34
Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India* JAYATI SARKAR AND SUBRATA SARKAR Indira Gandhi Institute of Development Research, Mumbai, India ABSTRACT Most of the existing evidence on the effectiveness of large shareholders in corporate governance has been restricted to a handful of developed countries, notably the UK, US, Germany and Japan. This paper provides evidence on the role of large shareholders in monitoring company value with respect to a developing and emerging economy, India, whose corporate governance system is a hybrid of the outsider-dominated market-based systems of the UK and the US, and the insider-dominated bank-based systems of Germany and Japan. The picture of large- shareholder monitoring that emerges from our case study of Indian corporates is a mixed one. Like many of the existing studies, while we find blockholdings by directors to increase company value after a certain level of holdings, we find no evidence that institutional investors, typically mutual funds, are active in governance. We find support for the efficiency of the German/Japanese bank-based model of governance; our results suggest that lending institutions start monitoring the company effectively once they have substantial equity holdings in the company and that this monitoring is reinforced by the extent of debt holdings by these institutions. Our analysis also highlights that foreign equity ownership has a beneficial effect on company value. In general, our analysis supports the view emerging from developed country studies that the identity of large shareholders matters in corporate governance. ß International Review of Finance Ltd. 2000. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA. International Review of Finance, 1:3, 2000: pp. 161–194 * We would like to thank Sudipto Bhattacharya, Sudipto Dasgupta, Jayendra Nayak, Sheridan Titman (editor), an anonymous referee, the participants of the conference on Financial Market Development in Emerging and Transition Economies, held at London Business School, September 2000, and the participants of the India and Southeast Asia Conference of the Econometric Society, held at Indian Statistical Institute, New Delhi, December 1998, for their helpful comments and suggestions. The usual disclaimer applies. An earlier version of this paper appeared as IGIDR Discussion Paper No. 153 in January 1999.

Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Embed Size (px)

Citation preview

Page 1: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Large Shareholder Activism inCorporate Governance in

Developing Countries: Evidencefrom India*

JAYATI SARKAR AND SUBRATA SARKAR

Indira Gandhi Institute of Development Research, Mumbai, India

ABSTRACT

Most of the existing evidence on the effectiveness of large shareholders incorporate governance has been restricted to a handful of developedcountries, notably the UK, US, Germany and Japan. This paper providesevidence on the role of large shareholders in monitoring company valuewith respect to a developing and emerging economy, India, whosecorporate governance system is a hybrid of the outsider-dominatedmarket-based systems of the UK and the US, and the insider-dominatedbank-based systems of Germany and Japan. The picture of large-shareholder monitoring that emerges from our case study of Indiancorporates is a mixed one. Like many of the existing studies, while we findblockholdings by directors to increase company value after a certain levelof holdings, we find no evidence that institutional investors, typicallymutual funds, are active in governance. We find support for the efficiencyof the German/Japanese bank-based model of governance; our resultssuggest that lending institutions start monitoring the company effectivelyonce they have substantial equity holdings in the company and that thismonitoring is reinforced by the extent of debt holdings by theseinstitutions. Our analysis also highlights that foreign equity ownershiphas a beneficial effect on company value. In general, our analysis supportsthe view emerging from developed country studies that the identity oflarge shareholders matters in corporate governance.

ß International Review of Finance Ltd. 2000. Published by Blackwell Publishers, 108 Cowley Road,Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.

International Review of Finance, 1:3, 2000: pp. 161±194

* We would like to thank Sudipto Bhattacharya, Sudipto Dasgupta, Jayendra Nayak, SheridanTitman (editor), an anonymous referee, the participants of the conference on Financial MarketDevelopment in Emerging and Transition Economies, held at London Business School, September2000, and the participants of the India and Southeast Asia Conference of the Econometric Society,held at Indian Statistical Institute, New Delhi, December 1998, for their helpful comments andsuggestions. The usual disclaimer applies. An earlier version of this paper appeared as IGIDRDiscussion Paper No. 153 in January 1999.

Page 2: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

I. INTRODUCTION

The effectiveness of large shareholders in the corporate governance of enterpriseshas been the subject of much theoretical and empirical research. The issue hasgained increasing importance in view of some recent evidence that concentratedownership of corporations is much more prevalent across countries than isgenerally believed, and that the `Berle and Means image' of the moderncorporation being widely held by dispersed shareholders `has begun to showsome wear' (La Porta et al. 1999).

The literature on large shareholders focuses on the extent to which theseshareholders, as compared to the dispersed ones, are in a better position to makecompany management accountable. A scanning of the existing theoretical andempirical literature, however, reveals conflicting predictions and evidence on therole of large shareholders in enhancing corporate value (see the survey by Shleiferand Vishny 1997, and the references therein). The benefits of large shareholdingare highlighted under the `convergence-of-interest' and the `efficient-monitoring' hypotheses, which argue that large shareholders are likely to bemore efficient than small and dispersed shareholders in monitoring companymanagement, as the former: (a) have substantial investments as well assignificant voting power to protect these investments; (b) are likely to mitigatethe collective action problem present among dispersed shareholders; and (c) arelikely to engage in relational investing and be more committed to a company inthe long run.1 The potential costs of large shareholders are set out in terms of the`conflict-of-interest' hypothesis and the `strategic-alignment' hypothesis, withthe former arguing that large shareholders may pursue non-profit maximizingobjectives for their personal benefits to the detriment of the minority, and thelatter arguing that different types of large shareholders, like institutionalinvestors and managers, may often find it mutually advantageous to collude ina way to reduce company value and hurt the interests of minority investors.2

Most of the theoretical arguments and evidence on the costs and benefits oflarge shareholders have, however, been restricted to a handful of countries,notably the US, UK, Germany and Japan,3 and not much is known outside thesecountries. In particular, very little information is available with respect todeveloping countries (Singh 1995; Shleifer and Vishny 1997), where corporateownership is found to be heavily concentrated (La Porta et al. 1999). The objectiveof this paper is to provide evidence from a developing and emerging economy,India, on the role of large shareholders in monitoring the performance of

1 For arguments under (a), see Berle and Means (1932), Jensen and Meckling (1976) and Shleiferand Vishny (1986); under (b), see Dodd and Warner (1983); and under (c), see Blair (1995) andBlack (1998).

2 For arguments under conflict of interest hypothesis, see Fama and Jensen (1983), Shleifer andSummers (1988), Schmidt (1996) and Burkart et al. (1997); for strategic alignment hypothesis,see Brickley et al. (1988) and Pound (1988).

3 For a comprehensive account on the literature on corporate governance, mainly with respectto the triad, namely Europe, USA and Japan, see Hopt et al. (1998) and Keasey et al. (1999).

162 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 3: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

company management. Other studies of corporate governance in the context ofdeveloping and transition economies include those with respect to Russia, theCzech and Slovak Republics, China, a cross-section of transition economies andIndia.4 The advantage of undertaking a case study of India to address some of thegovernance issues specific to developing and emerging countries lies in the factthat vis-aÁ-vis other countries in this group, market and non-market institutions ofcorporate governance in India have evolved over a sufficiently long time,5 arewell defined and stable, and are relatively advanced with respect to the range anddepth of existing statutes and the legal framework regulating corporate activities.India is also representative of many developing countries in terms of its relianceon external sources of finance, as well as the prevalence of insider-dominatedfamily businesses.

Several of the costs and benefits arising from the presence of large shareholdersas highlighted in developed country studies could be equally relevant in thecontext of developing countries like India. However, some of the institutionalspecificities of developing countries, such as a less developed capital market, a lessactive takeover market, a greater involvement of government-owned financialinstitutions in corporate financing, a higher dependence on external sources offinance and the absence of a well developed managerial market, could impact onthe costs and benefits of large shareholding in these countries in some uniqueways, so that mechanically extrapolating the experiences of corporategovernance systems in developed countries may not yield the necessary answers.

One unique feature of developing countries is that despite having relativelyunderdeveloped equity markets, low investor protection and weak bankruptcyregulations, these countries rely much more heavily on banks and equity marketsto finance long-term projects than developed countries do (Singh, 1995). Thus,both equity and debt governance are of particular significance in developingcountries. In this context, the role of lending institutions like developmentbanks, many of which are government controlled, acquires particular significancegiven that these institutions often hold blocks of equity alongside debt contracts.

Another key feature of many developing and transition economies is the highand often `generic tendency' towards insider control and ownership6 that coexistswith relatively weak external disciplining mechanisms like less developed capital

4 For Russia, see Blasi and Shleifer (1996); for the Czech and Slovak Republics, see Claessens et al.(1996) and Claessens (1997); for China, see Xu and Wang (1997); for a cross-section oftransition economies, see Gray and Hanson (1993) and Aoki and Kim (1995); for India, seeKhanna and Palepu (1998), Chhibber and Majumdar (1999) and Sarkar and Sarkar (1999).

5 For instance, the history of the capital market in India dates back more than hundred yearswith the establishment of the Bombay Stock Exchange in 1875, and the existing CompaniesAct 1956, which governs the activities of Indian companies, has its roots in the IndianCompanies Act of 1913.

6 Aoki (1995, p. 3) uses the quoted phrase to characterize the privatization programmes of manypost-communist transition economies where ownership rights have been conferred toenterprise insiders, like managers and employees, to make privatization more acceptable(Frydman et al. 1998).

ß International Review of Finance Ltd. 2000 163

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 4: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

and takeover markets and a relatively weak `rule of law'.7 These characteristicsworking together can have opposing effects on company value. On the one hand,from the agency cost perspective, the combination of concentrated insider controland weak external disciplining mechanisms can exacerbate expropriationincentives by insiders. On the other hand, from the Coase±Williamson transactioncosts perspective, concentrated insider ownership can generate significant benefitsin situations where implicit trust-based contracting is widespread. As Kester (1992)argues in the context of Japanese corporations, the success of such contractingdepends upon maintaining a relatively continuous and stable managementstructure that can be effectively promoted by insider ownership, with interlockingdirectorates, cross-holdings and family control. This transaction cost perspective isespecially relevant for insider-dominated systems like India where, in view of weakand cumbersome legal systems, explicit market-based contracting is hazardousand implicit trust-based contracting is the norm. Whether such contractingindeed leads to benefits that are large enough to outweigh the agency costsremains an open question and is one of the issues analysed in this paper.

Some of the issues addressed in this paper have also been analysed with Indiandata in some recent papers (Chhibber and Majumdar 1999; Khanna and Palepu1999). The former analyses the relation between foreign ownership and companyperformance using the accounting measures of rate of return on assets and therate of return on sales, but does not focus on the larger issue of the role of othermajor shareholders in corporate governance. The latter examines the relativeefficiency of domestic financial institutions and foreign financial institutions ingoverning business group-affiliated companies and non-group companies, and iscloser in focus to our research. However, we specifically consider two governanceissues of relevance that are distinct from those addressed in the earlier papers.

First, in analysing the impact of large shareholders on company value we buildinto our framework the possibility that effective monitoring by owners ofcompanies in countries like India may also depend on the level of ownership, as isoften found in research work with respect to the US. Some insights on this issue,but with respect to only foreign ownership, can be found in Chhibber andMajumdar (1999). By adopting a spline methodology akin to several papers in thecorporate governance literature, particularly with respect to the US, we inquire asto whether there is any significant non-linearity in the relationship betweenownership and company performance and whether the interests of the ownersand the company converge as ownership stakes increase. Our results for each typeof large shareholder indicate that the incentives for monitoring changesignificantly when ownership stakes rise beyond a particular threshold. Suchchanging incentives are likely to remain uncaptured or at best show up in a weakform under a linear specification, as is the case in Khanna and Palepu (1999).

