Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
i
University of Nigeria
Virtual Library
Serial No.
Author 1
Author 2
Author 3
Title:
Keyword:
Category:
Publisher:
Publication
Date:
Signature:
Description:
MUSA, OJONAGO DANIEL
PG/M.Sc./10/57458
AN EMPIRICAL INVESTIGATION OF THE BUOYANCY AND ELASTICITY OF
TAX: THE NIGERIAN EXPERIENCE
FACULTY OF THE SOCIAL SCIENCES
DEPARTMENT OF ECONOMICS
Chukwuma Ugwuoke
Digitally Signed by: Content manager’s Name
DN : CN = Webmaster’s name
O= University of Nigeria, Nsukka
OU = Innovation Centre
ii
AN EMPIRICAL INVESTIGATION OF THE BUOYANCY AND ELASTICITY OF
TAX: THE NIGERIAN EXPERIENCE
BY
MUSA, OJONAGO DANIEL
PG/M.Sc./10/57458
AN M.SC DISSERTATION SUBMITTED TO THE
DEPARTMENT OF ECONOMICS
FACULTY OF THE SOCIAL SCIENCES
UNIVERSITY OF NIGERIA, NSUKKA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE AWARD OF
MASTER OF SCIENCE (M.SC) DEGREE IN ECONOMICS
SUPERVISOR: PROF. N. I. IKPEZE
iii
FEBRUARY 2013
TITLE PAGE
An Empirical Investigation of the Buoyancy and Elasticity of Tax: The Nigerian Experience
iv
CERTIFICATION
This is to certify that Musa Ojonago Daniel, an M.Sc student of the University of Nigeria
Nsukka with registration number PG/M.Sc/10/57458 has successfully completed the research
required for the Award of Masters of Science Degree in Economics in the University of Nigeria
Nsukka.
Musa Ojonago Daniel Date
PG/M.Sc/10/57458
Supervisor Date
Prof. N. I. Ikpeze
v
APPROVAL
The research work titled: “An Empirical Investigation of the Buoyancy And Elasticity of Tax:
The Nigerian Experience” has followed due process and has been approved to have met the
minimum requirement for the award of the Master of Science degree in the Department of
economics - University of Nigeria, Nsukka.
Approved
Supervisor Date
Prof. N. I. Ikpeze
Head of Department Date
Dr. G. C. Aneke
Dean of Faculty Date
Prof. C. O. Ugwu
External Examiner Date
vi
DEDICATION
This work is dedicated to my wife Mrs Patience Ekojoka Daniel whose act of true love and
unreserved understanding permitted the easy completion of this work.
vii
ACKNOWLEDGEMENTS
My unalloyed thanks are due to God Almighty who is the core brain behind the success
gotten all through the pursuit of this higher degree. However it is must be mentioned here that
this research work would not have been realized without the direct or indirect contributions of
some persons.
Hence, my in-depth gratitude goes to my project Supervisor; Prof N. I. Ikpeze whose
directions and encouragements guided me in the course of this study, your type is rare to come-
by. Appreciation also go to all lecturers in the Department of Economics – University of Nigeria
Nsukka, who in one way or the other contributed to my academic success. Special mention is
made here of Mr. Jonathan E. Ogbuafor. Your painstaking devotion to this work will always be
remembered.
I am eternally grateful to my wife, Mrs Patience Ekojoka Daniel for the expression of
true love and absolute understanding during the pursuit of this program, I LOVE U MA. To my
parents, Late Mr Peter Musa and Mrs Medina Peter; receive the reward of your labour of love in
the bosom of the LORD. I remain grateful to my family members: Dr. and Mrs. Abraham Peter
Musa, Sunday Peter, Elijah Peter, Samson Peter, Isaac Peter, Joy Peter and others. I would not
forget in a hurry the care and love you showed me in the course of this study.
This list would not be complete, if I fail to acknowledge Yuni Denis, Job Ojonugwa, Mr
& Mrs Newman Yakubu, Mr & Mrs David Wayas, Moses Adams, Sharon Musa, all the
members of room 405 Odili hall (2012 occupants), Godwin Egwu, Abraham Daniel, my spiritual
fathers: Pastor Charles Edino, Evang Sunday Oguche, Revd. Ameh, J.S, Bishop Ken Uloh, Pst
Dave John, and Pst Sunday Omojo Akoh. Your fervent prayers, academic and financial support
saw me through this study.
May God richly bless you all for contributing to my success in life.
Musa, Ojonago Daniel
2013
viii
.
ABSTRACT
Many countries in the world have greatly sponsored their government expenditures with the aid
of tax revenue, and owe their developments to this internally generated revenue. The rate of
increase depends on the elasticity and buoyancy of tax and it’s on this premise that, this study
investigated the elasticity and buoyancy of tax in Nigeria in an attempt to ascertain its flexibility
and hence the possibility of increasing the tax base in Nigeria. The study used the Vector Error
Correction Model (VECM) with the dynamic OLS technique. The results of the study therefore
suggest that aggregate revenue is relatively elastic and significantly buoyant according to the
2004 tax reforms. And the results of the four major taxes tested showed that only PPT was found
to be relatively elastic while VAT, CED and CID were relatively inelastic. However the results
further suggest that, while VAT and CIT are not significantly buoyant according to the 2004 tax
reforms, PPT and CED are significantly buoyant. Finally, the study used the 2005 structural
break to establish that aggregate tax revenue dropped significantly after the boom period. The
study therefore concludes that tax in Nigeria is relatively flexible with respect to growth and
therefore more could be done to increase it.
Keywords: Tax, tax reform, elasticity, buoyancy, Nigeria.
ix
x
TABLE OF CONTENTS
Cover Page…………………………………………………………………………..…………....i
Title page…………………………………………………………………………….…………...ii
Certification Page …………………………………………………………………….………….iii
Approval Page ……………………………………………………………………….... ……….iv
Dedication …………………………………………………………………………….………….v
Acknowledgements …………………………………………………………………….…….......vi
Abstract………………………………………………………………………………………......vii
Table of Contents ...……………………………………………………………………….........viii
List of Tables …………………………………………...……………………………..………….x
List of figures…………………….……………………...……………………………..………….x
Appendices.………………….……………………...……………………………..…………..….x
CHAPTER ONE: INTRODUCTION
1.1 Background to the Study..……………………………………………………………………..1
1.2 Statement of the Problem….….…………………………………………………...……..........3
1.3 Objectives of the Study…..……………………………………………………………………4
1.4 Statement of Hypothesis..……..………………………………………………………………5
1.5 Significance of the Study…………….……………………………………………….............5
1.6 Scope of the Study …….………..…………………………………………………………….6
1.7 Limitation of the study…………………………………………………………………….…..6
1.8 Organisation of the study……...………………………………………………………………6
CHAPTER TWO: REVIEW OF LITERATURE
2.1 Conceptual framework…………………………………………………….…………………7
2.2 Theoretical literature………………….…………………………………….………..............8
2.2.1 Tax performance Assessment Approach……………………………….……………….…8
2.2.2 Tax Structure and Economic Development………………….. ………….………….…....9
2.2.3 Realistic Theories…………………………………………………………..…..................10
2.3 Analysis of the Nigerian tax structure …………………………………………...…………12
xi
2.3.1 Tax type and Tax Jurisdiction …………………………………………………................12
2.3.2 Main Taxes in Nigeria ………………………….…………………………………………14
2.3.3 Tax Reforms in Nigeria…………………………………...………………….……………15
2.3.4 Personal Income Tax (Ammendment) Act 2011 ( Pitam)……...…………………….…..16
2.4 Empirical Literature……………………………..…………..……………………………….20
2.4.1 Global Empirical Evidence………………...…………………………………...………….20
2.4.2 Domestic Empirical Evidence…………………………..………………….………………22
2.5 Limitations of previous studies ….........................................................................................24
CHAPTER THREE: METHODOLOGY AND DATA
3.1 Theoretical Framework……………………………………………………………………..25
3.2 Analytical Framework………………………………………………………………………27
3.3 Model Specifications……………………………………………………………………..…28
3.3.1 Estimation of Tax elasticity and Buoyancy………………………………………….…..29
3.3.2 Model Specification of the objectives of the study………………………………….……29
3.3.3 Model Specification for objective 3……………………………….………………………32
3.4 Justification of the Models Specified…………………………………………….................33
3.5 Estimation Technique and procedure…………………………………...............................34
3.6 Data Sources ……………………………………………………………………..................35
CHAPTER FOUR: DATA ANALYSIS AND EMPIRICAL RESULTS
4.1 Presentation of the Unit root test – ADF ……………..………..…….……………….........36
4.2 Regression results of Aggregate tax…………………………………………………………37
4.3 Regression results of the four major types of tax……………………..…………………....40
4.4 Regression results of aggregate tax revenue during and after the oil-boom in Nigeria …...43
4.5 Evaluation of Working 4.5.1 Hypotheses……………………………………………...……44
4.5.1 Test of Hypothesis 1………………………….……..……………………………………..44
4.5.2 Test of Hypothesis …………………….…………………………………………………..45
4.5.3 Test of Hypothesis 3………………….……………………………………………………45
CHAPTER FIVE: SUMMARY, CONCLUSION AND POLICY RECOMMENDATIONS
5.1 Summary of Findings …………………...…………………………………………………..46
xii
5.2 Conclusion ………………………….……..…………………………………………...........47
5.3 Policy Recommendations …………….…..………………………………………….………47
5.4 Suggestions for further Study……………..…………………………...................................49
References………………………………………………………………………………………..50
LIST OF TABLES
Table 2.1: Nigeria’s Federal, State and Local Tax Jurisdiction and Assignment……....……..…13
Table 2.2: Summary of Empirical Literature…………………………………………………….24
Table 1.2: Types of elasticity and designations................... …………………………..….…......27
Table 3.1: Unit Root results for all Variables……………...........................................................36
Table 2.1: Regression results for aggregate tax and its determinants...........................................38
Table 4.2: Regression results for VAT, CIT, PPT and CED…...........................................................41
Table 4.3: Regression results for Aggregate tax to determine the oil boom structural change ...44
LIST OF FIGURES
Figure 4.1: Scatter plot of Aggregate tax against its residual…….……………..........................37
Figure 4.2: scatter plots of CED, PPT, CIT and VAT and their residuals……...…….……..…..40
Figure 4.3: Zivot - Andrews Unit root test for oil revenue ………..……. …………………..…43
APPENDICES
APPENDIX 1: Unit Root test results…………………………………..…………………..……54
APPENDIX II: Regression Results …………….……………………………………………….60
xiii
CHAPTER ONE
INTRODUCTION
1.1 Background of the Study
Tax is a major source of government revenue all over the world. Governments use tax proceeds to
render their traditional functions such as the provision of public goods, maintenance of law and order,
defense against external aggression, regulation of trade and business to ensure social and economic
maintenance (Azubike, 2009). The economic effects of tax include micro effects on the distribution of
income and efficiency of resources use as well as macro effect on the level of capacity output,
employment, prices and growth (Musgrave and Musgrave, 2004).
Among all the sources of revenue to government, taxation is the most important one. Owing to the
inherent power of government to impose taxes, the government is assured at all times of its tax revenue
no matter the circumstances. With modifications as a result of different manifestos of opposing political
parties, the government’s ability to impose tax is unlimited. It is in this light of the significance of this
source of government revenue that its assessment, effectiveness and collectivity are paramount for
optimum benefit (Effiok, 2006).
The Nigerian tax system is lopsided, and dominated by oil revenue. The most veritable tax handles are
under the control of the federal government while the lower tiers are responsible for the less strong
ones—the federal government taxes corporate bodies while state and local governments’ tax
individuals. A major element contributing to this development was the prolonged military rule that had
ignored constitutional provision. During the early stage of the country’s economy, revenues were largely
derived from primary products. Between 1960 and the early 1970s, revenue from agricultural products
dominated, while revenue from other sources was considered as residual. Since the oil boom of 1973/4
to date, however, oil has dominated Nigeria’s revenue profile, thus, indicating that traditional tax
revenue has not assumed a strong role in the country’s management of fiscal policy. Instead of
transforming or diversifying the existing revenue base, fiscal management has merely transited from
one primary product-based economy to another, making the economy susceptible to fluctuations of the
international oil market.
The need to address this problem led to several taxes policy reforms. The tax policy reviews of 1991 and
2003, as well as the yearly amendments given in the annual budget, were geared towards addressing
this issue. But not much has been achieved. Perhaps to understand the importance of tax policy
xiv reforms, one needs to appreciate the urgency for such reforms. Why the need for tax policy reforms in
Nigeria? First, there is a compelling need to diversify the revenue portfolio for the country in order to
safeguard against the volatility of crude oil prices and to promote fiscal sustainability and economic
viability at lower tiers of government. Second, Nigeria operates on a cash budget system, where
proposals for expenditure are always anchored to revenue projections. This facilitates determining the
optimal tax rate for a given level of expenditure. Thus accuracy in revenue projection is vital for devising
an appropriate framework for sustainable fiscal management, and this can be realized only if reforms
are undertaken on existing tax policies in order to achieve some improvement.
Thirdly, the study groups on the review of the Nigerian tax system in 1991 and 2003 highlighted the
need to increase tax revenue and reduce expenditure as the major fiscal issues to be addressed. As such,
the primary objective of the committees was to optimize revenue from various sources within the
country. Finally, the necessity to improve the tax notification procedure was underscored in order to
facilitate effective evaluation of the performance of the Nigerian tax system and to promote adequate
planning and implementation. The quality of management associated with regular and a result-oriented
tax reform has a significant bearing on the overall macroeconomic performance and the distribution of
resources between public and private sectors as well as within the public sector.
Besides, Wagner’s law stipulates that public expenditure is a natural consequence of economic growth
(Demirbas, 1999). The magnitude of government surplus or deficit is probably the single most important
statistic measuring the impact of government fiscal policy on an economy (Siegel, 1979). Many
developing countries including Nigeria in their attempt to increase growth have increased public
expenditure but have not been able to match this increase with revenue mobilization through taxation
and this has resulted in huge fiscal deficit. In the case of Nigeria, tax mobilization as a source for
financing development activities has been a difficult issue primarily because of various forms of
resistance such as evasion, avoidance and corrupt practices attending to it. These activities are
considered as sabotaging the economy and are readily presented as reasons for the underdevelopment
of the country (Adegbie and Fakile, 2011). Again, the Nigerian tax system has been weak in its revenue
mobilization due to inadequate data on the tax base and heavy reliance on oil revenue. With the
volatility in oil prices and excruciating impacts of the recent global financial crises, taxation deserves
more attention now than ever in Nigeria (Adeleke, 2011).
