every year in our sample period, but it is important to note that while using this Sukuk fiscal policy economic growth and in savings/increments in profit payment may respond. This would change the primary surplus and thus may amplify the benefits achieved by these instruments in stabilizing the Debt/GDP ratio.

Primarily in years of sluggish or negative growth with debt sustainability as a major concern, governments



tend to move towards an increase in the primary surplus, effectively implying that fiscal policy is procyclical. In our sample countries, this is evident when we analyze the primary surplus during economic downturns. In the case of Malaysia and Indonesia it was evident during the Asian financial crisis when Malaysia with an average deficit of 6% in 1997–98 moved to a 6% surplus in the succeeding 2 years. This is a common occurrence in Islamic countries, which are primarily liquidity constrained and have to impose these policies to maintain their credibility in the international credit market.

5. Viability for investorsBy illustrating some of the ways GDP-linked Sukuk can benefit the Islamic countries, within our sample countries, the authors do realize that implementation of these instruments require a much more thorough and rigorous empirical work. In addition to the fine-tuning of these instruments, a major obstacle that can be foreseen is the creation of markets for new instruments globally. Past experiences with the introduction of innovative instruments in world credit markets has been met with skepticism and resistance.

Since GDP-linked Sukuk based on pure musharakah structure is a novel idea in the Islamic financial landscape, not much debate has happened on potential obstacles. Nevertheless, owing to the global economic environment, these instruments would have to be introduced in the world financial markets, which are dominated by the conventional investors. Due to this reason, we borrow some of the concerns of global investors from literature of conventional finance that seem to hold valid for these instruments as well.

A serious question is based on whether international investors would be willing to directly expose their investments to volatility of GDP in Muslim and emerging countries. In the understanding of the authors, international investors are already exposed to the economic growth of the country implicitly via the investments in the stock markets and the standard debt contracts. If a country goes into a recessionary phase, and debt sustainability levels are breached, the borrower country is likely to default or go into a restructuring program, which is a heavy legal and financial cost. The GDP-linked Sukuk tends to provide a buffer against the possibility of default and its costs by providing the investors a steady return that is bound to pay a higher amount in the long term as economies pick up.

Secondly, the issue of risk exposure and diversification while using GDP-linked Sukuk is an implementation issue. Studying the economic structure of different regions and considering that Muslim economies are spread far and wide across the globe, the economic cycles are not perfectly synchronized and have low or negative correlations. If a mass issuance of the GDP-linked Sukuk is done by Muslim countries, from an investor perspective a well-diversified portfolio can be constructed to hedge against over exposure to a single business cycle. Table 4, provides the correlation matrix for GDP growth rate amongst a selected few Islamic countries.

Our sample countries are highlighted in grey. It can be observed from the specific countries that most of them have a varying correlation in growth with each other. Malaysia and Indonesia, being in the same region and having similar economic fundamentals, tend to have a high correlation while these two countries with Iran have a very low correlation. Looking at the selected few countries, we observe that there are traces of negative correlation between Islamic countries as well, for instance, in the case of Bangladesh with Malaysia/Indonesia, and Nigeria with Malaysia. This bodes well from an investor perspective for diversification and builds a good case for these instruments.

6. ConclusionThis study focuses on the predicament of several Muslim countries that are debt-ridden and struggling to finance their debt. Most commonly, these countries turn to western financial intermediaries offering high interest rates to meet their debt requirements. Unfortunately, the reliance of Muslim nations on high-interest rate borrowing has largely been neglected in Islamic finance, with little to no research available on alternatives.

The development of financial instruments for sovereign financing has only just begun whereby a few seminal studies started to show how Shariah compliant instruments can help alleviate Muslim nations from their current predicament.

Owing to fact that 51 out 57 OIC member countries are debt-ridden, the authors propose a solution to this problem found within the ambit of Islamic finance. Stemming from conventional finance, the authors recommend the use of GDP-linked papers using the principles of Musharakah. The paper highlights the key benefits of using GDP-linked Sukuk arguing that such an instrument will allow for a

Table 5. Correlation of GDP growth rates amongst selected few Islamic countries.

Bangladesh Egypt Indonesia Iran Malaysia Nigeria Tunisia Turkey

Bangladesh 1.000 0.369 (0.125) 0.321 (0.127) 0.070 0.431 0.225Egypt 1.000 0.003 0.043 0.013 0.214 0.336 0.001Indonesia 1.000 0.101 0.778 0.122 (0.096) 0.177Iran 1.000 0.182 0.304 0.559 0.063Malaysia 1.000 (0.009) 0.027 0.296Nigeria 1.000 (0.020) 0.074Tunisia 1.000 0.142Turkey 1.000

Rizvi and Arshad


lower probability of default, which permits counter cyclical policies to boost the economy. Similarly, it leads to an ease of smoother tax rates and curbs excessive spending during times of economic growth.

Covering Malaysia, Indonesia, Iran and Turkey, the authors attempt to further emphasis the benefits of GDP-linked Sukuks. In the case of Malaysia, using GDP-linked sovereign paper would have provided an average profit payment of 5.99% as compared to a 6.54% interest payment that was paid by the Malaysian government over the 25-year period. This would provide more room for Malaysia to avoid procyclical fiscal measures. In addition, Malaysia would have saved nearly 14% of its national income.

Indonesia, on the other hand, only shows a difference of 30 basis points between the profit payments and interest rate payments on GDP-linked sovereign paper and borrowings. Nonetheless, the savings on its national income would have amounted to 5%. The case of Turkey showed that due to the volatile nature of the economy, GDP-linked Sukuk would have provided a savings of around 24% as a fraction of national income. The average payment per annum for GDP-linked paper is 4.77% compared to 6.83% average annual interest paid by Turkey. Lastly, Iran’s saving would have been 6% over national income while profit payments would have been only 4.07% compared to the 5.30% that was paid out to lenders.