7 For instance, the average value of the `rule of law' index from La Porta et al. (1998) shows thatwhile the average value for the group of developed countries, namely the US, the UK,Germany, Japan, Australia and Canada, is 9.5, the value for developing countries, namelyIndia, Pakistan, Brazil, Chile, Mexico and Turkey, is 4.9.

164 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 5: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Second, in analysing the role of domestic financial institutions in governance,we examine if there could be differences in monitoring among different types ofinstitutions. Domestic financial institutions in India are state-owned and includethe development financial institutions (DFIs) and the institutional investors, i.e.mutual funds and insurance companies. Although an aggregate analysis mayreasonably assume DFIs and institutional investors to act as a single block by virtueof being state-owned (Khanna and Palepu, 1999), it is important to recognize thatthe type of holdings, and hence the resultant incentives to monitor, may be quitedifferent between these two groups. DFIs in India, like those in the bank-basedgovernance systems of Germany and Japan, hold both equity and debt, whileinstitutional investors hold only equity. As Prowse (1990) argues in the context ofcomparing institutional investment patterns and corporate financial behaviour,dual holdings in Japan may have reduced the `inherent principal±agent conflict'between shareholders and debtholders compared to that in the US, wherefinancial institutions do not have such dual positions. Similarly, Mulbert (1998)argues in the context of Germany that banks which act as both lenders and largeshareholders are not in the same position as `pure' investors like funds, and willassociate corporate control with objectives that are different from those of welldiversified shareholders.8 We indeed find support for such arguments in ouranalysis, and in the process, provide some fresh evidence on the yet unresolvedempirical question of the joint impact of debt and equity holding by financialinstitutions on corporate governance.9 Further, our analysis of institutionalinvestors, separated from the DFIs, also enables us to comment specifically on therole of mutual funds in governance in developing countries, an issue that hasreceived considerable attention in the developed country literature (see, forexample, Black 1998) and is increasingly coming under scrutiny in India.

The rest of the paper is organized as follows. Section II describes the ownershipstructure of Indian corporates and the institutions of corporate governance inIndia. Section III describes the data source, sample, variables and estimationmethodology. Section IV presents the empirical findings and section V concludesthe paper.

II. OWNERSHIP PATTERN AND INSTITUTIONS OF CORPORATEGOVERNANCE IN INDIA

Formal institutions of corporate governance in India have been in place for alarge number of years, though corporate governance issues came to the forefrontonly following the adoption of the structural adjustment and globalization

8 Similarly, Miyajima (1995, p. 395) argues in the context of Japanese banks that `the purpose ofa bank's investment was not to maximize the dividend or to realize portfolio profitability inthe normal sense, but to maintain close relations with borrower companies . . . to prevent anopportunistic action by a large borrower.'

9 Some related empirical analyses in this area are by Edwards and Fischer (1994), Aoki and Kim(1995), Gorton and Schmid (1996) and Mulbert (1998).

ß International Review of Finance Ltd. 2000 165

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 6: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

programme by the government in July 1991. With India becoming moreintegrated with the world market and more companies tapping external sourcesof finance from the debt and capital markets, public concerns started becomingmore focused on effective investor protection, on transparency of operations inbusiness and industry and on the need to attain international standards fordisclosure of company information. Most importantly, the role of largeblockholders like banks and institutional investors started to come increasinglyunder scrutiny in view of their reportedly passive role in the corporategovernance of companies (see, for example, CII 1998).

A. Ownership pattern

Table 1 presents the ownership pattern of Indian corporates and the potentialchannels of governance based on the sample of companies chosen for ourempirical analysis. Our sample is drawn from the Corporate Information onMagnetic Medium, or the CIMM10 database, created by the Centre for Monitoringthe Indian Economy (CMIE). This database contains detailed information on thefinancial performance of companies in India compiled from their profit and lossaccounts, balance sheets and stock price data. The database also containsbackground information, including ownership pattern, product profile, plantlocation, new investment projects and so on for the companies. In our analysis,we consider only private and foreign manufacturing companies, although theIndian corporate sector is characterized by the coexistence of public companies aswell. Of the 3217 private and foreign manufacturing companies for whichfinancial data are available in the database for the year 1995±6,11 ownershipinformation for the year 1995±6 is available for 1567.12 These 1567 companiesconstitute the sample for our analysis.

Our sample is representative in that it includes nearly half the privatemanufacturing companies listed in the database and these closely mirror many ofthe population characteristics.13 Moreover, given that the coverage of companies

10 The CIMM database, originally developed on the MS-DOS platform, is now on the Windowsplatform and renamed PROWESS.

11 The year 1995±6 denotes the period from April 1995 to March 1996. April to March is thefinancial year for accounting purposes in India.

12 All publicly traded companies listed on the Bombay Stock Exchange (BSE) ± the major stockexchange in India ± need to furnish information on their equity ownership to BSE at regularintervals. However, since companies usually report this information at different points of time,CIMM reports the company-wise ownership data as of the latest date of reporting. In ouranalysis, we dropped companies for which these data were not available for 1995±6 eventhough the requisite financial information was available.

13 For the sample of 1567 companies and the population of 3217 companies, the mean values ofsome of the variables that describe company characteristics are respectively: sales (Rs. million),1138 and 1641; age (years), 19.6 and 19.6; advertising expense to sales ratio (%), 0.57 and 0.55;exports to sales ratio (%), 11.61 and 13.33; depreciation expense to sales ratio (%), 6.48 and6.20; debt±equity ratio (%), 92.37 and 80.21; and proportion of group companies (%), 32.50and 32.40.

166 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 7: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

in the CIMM database has expanded considerably over the years,14 our sample for1995±6 is likely to be more reflective of the Indian corporate sector compared tothose of earlier studies pertaining to the early nineties (Khanna and Palepu 1999;Chhibber and Majumdar 1999). Indeed, a statistical analysis of the distribution ofcompanies in the database based on three segmentation variables, i.e. size,ownership group and industry classification, shows that the characteristics of theCIMM database have changed significantly over the years, with stabilityemerging only from the year 1994±5 (Choudhury 1999). Our choice of the year1995±6, and not a later year, is guided by the fact that, at the time of our analysis,it was the year for which ownership information was available for the mostnumber of companies in the CIMM database.

As Table 1 shows, the principal holders of corporate equity are: (a) directorsand their relatives; (b) corporate bodies; (c) foreign investors; (d) the term lendinginstitutions, primarily composed of the three government controlled/promoted

Table 1 Pattern of Equity Ownership in Different Types of Companies in India

Type of company Mean equity holding (in %) by

Directors Corporate Foreign Financial Institutional Publicand bodies institutions investors

relatives

Private companies 8.2 33.5 9.3 4.2 10.3 34.5belonging tobusiness groups

Private stand-alone 21.3 18.5 7.4 3.2 3.2 46.4companies

Foreign companies 0.8 18.3 42.0 4.3 12.2 22.4belonging tobusiness groups

Foreign stand-alone 2.8 14.0 43.1 1.7 8.4 30.0companies

All sample 15.4 23.8 10.1 3.6 6.2 40.9companies

Percentage of 26.6 42.5 12.8 2.5 6.7sample companiesin which equityholdings is greaterthan 25%

Based on a sample of 1567 private manufacturing sector companies for which ownershipinformation for the year 1995±6 is available from the CIMM database provided by the Centre forMonitoring the Indian Economy (CMIE).

14 The coverage of companies in the CIMM database is based mainly on availability of the annualaccounts of companies rather than on any formal procedure. Over the years, CMIE has beenable to improve its accessibility to the company annual reports.

ß International Review of Finance Ltd. 2000 167

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 8: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

development financial institutions (DFIs)15 and the state finance corporations;(e) institutional investors, namely the government-owned mutual fund ± theUnit Trust of India ± and three government-owned insurance companies; and (f)the public. Of the six groups, the first five can be considered as large shareholdersor blockholders. As Table 1 also shows, privately owned manufacturingcompanies in India can be broadly classified into four principal ownershipgroups: (a) private companies belonging to business groups; (b) private stand-alone companies; (c) foreign companies belonging to business groups; and (d)foreign stand-alone companies.

Corporate bodies are on average substantial blockholders in private companiesbelonging to business groups, and 42.5% of all sample companies have equityownership by corporate bodies in excess of 25%. Directors and relatives hold onaverage 21.3% of equity ownership in private stand-alone companies and havemore than 25% equity ownership in 26.6% of sample companies. The relativelyhigh proportion of concentrated shareholding by directors and relatives andinter-corporate holdings is on account of the predominance of family-ownedbusiness, a feature that is typical of corporates in many developing countries.Foreign companies, both stand-alones and those belonging to business groups,almost definitionally have foreign investors as substantial blockholders. Amongthe other blockholders, institutional investors, nearly monopolized by the UnitTrust of India, have substantial holdings in group companies, both domestic andforeign. Financial institutions on average hold lower blocks of equity incomparison to institutional investors.

While cross-country comparisons are somewhat difficult given that thereporting of equity ownership data is not uniform, some broad comparisonswith the two main prototype governance systems in the world (Table 2) suggestthat the Indian corporate governance system is by and large a hybrid of the`outsider systems' of the US and UK, and the `insider systems' of continentalEurope and Japan. Equity holdings by non-financial corporations, which areprimarily inter-corporate cross-holdings, are much higher than in the US andUK and are more comparable to Germany and Japan. However, at the sametime, the participation of the small investor in corporate equity is at comparablelevels with the US, with India having the largest number of listed companies inthe world. In addition, while different types of financial institutions separatelyhold much smaller blocks in comparison to those in other countries, given thatnearly all of these institutions are government controlled and fall under theaegis of the Ministry of Finance, together they form a much bigger homogenousblock in India than that in other countries. The participation of institutionalinvestors like mutual funds and insurance companies which are nearly fully

15 The three major DFIs are the Industrial Development Bank of India (IDBI), with 72%government ownership, the Industrial Credit and Investment Corporation of India (ICICI),promoted by the government-owned mutual fund and insurance companies, with a combinedshare of 32%, and the Industrial Finance Corporation of India (IFCI), promoted by thegovernment-owned financial institutions and insurance companies, with a combined share of42%.

168 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 9: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

owned by the government in India is also significant, and comparable to theextent of their participation in Japan and Germany, but much less in scope thanthat in the US.

B. Institutions of governance

In India there are primarily three avenues through which the `exit' and `voice'options of debt holders and equity holders are intended to be institutionally

Table 2 Characteristics of Equity Ownership and Shareholder Rights in India andSelect Countries

Indiab US UK Germany Japan(1996) (1993) (1993) (1993) (1993)

Distribution of shareholding (%)All corporations 33.6 46 64 68 69Financial institutions 9.8 46 62 29 45Banks/lending institutions 3.6c 1 14 22Insurance companies 2.1 5 17 7 17Pension/investment funds ± 26 34 ± 1Mutual funds 4.1 11 7 8 3Others ± 4 3 ± 1

Non-financial corporations 23.8 ± 2 39 24

Individuals 40.9d 49 18 17 24

Foreign 10.1 5 16 12 7

Government ± ± 1 4 1

Others 15.4d ± 2 ± ±

Ownership by three topshareholders in the tenlargest non-financial firms (%)Mean 40 20 19 48 18Median 43 12 15 50 13

Anti-director rights of minority 5 5 5 1 4shareholdersa (scale of 6)

Per capita GNP ($) (1993) 300 24 740 18 060 23 560 31 490

a Anti-director rights measure how strongly the legal system protects minority shareholdersagainst dominant shareholders in the corporate decision-making process, including the votingprocess. For details, see La Porta et al. (1998).b Data based on the study sample of 1567 companies.c Includes commercial banks, government-owned/promoted term-lending institutions andfinancial corporations.d Individuals include top 50. Others include shares held by directors and relatives.