The literature suggests three issues that should guide decisions on the fiscal deficit profile for an
economy. The first relates to the usefulness of fiscal deficit as a tool for enhancing accelerated growth
and development. The second issue relates to the mode of financing the deficit. Some of the financing
xv options include the running down of government accumulated cash balance, net borrowing from the
banking system or from abroad, issuing of new currency as well as drawing down of foreign assets (Ariyo
and Raheem, 1990). Third, and most importantly, a fiscal deficit profile must be sustainable (Buiter,
1983). Otherwise, the country will become perpetually insolvent (Wickens and Uctum, 1990).
What has become of concern to economists and interested observers in recent times is the rising
magnitude of deficits by various governments. There is therefore a growing recognition that the
formulation and implementation of macroeconomic management proposals, most especially for
economic reforms, should explicitly recognize the implications of fiscal deficit on the economy. These
reforms should cover not only the size and financing pattern of government deficits but also the
structure of taxation and the level and composition of public expenditure (Chibber and Khalizadeh,
1988).
The findings of researchers in this field suggested the need for concern about the problem of fiscal
deficit in Nigeria. Some of them reported that fiscal deficit has become a recurring feature of Nigeria’s
fiscal policy with the absence of any identifiable macroeconomic objective to justify this deficit-prone
behaviour (Ariyo and Raheem, 1990). It was also reported that the fiscal deficit in Nigeria has become
unsustainable since 1980 (Ariyo, 1993). An accurate estimate of the optimal level of expenditure
requires knowledge of the buoyancy-total response of tax revenue to changes in national income and
discretionary changes in tax policy over time; and tax elasticity- automatic response of tax revenue to
GDP changes less the discretionary changes. It assists in identifying a sustainable revenue profile for the
country and also helps in determining appropriate modifications to the existing tax structure and rates
as well as areas for improving tax administration. Hence, this study has empirically investigated the
buoyancy and elasticity of tax in Nigeria.
1.2 Statement of the Problem
Fiscal deficit has become a recurring feature of public sector financing all over the world. Its widespread
use is partly influenced by the desire of various governments to respond positively to the ever-
increasing demands of the populace and to enhance accelerated economic growth and development
(Ariyo, 1993). This tendency toward deficit financing is more pronounced in developing countries where
the populace looks to the government for the satisfaction of most needs. However, the rising magnitude
of this deficit since 1980 in Nigeria is now of great concern. An appraisal of the budgetary process in
Nigeria shows that annual expenditure proposals are always anchored on projected revenue, thus, the
xvi accuracy of revenue projection which requires the knowledge of buoyancy and elasticity of tax is a
necessary condition for devising an appropriate framework for fiscal deficit management in Nigeria.
The federal government of Nigeria (FGN) commenced the implementation of a structural Adjustment
programme (SAP) in 1986 that amounted to significant structural changes in the macroeconomic
management framework for the country. One of the core objectives of the SAP is to enhance the degree
of self-reliance within the economy. Of equal importance is the need to diversify the economy’s revenue
base, in order to minimize the extent of dependence on oil as the major source of revenue. All these
have potential implications for the yield of non-oil tax revenue sources. For example, one major
consequence of SAP is the rekindled interest in export of cash crops such as cocoa. Ordinarily, this
should have resulted in a significant upsurge in revenue from export duties, but as part of the reform,
the FGN scrapped export duties as an element of the package of incentives meant to promote exports.
There were significant downward revisions in tax rates and import tariffs as well. The corporate tax rate
was reduced from 45% to 40% in 1987 in order to encourage re-investment activities by existing
organizations and to encourage new investments. Similarly, import duties on certain categories of
imports were reviewed. Among these was the elimination of duties on trucks and commercial vehicles
to ease the transportation problem in the country. Also duty exemptions were granted on items
required on some public projects. Generous tariff concessions were also allowed on machinery and raw
materials that could not be sourced locally, at least not in the short run. Several policies having
implications for the yield of specific tax sources were also initiated to mitigate the negative effect of SAP
on the populace. The introduction of SAP generated several changes in tax-related policies, so that any
growth in GDP during this period might not necessarily translate into higher tax yield. The determination
of the buoyancy and elasticity of individual main taxes therefore becomes an empirical question.
However, before the above economic reform, the advent of the oil boom in the 1973/74 fiscal year
encouraged over-reliance on oil revenue to the neglect of traditional revenue sources. As a result, some
non-oil revenue sources were either abandoned or became of less concern to the government, hence,
the need to assessing the optimal revenue derivable from these non-oil sources and to equally find out
the changes that this major experience had brought about.
The motivation of this study however, is to ascertain the elastic nature of tax and its buoyancy with
respect to tax reforms. Elasticity of tax in this study measures the proportion of change of tax, with
respect to a unit change in economic growth. It is the interest of this study to analyze the elasticity of
the aggregate tax and four of the major taxes. Also the need to improve non-oil proceeds has been on
xvii an increase over the years and so many tax reforms have been made to boast tax revenue. This study
therefore investigates to what extent the 2004 tax reform was buoyant to the tax base.
Based on this background the study seeks to answer the following research questions;
i. What is the tax buoyancy and elasticity of the total tax revenue in Nigeria?
ii. What is the tax buoyancy and elasticity of the main taxes in Nigeria namely; Company Income
Tax, Petroleum Profit Tax, Custom and Excise Duties and Value Added Tax (VAT)?
iii. What are the major structural changes in the tax revenue before and after the oil- boom in
Nigeria?
1.3 Objectives of the Study
The main objective of this study is to analyze empirically the buoyancy and elasticity of tax system in
Nigeria with respect to its efficiency in revenue mobilization in the attempt to address the unsustainable
fiscal deficit challenge facing the country.
The specific objectives are:
i. To ascertain the tax buoyancy and elasticity of the Total Tax Revenue in Nigeria.
ii. To ascertain the tax buoyancy and elasticity of the main taxes in Nigeria namely; Company Income
Tax, Petroleum Profit Tax, Custom and Excise Duties and Value Added Tax (VAT).
iii. To investigate the presence of structural change in tax revenue during and after the oil-boom in
Nigeria.
1.4 Statement of Hypotheses
The hypothesis of the study is formally stated in its null form while the alternative is implied.
Ho1: The total tax revenue in Nigeria is not significantly buoyant.
Ho2: Tax revenue of the main taxes in Nigeria with respect to GDP growth are not significantly buoyant.
Ho3: There is no significant structural change in tax revenue between the pre and post oil-boom
era in Nigeria.
1.5 Significance of the Study
xviii An accurate estimation of the optimal level of expenditure requires knowledge of the buoyancy and
elasticity of the tax system. The study will help the government to know if they are keeping track on tax
revenue mobilization with respect to GDP growth. Additionally, estimation of individual tax buoyancy
and elasticity would help the policy makers especially the fiscal authorities to identify those taxes which
are income elastic or otherwise and therefore aim at directing their efforts at the more elastic ones to
raise overall tax revenue and making sure that the inelastic ones become elastic. Besides, the estimation
of decomposed buoyancy into the pre and post oil- boom era helps to shed more light on the
weaknesses and strengths of the system that existed before and what pertains today. This also helps the
tax authorities in identifying issues that need improvement and restructuring in the Nigerian tax system.
Finally, to academicians, this work will serve as a reference point for further research and consequently
add to the existing stock of literature in public sector Economics.
1.6 The scope of the study
This study investigates empirically the buoyancy and elasticity of the Nigerian tax system (that is, the
efficiency of tax system in terms of its revenue mobilization capacity with respect to GDP) from 1980-
2011 bearing in mind the effect of the 2004 major tax reform in Nigeria to determine the buoyancy as
the tax base experienced changes. Four major taxes were considered namely; Company Income Tax,
Petroleum Profit Tax, Custom and Excise Duties and Value Added Tax (VAT). It also incorporates the
behavior of tax system before and after oil- boom in Nigeria. The core variable of interest are; National
income, Government expenditure, External Grant, Non-Tax Revenue, Budget Deficit and Inflation as the
control variable.
1.7 Limitations of the study
All research works generally record a number of limitations as hindrances in the course of the research
and this was not an exception. The study originally intended to cover six major taxes and this couldn’t
happen due to the unavailability of data for personal income tax and the late start of education tax.
Education tax started in the year 2000, thereby providing a very small degree of freedom with only 11
years covered. On the other hand personal income tax was merged with other types of tax and so
making it difficult to get data for personal income tax exclusively.
1.8 Organisation of the study
This study is organized into five chapters; the first chapter detailed the background of the study showing
the related works, policies and the existing debate. Then the statement of the problem that showed the
xix research and economic problem of the study that translates into the research questions and problems
of the study. The second chapter details the existing theoretical and empirical literature on the subject
matter. While, chapter three showed the analytical framework and model specification of the model
used and the presentation of results and evaluation of hypothesis are discussed in chapter four. The last
chapter captioned as chapter five, summarises the work, concludes proffer recommendations from the
findings and then proposes areas for further research.
CHAPTER TWO
LITERATURE REVIEW
2.1 Conceptual Framework
Every country in the process of formulating its budget undertakes revenue projections. When the
revenues turn out to be smaller than the budget expenditures, countries end up with deficit financing.
Since underdeveloped countries have few possibilities for prolonged external financing of budget
deficits, without causing too much disruption in the macroeconomic environment, each country must
decide how best to increase its internal tax revenues to meet its expenditure needs. One way that
countries raise additional revenue is by making discretionary tax measure changes. The best outcome
expected from such changes is that the tax system will automatically yield corresponding tax revenues
as income or GDP grows, on a sustainable basis.
Elasticity of Tax
Tax elasticity is the ratio of the percentage change in tax revenue to the percentage change in income or
GDP, assuming that no discretionary changes have been made in the tax rate or tax base. It is defined as:
xx
Where: ETY = Elasticity of tax revenue to income or GDP,
ΔT = Change in tax revenue, and ΔY = Change in income GDP.
Mirambo (2001), Elasticity measures the responsiveness of tax revenue to changes in national
income if the tax structure would have remained unchanged. To estimate elasticity of any tax
system, revenue series have to be corrected for the effects of discretionary changes in tax policy.
Mansfield (1972) argues that automatic growth in tax revenue alone, abstracting from discretionary
changes, is the elasticity of the tax. High tax elasticity, that is, a tax elasticity coefficient of one or more,
is said to be particularly desirable since it allows growth in expenditure to be financed by raising tax
revenue without recourse to the politically unpopular decision to raise tax rates.
Tsegaye (1993) says that a high elasticity may simply reflect the progressiveness of the tax
structure, showing positive ratios of tax revenues to increases in income. A high elasticity (that is
greater than unity) implies that the tax revenue increases faster than the income. This means if
the tax is meant to maximize revenue, the government could rely on more elastic taxes which do
not require frequent discretionary changes. It is therefore essential that the tax elasticity be equal
to or exceeds unity to maximize revenue.
Tax Buoyancy
The buoyancy of tax measures the responsiveness of tax revenue to changes in income without
controlling for the discretionary changes in tax policy. The discretionary changes are the changes which
result in more tax revenue from the same tax base. The sources of such changes are changes in tax
legislation or changes in the tax rate (Osoro 1993)
Jayasundera (1991) explains that the buoyancy of a tax system reflects the total response of tax revenue
to changes in national income as well as effects of discretionary changes in tax policies over time.
Matundu (1995) adds to the view of Jayasundera that a buoyancy coefficient which is greater than one
would imply that for every one percent increase in GDP, tax revenue increases by more than one
percent.
2.2 Theoretical Literature
xxi Theories that relates to the subject matter shall be treated basically under three major headings to
reflect and achieve the focus of the study. These are; Tax performance Assessment Approach, the
perspective of tax structure and economic development and Realistic theories.
2.2.1 Tax performance Assessment Approach
Hindrichs (1966) and Musgrave (1969) explained the role of various tax categories in determining tax
effort that expresses the ratio of the actual tax collected to potential tax and used as an indicator of how
much a country is utilizing its taxable capacity. According to the authors, the four main approaches to
assess tax performance are ability to give up approach, efficient resource use approach, ability to collect
approach and comparison with average performance (stochastic) approach. The commonly used
approach for measuring tax effort is to regress the tax to output ratio on a set of variables including the
major determinants of output (Bahl, 1971; Chelliah, 1971) that serves as proxies for tax handles. The
predicted tax ratio therefore gives the ratio that the country would have if it had made the average tax
effort. Thus, it becomes a measure of the taxable capacity of the country while the regression
coefficients act as the average effective rates on the base. The tax effort approach to measuring tax
performance is termed ―static in that it gives the potential for tax increase at a given point in time
through comparisons with other countries. However, in order to determine if a country has made efforts
at increasing tax revenue over a period – tax performance in the dynamic sense which measures the
sensitivity and response of the tax system with respect to income/GDP such as tax buoyancy should be
used.
It is important to obtain both the buoyancy and elasticity of the tax system because the responsiveness
of tax revenue to changes in GDP is of two types – automatic response to GDP change and the response
resulting from discretionary changes in the tax policy such as changes in the tax rate and/or base;
changes in the efficiency of tax administration; the introduction of new taxes and the abolition of
others, etc. Therefore, historical tax revenue series have to be refined by adjusting them to exclude
revenue changes attributable to discretionary measures. Estimating the income tax elasticity is useful
for displaying the extent to which tax system is responsive to changes in the tax composition and the
value of GDP (Teera, 2002). When the elasticity of major revenue sources remains low either due to low
base or evasion or avoidance, the governments raise additional resources through discretionary
measures. Then, the growth of tax revenue comes through high buoyancy rather than through elasticity.
The coefficient of elasticity depends on the level of tax base to changes in income. This makes it possible
to break up the value of elasticity into two components - the response of the tax base to a change in
xxii income, and the response of the tax yield to a change in the tax base of individual taxes through
decomposition of elasticity (Musgrave, 1959).
2.2.2 Tax Structure and Economic Development
A country’s tax system is a major determinant of other macroeconomic indexes. Specifically, for both
developed and developing economics, there exists a relationship between tax structure and the level of
economic growth and development. Indeed, it has been argued that the level of economic development
has a very strong impact on a country’s tax base (Hinricks, 1966, Musgrave, 1969), and tax policy
objectives vary with the stages of development. Similarly, the (economic) criteria by which a tax
structure is to be judged and the relative importance of each tax source vary over time (Musgrave,
1969). For example, during the colonial era and immediately after the Nigerian (political) independence
in 1960, the sole objective of taxation was to raise revenue. Later on, emphasis shifted to the infant
industries protection and income redistribution objectives.
In his discussion of the relationship between tax structure and economic development, Musgrave (1969)
divided the period of economic development into two, the early period when an economy is relatively
under-developed and the period when the economy is developed. During the early period, there is
limited scope for the use of direct taxes because the majority of the populace resides in the rural.
Because their incomes are difficult to estimate, tax assessment at this stage is based on presumptions
prone to wide margins of error.