The authors further highlight that using GDP-linked Sukuk would reduce the volatility in the Debt/GDP ratio for the countries. The maximum value of Debt/GDP ratio was considerably lower for all four countries, with Malaysia and Indonesia recording a reduction of nearly 400 bases and 300 bases respectively. Parallel trends apply for minimum levels of Debt/GDP ratio, where the minimum levels would have been much less than the actual Debt/GDP ratios that these countries experienced.

While this study lays out an introductory emphasis on the benefits of a GDP-linked Sukuk, it is not without limitations. The implementation of this study has been restricted to only four countries; further research calls for a much larger sample study. Furthermore, this study makes an assumption that all interactions of economic variables remain the same, while in reality a more detailed analysis would be better to see how using these instruments will change the behaviour of growth and other economic variables.

This preliminary study is an emphasis on the benefits of Islamic finance as an alternative for government financing. The empirical evidence provided by this study is promising and provides an avenue for further research. The plight of Muslim nations with respect to their debt structure has largely been ignored in Islamic finance. It is the opinion of the authors that such instruments are vital for the support and development of Muslim nations and should be further researched.

ReferencesBailey N. (1983) A Safety Net for Foreign Lending. Business

Week. January 10.

Barro RJ. (1995) Optimal Debt Management. Working Paper. NBER, 5327.

Caballero RJ. (2002) Coping with Chile’s External Vulnerability: A Financial Problem. MIT.

Calvo GA, Reinhart CM. (1999) When Capital Inflows Come to a Sudden Stop: Consequences and Policy Options. Center of International Economics, Department of Economics. University of Maryland.

Calvo GA, Izquierdo A, Talv E. (2003) Sudden Stops, The Real Exchange Rate, and Fiscal Sustainability: Argentina’s Lessons. NBER Working Paper.

Detragiache E, Spilimbergo A. (2001) Crises and Liquidity—Evidence and Interpretation. IMF Working Papers 01/2. International Monetary Fund.

Diaw A, Bacha OI, Lahsasna A. (2011) Public Sector Funding and Debt Management: A Case for GDP-Linked Sukuk. Paper presented at the 8th International Conference on Islamic Economics and Finance, 19–21 December, Doha, Qatar.

Diaw A, Bacha OI, Lahsasna A. (2012) Incentive Compatible Sukuk Musharakah for Private Sector Funding. ISRA International Journal of Islamic Finance. 4(1).

Dreze JH. (2002) Loss Reduction and Implicit Deductibles in Medical Insurance. CORE Discussion Papers 2002005. Université catholique de Louvain, Center for Operations Research and Econometrics (CORE).

Durdu CB, Mendoza EG. (2006) Are Asset Price Guarantees Useful for Explaining Sudden Stops? The Globalization Hazard-Moral Hazard Trade of Asset Price Guarantees. IMF Working Paper. WP/06/73.

Easterly W. (2001) The Effect of International Monetary Fund and World Bank Programs on Poverty. Policy Research Working Paper Series, 2517. The World Bank.

Froot K, Keen MJ, Stein, J. (1989) LDC Debt: Forgiveness, Indexation, and Investment Incentives. Journal of Finance. 44(5):1335–1350.

Haldane A, Quah D. (1999) UK Phillips Curves and Monetary Policy. Journal of Monetary Economics. 44(2):259–278.

Hasan MA, Siddiqui AN. (1992) Is Equity Financed Budget Deficit Stable in an Interest Free Economy? Journal of Islamic Economic Studies. 1(2).

Krugman P. (1988) Financing vs. Forgiving a Debt Overhang. Journal of Development Economics. 29:253–68.

Obstfeld M, Peri G. (1998) Regional Nonadjustment and Fiscal Policy: Lessons for EMU. Center for International and Development Economics Research (CIDER) Working Papers C98–096. University of California at Berkeley.

Shiller RJ. (1993) Aggregate Income Risks and Hedging Mechanisms, NBER Working Papers 4396. National Bureau of Economic Research, Inc.

Zamir I, Abbas M. (2012) An Introduction to Islamic Finance: Theory and Practice. 2nd Edition. Wiley Publications.

Zamir I, Abbas M, Hossein A, Noureddine K. (2011) Risk Sharing in Finance: The Islamic Finance Alternative. Wiley Publications.


Cite this chapter as: Aydin N S, FRM, Rainer M (2015). Concept and mathematics of Islamic valuation and financial engineering. In H A El-Karanshawy et al. (Eds.), Islamic banking and finance – Essays on corporate finance, efficiency and product development. Doha, Qatar: Bloomsbury Qatar Foundation

Concept and mathematics of Islamic valuation and financial engineeringNadi Serhan Aydin, FRM1, Martin Rainer2

1Organization of Islamic Cooperation (OIC), Ankara Centre – SESRIC, Institute of Applied Mathematics, Middle East Technical University, Ankara, Turkey, Email: [email protected] Institute, Institute of Applied Mathematics, Middle East Technical University, Ankara, Turkey, Email: [email protected]

Abstract - Starting from the fundamental principles, whether a designed contract implies a clean bay’ rather than riba is central to our discussion. We argue that riba and gharar may easily arise through neglect of risk or inappropriate valuation methods for value and risk of assets and financial instruments. This possibility, in turn, strongly necessitates an estimation of expected forward values, market risk, and default risk, which is consistent with Islamic principles of avoiding riba and gharar. Based on consistent estimatation of risk and return, the expected costs of risks should be quantified for all parties to the contract in order to judge, whether a contract is free from usury (riba) and evitable risk (gharar) and whether it the remaining inevitable risks are distributed fairly between the counter-parties, for example between the investor (rabbu l-mal) and the entrepreneur (mu-arib) within a mu-araba contract. Therefore, an unbiased quantitative estimation of the risks and returns is desirable so as to put Islamic principles to work.