Sources: Data on distribution of shareholding for US, UK, Germany and Japan from OECD (1995).Other data for all five countries pertain to the year 1993 and are sourced from La Porta et al.(1998), Tables 7 and 5.

ß International Review of Finance Ltd. 2000 169

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 10: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

ensured in order to make company management accountable. These are theCompanies Act 1956,16 the market for corporate control and the participation ofpublic financial institutions and institutional investors on the managementboards of companies.17

The Companies Act 1956, besides aiming to ensure that the interests ofcreditors and shareholders are adequately protected and that shareholder voice isadequately represented in the management of companies,18 has provided for agreater measure of government control over the affairs of joint stock companies.As many analysts have noted in their research on corporate governance, the mostimportant legal right that shareholders have is the right to vote on importantcorporate matters such as mergers and liquidations, as well as in the elections ofboards of directors. Among the key legal rights that Indian shareholders have areproportional voting rights and voting through proxies,19 and the right to removea director before the expiry of his or her period of office by ordinary resolution,subject to certain tenurial clauses like life-time employment. A comparative studyof minority shareholder rights across the world reveals that these rights in Indiaare on a par with some other common law countries, notably the US and the UK(Table 2), although the enforcement of these rights as reflected in the efficiency ofthe judicial system and the rule of law is somewhat weaker in India (La Porta et al.1998).

The exercise of the exit option by shareholders in case of an under-performingcompany in India is through the market for corporate control, which althoughnot highly active by US and UK standards, is more vibrant than the markets inJapan and Germany. Takeovers in India are currently regulated by the SubstantialAcquisitions of Shares and Takeovers Regulations, first promulgated in 1994 by theSecurities Exchange Board of India (SEBI), and then replaced by a morecomprehensive Act in 1997. Under current regulations, the acquisition of 10%of shares triggers a minimum public offer of 20%. The activity in the market forcorporate control has considerably improved in recent years due to capital marketreforms and development. With detailed disclosure requirements, greaterinvestor protection, screen-based trading and a move towards integrated markets,investors are becoming increasingly well informed about true companyperformance and are also able to respond quickly to any market offer. Thus,while 121 takeovers and mergers occurred between 1988 and 1992, 166

16 For a comprehensive account of the Companies Act, see Puliani and Puliani (1998).17 For more discussion of institutions of governance and recent developments, see Sarkar and

Sarkar (1999).18 Currently, a revised version of the Companies Act, the Companies Bill 1997, seeking to

streamline the existing Act by reducing 658 sections and 15 schedules to 458 sections andthree schedules, is awaiting the approval of the Parliament. The comprehensive revisions arebeing sought to provide greater flexibility, disclosure and self-regulation of Indian companiesoperating in a liberalized, dynamic and highly competitive environment.

19 According to Section 176 of the Companies Act 1956, `any member of a company entitled toattend and vote at a meeting of the company shall be entitled to appoint another person(whether a member or not) as his proxy to attend and vote instead of himself'. The sectionspecifies the conditions that are applicable to proxy voting.

170 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 11: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

companies were acquired between 1993 and 1997, with nearly 85% of theacquisitions taking place between 1995 and 1997.

Finally, a key feature of corporate governance mechanism is that, like inGermany and Japan, the DFIs and the institutional investors have a substantialpresence in company decision-making. Thus, the government owned Unit Trustof India generally has nominees on company boards by virtue of being asubstantial equity holder. Similarly, the DFIs, being either major equity or debtholders, also have their nominees typically represented in corporate boards.Almost all DFIs' debt contracts carry a covenant that they will be represented onthe board of the debtor company via a nominee director (CII 1998). Thus DFIsand institutional investors have significant voting rights as well as the potentialpower for `jawboning' company management. However, as has been seen manytimes in Germany and Japan, DFIs and institutional investors in India have beenperceived by and large to be passive in corporate governance. Attention to thisproblem has been particularly drawn by several committees, including theCommittee on Takeover Regulations appointed by the SEBI and the Committeeon Corporate Governance instituted by the Confederation of Indian Industry(CII).

III. PERFORMANCE MEASURES AND THE EMPIRICAL MODEL

The variables used in our empirical analysis can be grouped into three categories:(a) performance variables measuring company performance; (b) variables of interestdescribing the extent of equity ownership of different types of blockholders andthereby the ownership mix of the company's equity; and (c) control variablesdescribing the characteristics of the company which might also affect itsperformance.

A. Performance variable

We use two measures, namely market to book value ratio (MBVR), and a proxy forTobin's Q ratio (PQ-ratio), to measure company performance. MBVR is calculatedas the ratio of the product of the number of equity shares and the closing price ofthe share on the last day of the financial year to the book value of equity andreserves.20 Tobin's Q is defined as the ratio of market value of equity and marketvalue of debt to the replacement cost of assets. However, in India, as in manydeveloping countries, the calculation of Tobin's Q is difficult primarily because alarge proportion of the corporate debt is institutional debt that is not activelytraded in the debt market. Further, most companies report asset values tohistorical costs rather than at replacement costs. We therefore calculated a proxy

20 For each company we checked the closing price of the share during the last seven trading daysof the financial year to see if there were any unusual changes in the share price on the lasttrading day. In no case did we find any significant changes.

ß International Review of Finance Ltd. 2000 171

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 12: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

for Tobin's Q by taking the book value of debt and the book value of assets inplace of market values.21

Each of our two measures has its strengths and weaknesses. MBVR isempirically a cleaner measure than the PQ-ratio, has been used as an alternativeto Tobin's Q for developing country studies (e.g. the study by Xu and Wang 1997on China), as well as in other studies (Capon et al. 1996, p. 54), and is morealigned to the objective of the shareholders. However, it excludes debt. In thisrespect, PQ-ratio is a more comprehensive measure of company performance, andhas also been used for developing country studies (e.g. the study by Khanna andPalepu 1999). However, in the Indian context, PQ-ratio is also likely to be morenoisy given that institutional debt forms a significant proportion of externalfinance for Indian companies and that such institutional lending in many cases isguided by social objectives, so that market and book values of debt may divergesignificantly. We therefore use both these measures in our empirical analysis tocheck for the robustness of the results to alternative measures of performance.

We do not use accounting measures like return on equity and return on assetsto measure performance as companies in India do not follow uniform accountingstandards. In the US, standards are set by the Accounting Standards Board and it ismandatory for corporate bodies to conform to these standards. In India, theadoption of uniform International Accounting Standards Committee (IASC)standards by companies is not yet mandatory and it would suffice for a companyto disclose its accounting policies by way of a note in its audited accounts. Arecent ranking of the quality of accounting standards on a scale of 0±100 across49 countries gives India a score of 57, which is below the global average of 60.9and significantly below the US score of 78 (La Porta et al. 1998, Table 5).

B. Variables of interest

Our main variables of interest are the extent of equity ownership by differenttypes of blockholders. These are:

• DIR_S, the fraction of equity shares held by directors and relatives. This is theextent of shareholding by the managers of the company and as such reflectsthe extent of `direct' insider holdings.

• CORP_S, the fraction of equity shares held by Indian corporate bodies. Asdiscussed in section II, for group companies a large part of these holdings maybe holdings by other companies in the group. For such a company, a part ofCORP_S reflects the extent of `indirect' insider holdings. Unfortunately, a

21 Another measure of company performance can be obtained by dividing the market value of thecompany (calculated as the market value of equity plus the book value of debt) by total salesinstead of total assets. While this measure might have merit with respect to other countries,this was not the case for our sample of Indian companies, where our analysis revealed a verylow correlation of this measure with MBVR and PQ-ratio. This was in turn on account of theinefficient asset utilization by many small and young firms that our analysis revealed.

172 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 13: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

break-down of holdings by group and non-group companies is not available inthe CIMM database.

• IINV_S, the fraction of shares held by government-owned insurancecompanies and government-sponsored mutual funds, which reflects theextent of shareholding by institutional investors.

• DFI_S, the fraction of shares held by all government-owned developmentfinancial institutions, including banks and other state financial corporations.

• GOV_S, the sum of IINV_S and DFI_S, which reflects the extent of holdings bythe government.

• FOR_S, the fraction of shares held by foreign entities. This includes holdings byforeign corporations as well as holdings by foreign institutional investors.Again, a break-down between these two types of foreign entities is notavailable in the CIMM database.

C. Control variables

Apart from its ownership structure, the performance of a company is influencedby other external factors that operate through both the product and the capitalmarket. In the empirical literature it is customary to control for the effect of theseexternal factors to avoid any spurious relation between performance andownership structure. In keeping with earlier work we include standard measuresof leverage (LEVRG), size (LSALES), intangible assets (ADVINT), diversification(DIV), capital intensity (DEPINT) and age effects (AGE) in our analysis. Inaddition, all regressions are controlled for industry effects through theincorporation of industry-specific dummy variables (IND-DUMMY_i). A list ofcontrol variables along with their description and possible effects is given in theappendix.

D. The empirical model and estimation methodology

As stated in the introduction, we adopt a spline specification in estimating therelationship between company value and equity holdings by blockholders. Such aspecification allows for the effect of holdings to change at different thresholdpoints known as spline nodes.22 This approach has been used, among others, by

22 Spline is a commonly used technique in empirical analysis to allow for a piecewise linearrelation between two variables, and allows the slope of the regression equation to change atcertain points which are known as spline knots or spline nodes. The technique ensures that theregression line is continuous at the different spline nodes, which is unlikely to be the case ifone uses a slope dummy instead. Suppose one postulates that the relation between y and x ispiecewise linear, with the linear relation changing at two knots, say x1 and x2, with x1 < x2.Then, under the spline technique, three spline variables (the number of spline variables isalways one more than the number of knots) are defined as follows: spline1 � x1 if x � x1;� x ifx < x1; spline2 � x2 ÿ x1 if x � x2;� xÿ x1 if x1 < x < x2;� 0 if x � x1; spline3 � xÿ x2 ifx > x2;� 0 if x � x2. A piecewise linear relation between y and x is then estimated using a linearregression with all the three spline variables.

ß International Review of Finance Ltd. 2000 173

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 14: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Morck et al. (1988) and McConnell and Servaes (1990). We adopt a similarapproach for our estimation on the ground that a non-linear relationshipbetween company value and equity ownership is consistent with the theoreticalliterature on conflicting shareholder incentives that depend on the extent ofshareholding. We find no prima facie justification that would potentially rule outthe presence of such non-linearities in the context of an emerging economy likeIndia. In addition, adopting a spline methodology also enables us to comparemore meaningfully the results obtained in the context of an emerging economywith existing results obtained in the context of developed countries. Thus theempirical model that we estimate takes the form:

performance variable � f [spline(holdings by different types of blockholders),control variables] � error

We started by estimating the spline specification with respect to equity holdingsby directors and relatives and then progressively introduced other blockholdersinto the regression. One issue that arose with the spline specification was how tochoose the number of spline variables and the exact spline nodes. For this we firstestimated the MBVR regression by using a quadratic specification for holdings bydirectors and relatives, since the quadratic specification allows one to determineendogenously the threshold point. The estimated threshold point was found tooccur at 24%, with MBVR declining as directors' holdings increased from 0 to24% and MBVR increasing thereafter. Interestingly, the threshold point of 24%was just below the 26% percent mark that gives the directors, and for that matterany blockholder, the power to block a special resolution under the CompaniesAct of 1956.23 The turning point was also very close to the 25% mark, which hasbeen one of the specified spline nodes used in many empirical studies.