The early period of economic development is, therefore, characterized by the dominance of agricultural
taxation, which serves as a proxy for personal income taxation, and in Nigeria the various marketing
boards served as effective mechanisms for administering agricultural taxation. Agricultural taxation
substituted for personal income tax given the difficulty in reaching individual farmers and the inability to
measure their tax liability accurately. Further, the large percentage self-employment to total
employment makes effective personal income tax unworkable (Musgrave, 1969). This problem hereby
necessitates the use of the ability-to-pay principle, effectively limiting personal income taxation on the
wage income of civil servants and employees of large firms both of which account for an insignificant
proportion of the total working population.
During the early period of economic development, direct taxes in form of company income taxes cannot
be important because there are few home-based industries. The same principle applies to excise tax (an
indirect tax) on locally manufactured goods. Both will increase in relative importance as economic
development progresses, however, due to growth or non-static nature of the bases of these taxes.
xxiii Several retail outlets also make a sales tax system difficult to implement, and a multiple-stage sales tax
system even more so (Musgrave, 1967). Further, the rudimentary nature of the economy precludes
retail form of taxes.
At this stage also, taxes are difficult to collect because of the lack of skills and facilities for tax
administration (Hinricks, 1966). Given this, a complicated tax structure is not feasible and the amount of
revenue from personal income tax will depend on taxpayer’s compliance and the efficiency of the tax
collector. An important source of government revenue during the early stage of economic development
is the foreign trade sector because exports and imports are readily identifiable and they pass through
few ports. However, revenue from export and custom duties is not stable because of periodic
fluctuations in the prices of primary products. This tends to complicate plan implementation in many
developing countries (Massel et al., 1972).
Economic development brings with it an increase in the share of direct taxes in total revenue. This is
consistent with the experience of developed economies in which direct trades yield more revenue than
indirect taxes. For example personal income tax becomes important as the share of employment in the
industrial sector increases. Also, as the dominance of the agricultural sector decreases, sales tax may be
broadened because a great deal of output and income will go through the formal market as the
economy becomes more monetized. Musgrave (1969) noted that at this stage, taxes may be imposed on
firms or individuals, on expenditures or receipts, and on factor inputs or products among others. He
further argued that there would be a tendency to shift from indirect to direct taxes. His theory relates to
a normal development process, however, it does not consider a situation where the sudden emergence
of an oil boom provides an unanticipated source of huge revenue. Hence, this stereotype may not be
applicable to an oil-based economy like Nigeria. Nevertheless, the theory still represents a benchmark
against which country-specific empirical evidence may be compared.
2.2.3 Realistic Theories
In this sub-section economic theory that relates the dependent variable to independent variables is
analysed. In this regard we take tax revenue as the dependent variable and all the other variables that
influence it as independent variables.
xxiv
According to Wojciech (2003), tax elasticity and buoyancy in literature had been estimated or analysed
by regressing aggregate tax based revenue on gross domestic product (GDP) - a proxy for the tax base,
and incorporating a dummy variable singer (1968) or some other proxy to capture the exogenous
influences exerted by tax legislation on the tax net, the tax rate or the tax structure. What should be
noted is that besides discretionary changes in the tax net, rate and structure arising from legislative
innovations, there are other sources of exogenous influences on the tax yield, and hence on tax
elasticity and buoyancy.
External developments in open economies affect the tax base and hence the tax yield both directly and
indirectly. Ventures like the East African common market can affect tax revenue generation. `Studies
have revealed that a relatively large foreign trade sector tends to be related with a high tax level. It has
been argued that this relationship is due to the administrative ease of taxing imports and exports.
However, different authors have formulated the variables of foreign trade differently as M/Y, (M+X)/Y
and X/Y but (M+X)/Y was found to be superior because the ratio of its coefficient to its standard error
was the highest and its equation had the highest adjusted R2 (Joergen Lotz and Elliott 1970). From
economic analysis carried out by Joergen et al,(1970), it is stated that there is a positive relationship
between tax revenue and the country’s openness.
National income (GDP) is theoretically positively related to tax revenue. Similarly, an evaluation of tax
systems in developing countries reveals a positive relationship between national income and total tax
revenue. This finding supports the hypothesis that as countries develops, tax bases expand more than
proportionately to the growth in national income (Musgrave, 1984).
According to Akinlo (2006), external grant has a negative relationship with tax revenue. That implies
increment in external grants reduces effort to collect revenue. In recent years there has been a growing
interest in the possible linkages between high levels of development and taxation in Africa. It is assumed
that without aid, governments would be forced to raise more taxes or borrow from other sources.
According to the present findings, increase in development aid appears to be a source of disincentive to
making full use of domestic resources for revenue generation (Ayoki 2007).
xxv There is an inverse relationship between budget deficit and total tax revenue. When budget deficit
increases, expands the external public debt which suffocates internal investment and hence reduces tax
base and tax revenue generation (Tanzi 1981,1992).
Agbeyegbe ,Stotsky and Woldemariam (2005) say that theoretical studies show that real exchange rate
has a positive relationship with total tax revenue. For example the depreciation of Uganda shilling by say
one percent against Us dollar can increase overall tax by a certain percent level point of GDP.
Christodoulakis (1994) says that there is a negative relationship between total tax revenue and inflation.
Inflation reduces the value of tax revenue and tax rates cannot be adjusted automatically with reference
to changes in underlying inflation.
Bolnick (2002) in his article stated that literacy rate has a positive relationship with total tax revenue.
The more people are educated the more they learn the importance of tax and can easily comply with tax
payment. The government can achieve a significant rise in tax revenue by investing in mass education.
One of the millennium Development goals is to promote literacy. Each country is expected to advocate
for universal primary education. Each child has a right to education. The parents helped by the
government should make it a point that their children receive basic education. They should not stop at
basic education but their children should aim at attaining professionalism through attainment of tertiary
or university education.
Political stability influences the level of tax collected. Instability lowers tax revenue collected. Thus,
there is negative relation between political upheaval and total tax revenue ( Ayoki 2007).
2.3 Analysis of the Nigerian tax structure
This subsection presents an overview of tax type and jurisdiction, brief insight on some types of tax
within the scope of this study and issues that pertains to tax reforms in Nigeria.
2.3.1 Tax type and Tax Jurisdiction
The assignments of fiscal instrument in Nigeria were guided by constitutional provision. The federal
constitution gave the federal government exclusive power to collect levies like custom and excise,
company tax, education tax and mining rents, VAT etc. All these revenues (with the exception of
education tax) are paid into the federation account for distribution among the three tiers of government
xxvi in line with national constitution. The states and local governments are left with the powers to collect
other fees. The main types of tax revenue for the federal and sub-national government are listed in the
Table below. Nigeria’s local government’s autonomy to perform their functions is not absolute. They
retain their functions and fiscal relations with states and federal government.
Table 2.3: Nigeria’s Federal, State and Local Tax Jurisdiction and Assignment
Tax Legal Jurisdiction Collection Retention
Import duties Federal Federal Federation account
Excise duties Federal Federal Federation account
Mining rents and royalty Federal Federal Federal account
Petroleum profits tax Federal State State
Capital gains tax Federal State State
Personal income tax (other than listed
in 8)
Federal State State
Personal income tax: armed and
police forces, external affairs officers,
non-residents, residents of the
federal capital territory
Federal Federal Federal
Value added tax (sales tax before
1994)
Federal Federal/state Federal/state
Company tax Federal Federal Federal
Stamp duties Federal State State
Gift tax Federal State State
Property tax and ratings State State/local State/local
Licenses and fees Local Local Local
xxvii
Motor park dues Local Local Local
Motor vehicle State Local Local
Capital transfer tax (CTT) Federal State State
Pools betting and other betting taxes State State State
Entertainment tax State State State
Land registration and survey fees State State State
Market and trading license and fees State Local Local
Source: Anyanwu, 1995; Jimoh, 2003; Federal Republic of Nigeria Constitutions, 1963, 1979 and 1999.
2.3.2 Main Taxes in Nigeria
The Chartered Institute of Taxation of Nigeria (CITN): Nigerian Tax Guide and Statutes did a
comprehensive explanation of the main taxes in Nigeria out of which the targeted ones are shown
below.
Personal Income Tax: The legal basis for this tax is found in the provisions of the personal Income Tax
Decree [now Act]. 104 of 1993. Every taxpayer in Nigeria is liable to pay tax on the aggregate amount of
his income whether derived from within or outside Nigeria, the salaries, wages, fees, allowances, and
other gains or benefits, given or granted to an employee are chargeable to tax. The employers of labour
are deemed to be agents of the tax authority for the purposes of remitting taxes deducted from salaries
due to employees. However, residency of the taxpayer determines the extent of a taxpayer’s liability in
Nigeria. A person’s place of residence for this purpose is defined as a place available for his domestic use
in Nigeria on a relevant day, excluding hotels and rest houses. A person is deemed resident in Nigeria if
he resides in Nigeria for 183 days in any 12 month period, expatriates holding residence permits are
liable to tax in Nigeria even if they reside in the country for less than 183 days in any 12-month period.
Once residence can be established, the relevant tax authority of the territory is the tax authority in
which the taxpayer has his place of residence or principal place of business.
Company Income Tax: This tax is payable for each year of assessment of the profits of any company at a
rate of 30%. These include profit accruing in, derived from, brought into or received from a trade,
business or investment. Also companies paying dividends to its shareholders are first obliged to pay tax
on its profits at the companies’ tax rate. Generally, in Nigeria, Company dividends or other company
xxviii distribution whether or not of a capital nature made by a Nigerian is liable to tax at source of 10%.
However, dividends paid in the form of bonus share or scrip shares to individual shareholders are not
subject to tax. Also, where a company is a shareholder in another company then such dividends are
excluded from the profits of the company for the purposes of computation of the tax.
Petroleum Profit Tax: Nigerian law by virtue of the petroleum profit Tax Act requires all companies
engaged in the extraction and transportation of petroleum to pay tax. The taxable income of a
petroleum company comprises proceeds from the sale of oil and related substances used by the
company in its own refineries plus any other income of the company incidental to and arising from its
petroleum generations.
The taxable income of a petroleum company is subject to tax at 85%, but this percentage is lowered to
65.75% during the first 5 years of operation. Where oil companies operate under production sharing
contracts they will be liable to tax at a rate of 50%. There are however some concessions granted
petroleum companies known as, Capital Allowance and Petroleum Investment Allowance; the former is
deducted in arriving at the taxable income and entails expenditure on equipment, pipelines, and storage
facilities, buildings and drilling costs, these are referred to as qualifying assets. The applicable rate of
Capital Allowance for any year of 20% of the cost of the qualifying assets applied on a straight-line basis
for the first 4 years and 19% for the 5th year. The latter is regarded as an addition to capital allowance
and covers allowance in respect of new investments in assets for petroleum exploration; it is available in
the accounting period in which the assets are first used.
Education Tax: An education tax of 2% of assessable profits is imposed on all companies incorporated in
Nigeria. This tax is viewed as a social obligation placed on all companies in ensuring that they contribute
their own quota in developing educational facilities in the country.
Custom and Excise Duties: This is a source of revenue for the federal government. It is payable by
importers of specified goods. This tax is charged solely by the federal government and collected through
the Nigeria custom service. Excise duty was levied on a variety of locally produced goods until 1998
when the tax was abolished. It was however partially reintroduced, with effect from January 1, 1999.
The application law for customs and excise duties is the Customs and Excise Management Act.
Value Added Tax (VAT): This was introduced by the VAT decree No 2 of 1993, to replace the old sales
tax. It is a consumption tax levied at each stage of the consumption chain, and is borne by the final
consumer. It requires a taxable person upon registering with the Federal Inland Revenue Services to
charge and collect VAT at a flat rate of 5% of all invoiced amounts of taxable goods and services.
xxix VAT paid by a business on purchases is known as input tax, which is recovered from VAT charged on
company’s sales, known as output tax. If output exceeds input in any particular month the excess is
remitted to the Federal Inland Revenue Services (FIRS) but where input exceeds output the taxpayer is
entitled to a refund of the excess from FIRS though in practice this is not always possible.
2.3.3 Tax Reforms in Nigeria
The dependence on oil revenue by all tiers of government in Nigeria has made the federal government
to reform the existing tax laws. According to Alli (2009), the objectives of tax reforms in Nigeria includes:
to bridge the gap between the National Development needs and the funding of the needs; to ensure
taxation, as a fiscal policy instrument, to achieving improved service delivery to the public; to improve
on the level of tax derivation from non- oil activities, vis- a`- vis revenue from oil activities; efforts at
constantly reviewing the tax laws to reduce/ manage tax evasion and avoidance; and to improve the tax
administration to make it more responsive, reliable, skillful and taxpayers friendly and to achieve other
fiscal objectives.
The Nigerian tax reform has experienced series of reforms since 1904 to date. The effects of the various
reforms in the country is as follows: Introduction of income tax in Nigeria between 1904 and 1926; grant
of autonomy to the Nigerian Inland Revenue in 1945; the Raisman Fiscal Commission of 1957; formation
of the Inland Revenue Board in 1958; the promulgation of the petroleum profit tax ordinance No. 15 of
1959; the promulgation of Income Tax Management Act 1961; establishment of the Lagos State Inland
Revenue Department; the promulgation of the companies Income Tax Act (CITA) 1979; establishment of
the Federal Board of Internal Revenue Service between 1991 and 1992; and tax policy and
administration reforms amendment 2001 and 2004.
The government embarked upon the latest tax reform process by instituting study group on the Nigerian
Tax System, consisting of individuals from business, academia, and the government to study the present
tax laws and recommend the appropriate reform in general and their impact to the overall economy. As
a result of the reform, nine (9) bills on tax reforms where approved by the Federal Executive Council for
the consideration of the National Assembly and subsequently passed as Act. The Acts, are as
enumerated as follows: Federal Inland Revenue Service Act 2004; companies Income Tax Act 2004;
Petroleum Profit Tax Act 2004; Personal Income Tax Act 2004; Value Added Act 2004; Education Tax Act
2004; Customs, Excise Tariffs, etc (consolidation) Act 2004; National Sugar Development Act 2004; and
National Automotive Council Act 2004.
xxx The Chartered Institute of Taxation of Nigeria (CITN), established in 1982 and Chartered by Act No. 76 of
1992 to regulate tax practice and administration in the country, and to this extent a major stakeholder
in the Nigerian tax system submitted a memorandum on the proposed 2004 amendment. Their
memorandum objectives include: to strengthen the powers of the Accountant General of the federation
to monitor the revenue being generated by ministries, extra-ministerial departments and parastatals; to
enforce remittance of the revenues collected to the consolidated revenue fund or federation account;
to strengthen the oversight functions of the National Assembly in monitoring the revenue generated by
ministries, and others; to increase the penalty for under declaration of revenue generated from three to
five years.