We argue that Islamic principles, in particular the avoidance of riba¯ and gharar, should be applied to real economic value in the first place, and not a priori to a monetary value in terms of conventional currency. In order to reconcile monetary value with economic value, we propose a reference currency linked to an appropriate commodity basket, reflecting the common economic realities and needs of the respective monetary union. Based on this currency, real economic value can be computed in analogy with conventional financial engineering methods.

In order to reflect global economic needs and realities, a global reference currency should be linked to a basket of commodities including in particular the natural resources necessary to ensure both, sustainable survival of mankind and a sustainable living standard above poverty.

Referring to the recent financial crisis of the European Union, we argue that apart from the common economic realities and needs within a given socio-political union, such as the OIC countries, also the different realities and needs should be honoured appropriately. We propose a 3-level construction of reference currencies, reflecting the economic realities and needs globally – for each region, and for each country.

We compare conventional financial engineering, based on zero bond numéraires computed from fixed income forward contracts, with Islamic financial engineering based on numéraires computed from bay’u l-salam or/and forward contracts on the basis of the reference currencies relevant for the counter-parties of the contract.

We propose that contract valuation and risk management should be performed on the basis of Islamic financial engineering rooted on the reference currencies reflecting the economic realities and needs relevant for the counter parties. Considering the benefits of such a risk management for social economy, particularly when the Organization of Islamic Cooperation (OIC) countries are considered, we argue that an implementation of the described Islamic financial engineering

Aydin et al.


enables Islamic risk management making transparent expected return and risks and enables their fair distribution between the counter-parties.

Islamic financial industry following the guideline of such principles of Islamic financial eng-ineering would be able to contribute to sustainable development by (i) more risk-(and-return)-consciousness reflected in the participatory structure of financial contracts, and thus, (ii) encouraging Islamic financial institutions to innovate financial products consistent with real implementation Islamic principles, reflecting the real necessities of modern business and economy.

appears: Iceland had its own currency, which helped to soften the consequences, but Greece is currently trapped within the EUR monetary union. Recently enterprises and banks already simulate scenarios for the reintroduction of the Drachma1 in Greece. The point that was missed, when Greece entered 2001 for the second time2 into a monetary union, is that a common currency can be sustainable only within a political region that is operating sufficiently within its economic realities and needs. If some important factor such as economic productivity differs largely among the participating countries, an artificially introduced monetary union creates artificial fluxes and processes that may be in contradiction to the economic reality of some country. This recently became evident with Greece again. The idea that a monetary union of a region like Europe is a priori going to strengthen its political union has proven again to fail, because in the case of Greece, the EUR did not reflect Greek economic reality, with economic mismatch finally resulting in political instability. We will come back to this point below when we discuss the construction of reference currencies reflecting real value of economy.

On the background of financial crises in Europe, what can be learned from Islamic finance? On the Banque de France conference, March 4, 2011, Kenneth Rogoff of Harvard University commented: “Western policymakers and economists often portray Islamic financial systems, with their emphasis on shared risk and responsibility in lending, as less efficient than western systems that put no strictures on debt. Yet one can equally argue that Western financial intermediation is far too skewed towards debt, and as a consequence generate many unnecessary risks.”

In the Muslim world, conventional agreed interest loans are commonly objected for the fact that they avoid a sharing participation in the default risk related to the purpose of the loan. For the same reason, transfer of risk by selling it to another party is often objected. Within the Islamic finance community, this sometimes appears to create a far spread impression that a passive attitude towards risk would be more ethical rather than an active management of the risks involved within a project and related contracts. We would like however to argue that both, transparency about all risk involved, and clear agreements how they are to be shared should be part of any contract. Also we believe that for a single agent, entrepreneur or investor, it should be legitimate to use his portfolio of contracts with different counter-parties in order to minimize his own exposures to different types of risks. In fact, avoidance of gharar is a fundamental request according to Islamic discourse, e.g., numerous hadiths forbidding gharar sales. Having accepted this, it is becoming clear that, modern tools of risk

1. IntroductionRecent decades have witnessed an explosion in financial innovation and engineering of novel contracts. First of all, from the so-called “conventional banking system” emerged increasing the need to adapt to the challenges from the demand side. Secondly, particularly also within the so-called “Islamic finance industry,” the incentive to innovate around prohibited and disadvantaged transactions has been unfortunately high. A clash between abused financial engineering and Islamic principles seems to prevail, not only since the renaissance of Islamic finance, and not in the Muslim world only. In medieval Europe canonical law prohibited usury, i.e., riba. European merchants, however, used a combination of contracts called contractum trinius to circumvent prohibition of usury. The construction uses put-call parity to synthesize a conventional interest-bearing loan. During the recent renaissance of Islamic finance, this construction prevailed in practices of some so-called “Islamic” banks, when they include into murabaha contracts a combination of conditions similar to contractum trinius, thus effectively ensuring and concealing a risk-free profit, which might also be called “regulatory arbitrage.” Exceeding considerably the inflation rate, such a so-called “risk-free return” is nothing else but usury, i.e., riba. Such a combination of contracts in order to circumvent Islamic principles may be considered as product structuring that is a part of conventional financial engineering.

Here and below, we will be concerned however not so much with product structuring but instead with quantitative valuation of financial instruments. Careful computation of expected value, expected return, and expected risks is a prerequisite not only for risk management, but also already for thorough valuation of compliance of contacts with respect to regulatory requirements including compliance with Islamic law. A clean product structuring, therefore, depends on a serious valuation of value and risk as the more fundamental and more demanding section of financial engineering, which we will deal with subsequently.