However, though the quadratic specification endogenizes the turning point, itcannot allow the relation between company performance and extent of equityholdings to be different in more than two segments. Such a relation has beenfound in some empirical studies (Morck et al. 1988; McConnell and Servaes 1990).Thus, once the threshold point was determined, we replaced the quadraticspecification by a spline specification allowing for one as well as two spline nodes.We specified alternative values of the spline nodes around the threshold pointand undertook a specification search to determine both the number of splinesand the spline nodes that were most adequately supported by our data. Theoutcomes of the specification search suggested that an appropriate specificationfor the spline formulation with respect to holdings by directors and relatives wasone spline node, with the spline node set at 25%. We adopted this specification

23 Under the Companies Act of 1956, important decisions like changing the line of business,diversification, amalgamation, alteration of shareholder rights or reduction in share capitalhave to be approved by a special resolution which is deemed to be accepted if `the votes cast infavor of the resolution . . . are not less than three times the number of the votes, if any, castagainst the resolution by the members so entitled and voting (Section 189)'.

174 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 15: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

for all types of blockholders and for the PQ-ratio as well. Presumably one can fixthe spline nodes for all types of blockholders endogenously through specificationsearches. However, given that our analysis uses five different types ofblockholders, such an approach would greatly increase the dimensionality ofthe specification search. The fixation of the spline node at 25% for otherblockholders has good institutional support, as discussed above. A description ofthe spline variables is given in the appendix.

Finally, we attempted to tackle in our estimation the issue of reverse causalitythat arises in the analysis of performance and ownership. While reverse causationis an important theoretical issue, it has not received much attention in empiricalwork because ownership patterns in most countries have been observed tochange only very gradually over time. In this case the relation that one obtainsfrom a cross-section regression between ownership and performance, treatingownership as exogenous, can be interpreted as a causal relation from ownershipto performance. Thus, it is not surprising that very few empirical studies, thoughrecognizing the problem of reverse causality, have considered this issue in theirestimation. In our analysis, we addressed the issue of reverse causality specificallyfor foreign holdings, for which we have found rather discrete increases inholdings in the early 1990s, especially after the liberalization of the regulationsguiding foreign participation in India.

IV. EMPIRICAL FINDINGS

Regression results showing the relation between company value and the extent ofequity holdings by different types of blockholders are reported in Tables 3, 4 and6. Regression results with respect to the effect of dual holdings of debt and equityby financial institutions are presented in Tables 7 and 8.

All regressions are estimated by the ordinary least squares (OLS) after checkingfor the presence of influential observations and heteroskedasticity. While somepapers have approached the problem of influential observations by re-estimatingthe regressions by truncating the distribution of the dependent variable at thelow and the high ends of the distribution, we adopted the more formal procedureof using the DFFITS statistics suggested by Belsley et al. (1980). Once theinfluential observations were identified, we computed the bounded-influence(BI) estimates using the weighted least squares technique suggested by Welsch(1980).24 The BI estimates were not significantly different from the OLS estimatesand did not lead to any qualitative changes in the results. We carried out aspecification test for heteroskedasticity because ours is a cross-section regressionbased on a sample that includes companies that differ significantly in size. The

24 The DFFITS statistic identifies an influential observation by analysing the change in itspredicted value when this particular observation is excluded from the estimation. The one-stepbounded-influence is a weighted least squares estimator with weights being inverselyproportional to the degree of influence of the observation. See the cited reference for details.

ß International Review of Finance Ltd. 2000 175

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 16: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

specification test was done following the methodology suggested by White(1980). The value of the chi-squared statistics is reported for each regression in thetables. The specification tests did not detect any significant presence ofheteroskedasticity in the regression equations. The OLS estimates are, therefore,efficient.25

A. Managerial holdings

The estimates reported in column 1 of Tables 3 and 4 show the effect ofmanagerial holdings on company value measured in terms of MBVR and PQ-

25 The test-statistic suggested by White (1980, p. 823) is a test of the joint hypothesis that themodel's specification of the first and second moments of the dependent variable is correct. Aninsignificant test-statistic, therefore, indicates not only that the model is free fromheteroskedasticty, but also that the linear specification is correct.

Table 3 Relationship between MBVR and Extent of Equity Holdings by Directorsand Corporate Bodies

Column 1 Column 2 Column 3

Variable Estimate t-statistic Estimate t-statistic Estimate t-statistic

INTERCEPT 0.1269 0.28 0.0080 0.02 0.0301 0.07DIR_SP1 ÿ0.0082 ÿ1.67* ÿ0.0037 ÿ0.74 ÿ0.0068 ÿ1.14DIR_SP2 0.0126 2.42** 0.0146 2.76** 0.0144 2.48**DIR_SP1 � GROUP 0.0051 0.49DIR_SP2 � GROUP 0.0061 0.44COR_SP1 ÿ0.0005 ÿ0.09 0.0001 0.02COR_SP2 0.0135 3.27** 0.0187 3.02**COR_SP1 � GROUP ÿ0.0043 ÿ0.57COR_SP2 � GROUP ÿ0.0072 ÿ0.89FOR_SP1 0.0108 1.89* 0.0159 2.71** 0.0166 2.80**FOR_SP2 0.0543 8.17** 0.0565 8.34** 0.0551 8.04**LSALES 0.1034 3.65* 0.0914 3.17** 0.1004 3.30**AGE 0.0007 0.28 0.0004 0.15 0.0003 0.13EXPINT ÿ0.0005 ÿ0.28 ÿ0.0006 ÿ0.36 ÿ0.0006 ÿ0.36ADVINT 0.1627 5.84** 0.1607 5.79** 0.1573 5.65**DEPINT 0.0638 6.71** 0.0647 6.83** 0.0647 6.83**LEVRG1 0.0042 12.14** 0.0042 12.16** 0.0042 12.22**DIV 0.1997 0.37 0.2049 0.39 0.1697 0.32Industry dummy Included Included Included

Adj. R-squared 0.27 0.28 0.28F 22.59 21.65 19.15N 1567 1567 1567Chi-squared for 266.78 295.14 327.63HeteroskedasticityProb. > chi-squared 0.13 0.53 0.99

** Significant at 1% level; * significant at 5% level.

176 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 17: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

ratio, respectively. The estimates show that MBVR declines by 0.8% for every 1%increase in directors' holdings up to 25% and thereafter MBVR increases by 1.3%for every 1% increase in directors' holdings (column 1 of Table 3). With respect toPQ-ratio, company value declines by 0.3% per 1% increase in directors' holdingsup to 25% and thereafter increases by 0.4% per 1% increase in directors' holdings(column 1 of Table 4). These results, therefore, suggest that the relationshipbetween managerial ownership and company value is non-linear, a finding that isconsistent with those of the studies by Morck et al. (1988) and McConnell andServaes (1990) with respect to US companies. However, the nature of non-linearity is different, as we discuss below.

While McConnell and Servaes (1990) find company value first to rise and thento decline continuously after a threshold level (49.4% for the 1976 data and37.6% for the 1986 data), we find company value to increase beyond a thresholdpoint (25%). Our results, in this respect, are similar to those found by Morck et al.(1988), who also find company value to increase beyond a threshold level (also

Table 4 Relationship between PQ-ratio and Extent of Equity Holdings by Directorsand Corporate Bodies

Column 1 Column 2 Column 3

Variable Estimate t-statistic Estimate t-statistic Estimate t-statistic

INTERCEPT 0.3913 2.37** 0.3701 2.21* 0.3828 2.27*DIR_SP1 ÿ0.0030 ÿ1.68* ÿ0.0016 ÿ0.86 ÿ0.0031 ÿ1.41DIR_SP2 0.0039 2.02* 0.0043 2.23* 0.0044 2.07*DIR_SP1 � GROUP 0.0034 0.90DIR_SP2 � GROUP 0.0016 0.31COR_SP1 ÿ0.0014 ÿ0.71 ÿ0.0010 ÿ0.43COR_SP2 0.0049 3.24** 0.0062 2.76**COR_SP1 � GROUP ÿ0.0021 ÿ0.78COR_SP2 � GROUP ÿ0.0017 ÿ0.56FOR_SP1 0.0041 1.94* 0.0058 2.71** 0.0062 2.85**FOR_SP2 0.0188 7.75** 0.0193 7.78** 0.0187 7.46**LSALES 0.0533 5.16** 0.0505 4.80** 0.0529 4.76**AGE ÿ0.0006 ÿ0.70 ÿ0.0007 ÿ0.81 ÿ0.0007 ÿ0.82EXPINT 0.0007 1.16 0.0007 1.09 0.0007 1.07ADVINT 0.0745 7.32** 0.0738 7.27** 0.0725 7.13**DEPINT 0.0197 5.67** 0.0199 5.76** 0.0200 5.77**LEVRG1 0.0000 ÿ0.24 0.0000 ÿ0.23 0.0000 ÿ0.14DIV 0.0997 0.51 0.0966 0.50 0.0857 0.44Industry dummy Included Included Included

Adj. R-squared 0.21 0.22 0.22F 16.56 15.91 14.09N 1567 1567 1567Chi-squared for 204.83 232.92 284.62heteroskedasticityProb. > chi-squared 0.97 0.99 0.99

** Significant at 1% level; * significant at 5% level.

ß International Review of Finance Ltd. 2000 177

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 18: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

25% in their study). The one difference between our result and that of Morck et al.(1988) is with respect to the effect of managerial holdings below the 25% level.While the latter finds company value first to increase between the 0 and 5% levelof equity holdings and then to decrease between 5 and 25%, we find companyvalue to decline uniformly over the entire range. However, this negative impactthat we find ceases to be statistically significant after incorporating the holdingsby other blockholders in the regression. In contrast, the positive effect ofmanagerial holdings on company value beyond the 25% level continues toremain statistically significant at the 1% level even after incorporating the effectsof other blockholders (Table 6). It may be a matter of interest to note that thispositive effect, as measured by the PQ-ratio, is half of that found by Morck et al.(1988) over the comparable range with respect to their sample of US companies,suggesting that the `convergence of interest' effect is less strong in our sample ofIndian companies.

B. Foreign holdings

The coefficients of both the spline variables associated with foreign ownership(column 1 of Tables 3 and 4) are positive and highly significant, suggesting thatan increase in foreign holdings increases company value in terms of both MBVRand PQ-ratio at all levels of equity holding. However, the increase in value issignificantly higher once the extent of holdings crosses 25%, with the coefficientof the second spline variable being over four times the coefficient of the firstspline variable. Further, the beneficial effect of foreign ownership above the 25%level is significantly higher than that of directors' holdings, with MBVRincreasing by 5.4% for every 1% increase in foreign holdings, compared to1.3% for every 1% increase in directors' holdings. Similar results hold true for thePQ-ratio.

Given the 24% limit applicable to foreign institutional investment (FII) in theequity of one single company in 1995±6, the significantly higher effect of foreignownership that we find for the second spline dummy can be safely associatedwith holdings by foreign corporate bodies26 (termed foreign direct investment(FDI) in the government regulations). Our results therefore suggest that theincremental effect of foreign corporate holdings on company value is higher thanthat of FII, though a strict comparison should look at the relative effects of thesetwo groups in comparable ranges of ownership, i.e. below 24% holdings.Unfortunately, as pointed out above, this break-down is not available in ourdatabase. The large effect that we find for foreign corporate holdings above the25% level could arise from an increased likelihood of foreign technology transferto Indian companies once foreign holdings become substantial. The other factorthat could work in conjunction to explain the large effect is that the diversion ofbenefits by insiders becomes less feasible once foreign stakes, and hence thepossibility of stronger external monitoring, go up.