2.3.4 Personal Income Tax (Ammendment) Act 2011 ( Pitam)
An Act, Personal Income Tax (Amendment) Act, 2011 was enacted to amend the Personal Income Tax
Act, Cap. P8. Laws of the Federation of Nigeria (LFN) 2004 and related matters. The Bill seeks to amend
the provisions of the Principal Act, Personal Income Tax Act, Cap. P8. LFN 2004, by substituting the
existing Sixth Schedule with the new Sixth Schedule. It was signed into law on 14 June 2011 but only
announced during the budget presentation by the President on 13 December 2011.
Key Changes
The highlights of the New Personal Income Tax (Amendment) Act, 2011 are as follows:
· Gross income is defined to include benefits in kind, gratuities, superannuation and any other incomes
derived solely by reason of employment.
· Principal place of residence to include places where branch offices and operational site of companies
are situated.
· Operational sites are defined in the bill to include oil terminals, oil platforms, flow stations,
construction sites, etc with a minimum of 50 workers.
· Any individual irrespective of status who works in more than one state for at least 20 days in at least 3
months of every assessment year will be liable to tax in such a state.
· Full tax exemption to be granted on interest from bonds issued by Federal, State and Local
Governments and their Agencies, corporate entities and interest earned on short term securities.
xxxi · Interest for default in tax remittance to be charged at the prevailing minimum re-discount rate of the
Central Bank of Nigeria on an annual basis.
· Individual tax clearance certificates (TCC) to be required for any change of ownership of vehicles and
application for land title transfer or perfection.
· 1% bonus which was previously given to early filers of tax returns has been abolished.
· Amendment to Section 33 is a new insertion to introduce a presumptive tax regime to bring in those in
the informal sector to the Tax Net. The section highlights the fact that “There will be a consolidated tax
free allowance of N200,000 or 1% of gross income whichever is higher plus 20% of the gross income, the
balance of income shall be taxed as specified in the new graduated tax table.”
· Amendment to Section 37 is a new insertion to increase the Minimum Tax rate payable from 0.5% to
1% of gross income since the threshold has increased from N30,000 to N300,000 in conformity with the
tax table in the schedule to the Bill.
· Section 52 introduces stiffer criminal penalties for non-compliance. Penalty for late filing of returns is
now N500,000 for corporate bodies and N50,000 for individuals.
· Amendment to Section 73 is to create a refund mechanism. “Introduction of WHT refund mechanism.
The excess WHT will be refunded within 90 days or granted as future tax credits”
· Section 80 New PAYE filing requirement: It is mandatory for employers of labour to file returns of their
employees not later than 31 January of every year; showing all the emoluments paid to its employees in
the preceding year.
· Tax officers to apply for a warrant from the High Court before levying any distress on a taxpayer.
· “itinerant worker” previously defined to mean an individual who works at any time during a year of
assessment (other than as a member of the armed forces) for a daily wage or customarily earns his
livelihood in more than one place in Nigeria and whose total income does not exceed N 600; to be
modified as follows:
· “itinerant worker includes an individual irrespective of his status who works at any time in any state
during a year of assessment (other than as a member of the armed forces) for wages, salaries or
livelihood by working in more than one state for a minimum of twenty (20) days in at least three (3)
xxxii month of every assessment year. The relevant tax authorities are empowered to collect taxes from
itinerant worker.”
· Place of Residence for an itinerant worker has been amended thus: “in the case of an individual who
works in the branch office or operational site of a company or other body corporate, the place at which
the branch office or operation site is situated: Provided that operational site shall include Oil Terminals,
Oil Platforms, Flow Stations, Factories, Quarries, Construction Site with a minimum of fifty (50) workers
etc”.
· The relevant tax authorities are empowered to collect taxes from an itinerant worker.
Conditions for exemption from Personal Income Tax:
1. For any employment wholly or partly performed in Nigeria have been modified to require evidence
that such individuals are liable to tax in another country under the provisions of a double tax treaty.
2. Expatriates who meet the tax exemption conditions in (1) above are being exposed to tax in Nigeria if
the other country does not have a double tax agreement with Nigeria.
3. Where remuneration is borne by a fixed base of the non-resident employer in Nigeria, the individual
will be deemed to be liable to tax in Nigeria.
4. The 183 days residency rule under which a foreigner becomes liable to tax regardless of other
conditions has been modified to include periods of temporary absence or leave.
Income Tax Table:
Relief shall be granted thus;
· Higher of 1% of Gross
· Or a consolidated relief allowance on income at a flat rate of N200,000; Plus 20% of annual gross
Income.
The taxable income bands have been increased while the top tax rate was reduced marginally from 25%
to 24%.
After the relief allowance and exceptions had been granted in accordance with the above, the balance
of income shall be taxed as specified in the following tax table, subject to a minimum of 1percent of
Gross Income whichever is higher.
xxxiii · First N300,000 @ 7%
· Next N300,000 @ 11%
· Next N500,000 @ 15%
· Next N500,000 @ 19%
· Next N1,600,000 @ 21%
· Above N3,200,000 @ 24%
Tax Exempt Income:
The following deductions are tax exempt –
(a) National Housing Fund Conribution
(b) National Health Insurance Scheme
(c) Life Assurance Premium
(d) National Pension Scheme
(e) Gratuities
Implications to Tax Payers
· Reduced tax for Tax Payers in line with the National Tax Policy objective of reducing direct tax and
increasing indirect tax.
· Section 3: Temporary staffs (casual workers, interns, contract staff) are now specifically liable to PIT.
· Deletion of various obsolete provisions.
· Strengthening enforcement provisions in the law.
· Tax Exemption of pensions from chargeable income.
· Replacement of the old reliefs and allowances with consolidated and higher allowances.
· Amendments to interpretation and other relevant sections of the law relating to the Income Tax Table.
· Increase in penalty rates for violation of laws to deter non-compliance.
xxxiv
Comparative with the Old Regime
The new PITAM ensures that high income earners pay more tax and vice versa. The Bill seeks to bring
personal income tax administration up to date with existing realities and simplify compliance by tax
payers. It also replaces a complicated system of computing reliefs and allowances under the existing
laws with a simpler and more straightforward system. In line with the National Tax Policy and
furtherance to the ongoing reforms of the tax system, the amendments will shift focus from direct to
indirect taxation to serve as a tool for income redistribution and computation of taxable income, while
taxpayers would be expected to respond through a higher compliance rate.
2.4 Empirical Literature
A lot of empirical literature exists on tax revenue and reforms. However, this study narrows its focus on
the buoyancy and elasticity of tax. Some of the research been carried out on buoyancy and elasticity of
tax include; Osoro (1975); Ole (1975); Choudhry (1979); Mtatifikolo (1990); Njoroge (1993); Kusi
(1998);Milambo (2001); Graeser (2004); Twerefou el tal (2009); Milwood(2012) Omoruyi (1983) and
Ariyo (1997).
2.4.1 Global Empirical Evidence
Osoro (1993) study on the revenue productivity of the tax system in Tanzania for the period 1969-1990
showed a low elasticity for the total tax system, as well as for individual taxes. Elasticity for total tax
revenue was 0.76 with buoyancy of 1.06 which means that the Tanzania tax system was unproductive
over the study period. Elasticity of individual taxes were as follows: income tax, 0.76; company tax, 1.13;
sale tax, 0.79; PAYE, 0.66; import duty, 0.55. Only company tax had elasticity above unity, which means
that 1 percent increase in GDP was on average accompanied by 1.13 percent increase in revenue from
company tax. The other four taxes had elasticity below unity, meaning that these taxes lagged behind
GDP. The study concluded that the tax reforms in Tanzania had failed to raise tax revenues.
Ole (1975) estimated income elasticity of tax structure of Kenya for the period 1962-1972. Tax revenue
was regressed on income without adjusting for the unusual observations. The results showed that the
tax structure was income inelastic with an index of 0.81 for the period studied. After the study it was
recommended that the tax system required urgent reforms to improve its productivity. The results also
xxxv implied that Kenya’s tax structure was not buoyant and therefore the country would require foreign
assistance to close the budget deficit.
Choudhry (1979) estimated the elasticity of tax revenue of the United States, United Kingdom, Malaysia
and Kenya. The overall elasticity were 1.04 and 1.24 for the United States and United Kingdom
respectively. Malaysia and Kenya had slightly higher elasticity of 1.57 and 32 respectively. Estimation of
the buoyancy revealed that in the United States and the United Kingdom, revenue reducing
discretionary changes in income taxation contributed to the low buoyancy and elasticity, while in
Malaysia and Kenya, revenue increasing DTMs contributed to the comparatively higher buoyancy and
elasticity.
Mtatifikolo (1990) did a study on the performance of the tax system for the period since the major tax
reforms of 1973-1984. This study gives an estimate of the buoyancy and elasticity of the major taxes.
Mtatifikolo used the same method as one adopted by Thac and Lim (1984) as an indicator of the tax
effort of the government of Tanzania. The results showed buoyancy of 0.998 for the total tax system.
Buoyancies of individual taxes were as follows; PAYE, 0.97; Business income tax, 1.27; income tax, 1.17;
tax on import, 1.16; sales and excise tax, 1.16. The study revealed that having observed a low buoyancy
of the business income tax relative to the elasticity, this suggests substantial tax evasion and avoidance.
Njoroge (1993) studied the revenue productivity of tax in Kenya for the period 1972/73 to 1990/91. Tax
revenue was regressed on income after adjusting tax revenues for discretionary changes. The period of
study was divided into two to make it easier to analyze the effects of tax reforms on revenue from
various taxes. Income elasticity of total tax structure was found to be 0.67 for the period 1972 to 1981.
This meant that the government received a decreasing share of rising GDP as tax revenues. The elasticity
estimates for individual taxes were as follows: sales tax 0.6, import duties 0.45 and income tax 0.93. The
buoyancy for the overall tax system for the same period was 1.19, implying that the tax system was
quite buoyant. For the period 1982 to 1991, Njoroge (1993) found that the overall elasticity was 0.86
while buoyancy was 1.00. The study concluded that from a revenue point of view, the system did not
meet its target; hence it required constant review as the structure of the economy changes. However,
the results could not be relied upon because the study never took into account time series properties of
the data.
Kusi (1998) studied tax reform and revenue productivity of Ghana for the period 1970-1993. The results
showed a pre-reform buoyancy of 0.72 and elasticity of 0.71 for the period 1970-1982. The period after
xxxvi reform 1983-1993, showed increased buoyancy of 1.29 and elasticity of 1.22. The study concluded that
the reforms had contributed significantly to tax revenue productivity for the period 1983-1993.
Milambo (2001) used the Divisia Index method to study the revenue productivity of the Zambian tax
structure for the period 1981-1999. The results showed elasticity of 1.15 and buoyancy of 2.0, which
confirmed that tax reforms, had improved the revenue productivity of the overall system.
Graeser (2004) studied the tax elasticity of Botswana for 1990 – 2003 periods. He found that the overall
elasticity of the sales tax regime is 1.29, which according to him is a bit higher than expected. The study
showed that generally, broad consumption taxes usually show elasticity slightly less than one-perhaps
0.95 and part of this can be explained by the specific nature of the Botswana economy. For the period of
1990 to 2000, motor vehicle tax had elasticity of over +2.3, and for the last four years (2000-2003), there
was a reduction from elasticity of +2.3 to only +0.5.
Twerefou et al (2009) used the Dummy Variable Technique to control for effects of the Discretionary Tax
Measures on Historical Time Series Data for the period 1970-2007 to estimate the elasticity of the
Ghanaian tax system. Their findings revealed that the overall tax system in Ghana was buoyant and
elastic in the long run and buoyancy exceeded the elasticity, but in the short run the reverse was the
case. They also observed an improvement in both buoyancy and elasticity over the reform period (1985-
2007) as evidenced in pre-reform buoyancy and elasticity coefficient which were generally less than
unity but became greater than one after the reform.
Milwood, T.T (2012) Studied the relationship between GDP growth and the growth in tax revenue as
well as the responsiveness of taxes to fiscal policy in Jamaica. He estimated the buoyancy and elasticity
of tax revenues spanning the period March 1998 to December 2010, using the Divisia Index (DI)
approach. It was found that discretionary tax measures have had an overall impact on growth in total
revenue over the period. However, the automatic response of revenue to changes in the base was found
to be less than unity.
2.4.2 Domestic Empirical Evidence
Omoruyi (1983) did a comprehensive assessment of the productivity of the Nigerian tax system. He
evaluated the buoyancy of the tax system as defined by Sahota (1961) and Ghai (1966) for the period
1960 to 1979. He focused on both the indirect taxes such as import, export and excise duties, as well as
direct taxes such as personal income tax (federally collected) and petroleum profit tax. He discovered a
general satisfaction on the level of tax productivity in Nigeria during the period under review.
xxxvii Ariyo (1997) in his study of the productivity of the Nigerian tax system improved upon the one done by
Omruyi (1983) in the following respects. First, the study covered the period 1960-1990, and therefore
updates the analysis. Second, the study captured the impact of the structural changes in
macroeconomic management framework introduced since 1966. Third, Omoruyi (1983) disaggregated
his analysis in terms of decades (1960-1967, 1970-1980, etc). His research findings were as follows: For
the period covered by the study, there was an elasticity of 1.18 for government tax revenue relative to
GDP. The non-oil component, however, had a lower elasticity coefficient of 0.94, while the performance
of import duties (IMD) showed the same pattern. The cumulative effect of the oil boom PPT (petroleum
profit tax) was reflected with an elasticity of 2.60 and 1.51 in relation to GDP. He also found out that
company income tax was elastic with an elasticity coefficient of 1.21, which suggests an improved
efficiency in tax collection from this source over the years.
Table 4.2: SUMMARY OF EMPIRICAL LITERATURE
Author/Date Place of
Location
Nature of
Data
Methodology Findings
Osoro
(1993)
Tanzania Time series Proportional
Adjustment
Approach
Low elasticity for the total tax system, as
well as for individual taxes. The study
concluded that the tax reforms in Tanzania
had failed to raise tax revenues.
Ole (1975) Kenya Time series Dummy
Variable
Approach
The results showed that the tax structure was
income inelastic with an index of 0.81 for the
period studied.
Choudhry
(1979)
United States,
United
Kingdom,
Malaysia and
Kenya.
Time series Proportional
Adjustment
Approach
The overall elasticity were 1.04 and 1.24 for
the United States and United Kingdom
respectively. Malaysia and Kenya had
slightly higher elasticity of 1.57 and 32
respectively. Estimation of the buoyancy
revealed that in the United States and the
United Kingdom, revenue reducing
discretionary changes in income taxation
contributed to the low buoyancy and
elasticity, while in Malaysia and Kenya,
revenue increasing DTMs contributed to the
comparatively higher buoyancy and
elasticity.
Njoroge
(1993)
Kenya Time series Proportional
Adjustment
Approach
Overall tax system was buoyant and elastic.
Kusi (1998) Ghana Time series Dummy
Variable
Approach
The study concluded that the reforms had
contributed significantly to tax revenue
productivity for the period 1983-1993.