In Europe in particular, the growing public debt of several European Union (EU) member countries has surfaced already in several financial crises, which were escalated by salvation programs for financial institutions. According to IMF 2010, the public debt of Italy and Greece is significantly exceeding annual GDP, while others (like Portugal, Ireland, and Belgium) are following with over 90% of GDP. Continuing conventional public loan policy, it is likely that European public debt will exceed all possible salvation funds in the long run. Comparing the 2008 financial crisis of Iceland with the 2010 crisis of Greece, some important difference


Concept and mathematics of Islamic valuation and financial engineering

management and hedging should be applied appropriately using all the knowledge we currently have in order to reduce risk exposures. Quantitative evaluation of risks involved therefore is a necessary precondition in order to enable the counterparties to first obtain transparency about the existing risks and, on the basis of this information, to reach a fair agreement about the mutual distribution of risks. As far as Islamic finance is concerned, we currently perceive commonly loose interpretation of quantitative evaluation, and last but not least, avoidance of mathematics as being too sophisticated. However we’d like to remind that striving for knowledge is any Muslim’s duty, rather than condemning it plainly as misleading since the mathematics involved appears too complicated for many non-experts.

One historically grown reason for the adverse attitude of many Muslims towards financial mathematics and, in particular, the fair value approach comes from the conventionally common practices of discounting future cash flows with so-called “risk-free” interest rates often derived from forward rates of inter-bank markets and certain government bonds. Indeed these rates are deceptive. First because they are not really “risk-free” as their name suggests. Secondly, their level is considerably higher than sustainable, because they are indeed driven by and related to fixed interest loans. Moreover, the high level of these rates also contributes to push inflation rates up. Such objections against the common discount curves, based on the aforementioned interest rates, are fully legitimate. Below we will argue indeed for a more flexible and adequate construction of discount curves. We will also point out that the mathematical framework of fair value—based on relative prices with respect to some reference asset—does not require at all that the reference asset should be given by an artificial zero-coupon bond linked to “risk-free” discount rates, which, in turn, relate to fixed interest loans and the conventional inter-bank markets for forward rates. The mathematical framework of fair value works perfectly also with a universal commodity-linked currency or even equity as a reference asset.

2. Islamic finance for sustainable developmentThe role of financial institutions is to provide the capital for projects, in particular, for those projects that are useful or even vital for development of societies with sustainable infrastructures. Notwithstanding, many financial institutions (e.g., in Europe) that had been shaken by the last financial crisis, have been observed to refuse credits or offer credits only with unbearably high risk premiums. This attitude has severely threatened the existence of in particular local small- and medium-size family businesses and enterprises, since these suffer from additional discrimination by the traditional rating systems favoring large international players instead. As it is known, however, the economic and innovative power of the society is driven very much by locally rooted small and medium size enterprises, rather than the big transnational players. This is known to hold true in Europe, as well. And it is very likely to hold similarly for the group of OIC member countries that have for long been striving similarly for their sustainable development, while facing similar challenges along the way.

As a consequence, it is the task of each government to take care that the financial institutions follow up their duty of

providing the required capital for small and medium size companies in order for them to remain operational and continue making essential contributions for an innovative and vital economy. Following the financial crisis of 2007, new regulatory frameworks such as Basel III encourage credit institutes to impose more severe conditions and tougher rating conditions on entrepreneurs for credits. The regulatory requirements now became the pretence of credit institutes for harsher credit conditions, contradicting to generally decreasing interest rates in Europe, particularly in Germany. Government tried to intervene on this with several measures to enable and to push financial institutions to follow up their duty of providing liquidity for enterprises. However, by the time being, Europe’s financial industry reacts only very reluctantly. In this aspect, governments of most OIC countries should be in a better position not only because the influence of governments on the domestic financial sector has traditionally been more powerful, but much more also because the basis of understanding between governments and financial industry is derived from the common rules of Shariah. In this context, a very different culture of entrepreneurship has shaped the Islamic finance industry with the commonly accepted participatory means of financing, fixed income products being obsolete.

In the following, we will sketch some essential mathematical aspects of Islamic finance enabling financial engineering similarly easy and rigorous as in the conventional interested-based finance.

3. The myth of risk-free interest and fixed incomeIn this section we will argue that the notion of a risk-free fixed income is a myth kept up by the conventional banking sector, rather than an economic reality. Corresponding rates are set mainly by the agreement of an inter-bank market among conventional creditors. To calibrate a current credit contract to such rates may be questionable particularly for the situations where one or more counterparties of the contract have limited or no access to this inter-bank market.

After we have seen that fixed interest rates by themselves are not suitable as a basis for a risk-neutral measure, we conclude that a new basis for such a measure must be sought. This should be done by taking into account the position of the agent in the market, i.e., by carefully investigating her exposure to the various risks of the relevant markets and economies involved. Furthermore, credit spreads should be considered more flexible, i.e., not constant, and not necessarily always positive. A negative credit spread for some period would reflect the possibility that the credibility of the considered enterprise is in fact better than that of the reference.

It is one main issue in Islamic finance that a supposed risk-free interest rate should be close to zero. The reference rates of some countries like Switzerland and Japan have already come very close to this, since many years. Also in Europe, after the financial crisis of late 2000s, interest rates dropped drastically. This indicates that the economical reality in interest-based financial markets might essentially honour – sooner or later – the fact that there is no free lunch in any market and that the assumption of a risk free return is a myth.

Aydin et al.


4. Inflation of currencies, time-value and time-less valueTo the extent that inflation of any currency is inevitable, a time-value that compensates the expected inflation rate should be disputable. If we decide that full participation in the inflation risk is not bearable to the investor, at least a partial compensation of hedging against inflation risk should be admitted. On the other hand, one might argue that investor and entrepreneur perhaps share some common risks, such as the risks of everyday life, and accordingly, the inflation risk – as it is inevitable to both parties – should also be shared among them rather than being put on one party’s shoulders.