26 We would like to thank an anonymous referee for pointing this out.

178 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 19: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Our findings with regard to foreign holdings are largely in line with those ofsome earlier studies with respect to developing countries (Khanna and Palepu(1999) and Chhibber and Majumdar (1999) in the case of India; Vendrell-Alda(1978) in the case of Argentina; and Willmore (1986) in the case of Brazil).Further, our result that foreign ownership is non-linearly related to performanceis consistent with that obtained by Chhibber and Majumdar (1999). However, ourresults differ from those of Chhibber and Majumdar (1999) in terms of the precisenature of non-linearity. While Chhibber and Majumdar (1999) find foreignownership to display superior performance only when property rights devolve toforeign owners at sufficiently high levels of holdings (51% in their study), we findforeign holdings to increase company value even at low levels of holdings. Thisdifference in result could be due to the fact that Chhibber and Majumdar (1999)do not consider the combined effect on company value of holdings by largeshareholders other than foreign holders and use accounting rates of return inevaluating company performance.

To address the issue of reverse causality with respect to foreign holdings, welooked at the changes in such holdings between the years 1992 and 1995 for theforeign companies in our sample (we had 92 such companies), and examined ifthe increase in holdings, as well as the extent of such an increase, tended to occurin the better performing companies. The choice of 1992 as the comparison yearwas dictated by the fact that this was the year for which comparable ownershipinformation was available for the largest number of foreign companies. Results ofthis analysis are presented in Table 5 for the 69 companies for which ownershipdata were available for both 1992 and 1995. Panel A of the table shows thatbetween 1992 and 1995, the share of foreign equity increased in two-thirds of thecompanies, with the average increase being just over 9%, and decreased in theremaining one-third, with the average decrease being just under 3%. Did theincrease in foreign equity occur in the better performing companies of 1992? Thelast three columns in panel A show that the performance of the increasing sharecompanies as judged by return-on-net-worth and return-on-sales was not anyhigher than that of the decreasing share companies. In fact, the return-on-sales ofthe increasing share companies was actually lower. However, in terms of themarket indicator MBVR, the increasing share companies were the betterperforming companies of 1992.

Given this mixed evidence, we then looked at the increasing share companiesand examined if the extent of increase in equity holdings was positively related tocompany performance. Based on each of the three performance criteria, wecategorized the companies into two groups: one consisting of those whoseperformance was higher than the average performance in 1992, and the otherconsisting of companies whose performance was lower than the averageperformance in 1992. We then calculated the extent of increase in foreignholdings for these two groups of companies. Panel B gives the results of thisexercise. The results indicate that the extent of increase in foreign equity wasactually greater in the `lower than average' companies irrespective of whicheverperformance indicator we use. Thus, in terms of the market-to-book-value

ß International Review of Finance Ltd. 2000 179

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 20: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Table 5 Foreign Holdings and Reverse Causality

Panel A: all companies (69)

Company groups Sample Mean foreign equity holding (%) Mean value of performancebased on behavior size indicators in 1992of holdings

In 1992 In 1995 Difference Market to Return on Return onbook value sales net worth

Increasing share companies 46 43.52 52.67 9.15 6.26 0.05 0.15All companies 69 41.31 46.29 5.16 5.39 0.06 0.15Decreasing share companies 23 36.36 33.52 ÿ2.84 3.63 0.08 0.15

Panel B: increasing share companies (46)

Company groups based on Sample Mean value of performance Mean foreign equity holding (%)1992 performance size indicator in 1992

In 1992 In 1995 Difference

Market to book valueHigher than average 21 8.89 42.78 50.97 8.19All companies 46 6.26 43.52 52.67 9.15Lower than average 25 4.06 44.13 54.10 9.96

Return on salesHigher than average 21 0.08 42.42 49.64 7.23All companies 46 0.05 43.52 52.67 9.15Lower than average 25 0.02 44.44 55.21 10.77

Return on net worthHigher than average 24 0.24 42.89 49.11 6.21All companies 46 0.15 43.52 52.67 9.15Lower than average 22 0.06 44.20 56.56 12.36

18

Intern

ation

al

Rev

iewo

fFin

ance

Ltd

.2000

Intern

ation

al

Review

ofFin

an

ce

Page 21: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

indicator, while foreign holdings increased by 8% in the `higher than average'companies, it increased by 10% in the `lower than average' companies. Thisdifference becomes more significant when judged against the fact that theaverage value of the performance indicator for the increasing share companieswas more than twice the value of that for the `lower than average companies'.Similar findings occur with respect to the accounting indicators with greaterdegree of significance. Thus our analysis does not indicate that increases inforeign equity holdings occurred only in the better performing companies.

C. Corporate holdings

To find out the effect of equity holdings by corporate bodies on company value,we introduced the spline variables COR_SP1 (spline up to 25%) and COR_SP2(spline above 25%) in the regression. The results are reported in column 2 ofTables 3 and 4. The coefficient of COR_SP1 and the associated t-statistics in thetables suggest that, similarly to managerial holdings, holdings by corporatebodies do not affect company value as long as the holdings are below 25%.However, as the positive and statistically significant coefficient of COR_SP2shows, corporate bodies tend to influence company value positively once theirholdings cross 25%. For every 1% increase in corporate holdings, MBVR increasesby 1.4% and PQ-ratio by 0.5%. Both these effects are statistically significant at the1% level and are similar in magnitude to the effects observed for managerialholdings.

The estimates with respect to the effect of corporate holdings on companyvalue reported above, as well as the effect of managerial holdings reported earlier,were obtained by treating business group and stand-alone companiessymmetrically. However, in a country like India, with a significant number ofcross-holdings among group companies, the effect of insider ownership, i.e. theimpact of managerial holdings and corporate holdings, might be different forgroup and stand-alone companies. Insider ownership for group companies isdifficult to determine because managers in group companies may, apart fromexerting direct control over the company through their equity holdings, exerciseindirect control through cross-holdings in other group companies via `controlchains' or `pyramids'.27 In such cases, the low level of managerial holdings ingroup companies relative to stand-alones would effectively understate theinfluence of managerial ownership on company value. A similar logic wouldhold with respect to capturing the effect of corporate holdings in groupcompanies wherein a significant proportion of the holdings presumably consistof holdings by other companies in the group.

27 According to La Porta et al. (1998), a shareholder has x% indirect control over firm A if (1) itdirectly controls firm B, which in turn controls x% of firm A, or (2) it directly controls firm C,which in turn controls firm B or a sequence of firms leading to firm B through a control chain.A firm's ownership structure is a pyramid if it has an ultimate owner and there is at least onepublicly traded company between it and the ultimate owner. Corporate bodies are oftenpublicly traded companies and/or group companies.

ß International Review of Finance Ltd. 2000 181

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 22: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

To explore this issue of insider ownership we interacted a dummy variable calledGROUP (which equalled 1 for a group company and 0 otherwise) with the splinevariables with respect to managerial and corporate body holdings. Under thisspecification, the coefficients on the interaction terms pick up the difference, ifany, in the effect of managerial and corporate holdings between group and stand-alone companies. The results of this estimation are presented in column 5 of Tables3 and 4. With respect to both MBVR and the PQ-ratio, the estimates associated withthe interaction terms are statistically insignificant for holdings less than 25% aswell as for holdings of more than 25%. This is true for both managerial holdingsand corporate body holdings. In all cases, the associated t-statistic is much belowthe value implied by the conventional levels of significance. The results thusindicate that there are no significant differences between the effects of insiderownership and of corporate holdings on group and stand-alone companies.

The absence of any statistically significant differences that we find in ouranalysis is in line with the findings of the study by Khanna and Palepu (1999) ofthe absence of any group effects in their sample. However, our non-linearspecification clearly highlights that the effect of managerial holdings andcorporate bodies depend on the level of equity ownership. Thus while Khannaand Palepu (1999) report that managerial ownership has no significant effect oncompany value, we find a positive and significant effect beyond holdings of 25%.With regard to corporate bodies too, while the above-mentioned study finds apositive and significant relationship with company value, we find the existenceof such a relationship only beyond the threshold level of 25%.

D. Financial institutions and institutional investor holdings

We now introduce blockholdings by government institutions into theregression.28 Results of this regression are reported in columns 1 and 3 of Table6. The coefficients of the spline variable GOV_SP1 (spline up to 25%) suggest thatgovernment holdings have a negative impact on corporate value, measured byMBVR and PQ-ratio, but the effect is statistically significant only with respect tothe MBVR measure. For higher levels, the effect of government holdings, thoughpositive, is insignificant both with respect to MBVR and PQ-ratio, as is reflected inthe coefficients of GOV_SP2 (spline above 25%). These results point to aninteresting contrast with the results obtained with respect to the otherblockholders, namely that government financial institutions are passivemonitors even when they have substantial stakes in a company.

While the result that financial institutions as a group in India are passivemonitors is consistent with the conclusion arrived at by Khanna and Palepu(1999), it is worthwhile to undertake a more disaggregated analysis given the

28 Since our earlier analysis indicated the absence of any statistically significant differences in theeffect of managerial and corporate holdings between group and stand-alone companies, wedropped the interaction terms when we introduced holdings by institutional investors andfinancial institutions.

182 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 23: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Table 6 Relationship between MBVR and PQ-ratio, and Extent of Equity Holdings by Institutional Investors and FinancialInstitutions

Based on MBVR Based on PQ-ratio

Column 1 Column 2 Column 3 Column 4

Variable Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic

INTERCEPT ÿ0.0098 ÿ0.02 ÿ0.0528 ÿ0.12 0.3651 2.18* 0.3485 2.08*DIR_SP1 ÿ0.0044 ÿ0.86 ÿ0.0051 ÿ1.00 ÿ0.0014 ÿ0.74 ÿ0.0016 ÿ0.85DIR_SP2 0.0134 2.51** 0.0134 2.51** 0.0044 2.23* 0.0043 2.21*COR_SP1 ÿ0.0002 ÿ0.03 0.0000 ÿ0.01 ÿ0.0013 ÿ0.65 ÿ0.0012 ÿ0.62COR_SP2 0.0125 2.97** 0.0121 2.87** 0.0050 3.27** 0.0049 3.18**GOV_SP1 ÿ0.0102 ÿ1.79* ÿ0.0003 ÿ0.16GOV_SP2 0.0087 0.79 0.0031 0.76IINV_SP1 ÿ0.0151 ÿ2.24* ÿ0.0024 ÿ0.96IINV_SP2 0.0084 0.44 0.0026 0.37DFI_SP1 ÿ0.0047 ÿ0.62 0.0014 0.51DFI_SP2 0.0437 1.73* 0.0139 1.51yFOR_SP1 0.0166 2.81** 0.0169 2.87** 0.0060 2.77** 0.0061 2.84**FOR_SP2 0.0549 7.98** 0.0544 7.92** 0.0194 7.72** 0.0192 7.65**LSALES 0.1092 3.51** 0.1156 3.70** 0.0495 4.35** 0.0520 4.54**AGE 0.0009 0.37 0.0017 0.68 ÿ0.0007 ÿ0.83 ÿ0.0004 ÿ0.46EXPINT ÿ0.0006 ÿ0.33 ÿ0.0005 ÿ0.28 0.0007 1.10 0.0007 1.15ADVINT 0.1604 5.78** 0.1619 5.83** 0.0738 7.26** 0.0743 7.31**DEPINT 0.0642 6.77** 0.0644 6.81** 0.0200 5.76** 0.0200 5.78**LEVRG1 0.0042 12.25** 0.0042 12.09** 0.0000 ÿ0.26 0.0000 ÿ0.38DIV 0.2355 0.44 0.2884 0.54 0.0955 0.49 0.1163 0.60Industry dummy Included Included Included Included

Adj. R-squared 0.28 0.28 0.22 0.22F 20.38 19.35 14.89 14.14N 1567 1567 1567 1567Chi-squared for 298.72 353.37 256.63 405.17heteroskedasticityProb. > chi-squared 0.98 0.98 0.99 0.54

** Significant at 1% level; * significant at 5% level; y significant at 7% level.Since the coefficients of the interaction terms in column 3 of Tables 3 and 4 are all insignificant, the interaction terms are dropped in theregressions involving institutional investors and financial institutions.