Milambo
(2001)
Zambian Time Series Divisia Index The results showed elasticity of 1.15 and
buoyancy of 2.0, which confirmed that tax
reforms, had improved the revenue
productivity of the overall system.
Twerefou et
al (2009)
Ghana HTSD 1970-
2007
Dummy
Variable
Techique
The overall tax system in Ghana was buoyant
and elastic in the long run and buoyancy
exceeded the elasticity, but in the short run
the reverse was the case.
xxxviii Milwood,
T.T (2012)
Jamaica Time Series Divisia Index
(DI)
It was found that discretionary tax measures
have had an overall impact on growth in total
revenue over the period. However, the
automatic response of revenue to changes in
the base was found to be less than unity.
Omoruyi
(1983)
Nigeria Time Series Aggregative
measure
A general satisfaction on the level of tax
productivity in Nigeria during the period
under review.
Ariyo (1997) Nigeria Time Series Autoregressive
logarithm
approach
For the period covered by the study, there
was an elasticity of 1.18 for government tax
revenue relative to GDP. The non-oil
component, however, had a lower elasticity
coefficient of 0.94, while the performance of
import duties (IMD) showed the same
pattern. The cumulative effect of the oil
boom PPT (petroleum profit tax) was
reflected with an elasticity of 2.60 and 1.51
in relation to GDP. He also found out that
company income tax was elastic with an
elasticity coefficient of 1.21, which suggests
an improved efficiency in tax collection from
this source over the years.
2.5 Limitations of Previous Studies
A lot of research works have been carried out by several authors in the subject matter.
Osoro(1993),Mtatifikolo(1990),Njoroge(1993),Kusi(1998),Milambo(2001),Twerefou(2009),
Milwood(2012), Omoruyi(1983) and Ariyo(1997) worked on both buoyancy and elasticity of tax while
Ole(1975), Choudhry(1979) and Graeser(2004) worked on only elasticity of tax revenue for their various
countries.
The work done by Twerefou et al 2009 for Ghana and Milwood 2012 for Jamaica are the most recent
studies in this field. Apart from the country difference, the former used the OLS estimation technique
without considering its possible shortfalls. This present study shall adopt the standard OLS estimation
procedure which was modified into Dynamic OLS (DOLS) and was incorporated in vector error correction
model (VECM) model. According to the proposition of Sobel and Holcombe (1996), these ideas correct
for the shortfalls of the OLS model. The latter only mentioned VECM but only modeled ordinary OLS by
building ratios and not VEC Model. This present work shall observe all the procedures in modeling VECM
and also study the buoyancy and elasticity of tax across a major economic experience (pre and post oil-
boom era).
From the review of the empirical studies on the subject matter, much has been written for other
countries of the world, but such a study is still very scanty in Nigeria. The studies by Omoruyi (1983) and
Ariyo (1997) which according to my findings are the most comprehensive ones for Nigeria are deficient
xxxix (on policy relevance) to the extent of how stale their data sets are. More so, so many economic
development had taken place after their work especially that of latest (2004) tax reform, hence, this
present study shall make use of recent data and update the analysis to accommodate the present
economic realities. Besides, both studies did nothing on some main taxes like Value Added Tax (VAT)
and education tax which the present study shall incorporate in its findings.
Finally, most researchers on this topic have overlooked the time series properties of the data used;
hence, their results cannot be relied upon for policy making. In this present study, the time series
properties of the data would be observed jealously by the researcher.
CHAPTER THREE
METHODOLOGY AND DATA
3.1 Theoretical Framework
Every economic research that wants to avoid biasness of both specification and estimation procedure
must pay serious attention to theories. Hence, we considered very importantly, the review of some
economic growth theories as an establishment of the best direction to guide the model of this study.
Obviously, before the growth theory proposed by Romar, there were other growth theories which
thrived. Solow growth theory was one of such theories which was then in vogue. The Solow growth
theory was also known as the exogenous theory because it professed technology as an exogenous factor
which determines growth. One of the basic assumptions of the Solow model is the diminishing returns
to labour and capital and constant returns to scale as well as competitive market equilibrium and
constant savings rate. However, what is crucial about the Solow model is the fact that it explains the
long run per capita growth by the rate of technological progress, which comes from outside the model.
The endogenous growth theory or new growth theory was developed as a reaction to the flaws of the
neoclassical (exogenous) growth theory. Romar endogenous growth theory was first presented in 1986
in which he takes knowledge as an input in the production function. The theory aimed at explaining the
long run growth by endogenizing productivity growth or technical progress.
The major assumptions of the theory are:
1. Increasing returns to scale because of positive externalities.
2. Human capital (knowledge, skills and training of individuals) and the production of new
technologies are essential for long run growth.
xl
3. Private investment in Research and Development is the most important source of technological
progress.
Here in this study, we used the concept of Divisia Index (DI) for tax yields which is analogous to the
endogenous growth theory of total factor productivity. In the New growth theory, the revenue
generation (Tax) affects the long run economic growth because in this framework, a higher level of tax
revenue and capital formation allows for greater investment in human capital and Research and
development. The model predicts that the economy can grow forever as long as it does not run out of
new ideas or technological advancement.
Just like the exogenous growth theory, the endogenous growth theory professes convergence of nations
by diffusion of technology. That is, a situation where poor countries manage to catch up with the richer
countries by gradual imitation of technology by poorer countries. Romar states this production function
of a firm in the following form:
Y = A(R) F(Ri, Ki, Li)
Where:
A - Public stock of knowledge from research and development (R),
Ri - Stock of results from the stock of expenditure on research and development.
Ki - Capital stock of firm i
Li - Labour stock of firm i
The Ri actually represents the technology prevalent at the time in firm i. Any new research technology
spill over quickly across the entire nation. Technological progress (advancement) implies the
development of new ideas which resemble public goods because they are non-rival.
In accordance with the New Growth theory, we established a relationship between the growth in
revenue and the tax yield curve whereby increases in the tax bases cause growth in revenue resulting in
the movement along the tax yield curve. Tax elasticity is the term used to represent the movement
along the tax yield curve since it is an aggregate measure of the automatic growth in revenue. Technical
change is assumed to cause changes in factor productivity over and above those caused by changes in
factor inputs. As such, it would be expected that there would be a shift in the aggregate production
function. Also, a discretionary tax change causes the aggregate tax function to shift since it alters the
given tax system. The presence of a discretionary tax change (a technical change) results in a change in
xli the tax yield (output) due to the combined effect of growths in the tax bases (factor inputs) and a shift in
the curve caused by such a change. If there is no technical change or discretionary measure the given
structure remains unchanged.
As noted in Toni-Anne and Milwood (2012), the Divisia Index (DI) is similar to the index of technical
change which is the ratio of an index of total productivity to an index of factor productivity. It can be
interpreted that the percentage change in total productivity caused by technical change equals the
percentage change in output divided by the percentage change in factor inputs. It follows that the DI for
discretionary change is the percentage change in total tax yield divided by the percentage change in
total tax yield due to the built-in response to an increase in the bases. The appropriateness of the index
is due to the invariance property. If no technical change occurs, there will be no change in the index and
the growth in total factor productivity is due entirely to increases in inputs. Similarly, if there is no
discretionary tax change, the growth in revenue is entirely due to the growth in tax bases. Thus, a
change in the index should reflect the overall revenue effects of discretionary measures.
To fully incorporate the idea of DI from an aggregate tax function, invariance property which is
necessary and sufficient conditions for DI must be derived. The continuous differentiable aggregate tax
function for DI is given as:
f (x1(t),.....xk(t)).
Where the function f is linear homogeneous, and this is necessary condition for constant returns to
scale. The aggregate tax function condition is crucial to the existence of a relationship between tax
yields and the tax bases and essentially elasticity and buoyancy since if no underlying aggregate function
existed, there would be no way to determine these relationships. The “continuously differentiable”
character of an aggregate function ensures the regularity of such a function which would otherwise lead
to erratic behaviour of the tax yield. Thus, in order to justify the endogenous growth model, national
income and non – tax revenue will be used alongside with various tax sources in the specification of the
model.
3.2 Analytical Framework
In consonance with Endogenous Growth Model and the theoretical notions of revenue generation and
production functions reviewed above, this study will justify the short-run and long-run tax revenue
homogeneity condition by purporting that the country’s tax ratios have been increasing over time and
these have done so without erratic movements.
xlii In doing this, the study incorporates the vector error correction model (VECM) used by Bilquees (2004)
and the Dynamic OLS (DOLS) model proposed by Sobel and Holcombe (1996) and incorporated in Toni-
Anne and Milwood (2012) which both correct for the shortfalls of the OLS technique. The shortfall of the
OLS coefficient estimates relate to asymptotically biased coefficient estimates and the inconsistency of
the standard errors in the presence of nonstationary variables in a long-run co-integrating model.
However, the strategies of empirical modeling of the Long Run (LR) presents that carefulness should be
taken while modeling fiscal policy variables. For instance, economic theory suggest a specific nature
(both in signs and magnitudes) of the relationship among economic and fiscal policy indicators, and such
a theoretical intuition is applied and tested on the data normally by standard econometric procedures
such as OLS and IV regressions, that are relatively inadequate. Firstly, the dynamic nature of some
theories denies marginal effects to be instantaneous and full, and secondly theory does not always
suggest a stiff form of the relationships among variables and if it does, it is unlikely to hold constantly
(Jarir Ajluni, 2005). The theory of the Long Run (LR) allows for temporary deviations or disturbances
from a defined ‘equilibrium’ that variables would eventually converge into when all variation has taken
place.
Nelson and Plosser (1982) stated that most macroeconomic time series, such as the ones defined above,
are non – stationary series, integrated of order one, (I(1)), and then revolutionary finding of Granger and
Newbold (1974) of “spurious regression” problem makes standard econometric procedures clearly
inadequate. Modern developments in the literature; the evolution of co-integration in the work of Engle
and Granger (1987) and then Johansen (1988; 1991 and 1992) enable the applied researcher to properly
investigate theory using actual data.
On this basis, this study specified a Vector Error Correction Model (VECM) and Dynamic OLS (DOLS)
model which can be incorporate in the DI approach in order to estimate the buoyancy and elasticity of
tax revenues. This method seeks to separate the effect on total revenue of the (i) discretionary
measures and (ii) the built-in response of tax revenues to the growth in GDP. First, the effects of
discretionary tax measures are removed from total revenue growth using an index that isolates the
automatic growth in revenue. Next, the buoyancy is estimated with respect to GDP by a standard
regression technique. Finally, the estimated buoyancy is adjusted by a transformation of the index to
determine the elasticity of the tax yield. In the literature, the more popular methods of separating
discretionary measures from the built-in responsiveness of tax revenue to growth include the
proportional adjustment method, the constant rate structure method and the dummy variable method.
xliii
3.3 Model Specifications
3.3.1 Estimation of Tax Elasticity and Buoyancy
This study captures elasticity and buoyancy based on their definitions as stated in section 2.1. Tax
elasticity is defined as “the ratio of the percentage change in tax revenue to the percentage change in
income or GDP”, this therefore implies that tax elasticity is the coefficient of GDP/Income on tax
revenue, both for individual taxes and for the total tax revenue. “In empirical works an elasticity is the
estimated coefficient in a linear regression equation where both the dependent variable and the
independent variable are in natural logs. Elasticity is a popular tool among empiricists because it is
independent of units and thus simplifies data analysis” (Wikipedia 2013).
The coefficient of the GDP describes elasticity and it is defined thus
Table 3.1: Types of elasticity and designations
Alternative Coefficient (E)
Perfectly Elastic E= ∞
Relatively Elastic (more than proportionate change) 1 < E < ∞
Unit Elastic (equal proportional change) E=1
Relatively Inelastic (Less than proportionate change) 0 < E< 1
Perfectly Inelastic E = 0
While Tax Buoyancy according to section 2.1, is defined as “a measurement of the responsiveness of tax
revenue to changes in income without controlling for the discretionary changes in tax policy. The
discretionary changes are the changes which result in more tax revenue from the same tax base. The
sources of such changes are changes in tax legislation or changes in the tax rate” which suggests that the
dummy variable for tax reforms of 2004 captures discretional changes in the tax base, hence the
buoyancy.
3.3.2 Model Specification of the objectives of the study
In line with the major focus of this study which is to investigate the buoyancy and elasticity of tax system
in Nigeria with respect to its efficiency in revenue mobilization vis-à-vis the changes in National Income
(NI), thus, we assume aggregate revenue to be a homogenous function of NI. Putting this in a more
sophisticated procedure where economic theory suggests equilibrium relationship among aggregate
revenue and the national income; there are several forms of these equilibrium relations to hold using
these variables, hence a motivation for a system of equations. Also in recognition of the deterministic
xliv trend (denoted with h) of fiscal policy instrument, we formulate the model for the estimation of
objective 1, that is, the buoyancy and elasticity of tax estimate with respect to national income in the
following standard regression technique equation;
t 10 11 12 13 14 15 16 17 1 + + ....(1)k
t t t t t t t t tARV Z NI GX NTR XG INF D h
Where
ARV = Aggregate Tax Revenue
NI = National Income (GDP)
GX = Government Expenditure
NTR = Non-Tax Revenue
XG = External Grant
INF = Inflation rate
Dt = Dummy variable for tax reforms (where 0= 1980-2003 and 2004-2011)
Z = total base of tax k at time t;
= the elasticity of tax base with respect to national income
h = deterministic trend, since most fiscal policy instrument are trending;
= stochastic error terms.
Equation 1 estimates the buoyancy and elasticity of aggregate tax revenue with respect to national
income. However, estimating the individual taxes revenue in terms of buoyancy and elasticity (as given
in objective 2), we develop the following equations.
t 10 11 1 12 13 14 15 16 17 18 1 + .......(2)k
t t t t t j tPIT Z NI EP TW INF RER BD D
t 20 21 1 22 25 26 27 2 + ..............................(3)k
t t t j tCIT Z NI INT CDB D
t 30 31 1 32 33 35 36 37 3 + ............................(4)k
t t t t j tPPT Z NI BD RER RCO D
t 40 41 1 42 43 45 46 47 3 + ...............................(5)k
t t t t j tED Z NI BD RER TPR D
xlv
t 50 51 1 52 53 54 55 56 5 + ...........................(6)k
t t t t t j tCED Z NI BD RER IMP D
t 60 61 1 62 63 64 65 66 67 6 + .................(7)k
t t t t t j tVAT Z NI BD BOT INF CS D
All equations (1 to 7) were linearized to ascertain the elasticity given that elasticity in regression
coefficients are the estimated coefficients where both the dependent variable and the independent
variable are in natural logs.