Real contracts usually involve several cashflows and/or depend on asset values at different times. A fair contract evaluation requires the ability to compare cashflows and/or asset values at different times. The nominal value of assets is usually measured in units of a certain currency. However, the real economic value of this currency may changes with time, e.g., due to inflation. For our fair valuation, we are nevertheless interested in the intrinsic real value of the currency, and of our assets. E.g., in the face of inflation, this real value of assets might be measured by inflation-adjusted prices. These are obtained by adjusting future cash flows by discount factors and account just for the expected inflation of the currency.

Viewed superficially, discount factors might be objected for introducing undesirable time-value to cash flows. However, if they are chosen just such to compensate for the time-dependency of the nominal value of the currency, they in fact may yield in fact just the desired timeless real value standard. We would like to emphasize that the timelessness of value as suggested by Islam has to be requested for the real value rather than the nominal value in the first place.

Islam advocates in fact the use of time-independent measures for value. At times of the prophet (pbuh) gold was much more rare than today, since there was no mining industry yet. It was a fairly good inflation-free currency. Its value was stable and hardly to be influenced. Only exceptional political events such as conquests and sieges could trigger sudden regional gains and losses of huge gold treasures, which then could indeed change locally the real value. Except such extreme events, the real value of gold was stable. Today, however, developments of the mining industry, demands from high technology, and different market pressures apply almost continuously on the value of gold in different directions. Although gold might be still more stable than most paper money, its real value has become much more volatile than at times of the prophet (pbuh).

Hence the challenge is to find the reference assets, which represent a timeless stable standard of value, similar to gold in previous times.

5. The gold of 21st century: Back to commodity-based currencyFrom the very early times of Islam, up to the 20th century, the commodities of gold and silver have played an important role in defining the modern currencies. The histories of the GBP and USD in this respect and, in particular, their

successive decoupling from their reference commodities simultaneously with their devaluation is described, e.g., in El Diwany (2010).

Within the EU, after some period of fixed cross currency rates, the Euro was introduced. Less known however is that an early predecessor of the Euro was defined already in the early 1930s. The universal European currency intended as a “currency for peace” was called “l’Europa” (Le Fédériste (1933)). It was defined as a basket of several valuable commodities. Much later, Lietaer (2001), introduced a global Trade Reference Currency (TRC), dubbed also as “Terra,” comprising a basket of a dozen internationally circulated currencies.

One advantage of the general concept of a basket of commodities underlying to a reference currency is the increased stability. With its currency linked to a basket of commodities and its monetary authority backed 100 percent by a sufficiently large reserve of these commodities, artificial depreciation (appreciation) of a country’s local currency (domestic prices) would come to an end, together with an improvement in its immunity to the associated monetary and fiscal challenges facing its economy.

An inevitable effect of any commodity-based currency is the increase of efforts for production of or mining for the underlying commodities. Taking into account the challenges of 21st century and beyond, previous choices of baskets did not yet account sufficiently for the aspect of sustainability and desirability of the production of the commodities chosen for the basket.

According to Islam, at times of the Prophet Muhammad (pbuh), exploration and production of gold was still desirable as the most precious commodity known at that time. It was not yet challenged by an excessive mining industry with all its social and ecological problems. From our current modern point of view, in our opinion, the negative effects related to the intensive industrial production, such as ecological damages and exploitation of workers, cannot be ignored. In essence, what is true for gold and other metals is also true for oil and gas as well as agricultural commodities. For any commodity, consideration should me made whether and to what extent its production is desirable. So from an Islamic perspective, we would like to argue in favor of a historically conscious reading that the “gold” mentioned in the Holy Qur’an should be read just as synonymous for the “most precious commodity.” If we ask ourselves – from the perspective of the global challenges ahead of us – what are these “most precious commodities”, we might identify such commodities as cleanly produced renewable energy, clean drinking water, and agricultural products produced according to ecological standards.

6. Euro lesson: A multi-dimensional currency system reflecting diversity and community appropriatelyThe Euro currency was introduced well-intended, with the idea to enhance the political and economical solidarity between the countries involved and the ultimate goal to make them together more competitive in the world markets arena. However, the recent Euro crisis has shown that as long as the economic foundations of the constituents are basically



too different, the hard entry of one or more countries into a currency union may be a problem. If productivity of the countries highly differs, the stronger countries have to aid the weaker ones in order to avoid a potential collapse in the common currency. As a result, the political disputes going along with this process not only put the common currency union at risk, but even worse may damage the political union, as such concerns have been central to the Euro crisis.

For example, the EC currently is split about whether to introduce Euro bonds or not. The introduction of Euro bonds would be consequent since there exists already a common currency. The opponents however argue in view of the different economic realities within the member countries. The dilemma is that, any currency has to reflect the economic realities, but the EUR member countries now discover that they will not be able to sufficiently converge on their economic realities.