ßIn

ternatio

nal

Rev

iewo

fFin

ance

Ltd

.2000

18

3

Large

Shareh

older

Activism

inC

orpora

teG

overnan

cein

Develop

ing

Cou

ntries

Page 24: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

differing incentives of the two types of government institutions that constitutethis group. Recall that blockholdings by government institutions are of two types.First are the holdings by institutional investors, namely the governmentsponsored mutual funds and insurance companies, and second are the holdingsby development financial institutions. Institutional investors have only equityholdings, while development financial institutions typically have both equityand debt holdings in companies. To capture any potential differences in effect ofthe two types of financial institutions, we re-estimated the regression bydisaggregating government blockholdings into the fraction held by institutionalinvestors and that held by development financial institutions. The results arereported in columns 2 and 4 of Table 6.

With respect to institutional investors, company value tends to decline asshare of institutional investors increases from 0 to 25%. This effect is statisticallysignificant with respect to MBVR, with MBVR declining by 1.5% with every 1%increase in institutional holdings. For holdings above 25%, institutional investorshave a positive but insignificant effect on company value as measured by bothMBVR and PQ-ratio, a result that is similar to that obtained with respect togovernment holdings as a whole. These results contrast sharply with the strongpositive effects found by McConnell and Servaes (1990) with regard to UScorporates and Xu and Wang (1997) with respect to Chinese enterprises, but areconsistent with the conclusions of the survey by Black (1998) with regard to theinactivism of institutional investors in general in the US.

With respect to development financial institutions, company value remainsunchanged as the share of these institutions increases from 0 to 25%. However,company value, measured by both MBVR and PQ-ratio, tends to rise significantlyas the share increases beyond the 25% level, The magnitude of the effect is muchhigher than those observed with respect to managerial and corporate bodyholdings, and is similar to that observed with respect to foreign holdings. Ourresults suggest that, unlike institutional investors, development financialinstitutions have a strong monitoring effect on the company once they havesubstantial stakes in it. Our result contrasts with the finding by Xu and Wang(1997) in the case of Chinese enterprises that holdings by financial institutionsuniformly reduce company value.

E. Dual holdings of debt and equity

As discussed above, one of the features of development financial institutions inIndia is that they typically hold both equity and debt in companies. Does debtholdings by these financial institutions influence their incentive to monitorcompanies, and if it does, what is the nature of this influence? Existing theoreticaland empirical literature on the incentives for corporate monitoring under dualholdings (i.e. `split-financing') reveals several costs and benefits, but the empiricalevidence on the subject is so far inconclusive, and restricted primarily to Germancompanies (see, for example, Mulbert (1998) for a detailed discussion). One of thebenefits with banks as blockholders is that the stock market can interpret a loan

184 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 25: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

approval as a positive signal and thereby respond favourably with a positive shareprice reaction. This in turn generates additional stock return to all shareholders,including the banks. Another important benefit is the mitigation of informationalasymmetries, since the company has the incentive to submit better information tothe bank with the expectation that, as a blockholder, a bank is less likely to misuseconfidential information. Among the costs of dual holdings, the most commonlymentioned is the possibility that banks, in spite of being blockholders, wouldseldom be in a position to exercise the `exit' option as a monitoring device becauseby doing so they would have to forego the benefits of `relational investing'.

To draw some basic insights on this issue of governance under dual holdings,we created a dummy variable which we coded as 1 if the share of developmentfinancial institutions in total long-term debt exceeded a cut-off point, andinteracted this dummy variable with the spline variable relating to equityholdings by these institutions. The interaction was done for the first splinevariable only (up to 25% holding) because it is in this range of equity ownershipthat we have found development financial institutions to be particularly passive,so that the incentive for debt-monitoring is likely to be strong. We then estimatedthe regression for different cut-off points. The results for six cut-off points (40, 45,50, 55, 60 and 65%) are reported for MBVR and the PQ-ratio in Tables 7 and 8,respectively. The coefficient of the interaction variable is positive in all theregressions, with the t-statistic becoming significant at the 50 and 55% cut-offpoints for MBVR and beyond the 45% cut-off point in the case of the PQ-ratio.Further, the t-statistic assumes the highest level of significance at the 55% cut-offpoint for both the MBVR and the PQ-ratio.

These results indicate that the effectiveness of development financialinstitutions in monitoring companies in which they have low equity stakesdepends significantly on their debt holdings. Where debt holdings are low orabsent, development financial institutions, like institutional investors, appear tobe poor monitors. However, as the level of debt holdings increases, developmentfinancial institutions seem to exert a positive and significant impact on companyvalue, suggesting that these institutions step up their controlling and monitoringactivities with higher levels of debt.

While a more detailed theoretical and empirical analysis is needed tounderstand the underlying dynamics of our result, the interesting insight ofour analysis is that the net benefits of `split-financing' depend on the levels ofdebt for any given level of equity, and are positive only beyond a threshold levelof debt. Thus, conflicting empirical evidence on the governance value of dualholdings that are present in the literature with respect to German firms (see, forexample, Nibler 1995; Gorton and Schmid 1996)29 can be explained consistentlyin terms of our result.

29 Gorton and Schmid (1996) analysed two samples of German firms for the years 1974 and 1985and show that while there existed a uniquely significant positive influence of bank holdingson the value of the firm in the 1974 sample, no such unique relation was found for the 1985sample, and monitoring by bank blockholders was similar to that of other blockholders. Thelatter result finds support in the study by Nibler (1995).

ß International Review of Finance Ltd. 2000 185

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 26: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Table 7 Relationship between MBVR and Extent of Debt Holdings by Financial Institutions

Share of debt > 40% Share of debt > 45% Share of debt > 50% Share of debt > 55% Share of debt > 60% Share of debt > 65%

Variables Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic

INTERCEPT ÿ0.1811 ÿ0.47 ÿ0.1643 ÿ0.43 ÿ0.1640 ÿ0.42 ÿ0.1657 ÿ0.43 ÿ0.1654 ÿ0.43 ÿ0.1674 ÿ0.43DIR_SP1 ÿ0.0055 ÿ1.06 ÿ0.0057 ÿ1.10 ÿ0.0058 ÿ1.12 ÿ0.0057 ÿ1.12 ÿ0.0057 ÿ1.11 ÿ0.0056 ÿ1.09DIR_SP2 0.0136 2.55** 0.0137 2.56** 0.0137 2.58** 0.0137 2.58** 0.0138 2.59** 0.0137 2.56**COR_SP1 0.0002 0.03 0.0003 0.06 0.0003 0.06 0.0004 0.07 0.0004 0.07 0.0003 0.06COR_SP2 0.0121 2.88** 0.0121 2.86** 0.0120 2.83** 0.0120 2.84** 0.0120 2.85** 0.0120 2.85**IINV_SP1 ÿ0.0153 ÿ2.27* ÿ0.0153 ÿ2.28* ÿ0.0152 ÿ2.26* ÿ0.0151 ÿ2.25* ÿ0.0150 ÿ2.23* ÿ0.0149 ÿ2.21*IINV_SP2 0.0086 0.45 0.0084 0.44 0.0082 0.43 0.0082 0.43 0.0081 0.42 0.0078 0.41DFI_SP1 ÿ0.0153 ÿ1.08 ÿ0.0192 ÿ1.41 ÿ0.0205 ÿ1.66* ÿ0.0203 ÿ1.68* ÿ0.0184 ÿ1.58 ÿ0.0171 ÿ1.50DFI_SP2 0.0452 1.79* 0.0477 1.88* 0.0466 1.85* 0.0460 1.82* 0.0450 1.79* 0.0442 1.76*INTERACTION 0.0131 0.90 0.0180 1.29 0.0212 1.65* 0.0215 1.71* 0.0196 1.56 0.0182 1.47FOR_SP1 0.0169 2.88** 0.0169 2.87** 0.0169 2.88** 0.0170 2.89** 0.0169 2.88** 0.0169 2.87**FOR_SP2 0.0543 7.90** 0.0543 7.90** 0.0542 7.89** 0.0543 7.91** 0.0544 7.93** 0.0544 7.92**LSALES 0.1168 3.73** 0.1175 3.75** 0.1186 3.79** 0.1186 3.79** 0.1179 3.77** 0.1183 3.78**AGE 0.0017 0.70 0.0018 0.72 0.0018 0.72 0.0018 0.73 0.0018 0.71 0.0017 0.70EXPINT ÿ0.0005 ÿ0.30 ÿ0.0005 ÿ0.31 ÿ0.0006 ÿ0.33 ÿ0.0006 ÿ0.33 ÿ0.0006 ÿ0.34 ÿ0.0006 ÿ0.34ADVINT 0.1638 5.93** 0.1648 5.97** 0.1649 5.97** 0.1649 5.97** 0.1644 5.96** 0.1642 5.95**DEPINT 0.0647 6.83** 0.0647 6.84** 0.0646 6.83** 0.0645 6.82** 0.0645 6.82** 0.0645 6.82**LEVRG1 0.0042 12.02** 0.0042 12.02** 0.0042 12.04** 0.0042 12.05** 0.0042 12.04** 0.0042 12.04**DIV 0.4035 0.85 0.3787 0.80 0.3864 0.78 0.3774 0.80 0.3795 0.80 0.3792 0.80INDUSTRY dummy Included Included Included Included Included Included

R-squared 0.28 0.28 0.28 0.28 0.28 0.28F 19.37 19.41 19.45 19.45 19.44 19.43N 1567 1567 1567 1567 1567 1567Chi-squared for 614.37 412.80 426.03 418.51 392.88 386.34heteroskedasticityProb. > chi-squared 0.00 0.79 0.62 0.73 0.94 0.96

** Significant at 1% level; * significant at 5% level.

18

Intern

ation

al

Rev

iewo

fFin

ance

Ltd

.2000

Page 27: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Table 8 Relationship between PQ-ratio and Extent of Debt Holdings by Financial Institutions

Share of debt > 40% Share of debt > 45% Share of debt > 50% Share of debt > 55% Share of debt > 60% Share of debt > 65%

Variables Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic Estimate t-statistic