Where
PIT = Personal Income Tax
CIT = Company Income Tax
PPT = Petroleum Profit tax
ET = Education Tax
CED = Custom & excise Duties, and
VAT = Value Added Tax
INF = Inflation rate
BD = Budget deficit
RER = Real exchange rate
BOT = Balance of trade
EP = Employed Population
TW = Total Wage
CDB = Cost of doing Business
RCO = Revenue from crude oil
TPR = Total production revenue
IMP = Import
CS = Company sales
xlvi Dt = Dummy variable for discretionary changes (tax reforms).
Z = the tax base of tax k at time t.
However for the purpose of normalization and identification of the co-integrating relationships among
the variables, the error estimates of equations 1 can be derived as;
1 t 10 11 12 13 14 15 16 17 - + + ....(8)k
t t t t t t t t jARV Z NI GX NTR XG INF D h
By incorporating the Dynamic OLS (DOLS) with lags of both the dependent and explanatory variables,
also a deterministic trend is usually included, and then it is directed to obtain disequilibrium deviations.
The buoyancy and elasticity equations can be adjusted by a transformation of the index to determine
the elasticity of the tax yield by ensuring that the optimal lag order is determined. In the case of
equations 1 above we would define a vector, where all variables are in lags.
t 10 t-1 11 1 12 1 13 1 14 1 15 1
16 1 17 1 1 1
+ + +
( ) ....................................................................................(9)
k
t t t t t
t i j t t t
ARV ARV Z NI GX NTR XG
INF D D h
However, the identification of lag length is not without a problem, but in order to solve this problem, we
examine the possible main lag selection statistics (criterions), which are the Final Prediction Error (FPE),
the Akaike Information Criterion (AIC), Schwarz Bayesian Criterion (SBC) and the Hannan-Quinn
Information Criterion (HQIC). Here, AIC and SBC are the most popularly applied, which usually the SBC
favours a lower lag order while the AIC is the typical standard for determination of the lag order and is
adopted for this model. To achieve this, we first of all, calculate the above criterion and determine the
log likelihood estimate of the underlying model as;
1
1
1 T[ln ln(2 ) ]
2
n
t
i
LL Y K K
--------------------------------------------------------------------------- (10)
Where T is the total number of observations, K is the number of variables (Equations), 1
1
n
t
i
Y
is the
maximum likelihood estimate of !( )t tE the vector of residuals in equations (9) that has the dimension
K x 1. The LL can be obtained after fitting the model at any lag order level. The procedure is to run a
number of model estimation with lag order from one to a defined maximum subject to the data
frequency and the sample size. In quarterly data for example, a lag order of more than 12 (three years)
xlvii would be too high for a sample of 100 observations, a lag order of 8 would be more suitable as a
maximum lag order, the statistics provided below are different lag order equations;
1 2 p pLR LL LL ----------------------------------------------------------------------------------------- (10a)
2 2 PpLL
AICT T
------------------------------------------------------------------------------------ (10b)
k
R
T mFPE
T m
--------------------------------------------------------------------------------------- (10c)
ln( )2 P
LL TSBC p
T T
-------------------------------------------------------------------------------- (10d)
2ln(ln )2 p
LL THQIC P
T T
------------------------------------------------------------------------- (10e)
With the specified lags order equations, our previous equations (9) would now be augmented by p
numbers of lags and inclusion of a deterministic trend as given in equation (11) below:
1 11t t-i1
+p
i t
i
ARV NI t
-------------------------------------------------- (11)
Following the Engle – Granger representation theorem, if the I ~ I(1) non-stationary variables are co-
integrated then an Error Correction Model represents these variables. In first differences the VECM
would take the form of:
01 11t t-i t-i1
+p
i t
i
ARV ARV ARV t
-------------------------- (12)
Where t-i
ARV is obtained since 1t t t-1
ARV AVR AVR
And is a square matrix with dimensions equal to the number of variables in the system and is known
as Matrix of Long – Run Multipliers, the “dynamic” or “impact parameters” matrix. It is clear that under
the presence of co-integration, the “stationary VEC” in differences is miss-specified, the difference is
obvious in the lack of the matrix of Long – Run multipliers.
xlviii To solve this problem we adopt Co-integration Ranks by applying Johansen reduce rank tests to estimate
the number of LR relations. Johansen (1995) showed that the number of co-integration relations is
determined by the rank of the dynamic matrix with the dimension of k x k and has a reduced rank in
the presence of co-integration which has a rank of zero if there is no co-integration, hence the rank r of
is 0 r m, m < k.
Rank of a matrix is the number of linearly independent columns or rows. The linearly independent
relations are the LR relationships among our variables expressed in the rank of the dynamic LR matrix,
hence the importance and convenience of the Johansen approach. This is how Johansen procedure is
clever. To determine this ranks, Johansen introduced the well-known ‘Trace’ and ‘Maximal Eigenvalue’
statistics of characteristic roots test for the number of co-integration relations and reported their critical
values obtained using a Monte Carlo simulation. This will automatically be run in the STATA version 11.0
adopted for this analysis.
3.3 Model Specification for Objective 3
In order to estimate the 3rd objective of this study, we incorporate the presence of a technical change,
which results in a change in the tax yield (output) due to the combined effect of growth in the tax bases
(factor inputs) and a shift in the curve caused by such a change. This can be derived with dummy model
technique in the following way:
t 10 11 12 13 14 15 16
17
+ + ( )
.....(13)
k
t t t t t i t
t t
ARV Z NI GX NTR XG D
INF
Where the dummy variable Di is defined such that D = 1, for the period of oil boom, 0 for period before
oil boom. In this model the estimation of aggregate tax revenue after the oil boom is given as;
t 10 11 12 13 14 15 16
17
, 1 + + ( )k
t t t t t i t
t t
E ARV D Z NI GX NTR XG D
INF
-------------------------
(14)
All the variables remained as defined above. Note that, if there is no technical change or discretionary
measure the given structure of tax revenue remains unchanged, and the estimate for structural model
given before the oil boom is given as;
xlix
t 10 11 12 13 14 15 16, 0 + +k
t t t t t tE ARV D Z NI GX NTR XG INF -------------------------
(15)
With the help of equations 14 & 15 above, the study will ascertain if there are changes in the structure
of tax revenue, and if there is, the models will direct us on the sources of the structural changes.
To capture the boom period dummy, the study used the Zivot Andrews unit root test that has the
potential of detecting the most severe structural break during the period of study.
3.4 Justification of the Models Specified
The study adopted the standard OLS estimation procedure which was modified into Dynamic OLS (DOLS)
and was incorporated in vector error correction model (VECM) model. According to the proposition of
Sobel and Holcombe (1996), these ideas correct for the shortfalls of the OLS model. The shortfall of the
OLS coefficient estimates relate to asymptotically biased coefficient estimates and the inconsistency of
the standard errors in the presence of non-stationary variables. Again, the Divisia Index (DI) approach
was introduced to estimate the buoyancy and elasticity of tax revenues.
As rightly noted in Toni-Anne and Milwood (2012), this method seeks to separate the effect on total
revenue of the (i) discretionary measures and (ii) the built-in response of tax revenues to the growth in
GDP. With DOLS and VEC, the effects of discretionary tax measures are removed from total revenue
growth using an index that isolates the automatic growth in revenue. Next, the buoyancy is estimated
with respect to GDP by a standard regression technique. Lastly, the estimated buoyancy is adjusted by a
transformation of the index to determine the elasticity of the tax yield. And this methods will separate
discretionary measures from the built-in responsiveness of tax revenue to growth include the
proportional adjustment method, the constant rate structure method and the dummy variable method.
Besides, the variance decomposition as an aspect of VAR/VECM is one of the most popular techniques
for capturing the impulse responses and transmission of shocks among the variables.
3.5 Estimation Technique and Procedure
In carrying out the procedures explained above, the following technique will be used in line with the
models.
3.5.1 Procedure
The estimation commences with an extensive unit root test to confirm the stationary states of the
variables that entered the model using both the Augmented Dickey-Fuller (ADF) and Phillips-Perron
l (PP) tests. Both tests of unit root are used in order to guarantee that our inferences regarding the
important issue of stationarity are not likely driven by the choice of the testing procedure used.
The testing procedure for the ADF test is as follows:
tptpttt XXXtX ...1110-------------------------------------------------------- (14)
Where, 0 is a constant, t is the coefficient on a time trend and p is the lag order of the autoregressive
process and is the difference operator. The unit root test is then carried out under the null
hypothesis = 0 against the alternative hypothesis of < 0. We compare the computed value of the
test statistic with the relevant critical value for the test. For instance, if the computed test statistic is
greater (in absolute value) than the critical value at 5% or 1% level of significance, then the null
hypothesis of = 0 is rejected and thus no unit root is present, otherwise, it is accepted.
The test for stationarity is first conducted at level, however, if the variables are not stationary at level;
we then difference them and test for the stationarity of the differenced variables. Supposing the
variables are stationary at first difference, we conclude the variables are integrated of order one (i.e.,
I(1)).
3.5.2 Co-integration Tests
Having confirmed the stationary properties of the variables, we proceed to determine the existence of
a long-run relationship among these variables. A cointegrating relationship exists between series, if
there is a stationary linear combination between them. To ensure a robustness check of the
cointegration estimation, we shall use both the Engle-Granger approach and the Johansen maximum
likelihood procedure. Using the Engle-Granger procedure of cointegration test, we first regress the
dependent variable on its various determinants to obtain the estimated coefficients, then estimate the
residuals from this regression and save it and finally, test if the saved estimated residual series is
stationary or not. However, since the Engle-Granger approach suffers problem of normalization, the
multivariate Johansen procedure, which uses maximum-likelihood method of estimation and does not
suffer normalization problem (Gujarati, 2005) shall be fully utilized.
3.6 Data Sources
The study will solely rely on secondary data, which shall be sourced from CBN statistical Bulletin,
National Bureau of statistics (NBS) and Federal Inland Revenue Service (FIRS) of Nigeria etc.
li
CHAPTER FOUR
PRESENTATION OF RESULTS
4.1 Presentation of the Unit root test – ADF
The study used the augmented dickey fuller technique to test for unit root of both the dependent and
explanatory variables used in the model. The results are therefore presented on the Table 4.1.
lii
Table 4.1: Unit Root results for all Variables
Variables Trend/Or not No of
lags
Order of
stationarity
Stationary critical
value
Gross Domestic Product (GDP) Trend 0 I(2) 1%
Aggregate tax revenue No trend 4 I(2) 5%
Government expenditure Trend 8 I(2) 5%
Oil revenue Trend 8 I(0) 5%
External grant (ODA) No Trend 0 I(0) 5%
company income tax Trend 2 I(1) 1%
VAT Trend 0 I(1) 1%
Exchange rate No trend 4 I(2) 5%
Petroleum income tax Trend 0 I(1) 1%
Prime lending rate No trend 0 I(1) 1%
Inflation rate No trend 0 I(1) 1%
Custom duties tax Trend 1 I(1) 1%
Import Trend 0 I(2) 1%
Terms of trade Trend 0 I(1) 1%
Lag of total tax Trend 0 I(1) 1%
Lag of company income tax Trend 2 I(1) 1%
Lag of petroleum tax Trend 0 I(1) 1%
Lag of custom ex duties No Trend 0 I(1) 1%
Lag of VAT trend 0 I(1) 1%
Residual for total tax Trend 0 Not stationary
Residual for CIT Trend 0 Not stationary
Residual for VAT Trend 0 Not stationary
Residual for PPT Trend 0 Not stationary
Residual for CED Trend 0 Not stationary
The above table shows us that most of the variables had unit roots and hence were not stationary,
thereby requesting for differencing to make them stationary. However Oil revenue and external grant
were stationary at level form. While the others were I(1) and I(2) process. The fact that there exists unit
liii root for most of the variables is a necessary condition for co-integration. Co-integration is aimed at
testing the long –run relationship of the variables, wherein if it is validated then it means there is a
problem that needs to be corrected with the error correction model. Surprisingly in this study, the co-
integration test which is the unit root test of the residuals of each regression suggests that there exist no
co-integration in any of the regressions. This therefore implies that the sufficient condition for an error
correction model is not satisfied; therefore we conclude that there exists no long-run relationship
amongst the variables.
4.2 Regression results of Aggregate tax
To achieve objective one, which is to investigate the elasticity and buoyancy of aggregate tax in Nigeria
with respect to GDP, we ran the regression as specified in equation 1. However the study first of all tests
for linearity of the dependent and explanatory variables as prescribed by the assumptions of the
classical linear theory. To be able do this, the researcher drew scatter plot of the dependent variable –
aggregate tax against its residual and the outcome is shown in figure 4.1.
Figure 4.1: Scatter plot of Aggregate tax against its residual
0
2.0
e+
06
4.0
e+
06
6.0
e+
06
8.0
e+
06
tota
lta
x
0 2000000 4000000 6000000 8000000Fitted values
liv The scatter plot above shows a 45 degree line pattern of the dependent variable and the residual hence
suggesting that there exist a linear relationship between aggregate tax and its explanatory variables
therefore validating the assumption of linearity for the regression whose results are stipulated below.
Table 5.2: Regression results for aggregate tax and its determinants
Variables Coefficients of determinants
lag1totaltax 0.295**
(2.89)
realgdp 1.124
(0.38)
govexp 0.0745
(0.23)
oilrevenue 0.447 *
(2.23)
extgrantna~a -53520.98
(-0.45)
inflation -1858.5
(-0.30)
taxrefdummy 3219970.5***
(6.89)
_cons -257965.9
(-0.40)
N 31
R square 0.9868
Durbin Watson 2.683756
F- porbability 0.0000
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001
The regression result above suggests that the overall significance of the model is good as the F
probability is very low (0.0000), and the R2 is equally very high (0.9868) suggesting that 98.68% of the
dependent variable are explained by the explanatory variables. Also there exists no positive or negative
strong auto-correlation in the regression since the Durbin Watson statistic (2.63) falls in the zone of
indecision.
The main objective from this result is to ascertain the elasticity of aggregate tax with respect to the
gross domestic product which is given by the coefficient of GDP (having linearized aggregate demand
and GDP). The degree of responsiveness of tax with respect to GDP is therefore given as 1.124.
According to table 3, 1.124>1 implies that the responsiveness of tax with respect to a change in tax is
relatively elastic. Though the elasticity is close to 1, the study opines that policies aimed at increasing
lv the tax base can seriously consider increasing aggregate tax to some extend since it is relatively elastic.
This is important to Nigeria, given that over the years efforts have been made to improve on internally
generated revenue. Aggregate tax can therefore accommodate an increase at least to some extent.
To ascertain the buoyancy of tax this study considered the 2004 tax reforms and stipulated a dummy to
capture the structural break. The dummy variable is significant up to 99.9% confident interval thereby
suggesting that aggregate tax was significantly buoyant to the tax reforms of 2004. This also implies that
many more vision designed policies like such could be designed to improve the tax base of the economy.