In our opinion, the choices for countries that belong to a certain political entity (e.g., the EU) to either share a common currency 100 percent or withdraw from the monetary partnership, are insufficient in number in the face of the current economic and political realities. The currently practiced “0 or 1” hard entries into a currency system apparently ignore the risk that a similar hard exit from it may endanger the whole system. Therefore, we propose a structured system of world currencies, starting from a global reference currency Cw. It is defined by a basket of commodities S1 … Sn whose productions are globally desirable, i.e.:

Cw =

=∑ w Si ii

n

1

(1)

Similarly, we may consider a basket Cr of commodities, the production of which is desirable just within a certain region, e.g., the political entity of the OIC member countries. Cr acts as the currency of this region. ΔCr: = Cr - Cw accounts for structural differences of the region from the rest of the world. It may also give additional (non-negative) weight to some of the commodities in Cw, e.g., clean drinking water, which is already within the global world basket. It may also contain new commodities specific for the region. Hence:

ΔCr = +

= = +∑ ∑w S w Si

ri

i

n

ir

ii n

m( ) ( )

1 1

(2)

Finally, we admit an extra basket for individual countries in order to take into account their particular situation, differing from that of their region as well as the world at large:

ΔCc = +

= = +∑ ∑w S w Si

ci

i

m

ic

ii m

l( ) ( )

1 1

(3)

The currency Cc of any country will be composed from a worldwide defined fraction bw of the world currency, the regionally-defined fraction br of its regional commodity basket, and the remaining local country fraction (1-bw-br) of its local commodity basket, i.e.:

C C C Cc w w r w c= + + - -β β β βr rΔ Δ( )1 (4)

The fraction bw should be negotiated on a world conference, common for all regions. It should be as high as possible in order to meet global challenges. However, the case where bw < 1 is desirable in order to admit the differences according to the particular situation of different regions of the world.

Similarly, the fraction br should be negotiated on a regional conference, common for all countries. It should be as high as possible in order to meet common regional challenges. However, br < 1 is desirable in order to admit in this case the differences according to the particular situation of the different countries of the region.

By construction, the higher the value of bw, the more strongly any two world currencies are correlated. The same is true for the currencies of the countries within a region: the higher the value of bw + br, the stronger the correlation between the currencies of any two countries within a region will be. In particular, the negotiable weight br may give to a region, e.g., Europe, the flexibility to regulate a partial entry into a currency union with some weight br, realistically reflecting the degree of economical unification.

7. Fighting usury: The risk-neutral fair valueThe value of an asset can equivalently be quoted in the form of a price or in the form of a rate, which is relative to a reference price. Hence, a quoted rate per se does not yet imply riba, while, vice versa, quoted prices may conceal riba. The form of the quotation, whether it is given price St, or rate, Rt, is irrelevant to the question of usury.

The equivalence relation is:

S tS

R t( )( )

( )0

1= + ⋅t (5)

When St is unknown and subject to risk, it will be stochastic and one has to consider its expected value EQ[St] under a certain risk measure Q. The conventional present value (at time 0) of the stochastic variable S is then given by:

S

tR tE S tQ( )

( )[ ( )]0

11

=+

(6)

The spot rate R(t) may equivalently be converted into a discount factor, corresponding to the initial value P(0, t) of a zero-coupon bond with a nominal value of P(t, t) = 1 and time-to-maturity t, i.e.:

P t

tR t( , )

( )0

11

=+

(7)

The relation (6) above may then be written as:

SP t

E S t

P t tE

S tP t t

QQ

( )( , )

[ ( )]

( , )( )

( , )0

0= =

(8)

Aydin et al.


This is simply the martingale expression for the relative price of S with respect to the zero-coupon bond P(·,t), whose price is inversely proportional to the spot rate R(t).

In order to determine whether a certain quoted price or rate implies usury, it is rather important to have at hand an independent method to determine a fair price with a risk-neutral position. The expected price should be computed under a measure Q, which in particular compensates any risk that is out of the responsibility of the considered counterparty. The computed value is supposed to match as close as possible the realistic economic expectation.

Conventional finance usually assumes depreciation of cash flows, e.g., according to S(t), in a certain currency during time. It further assumes that the appropriate way to account for this is to discount with a particular spot rate curve R(t) derived from forward rates F t F t T Ti i i( ) ( ; , )= -1 corresponding to quoted interest rates from certain inter-bank markets (e.g., LIBOR).

From an Islamic finance perspective, the questionable point here is the market from which the quotes are taken. The method of calibration of the expected spot value to the forward curve by itself agrees with the principle of fair value.

Even within the still very conventional setting of R(t) and P(·,t) and relations (6) and (8), there is already the flexibility to adjust the rate R(t) (and equivalently the zero-coupon bond) to the any expected time-dependent performance. It is not at all required that it is linked to the conventional inter-bank markets for deposits or swaps.

If it is linked to performance of equity, commodity, or (as we will consider below) to currency, it may be consistent with principles of Islamic finance. Also compensation of inflation of a currency may be considered as legitimate under circumstances that the counterparty cannot bear to be exposed to inflation risk, e.g. because it is too small and does not have the possibility to protect itself. In this case R(t) could be linked to inflation rate.

Concerning relation (2) above, remarkably, in mathematical finance the meaningful quantities are mostly relative. In particular, price dynamics is investigated mainly with using the relative price ˆ: /S S S= 0 with respect to the price S0 of a reference asset. Several desirable properties for the reference asset are usually postulated. Best stability of its value is one of them. The more the price is stable, the lower its volatility, i.e., the less risk of change is connected to the asset price. In the past, the commodity gold was considered as the most stable physical asset with the least volatility of its price.

Note, however, that considering “risk-free” zero bonds or rates is unrealistic.3 Therefore, we propose, instead of the zero-bond linked to a deterministic rate R(t) to consider another reference asset instead, reflecting the stochastic risk. We propose the commodity-linked currency (4) as a natural reference asset for this purpose. With this choice, relation (8) will be replaced by:

S S tt

( )( )

( )( )

00C

ECcountry

Qcountry

=

(9)

Note that this relation holds, independent of the auxiliary currency units within which S and Ccountry are evaluated. This invariance is an advantage of relative prices. In general, the relative price in (8) will be stochastic. However, the relative price of Ccountry to itself becomes trivially constant equal to 1.