INTERCEPT 0.3148 2.23* 0.3252 2.30* 0.3244 2.29* 0.3237 2.29* 0.3237 0.29* 0.3228 2.28*DIR_SP1 ÿ0.0018 ÿ0.96 ÿ0.0019 ÿ1.01 ÿ0.0019 ÿ1.00 ÿ0.0019 ÿ1.00 ÿ0.0018 ÿ0.98 ÿ0.0018 ÿ0.95DIR_SP2 0.0044 2.28* 0.0045 2.30* 0.0045 2.30* 0.0045 2.30* 0.0045 2.31* 0.0044 2.28*COR_SP1 ÿ0.0011 ÿ0.58 ÿ0.0010 ÿ0.53 ÿ0.0011 ÿ0.54 ÿ0.0010 ÿ0.53 ÿ0.0010 ÿ0.53 ÿ0.0011 ÿ0.55COR_SP2 0.0049 3.20** 0.0049 3.17** 0.0048 3.14** 0.0048 3.14** 0.0049 3.16** 0.0049 3.16**IINV_SP1 ÿ0.0025 ÿ1.02 ÿ0.0025 ÿ1.04 ÿ0.0024 ÿ0.99 ÿ0.0024 ÿ0.98 ÿ0.0023 ÿ0.96 ÿ0.0023 ÿ0.93IINV_SP2 0.0027 0.39 0.0026 0.37 0.0025 0.36 0.0025 0.35 0.0024 0.35 0.0023 0.33DFI_SP1 ÿ0.0055 ÿ1.06 ÿ0.0071 ÿ1.44 ÿ0.0059 ÿ1.31 ÿ0.0060 ÿ1.36 ÿ0.0049 ÿ1.14 ÿ0.0041 ÿ0.99DFI_SP2 0.0149 1.61 0.0163 1.76* 0.0153 1.66* 0.0150 1.63 0.0145 1.58 0.0142 1.54INTERACTION 0.0085 1.59 0.0106 2.07* 0.0098 2.05* 0.0102 2.17* 0.0089 1.94* 0.0081 1.79*FOR_SP1 0.0062 2.86** 0.0061 2.85** 0.0061 2.86** 0.0062 2.86** 0.0061 2.85** 0.0061 2.84**FOR_SP2 0.0192 7.63** 0.0191 7.62** 0.0191 7.61** 0.0192 7.64** 0.0192 7.66** 0.0192 7.65**LSALES 0.0528 4.61** 0.0532 4.65** 0.0534 4.66** 0.0534 4.67** 0.0531 4.64** 0.0533 4.65**AGE ÿ0.0004 ÿ0.43 ÿ0.0004 ÿ0.40 ÿ0.0004 ÿ0.41 ÿ0.0004 ÿ0.40 ÿ0.0004 ÿ0.42 ÿ0.0004 ÿ0.44EXPINT 0.0007 1.12 0.0007 1.10 0.0007 1.09 0.0007 1.08 0.0007 1.07 0.0007 1.08ADVINT 0.0749 7.42** 0.0755 7.48** 0.0753 7.46** 0.0753 7.46** 0.0751 7.44** 0.0750 7.43**DEPINT 0.0202 5.84** 0.0202 5.85** 0.0201 5.81** 0.0201 5.81** 0.0201 5.81** 0.0201 5.81**LEVRG1 ÿ0.0001 ÿ0.47 ÿ0.0001 ÿ0.49 ÿ0.0001 ÿ0.45 ÿ0.0001 ÿ0.45 ÿ0.0001 ÿ0.45 ÿ0.0001 ÿ0.44DIV 0.1420 0.82 0.1275 0.74 0.1263 0.73 0.1300 0.75 0.1316 0.76 0.1317 0.76INDUSTRY dummy Included Included Included Included Included Included

R-squared 0.22 0.22 0.22 0.22 0.24 0.24F 14.23 14.30 14.30 14.32 14.28 14.26N 1567 1567 1567 1567 1567 1567Chi-squared for 663.90 502.63 427.73 422.03 411.71 380.01heteroskedasticityProb. > chi-squared 0.00 0.02 0.63 0.65 0.82 0.98

** Significant at 1% level; * significant at 5% level.

ßIn

ternatio

nal

Rev

iewo

fFin

ance

Ltd

.2000

18

7

Page 28: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

F. Control variables

Finally, we discuss briefly the importance of the various control variablesincluded in our analysis. Four, out of the seven, control variables are significant inall the regressions. These are the size variable, LSALES; the leverage variable,LEVRG; the advertising intensity variable, ADVINT; and the depreciationintensity variable, DEPINT. The coefficients associated with all these variablesare positive, with very high t-values. The positive sign of the size variable isconsistent with the argument that companies might acquire market power as wellas achieve significant economies of scale with an increase in size. The positivesign of the leverage variable is consistent with the tax argument advanced byModigliani and Miller (1963). It also supports the signalling argument advancedby Ross (1977) that a more efficient management may signal its expertise bycommitting to high fixed payments. The positive sign of the advertising intensityvariable supports the hypothesis that higher advertising expenditure tends toincrease the intangible assets of companies, which in turn are likely to increasecompany value. The positive sign of the depreciation intensity variable indicatesthat companies which are more capital-intensive and, therefore, probably moretechnologically advanced, tend to have higher market value.

G. Time stability of regression results

Since we have ownership data for only one year, the time stability of ourregression results is open to question. To address this issue in a simple manner wefollowed the approach adopted by Morck et al. (1988) and re-estimated theregressions by using the 1994±5 values of the performance variables and all thecontrol variables, but the 1995±6 values of the equity ownership variables. Theregressions were estimated using all the companies in our sample for whichfinancial information was available for the year 1994±5 (1272 companies). Sincechanges in ownership structure are likely to occur only slowly, these regressionsshould give us some indication about the time stability of our results.

The coefficient estimates based on the 1994±5 data reconfirmed our inferencesregarding the effect of directors' holdings, corporate bodies' holdings and foreignholdings on company value. However, there were some differences with respectto government holdings. While our results based on 1995±6 data suggested anegative but insignificant effect of government holdings on company value forrelatively low levels of equity ownership (i.e. less than 25%), the results based on1994±5 data indicated a negative and significant effect. The coefficient associatedwith the second spline dummy (ownership of more than 25%), which waspositive but insignificant in the 1995±6 regression, changed to negative butinsignificant in the 1994±5 regression. These results suggested someimprovement in the effectiveness of government ownership in enhancingcompany value over time. When we looked deeper into the role played by theinstitutional investors and the lending institutions in bringing about thisimprovement, we observed that this improvement was largely due to the role

188 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 29: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

of the domestic financial institutions. The fact that some of these institutionshave themselves accessed the capital market and have become publicly tradedduring this period could be one of the reasons behind their taking a more activeinterest in the governance of the companies.

V. CONCLUSIONS

Our empirical analysis of the activism of different types of large shareholders incorporate governance in developing and emerging countries like India has soughtto add to the sparse evidence that exists with respect to these countries. On theone hand, the picture that emerges from our study is consistent with someconclusions arrived at with respect to the outsider-dominated market-basedsystems of US and UK, and, on the other hand, it has some commonality withelements of the insider-dominated bank-based systems of Germany and Japan.

The common thread running through all our results is that the relationshipbetween each of the different types of shareholders and company value ispiecewise linearly related. This is consistent with the results of some of theexisting studies with respect to the US regarding the effect of managerial holdingson company value, and with respect to India regarding the relationship betweenforeign ownership and company performance. Beyond the threshold of 25%,barring institutional investors, we find the relationship between shareholdingand company value to be positive and statistically significant. An implication ofthis result is that concentrated ownership increases company value and ourevidence provides support for the `convergence of interest' hypothesis ratherthan the `entrenchment' and `conflict-of-interest' hypotheses. Put simply, thebenefits of concentrated ownership by directors and relatives, corporate bodiesand financial institutions outweigh its costs when ownership becomessubstantial. Of particular interest is our evidence of the absence of expropriationby insiders at any level of holdings. One could reasonably expect suchexpropriation to be present in developing countries like India that arecharacterized by relatively weak enforcement of legal rights of minorityshareholders. Our result suggests that other blockholders may have a positiverole in mitigating managerial agency problems in the presence of weak legalprotection.

The positive relationship between managerial holdings and company valuebeyond a certain threshold in India is consistent with the fact that many Indiancorporates, both small and large, are typically family dominated, where managersare de facto owners of enterprises. One could, therefore, expect them to maximizethe surplus generated by corporate assets once their ownership stakes becomesubstantial. A similar result holds for corporate bodies. Given the largelyanecdotal evidence in India that a significant component of corporate holdingsare cross-holdings, our results provide some support for the contention (Kester1992) that significant benefits might flow from lower implicit contractual costs ininsider-dominated systems such as that in India. Our analysis of any potential

ß International Review of Finance Ltd. 2000 189

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 30: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

differences in the effect of managerial holdings and corporate bodies across groupand stand-alone companies in India reveals that differences, where these exist,are not statistically significant.

The other point of consistency with many of the existing studies is with respectto the passivity of institutional investors as mutual funds and insurancecompanies, and the passivity of financial institutions at relatively lowconcentrations of equity holdings. We clearly find no evidence of a positiveand significant relationship between holdings of institutional investors andcompany value at any level of equity ownership, which is consistent with theconclusion arrived at in a recent survey of the role of institutional investors in UScompanies (Black 1998). The passivity of institutional investors at all levels ofequity ownership strengthens the profile of institutional nominees drawn up inseveral accounts in the literature ± nominees who seldom use the voice option asthey have little expertise in the specifics of the company they monitor, and whohave no risk of loss to bear if the value of an investment declines and no reward togain if the value increases.

Our study also finds support for the efficiency of the German/Japanese modelof bank-based governance. Our results suggest that financial institutions startmonitoring the company once they have substantial equity stakes in it. Ourresults also indicate a distinct improvement in the role of financial institutions inenhancing company value when their equity holdings are reinforced byconsiderable debt holdings. These findings, together with the finding thatconcentration of insider ownership beyond a threshold level increases companyvalue, lend credence to the viewpoint that insider-dominated, bank-basedsystems of governance could be effective in the context of transitional andemerging economies.

Our analysis of the role of government-controlled institutional investors anddevelopment financial institutions in monitoring company value throws somevaluable light on to the scarcely analysed question of whether governmentownership of such institutions necessarily impairs their ability to monitorcompanies effectively. While institutional investors and development financialinstitutions are both government-controlled, the impact of these institutions, assuggested by our analysis, is rather conflicting, with the former having nopositive impact on company value, but the latter increasing company valuebeyond a certain level of equity holding. These two results taken together tend tosuggest that government ownership may not necessarily be an impeding factor ineffective monitoring provided there are counteracting factors to neutralize theagency problems associated with government ownership. One suchcounteracting factor could be our finding that significant debt holdings bylending institutions reinforces their incentives to monitor and exercise theirvoice option more effectively relative to institutional investors.

Finally, our analysis also highlights the beneficial effect that foreign equityownership can have on the governance of developing country corporates. Thisresult is in line with that of existing studies with respect to other developingcountries that find companies with foreign ownership to have higher valuation.

190 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 31: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

In our analysis, foreign holdings increase company value at all levels of equityholdings, with the effect being significantly higher once the extent of holdingsbecome substantial. In general, our analysis supports the view emerging fromdeveloped country studies that the identity of large shareholders matters incorporate governance.

Jayati Sarkar and Subrata SarkarIndira Gandhi Institute of Development ResearchGeneral Vaidya MargGoregaon (East) Mumbai ± 400 [email protected]@igidr.ac.in

REFERENCES

Aoki, Masahiko (1995), `Controlling Insider Control: Issues of Corporate Governance inTransition Economies', in Masahiko Aoki and Hyung-Ki Kim (eds), CorporateGovernance in Transitional Economies: Insider Control and the Role of Banks. Washington,DC: EDI, World Bank.

Aoki, Masahiko, and Hyung-Ki Kim (eds) (1995) Corporate Governance in TransitionalEconomies: Insider Control and the Role of Banks. Washington, DC: EDI, World Bank.

Belsley, D., E. Kuh and R. Welsch (1980) Regression Diagnostics: Identifying Influential Dataand Sources of Collinearity. London: Wiley.