Apart from the main objective this result further show that the previous’ year’s tax is a positively
significant determinant of current year’s aggregate tax. This is expected given that, Nigeria like many
other African/developing countries produces their budgets based on last year’s income, and so targets
are often set just a little above the previous year’s figure and not based on the country’s needs and
wants.
Also, government expenditure and real GDP are not significant while oil revenue is a significant
determinant of aggregate tax in Nigeria. However they are all positively and directly related to aggregate
tax. This implies that government expenditure is not focused on taxable institutions and organizations
which should ordinarily turnaround to generate revenue through tax. In addition gross domestic product
also is not significantly based on tax in Nigeria due to the country’s overdependence on oil. While Oil
revenue is shown to be a significant determinant of aggregate tax since most of the oil revenue mainly
due to the large use of oil revenue in driving the economy. As Nigeria is still a major importer of refined
crude and so oil revenue can better stimulate aggregate tax if well harnessed.
Furthermore, inflation and external grants are equally not significant but have a negative and inverse
relationship with the aggregate tax of Nigeria. This implies that as tax inflation increases aggregate tax
drops though this is not significant. This could be attributed to the fact that as inflation rises, some firms
may close down hence reducing the amounts that would have been collected from them as tax. Also
external grant or official development assistance tends to have an inverse relationship with aggregate
tax and could be explained by the fact that most of these grants are usually not taxed, and some of
these grants (such as health aids) reduces the private firms that would have sold the health fascilities
and hence the tax that would have been gotten. A good example is the huge sums of money that have
been pumped into Africa and Nigeria in particular to eradicate malaria which is most times converted as
mosquito nets and hence relegates traders that are involved in mosquito nets.
lvi 4.3 Regression results of the four major types of tax
The second objective of this study is aimed at investigating the elasticity and buoyancy of the four major
taxes as designed by the scope. This section will use the same procedure as above, but only this time it
will be the different main taxes. However we test for the linearity of all the four taxes as dependent
variables and their residuals.
Figure 4.2: Scatter Plots of CED, PPT, CIT and VAT and their residuals
Figure 4.2 represents the scatter plots of the four major taxes under consideration which include; the
custom and excise duties, company income tax, petroleum profit tax and value added tax against their
various residuals. The four graphs depicts a 45 degree line pattern sloping from left to right which is
what is normally expected of a linear trend. This therefore suggests that all the taxes tested above are
linearly related to their explanatory variables and therefore validates the linearity assumption of the
classical linear theory. Based on this result therefore the researcher can further regress the equations
while testing for the other assumptions of the classical linear theory.
The table below therefore portrays the result of the regressions for all four taxes mentioned above. The
table shows their coefficients, t-values, R squares and the F probabilities.
0
1000
0020
0000
3000
00
cust
oman
dexc
ised
utie
s
0 100000 200000 300000Fitted values
0
1000
0020
0000
3000
0040
0000
com
pany
inco
met
ax
0 100000 200000 300000 400000Fitted values
0
2000
00400
00060
00008
0000
0 1.0e
+06
petr
oleu
mpr
ofitt
ax
0 500000 1000000Fitted values
0
1000
0020
0000
3000
0040
0000
vat
-100000 0 100000 200000 300000Fitted values
lvii
Table 4.3: Regression results for VAT, CIT, PPT and CED
(1) (2) (3) (4)
Variables VAT CIT PPT CED
Terms of trade 0.0218
(1.42)
Interest rate -4610.6 -1534.2
(-1.28) (-1.40)
Tax reform dummy -104694.1 19756.0 -335894.0* -69161.4*
(-2.01) (0.69) (-2.48) (-2.22)
Real GDP 0.796*** 0.291* 2.159*** 0.447*
(6.48) (2.65) (3.86) (2.23)
Lag1CIT 0.589***
(4.77)
lag1PPT -0.145
(-0.66)
Real exchange rate 1181.9 146.4
(1.96) (1.09)
Oil revenue 0.0524
(2.05)
lag1CED 0.425
(1.60)
import -0.00179
(-0.39)
_cons -130461.4 -39863.1 -518434.8** -97773.3
(-1.90) (-1.59) (-3.53) (-2.03)
N 18 31 31 31
Durbin Watson 1.993072 2.127995 1.479792 2.308298
R square 0.9608 0.9664 0.9149 0.9442
F probability 0.0000 0.0000 0.0000 0.00000
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001
The result above suggests very significant models for all the taxes, given that the F probability remains
0.0000 for all four of them therefore implying that the overall model is highly significant. Also their R2 is
very high as all of them are above 91% which implies that they are all highly explained by the
determinants. The Durbin Watson statistic shows that there exist no autocorrelation for VAT and CIT
while PPT and CED falls in the zone of indecision. This therefore implies that none of them are either
positively or negatively auto correlated.
The major objective however is to ascertain the elasticity and the buoyancy of these taxes. Surprisingly
unlike the aggregate tax, three of the taxes have their GDP coefficients lying between zero and one. This
therefore suggests that VAT, CID and CED are all relatively inelastic and implies that they have a little
proportionate change with respect to a unit change in GDP. Hence policies designed to increase or
decrease these tax bases should be neglected, this is because the result proposes that a change in GDP
will not over-change the tax volume of VAT, CID and CED. On the other hand petroleum profit tax (PPT)
lviii is relatively elastic, given that the coefficient of GDP is as high as 2.159 which is higher than the elasticity
of the aggregate tax. This implies that PPT is highly flexible and hence responds with a more than
proportionate change to a unit change in GDP. Nevertheless what is most interesting and of course
expected is that, all taxes tend to increase with increase in GDP.
The test of buoyancy for all four taxes shows that VAT and CIT are not buoyant with respect to the 2004
tax reforms as their p-values are higher than 0.05 even though the dummy for VAT has a t-value of -
2.01. While PPT and CED appear to be buoyant with their dummy values for structural break is
significant at 95% confidence interval. It is worth noting however that while CIT improved after 2004
(had a positive impact), the other taxes- VAT, PPT and CED did not.
Inferring from the determinants of VAT the result suggests that, terms of trade and interest rate are not
significant determinants of VAT, while real GDP is significant at 99.9% confidence interval. As expected,
interest rate has an inverse relationship with VAT. This could be explained by the fact that as interest
rate (prime lending rate) falls more firms and industries will be motivated to produce (add value in all
forms) and hence improving the VAT. On the other hand, terms of trade is positively related to VAT,
implying that as terms of trade increases VAT equally increases. While the regression results of company
and income tax (CIT) shows that, interest rate is still negatively related and not significant in determining
CIT. However real GDP and the lag of CIT are significant in determining CIT. This therefore implies that
the CIT of the previous year is significant in determining current CIT which is equally expected as most
targets are usually made based on the previous year’s amount. Real GDP is equally significant and
positive insinuating that an increase in GDP leads to an increase in CIT which is a good sign.
The regression results of petroleum profit tax (PPT) suggest that its lag, real exchange rate and oil
revenue are all not significant determinants of PPT. However while real exchange rate and oil revenue
are positively related, the lag of PPT is negatively related. The negative relationship of the lag may be
explained by the volatility that exists in the oil market which might be later transferred to the tax
collected from its proceeds. While real GDP and the tax reform dummy are significant determinants of
petroleum profit tax.
The regression results of the custom and excise duties (CED) show that real exchange rate, imports and
its lag are all not significant determinants, while real GDP and the dummy are both significant
determinants at 95% confidence interval. However, real GDP, real exchange rate and the lag of CED are
positively or directly related to CED, implying that as they increase, custom and excise duties increase as
well. What is most surprising is that there exists an inverse relationship between imports as against
lix custom and excise duties. This implies that as imports increase custom and excise duties drop which
shouldn’t be the case. This could only be attributed to the huge corruption that exists in the importation
of oil that constitutes a large amount of the country’s import.
4.4 Regression results of aggregate tax revenue during and after the oil-boom in Nigeria
To ascertain the significance of the structural break point this research first of all runs a Zivot Andrews
unit root test on oil revenue to identify the structural break point and hence the oil boom dummy
period.
Figure 4.3: Zivot Andrews Unit root test for oil revenue
Figure 4.3 represents the Zivot Andrews unit root graph results and indicates that the structural break
point was 2005, it is also visible that the boom period experienced before 2005 and a depression on oil
revenue was witnessed in 2005. The dummy period shall therefore be designed as 1 for 1980-2004 and
0 assigned to 2005-2011. The same regression for aggregate tax to ascertain if there was a significant
structural break is therefore presented below.
Table 4.4: Regression results for Aggregate tax to determine the oil boom structural change
(1)
Variables Totaltax
lag1totaltax -0.480
-3-2
-10
Bre
akp
oin
t t-
sta
tistics
1980 1990 2000 2010year
Min breakpoint at 2005
Zivot-Andrews test for oilrevenue, 1987-2006
lx (-1.93)
Real GDP 16.97**
(3.17)
Government expenditure 0.0481
(0.13)
External grant naira -407403.5
(-2.04)
Inflation -15186.4
(-1.60)
Oil boom dummy -6071376.7***
(-4.64)
_cons 11831272.2*
(2.44)
N 31
R square 0.9687
F probability 0.0000
Durbin Watson 2.723316
t statistics in parentheses
* p<0.05, ** p<0.01, *** p<0.001
The overall significance of the model is good as the F probability is as low as 0.000, with a high R2 of
96.87%. Also there exist no positive or negative auto correlations in the model. The oil boom dummy
shows that it is significant at 99.9% confidence interval and has an inverse relationship with aggregate
tax. This implies that after the boom period, aggregate tax fell significantly as compared to the boom
period. This therefore conjectures that oil revenue was a significant contributor to aggregate tax
especially in boom periods.
4.5 Evaluation of Working Hypotheses
4.5.1 Test of Hypothesis 1
H0: The total tax revenue in Nigeria is not significantly buoyant.
H1: The total tax revenue in Nigeria is significantly buoyant.
DECISION:
The p-value of the slope of the tax reform dummy variable is 0.000 and a t-value of 6.89 hence
significant at 99.9% confidence interval. The researcher therefore rejects the null hypothesis that total
tax revenue in Nigeria is not significantly buoyant. Hence, we conclude that the 2004 tax reform was
significantly in improving the tax base of aggregate tax implying that aggregate tax could be significantly
buoyant in Nigeria.
4.5.2 Test of Hypothesis 2
lxi Ho: Tax revenue of the main taxes in Nigeria with respect to GDP growth are not significantly buoyant.
H1: Tax revenue of the main taxes in Nigeria with respect to GDP growth are significantly buoyant.
DECISION:
The p-value for VAT and CIT are 0.066 and 0.494 respectively (both less than 0.05) and therefore not
statistically significant, hence we do not reject the null hypothesis for both taxes. While the p-value for
PPT and CED are 0.020 and 0.036 respectively, hence significant and the researcher does not reject the
null hypothesis. This therefore implies that while VAT and CIT are not buoyant according to the 2004 tax
reforms, PPT and CED are significantly buoyant.
4.5.3 Test of Hypothesis 3
Ho: There is no significant structural change in tax revenue between the pre and post oil-boom
era in Nigeria.
H1: There exists a significant structural change in tax revenue between the pre and post oil-boom
era in Nigeria.
DECISION:
The oil boom dummy period insinuated a significant structural break period at 99.9% confidence interval
as the p-value is 0.000, implying that we reject the null hypothesis. We therefore conclude that there
was a significant drop in aggregate tax revenue after the boom period.
CHAPTER FIVE
SUMMARY, CONCLUSION AND RECOMMENDATIONS
5.1 Summary
lxii Many countries in the world have greatly sponsored their government expenditures with the aid of tax
revenue, and owe their developments to this internally generated revenue. Nigeria started off on a good
foot with the huge agro-production after her independence. However the discovery of crude attracted
most investors to the oil sector thereby neglecting the other potential sectors that would have stirred
up the industrial production and then increase tax revenue. The motivation of this study was therefore
to find out how flexible tax is in terms of the elasticity and buoyancy, in an attempt to improve the tax
base and generate more income. The study therefore investigates the elasticity and buoyancy of
aggregate and four other major taxes with respect to economic growth in Nigeria. Also, the study
analyzed the flow of oil revenue during and after the boom period.
To be able to capture these objectives the study uses the 2004 tax reforms to establish a dummy period
and ascertain the buoyancy, while using the logged coefficient of GDP to aggregate tax and major taxes
in order to determine the Growth elasticity of tax. For the third objective the study uses the Zivot
Andrews structural break test to ascertain the most severe structural break of the oil revenue trend, and
then uses simple OLS regressions with dummy effects to analyze the trend of aggregate tax with respect
to the periods during and after the boom period.
The results of this research are far enriching as it suggest that aggregate revenue is relatively elastic.
This implies that an increase in the GDP will increase tax to a more than proportionate level. Also the tax
reform of 2004 was highly significant on aggregate tax and therefore infers that aggregate tax is
relatively buoyant. The same test was done on four major taxes namely; value added tax, custom and
excise duties, petroleum profit tax and company income tax. Out of the four taxes, only PPT was found
to be relatively elastic while the other three were found to be relatively inelastic. However the results
further suggest that, while VAT and CIT are not significantly buoyant according to the 2004 tax reforms,
PPT and CED are significantly buoyant. The Zivot Andrews structural break unit root test was then used
to ascertain the most severe structural break on the oil revenue trend and found it to be 2005. Then the
study used the 2005 structural break to establish the dummy period during and after the oil boom. With
the aid of the dummy the study regressed aggregate tax revenue against its determinants and the
results showed that there was a significant drop in aggregate tax revenue after the boom period. The
boom period therefore significantly contributed to aggregate tax revenue in Nigeria.
5.2 Conclusion
Nigeria’s potential of crude and other natural resources is uncontestable, however climate change has
raised the debate on how sustainable oil revenue can be and hence questioned its overdependence. It is
lxiii on this premise that this study investigated the elasticity and buoyancy of tax in an attempt to ascertain
its flexibility and hence the possibility of increasing the tax base. The study therefore suggested that
aggregate revenue is relatively elastic and significantly buoyant according to the 2004 tax reforms. The
results of the four major taxes tested showed that only PPT was found to be relatively elastic while VAT,
CED and CID were relatively inelastic. However the results further suggest that, while VAT and CIT are
not significantly buoyant according to the 2004 tax reforms, PPT and CED are significantly buoyant.
Finally, the study used the 2005 structural break to establish that aggregate tax revenue dropped
significantly after the boom period. The study therefore concludes that tax in Nigeria is relatively flexible
with respect to growth and therefore more could be done to increase it.