8. Conventional versus Islamic valuationLet us give first a simple description of a forward contract and its conventional valuation. A forward contract agrees to a fixed price: the fair forward price F to be paid at delivery time T in exchange for an asset S (say a commodity) having present value S(0) and unknown value S(T) at delivery. A fair value forward contract on some traded asset S should always have value 0 at time 0, the time of contraction. Using relative prices w.r.t. zero bonds having price P(t, T), conventional financial engineering demands:

0

00

00

= - =S

P TF

P T TF

SP T

( )( , ) ( , )

,( )

( , ).or equivalently (10)

Hence P(0, T) acts as a discount factor. Note that (10) implies a nominal value concept with respect to some conventional currency but tries to compensate this by taking zero bonds as numéraire. Above, (10) is postulated directly, with deterministic P(0, T). Note that, for independent unknown stochastic prices S(t) of the asset and B(t, T) of the zero bond, it holds

E

( ) EE

EP

P

P

PS tB t T

S tB t T

S tP t T( , )

[ ( )][ ( , )]

[ ( )]( , )

,

= = (11)

with deterministic prices P(t, T) = EP[B(t, T)], from which (10) follows immediately, assuming F = EP[S(T)]. The independency assumption may hold for equity or commodity assets, when it appears plausible that such assets develop essentially independent from inter bank markets determining B(t, T). When asset prices are non-trivially correlated with the stochastic discount factors, the above derivation of (10) has to be replaced by:

F S T

S TB T T

S tB t T

SP T

= =

=

=E E

( )E

( )P P P[ ( )]

( , ) ( , )( )

( ,0

0 )), (12)

whence again the fair price of the future contract is simply S(0)/P(0, T).

Let us now consider the valuation of a forward contract within an Islamic financial engineering framework relative to a reference currency C. According to (9) above, we replace (12) by:

E

( )E

( )Q Q

S TC T

S tC t

SC( ) ( )

( )( )

.

=

=00

(13)

Unlike (12), where B(T, T) = 1 by definition of the zero bond, here C(T) is in general still unknown, stochastic. Furthermore the stochastic forward price S(t) and the stochastic currency C(t) are unlikely to become either independent or fully dependent at maturity. Then only



value that can be estimated is their relative value. The fair relative value w.r.t. to the given currency is constant. For the future contract hence the fair price should be specified as relative to the reference currency, simply as S(0)/C(0). Remarkably, the fair price is quite analogous to the conventional future price, which also can be viewed as the relative price w.r.t. a zero bond.

Important from the perspective of Islamic principles is that, in the conventional case the zero bond numeraire is a pure interest based asset, which is linked essentially to inter bank markets, while in the Islamic case the numeraire is a currency, which we propose to be linked to a commodity basket supposed to reflect a more appropriate and relevant economic reality.

Now let us consider the special case of full dependence, S(t) = a · C(t), i.e., the traded asset is proportional to the reference currency (with constant factor a), and hence proportional to the commodity basket defining C. Then (13) yields:

E

( )(0)

E [ (T)]

E [ (T)] E [C(T)]QQ

Q Q

S TC T

SC

aS

CF

( )( )

,

= = = =0

(14)

and therefore F = a · EP[C(T)].

Comparing the conventional with the Islamic approach, a practically important difference is that the zero bond prices are computed conventionally purely deterministically from available quotes of money market, future, and swap prices; while for the reference currency and the commodity basket their forward prices are currently not easily available from quotes. Their estimation is however possible via calibrated stochastic processes. When the asset S(T) is non-trivially correlated with the reference C(T), then stochastic simulat-ion of their relative value may be necessary in order to determine the expected value in (13). This shows the need of an analogous stochastic forward rate model for commodities in analogy to the conventional LIBOR market model.

We finally emphasize that, the numeraires in (12) and (14) are not equivalent numeraires, i.e., they are not resulting from transformation of each other by the change of numeraire theorem.4 They are rather the different choices for the primary numeraire, defining the martingale measure P, respectively Q, i.e., defining what is meant by fair value.

This fair value according to (14) becomes more than another arbitrary definition exactly when the reference currency C reflects a real value linked to the real economy. Only in this way the Islamic choice (14) becomes really superior to the conventional choice (13). Otherwise, “value” would remain just a conventional concept. This is the case in the conventional case, since the interbank interest markets behind zerobond curves for conventional currencies do in most cases not really reflect the economic reality to which the counterparties are exposed. This becomes evident particularly in situations, when the finance industry meets some of their homemade crises. Then, the conventional zero curves tend to exhibit artefacts from the conventional banking sector. In particular the bias towards short maturities introduces basis spreads into the forward rates traded and quoted in the conventional sector. Since

the 2007 financial crisis, it has been realized now within the conventional banking sector that, because of the basis spread included in quoted forward prices, the discount curve which is supposed to yield fair arbitrage-free prices, can no longer be equal to the (inter bank) market forward curve, rather a 2 curve approach becomes necessary.5 The interpretation is that the conventional inter bank markets are biased, rather than risk-neutral.

9. Risk profiles and equivalent martingale measuresLet us now consider the dynamics of the relative price,

( ) ( )/ ( )S t S t S t= 0 . Since a future price S(t), t > 0, is subject to the risk of change, its expected value EQ ( )S t is of particular importance, since it is the most objective value one can assign to them. The expected value EQ depends on the measure Q related to the risk of change. Under the assumptions of no arbitrage and a complete market, there exists a unique measure Q with a risk-neutral expectation, i.e.:

EQ ( ) ( ).S t S= 0 (15)

This is called the martingale measure.

However, if the market is not complete, e.g., due to illiquidity, extreme events, or other reasons, the martingale measure Q and the derived arbitrage-free value are no longer unique. In the case of an incomplete market, it makes even more sense to consider also alternative reference assets. In particular, the reference curve R(t) may be chosen according to a particular tailor-made dynamical risk-aversion profile, agreed by the counterparties of a contract. A particular risk aversion profile corresponds to a certain utility function, which in turn may be used to select a particular martingale measure among several equivalent ones. One might even consider time periods within which R(t) becomes negative.