Berle, Adolf, and Gardiner Means (1932), The Modern Corporation and Private Property.New York: Macmillan.

Black Bernard (1998), `Shareholder Activism and Corporate Governance in the UnitedStates', in Peter Newman (ed.), The New Palgrave Dictionary of Economics and Law.London: Macmillan.

Blair, Margaret (1995), Ownership and Control: Rethinking Corporate Governance for theTwenty-first Century. Washington, DC: The Brookings Institution.

Blasi, Joseph, and Andrei Shleifer (1996), `Corporate Governance in Russia: an InitialLook', in Roman Frydman, Cheryl Gray and Andrei Rapaczynski (eds), CorporateGovernance in Central Europe and Russia. Volume 2: Insiders and the State. Budapest:Central European University Press.

Brickley, James A., Ronald C. Lease and Clifford W. Smith Jr (1988), `OwnershipStructure and Voting on Antitakeover Amendments', Journal of Financial Economics,20(1/2), 267±91.

Burkart, M., Denis Gromb and Fausto Panunzi (1997), `Large Shareholders, Monitoringand Fiduciary Duty', Quarterly Journal of Economics.

Capon, Noel, John Farley and Scott Hoenig (1996), Toward an Integrative Explanation ofCorporate Financial Performance. Boston: Kluwer Academic Publishers.

Chhibber, P. K., and S. K. Majumdar (1999), `Foreign Ownership and Profitability:Property Rights, Control, and the Performance of Firms in Indian Industry', Journal ofLaw and Finance, 42(1), 209±39.

Choudhury, Mita (1999), `Potential Selectivity Bias in Prowess ± a Company-levelDatabase on Indian Economy', mimeo, Indira Gandhi Institute of DevelopmentResearch.

CII (1998), Desirable Corporate Governance: a Code. Confederation of Indian Industry.

ß International Review of Finance Ltd. 2000 191

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 32: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Claessens, Stijn (1997), `Corporate Governance and Equity Prices: Evidence from theCzech and Slovak Republics', Journal of Finance, 52(4), 1641±58.

Claessens, Stijn, Simeon Djankov and Gerhard Pohl (1996), Ownership and CorporateGovernance: Evidence from the Czech Republic. Washington, DC: World Bank.

Dodd, Peter, and Jerold Warner (1983), `On Corporate Governance: a Study of ProxyContests', Journal of Financial Economics, 11, 401±39.

Edwards, Jeremy, and Klaus Fischer (1994), Banks, Finance and Investment in WestGermany since 1970. Cambridge: Cambridge University Press.

Fama Eugene, and Michael Jensen (1983), `Agency Problems and Residual Claims',Journal of Law and Economics, 26, 327±49.

Frydman Roman, Cheryl Gray, Merek P. Hessel and Andrzej Rapaczynski (1998), `WhenDoes Privatization Work? The Impact of Private Ownership on CorporatePerformance in Transition Economies', Economic Research Report 98-32, C. V. StarrCenter for Applied Economics, New York University.

Gorton, Gary, and Frank Schmid (1996), `Universal Banking and the Performance ofGerman Firms', Working paper 5453, National Bureau of Economic Research,Cambridge, MA.

Gray Cheryl, and Rebecca J. Hanson (1993), `Corporate Governance in Central andEastern Europe', Policy research working paper 1182, World Bank, Washington, DC.

Hopt, Klaus J., Hideki Kanda, Mark J. Roe, Eddy Wymeersch and Stefan Prigge (eds)(1998), Comparative Corporate Governance ± the State of the Art and Emerging Research.Oxford: Clarendon Press.

Jensen Michael, and William Meckling (1976), `Theory of the Firm: ManagerialBehavior, Agency Costs and Ownership Structure', Journal of Financial Economics, 3,305±60.

Keasey, Kevin, Steve Thompson and Mark Wright (eds) (1999), Corporate Governance.Cheltenham: Edward Elgar.

Kester, W. C. (1992), `Industrial Groups as Systems of Contractual Governance', OxfordReview of Economic Policy, 8(3), 24±44.

Khanna, Tarun and Krishna Palepu (1999), `Emerging Market Business Groups, ForeignInvestors and Corporate Governance', in Randall Morck (ed.), ConcentratedOwnership. Chicago: University of Chicago Press.

La Porta Rafael, Florencio Lopez-de-Silanes and Andrei Shleifer (1999), `CorporateOwnership Around the World', Journal of Finance, 54(2), 471±517.

La Porta Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer and Robert W. Vishny(1998), `Law and Finance', Journal of Political Economy, 106(6), 1113±55.

McConnell, John, and Henri Servaes (1990), `Additional Evidence on Equity Ownershipand Corporate Value', Journal of Financial Economics, 27, 595±612.

Miyajima, Hideaki (1995), `The Privatization of Ex-Zaibatsu Holding Stocks and theEmergence of Bank-centered Corporate Groups in Japan', in Masahiko Aoki andHyung-Ki Kim (eds), Corporate Governance in Transitional Economies: Insider Control andthe Role of Banks. Washington, DC: EDI, World Bank.

Modigliani, Franco, and Merton Miller (1963), `Corporate Income Taxes and the Cost ofCapital: a Correction', American Economic Review, 48, 261±97

Morck, Randall, Andrei Shleifer and Robert W. Vishny (1988), `Management Ownershipand Market Valuation: an Empirical Analysis', Journal of Financial Economics, 20, 293±315.

Mulbert, Peter O. Trier (1998), `Bank Equity Holdings in Non-financial Firms andCorporate Governance: the Case of German Universal Banks', in Klaus Hopt et al.(eds), Comparative Corporate Governance ± the State of the Art and Emerging Research.Oxford: Clarendon Press.

Nibler, Marcus (1995), `Bank Control and Corporate Performance in Germany: theEvidence', Research paper no. 48, Faculty of Economics and Politics, University of

192 ß International Review of Finance Ltd. 2000

International Review of Finance

Page 33: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

Cambridge.OECD (1995), Financial Market Trends, no. 62, December. Paris: OECD.Pound John (1988), `Proxy Contests and the Efficiency of Shareholder Oversight', Journal

of Financial Economics, 20(1/2), 237±65.Prowse, Stephen D. (1990), `Institutional Investment Patterns and Corporate Financial

Behavior in the United States and Japan', Journal of Financial Economics, 27, 43±66.Puliani, Ravi, and Mahesh Puliani (1998), Manual of Companies Acts, Corporate Laws and

SEBI Guidelines, 3rd edn. New Delhi: Bharat Law House.Ross, Stephen A. (1977), `The Determination of Financial Structure: the Incentive-

signaling Approach', Bell Journal of Economics, 8, 23±40.Sarkar, Jayati, and Subrata Sarkar (1999), `The Governance of Indian Corporates', in Kirit

S. Parikh (ed.), India Development Report 1999±2000. New Delhi: Oxford UniversityPress.

Schmidt, Klaus (1996), `The Costs and Benefits of Privatization', Journal of Law,Economics and Organization, 12, 1±24.

Shleifer, Andrei, and Lawrence Summers (1988), `Breach of Trust in Hostile Takeovers',in A. J. Auerbach (ed.), Corporate Takeovers: Causes and Consequence. Chicago:University of Chicago Press.

Shleifer, Andrei, and Robert W. Vishny (1986), `Large Shareholders and CorporateControl', Journal of Political Economy, 94, 461±88.

Shleifer, Andrei, and Robert W. Vishny (1997), `A Survey of Corporate Governance',Journal of Finance, 52(2), 737±83.

Singh, Ajit (1995), `Corporate Financial Patterns and Industrializing Economies: aComparative International Study', Technical paper, International FinanceCorporation.

Vendrell-Alda, J. L. M. (1978), Comparing Foreign Subsidiaries and Domestic Firms: aResearch Methodology Applied to Efficiency in Argentine Industry. New York: Garland.

Welsch, R. E. (1980), `Regression Sensitivity Analysis and Bounded-influenceEstimation', in J. Kmenta and J. B. Ramsey (eds), Evaluation of Econometric Models.New York: Academic Press.

White, H. (1980), `A Heteroscedasticity-consistent Covariance Matrix Estimator and aDirect Test for Heteroscedasticity', Econometrica, 48, 817±38.

Willmore, L. N. (1986), `The Comparative Performance of Foreign and Domestic Firmsin Brazil', World Development, 4, 499±517.

Xu, Xiaonian, and Yan Wang (1997), `Ownership Structure, Corporate Governance, andCorporate Performance: the Case of Chinese Stock Companies', Policy researchworking paper 1794, World Bank.

APPENDIX: LIST OF CONTROL AND SPLINE VARIABLES

Control variables

LEVRG is defined as the ratio of long-term debt to total equity plus reserves. Thisvariable captures the effect of corporate tax shields. In India, until recently,returns to equity were subjected to double taxation, which made debt financerelatively less costly than equity finance, ceteris paribus.

LSALES is defined as the logarithm of sales. This variable reflects the effect ofunobserved factors that are related to size. In the product market, size reflectspossible entry barriers that might result from economies of scale. Size also reflectsthe extent of market power of a company. In the capital market, size reflects the

ß International Review of Finance Ltd. 2000 193

Large Shareholder Activism in Corporate Governance in Developing Countries

Page 34: Large Shareholder Activism in Corporate Governance in Developing Countries: Evidence from India

higher ability of large companies to finance investment projects from internalsources as well as their higher ability to raise additional resources through theissue of new equity.

IND-DUMMY_i (one for each industry) is an industry dummy variable. IND-DUMMY_i equals 1 if the observation relates to industry i, and 0 otherwise. Thedummy variable controls for industry-specific effects. Among other things, thisvariable controls for differences in growth opportunities and the riskiness ofdifferent industries.

DIV is a dummy variable. DIV equals 1 if the company is diversified, and 0otherwise. A company is diversified if no single product (based on level 2 SICcode) contributes more than 40% to its total sales. DIV captures, albeitimperfectly, the effect of company-specific risk on performance.

AGE is defined as the number of years to date (1996) since incorporation. Thiscontrols for life-cycle effects as profits of older and mature companies may beenhanced owing to reputation-building and learning efforts. Older companiesmay be subject to the entrenchment of management and lack the ability torespond swiftly to changes in the environment.

EXPINT is export intensity, defined as the proportion of exports to total sales.It controls for the effects of exposure to international competition.

ADVINT is advertising intensity, defined as the ratio of advertising expenditureto sales. This partly captures the effect of intangible assets. Companies that incurhigh advertising expenditure may be successful in building up brand image and,with it, entry barriers for its market.

DEPINT is depreciation intensity, defined as the ratio of depreciationexpenditure to sales. This controls for the technology used by the company.The higher the capital intensity of the company's technological process, thehigher will be the ratio of deprecation to total sales.

Spline variables

DIR_SP1 is the first spline variable with respect to holdings by directors andrelatives. DIR_SP1 � 25 if DIR_S � 25, � DIR_S if DIR_S < 25.

DIR_SP2 is the second spline variable with respect to holdings by directors andrelatives. DIR_SP2 � DIR_S ÿ 25 if DIR_S > 25, � 0 if DIR_S � 25.

The spline variables for holdings by corporate bodies (COR_SP1, COR_SP2),foreign entities (FOR_SP1, FOR_SP2), government-owned institutions (GOV_SP1,GOV_SP2), institutional investors (IINV_SP1, IINV_SP2) and developmentfinancial institutions (DFI_SP1, DFI_SP2) are similarly defined.

GROUP is a dummy variable. GROUP equals 1 if the company belongs to abusiness group, and 0 otherwise.

194 ß International Review of Finance Ltd. 2000

International Review of Finance