The important issue from the Islamic perspective is the mutual agreement about the risk profile implied by the choice of R(t).

Determining R(t) rather freely according to a mutually agreed risk-profile, rather than determining it from the corresponding reference currency, might be interesting particularly for direct contracts between investor and entrepreneurs, where both of them wave developed a common understanding of the risks involved in the project, and correspondingly have consciously both agreed on a particular non-standard form of the R(t) curve. Because for the project under consideration, this choice is more suitable than the plain comparison with the value of the currency.

The last mentioned alternative approach of direct contractual agreement on a risk-profile and corresponding R(t) requires outmost transparency about the risks involved in the contract. Therefore, in cases where there is no detailed understanding and clarity about risks and the counter party attitudes appear to be rather indifferent, then it might be preferable to stick to a standard choice for the reference asset, namely our proposed commodity-linked currency.

Aydin et al.


10. Consequences for Islamic derivatives as hedging instrumentsWe exemplify the consequences of our modified evaluation framework for the design of Islamic derivatives. It is commonly agreed that such derivatives should be permitted only for the purpose of hedging.

For example, in Jobst 2010 three examples of synthetic instruments from asset-based investment finance have been given. The valuation nevertheless has been based on conventional measures of value using risk-free discount factors. As a consequence, the call-put parity could be used to synthesize a conventional loan.

If the present value would be computed using our proposed reference asset-based currency, the value of a certain amount of this currency would be constant without any interest. Call- put parity can no longer be used to synthesize a conventional interest-based loan. Hence motivation for such constructions will become void.

11. Concluding remarksWe have shown that the mathematical framework of risk-neutral valuation and arbitrage free contract prices may be applied for Islamic finance, similarly as in conventional finance, provided some modification on some conventional inputs. In particular, the conventional discount rates drawn from inter-bank markets forward interest rates have to be replaced.

A rather mild modification for this goal consists in replacing them by performance rates of underlying equity or commodity. Shifting the reference is from fixed income interest to equity or commodity, compatibility with Islamic principles might be achieved.

As a more challenging method, we propose to replace the conventional interest-based zero rate curve by an alternative reference asset. We propose in particular a commodity-linked multi-dimensional reference currency system. Globally negotiated fraction bw of the world part, and regionally agreed fraction br may provide easy political instruments to regulate the degree of required homogeneity of currencies globally or within a defined region such as Europe respectively.

The most challenging approach is nevertheless the one that leaves also the most contractual freedom and responsibility at the same time to the contracting counterparties. Provided all counterparties have both, the necessary capability for risk analysis and commitment for risk transparency then they might agree on a particular zero rate curve freely, according to their commonly-agreed risk aversion profile. The zero rates in this case may be linked closely to their common judgement of risk for the project underlying to the contract. The zero rates in this case may be even negative, meaning that a later cash flow is strategically preferable to an earlier one.

Notes1. From 1831 to 2001 the Greek currency was the

Drachma.2. From 1869 to 1914 the Drachma was effectively coupled

100% to the union monétaire latine.3. Even a conventional fixed income market in fact is not

risk-free, but suffers in particular also from the risk of changing market interest rates.

4. See e.g., Brigo & Mercurio 2007.5. See e.g., Bianchetti 2009 for the theoretical background,

and Ametrano & Bianchetti 2009 for practial implementation of the 2 curve approach.

ReferencesAmetrano FM, Bianchetti M. (2009) Bootstrapping the

Illiquidity: Multiple Yield Curves Construction for Market Coherent Forward Rate Estimation. SSRN Working Paper.

Anonymous. (1933) L’Europa – M nnaie de la Paix. Le Fédériste. 1 January.

Al-Bashir M, Al-Amine M. (2005) Commodity Derivatives: An Islamic analysis. In: Iqbal M, Khan T. (2005) (Eds.). Financial Engineering and Islamic Contracts. Palgrave MacMillan. New York. 58–98.

Brigo D, Mercurio F. (2007) Interest Rate Models–Theory and Practice, 2nd Edition. Springer Verlag, Heidelberg.

Bianchetti M. (2009) Two Curves, One Price – Pricing and Hedging Interest Rate Derivatives Using Different Yield Curves for Discounting and Forwarding. SSRN Working Paper.

El Diwany T. (2010) The Problem with Interest, 3rd Edition. Kreatoc, Ltd. London.

Hamzawy A. (2003) Hatar al-’Aulama: Die Arabische Globalisierungsdebatte – Eine Neuauflage der Kontroverse über Moderne und Authentizität? Welt des Islams. 43(2):173–213.

IMF. (2010) Data and Statistics. World Economic Outlook Data Base. July 2010. Available at: http://www.imf.org/external/pubs/ft/weo/2010

Jobst A. (2010) Derivatives in Islamic Finance. Paper presented at SSRN. Preprint.

Latif AM, Janahi AR. (2005) Shariah Alternatives to Government Bonds. In: Iqbal M, Khan T (Eds.). Financial Engineering and Islamic Contracts. Palgrave MacMillan. New York. pp. 99–122.

Lietaer B. (2001) The Future of Money. Random House. London.

Qutb S. (1949) Al-Adala al-Ijtima’iyya fi’l-Islam. Translation by Hardie JB (1980). Social Justice in Islam. Octagon Press. New York.

Rogoff K. (2011) Summary of Comments. Banque de France Conference. March 4.

Schacht J. (1982) An Introduction to Islamic Law. Clarendon Press. Oxford.

Documents

Developing Inclusive and Sustainable Economic and Financial … · 2016. 5. 18. · Developing Inclusive and Sustainable Economic and Financial Systems Cite this chapter as: El-Karanshawy