5.3 Policy Recommendations
i) Policy implications from the study abound based on the findings. The results of the first objective of
the study opine that aggregate tax is relatively elastic. This is important to policy makers as it shows that
aggregate tax has the potential to be increased to a more than proportionate level with any change in
economic growth. This therefore means that every policy that is aimed at improving gross domestic
product also increases tax to a larger extent. However it is worth noting that real GDP is not a significant
determinant of aggregate tax. Hence the Nigerian aggregate tax system is good to a considerable extent,
and so policies could be geared toward increasing this elasticity even further by strengthening the
private sector to generate more tax revenue. The 2004 tax reform was very significant and portrays a
buoyant tax system. Policies like this could be structured to increase the tax base and the reforms could
be strengthened to make it even more buoyant.
ii) The previous year’s tax is significant at 95% confidence interval in determining the aggregate tax
revenue of the current period. This reflects the governments of most African countries as their budgets
are mostly targeted just above the previous budget, without ascertaining the needs and wants of the
economy and looking for means to increase the tax base especially in an elastic aggregate tax system
like that of Nigeria as determined in this study. This study opines that aggregate tax is elastic to
economic growth. This suggest that policy makers could set higher targets for aggregate tax and depend
on its elastic nature to increase tax and not necessarily depending on the incremental budget plan.
iii)The other determinants considered were government expenditure, oil revenue, external grant and
inflation. Government expenditure was not significant in determining it, though positively related. This
means that an increase in government expenditure will increase aggregate tax and so any policy aimed
at improving government expenditure also increases aggregate tax. Oil revenue has a significant positive
lxiv impact on aggregate tax revenue and this implies that in as much as the country depends on oil, oil
revenue still has a significant impact on aggregate revenue. External grant is negatively related to
aggregate tax and this suggests that those sectors, on which external grants are made, are tax
generating sectors. This might be predominant when the external grant is rather direct, that is in the
form of finished products. Nevertheless the researcher notes that it is not significant. Inflation is also
negatively and insignificantly related to aggregate tax revenue. This is expected as inflation might reduce
output in the industry thereby reducing taxable companies and hence aggregate tax.
iv) The Value added tax (VAT) has an elasticity of 0.796 with economic growth, implying that there is a
less than proportionate change with VAT with every change in economic growth. This therefore implies
that favorable growth will lead to an increment in VAT though not in the same proportion. The fact that
it’s a very significant determinant of VAT further lays emphasis on its relevance on VAT. This should
however be the case because it is expected that economic growth breathes value addition and hence
value added tax. The tax reform dummy suggest that after the 2004 tax reforms, VAT declined implying
that the buoyancy was low, though it is not significant at 95% confidence interval. This however portrays
that the 2004 tax reform did not lead to an improvement in the quantity of VAT. Interest rate being one
of the determinants of VAT, suggests a negative and non-significant effect on VAT. This implies that VAT
increases with a fall in interest rate (lending rate), which therefore implies that the government could
increase VAT by reducing interest rate though interest rate is not significant as the results suggest. On
the other hand, the term of trade is a positive non-significant determinant of VAT. This is however
expected, and the government should note this positive relationship that term of trade has with VAT
and therefore use it to improve VAT.
The elasticity of company income tax (CIT) is very low, and insinuates that an increase in economic
growth (proxy by GDP) leads to a less than proportionate change in the CIT. real GDP however
significantly determines CIT at 10% significant level. Therefore must be considered considerably when
drafting policies to induce CIT, nevertheless the elasticity is just 0.291 that implies a very slow respond
to change in GDP. The tax reform dummy is equally not significant showing that there was no significant
change after the 2004 tax reform on CIT, but there was a positive improvement as testified by the
positive sign of the coefficient of the dummy. Though, it was not significant and hence not buoyant. Tax
reforms should therefore be restructured to have significant impacts on the tax base. The other
determinants indicate that interest rate is not a significant determinant on CIT, and that the lag of CIT is
very significant in determining CIT of the current year. This therefore shows that CIT depends greatly on
the past CIT’s.
lxv Petroleum profit tax has the highest elasticity with respect to VAT, CIT, and CED. The elasticity of 2.159
suggests that an increase in economic growth increases PPT to a more than proportionate level. This
could be used to the advantage of the state and federal governments to improve petroleum profit tax.
The buoyancy was equally high as the 2004 tax reform dummy was significant though it had a negative
impact on PPT. The lag of PPT did not significantly determine it at 5% significant level, just like real
exchange rate and oil revenue. But unlike the lag of PPT that had a negative relationship with PPT, real
exchange rate and oil revenue had a positive relationship.
v) Custom and excise duties (CED) record an elasticity that is less than 1, hence suggesting that for every
change in real GDP, CED changes by a less than proportionate amount. This means that CED responds
slowly to changes in real GDP and must be noted in designing policies aimed at improving CED or as a
consequence of every unit growth of GDP. The fact that GDP is significant also shows its contribution to
CED. Also, just like the case with PPT, the 2004 tax reforms had a negative significant effect on CED.
Hence better policies should be formulated bearing in mind that CED is inelastic with respect to GDP.
The other determinants; real exchange rates, lag of CED, and imports are all not significant determinants
of CED.
5.4 Suggestions for further Research
The study has gone a long way to determine the elasticity and buoyancy of aggregate tax as well
petroleum profit tax, custom and excise duties, company income tax, and value added tax. Given that a
study of this nature has not been carried out in Nigeria, it would therefore raise a debate in this sector
and proffer solutions to improve the tax volume. The elasticity of other taxes could be equally
investigated and other methods could be used to verify the elasticity and buoyancy of tax, to affirm the
results gotten from this study. In addition the elasticity of tax with respect to other indices could also be
investigated in order to know whether tax increases or decreases as such indices change, and to what
extent.
Also having established the elasticity and buoyancy of the afore mentioned taxes research on tax an
related fields can be made on the premise that aggregate tax in Nigeria is relatively flexible and buoyant
while of all the major taxes mentioned only petroleum profit tax is elastic ant the other three are
inelastic. Other reforms and policies on tax could equally be tested for buoyancy such as this study deed
for the 2004 reforms.
REFERENCES
lxvi Adegbie, F. F and Fakile, A.S (2011) “Company Income Tax and Nigeria Economic Development”
European Journal of Social Science. Vol. 22 No.2.
Adeleke, S. (2011) “Taxation, Revenue Allocation and Federalism in Nigeria: Issues, Challenges and Policy
options” Economic Annals, Vol. LVI, No.189/April-June.
Agbeyegbe, T.D., Stotsky J. and Asegedech W. (2005). Trade liberalization, exchange rate changes, and
tax revenue in Sub-Sahara Africa, Department of Economics, Hunter College and the Graduate
Center, City University of NY, NY, United States.
Akinlo A.E. (2006) “Macroeconomic factors and total productivity in sub-Sahara African countries,
International Research” Journal of Finance and Economics ISSN 1450-2887 Issue 1.
Anyanwu, J.C. (1995) “Revenue Allocation and Stable Fiscal Federal in Nigeria” Journal of Economic
Management, 2(2): 1-28.
Ariyo, A. (1993) “An Assessment of the Sustainability of Nigeria’s fiscal Deficit: 1970-1990” Journal of
African Economics, Vol. 2, no. 2:263-282.
Ariyo, A. (1997) Data generating procedures: A Comparative Analysis. Paper Presented at a workshop on
Quality of African Data Organized by the African Economic Research Consortium. Nairobi, Kenya,
11-12 February.
Ariyo, A. (1997) Productivity of the Nigeria Tax System. Africa Economic Research Consortium Research
Paper, No. 67.
Ariyo, A. and M.I. Raheem (1990) Deficit Financing and Economic Development: Empirical Perspectives
from Nigeria. Project Report presented to the African Economic Research Consortium. Abidjan,
December.
Ariyo, A. and M.I. Raheem (1991). Effects of Fiscal Deficit on some Macroeconomic Aggregate in Nigeria.
Final Report Presented at the workshop of the African Economic Research Consortium Nairobi,
Kenya, 7-11 December.
Ayoki, A. (2007). Tax performance in poor countries. Country Reprot Uganda. /PRA working paper No.
21.
Azubike, J. U.B (2009). Challenges of tax authorities, tax payers in the management of tax reform
processes. Nigeria Account 42 (2): 36-42.
lxvii Baily, M.J. (1980) “Inflation and Tax-Induced Resource Misallocation” National Tax Journal, Vol. 33: 375-
278.
Bolnick, B.R. (2002) Demographic effects on tax ratios in developing countries, Duke University, Durham,
NC 27706, USA.
Buiter, W. (1983) The Theory of Optimum Deficits and Debt. NBER working paper NO. 1232.
Chibber, A. and J. Khalizadeh-Shirazi (1988) Public Finance in Adjustment programs. PPR working paper
No. 128.
Choudhry, N.N. (1979) Measuring the Elasticity of Tax Revenue: A Divisia Index Approach. IMF Staff
Paper, Vol. 26.
Christodoulakis, N.M. (1994) Tax collection lags and the revenue maximizing inflation, Athens University
of Economics and Business, 76, Patission Street, Athens 10434, Greece.
Demirbas, S., (1999) Cointegration Analysis – Causality Testing and Wagner’s Law: The case of Turkey,
1950-1990, paper presented at the Annual Meeting of the European public choice society,
Lisbon, April 7-10.
Effiok, S.O (2006) “Fiscal Federalism, Tax Reforms and Productivity: A case for direct and indirect taxes in
Nigeria” Journal of Research in National Development. Vol. 4(1) pp.60-73.
Feldstein, M. (1980) “Fiscal Policies, Inflation and Capital Formation” American Economic Review, Vol.
70, September: 636-50.
FGN. The Constitutions of the Federal Republic of Nigeria. 1963 Edition, 1979 Edition, 1999 Edition.
Grasser H.N. (2004). Elasticity and Buoyancy of Tax System in Liberia. M.A. Dissertation Submitted to the
University of Botswana, Gaborone.
Gujarati D.N and Porter D.C (2003) Basic Econometrics. McGraw International Edition.
Hinricks, H (1966). A General Theory of Tax Structure Change During Economic Development. Cambridge,
Mass: Harvard Law School.
Jayasundera P.B. (1991). Buoyancy and Elasticity of Taxes, Report of the Taxation Commission, 1990,
Printed on the order of the Government of Sri Lank.
lxviii Jimoh, A (2003). Fiscal Federalism: The Nigerian Experience. Paper delivered at the meeting of Fiscal
Policy and Growth in Africa, Fiscal Federalism, Decentralization and the Incidence of Taxation,
Ad-Hoc Expert Group Meeting 7-9 October UNCC, Adis Ababa, Ethiopia, Economic Commission
for Africa.
Joergen R. Lotz and Elliot Morss (1970) “Economic Development and Cultural Change” Journal of
Economics, 18, No. 3, 328-341, University of Chicago press.
Koutsoyiannis, A. (1976) Theory of Econometrics. London: The Macmillan.
Kusi, K.N. (1998). Tax Reform and Revenue Productivity is Ghana. Africa Economic Research Consortium
Research Paper, N. 74.
Landau, D. (1983) “Government Expenditure and Economic Growth: A cross-country study” Southern
Economic Journal, Vol. 49, No. 3, 783-92.
Mansfield, C.Y. (1972). Elasticity and Buoyancy of a Tax System – A Method Applied to Paraguay, IMF
Staff Papers, Vol. 19.
Massell, B.F., Pearson and J.B. Fitch (1972). “Foreign Exchange and Economic Development: An Empirical
Study of Selected Latin American Countries”. Review of Economics and Statistics, May: 208 –
212.
Milambo (2001). Elasticity and Buoyancy of Tax System, Illustrations from Sub-Sahara Africa, Bergen:
Chr. Michelsen Institution.
Mtatifikolo F.P. (1990) “An Economic Analysis of Tanzania’s Tax Performance Experiences since the 1973
Tax Art” Eastern Africa Economic Review.
Musgrave, R.A (1969). Fiscal System. (New Haven): Yale.
Musgrave, R.A. and P.B. Musgrave (2004). Public finance in theory and practice. Tata McGraw Hill, New
Delhi, India.
Njoroge (1993). Revenue Productivity of Tax Reforms in Kenya. An African Economic Paper No. 23.
Ole (1975). Income Elasticity of Tax Structure of Kenya for Period From 1962/63 – 1972/75. An Africa
Economic Policy Paper, N. 65.
lxix Omorugi; S.E. (1983) “Growth and Flexibility of Federal Government Tax Revenue: 1960 – 1979”
Economic and Financial Review, Vol. 2, No. 1:11- 19.
Osoro N.E. (1993). Revenue Productivity Implications of Tax Reform in Tansania, Research Paper No. 20,
AERC, Nairobi.
Pindyck, R.S. and D.L. Rubinfeld (1981). Econometric models and Economic forecasts, 2nd ed., (Tokyo:
McGraw-Hill).
Prest, Allan R. (1962) “The sensitivity of yield of personal income tax in the United Kingdom” Economics
Journal, Vol. 52, September, 576-96.
Siegel, J.J. (1979) “Inflation-Induced Distortion in Government and Private Saving Statistics” Review of
Economic and Statistics, Vol. 61:83-90.
Singer (1968) “Techniques for data analysis” Journal of African Economics, No. 7.
Thornton, D.L. (1990). Do Government Deficits Matter? Review, The Federal Reserve Bank of St. Louis,
Vol. 72, no. 5: September/October: 25-39.
Tsegaye (1993). Buoyancy and Elasticity of Tax System, IMF Staff papers Vol. 31.
Wickens, M.R. and M. Uctum (1990). National Insolvency: A Test of the US inter-Temporal Budget
Constraint. Centre for Economic Policy Research, Discussion Paper Series, No. 437, Augus.
Wojciech Kopczuk (2003). Tax Bases, Tax Rates and the Elasticity of Reported Income. Department of
Economic, University of British Columbia, New York.
Wonnacott, R.T. and T.H. Wonnacott (1970) Econometrics. 2nd. Ed. New York: John Wiley and Sons.
Yagub M. (1994) “The causes and Consequences of Budget Deficit in Pakistan” Economic Review.
Yellen, J.L. (1989) “Symposium on the Budget Deficit” Journal of Economic Perspectives, Vol. 3, no.2:73-
93.
Daniel, K.T, Abel, F., E, O. A., and Emmanuel, E. A (2009) “Buoyancy and Elasticity of Tax” Journal of
Monetary and Economic Integration Vol.10,No.2;pp.37-70.
Bahl, R.W. (1971) A Regression Approach to Tax Effort and Tax Ratio analysis, IMF Staff Papers, Vol. 18,
pp. 570-612.
lxx Musgrave, R.A. (1959) The Theory of Public Finance New York: McGraw Hill.
Teera, J.M. (2002) Tax Performance: A comparative study, Working Paper 01-2 centre for Public
Economics. University of Bath.
Gujarati, D. N. (2005). Basic Econometrics. (Forth Edition); New York. McGraw Hill
Jarir Ajluni,(200)“Modelling The Long Run”. Applying the Cointegration Methodology: Notes in
Macroeconomics, pp. 6 – 7.
Johansen, S., (1988) “Statistical Analysis of Cointegration Vectors” Journal of Economic Dynamics and
Control, 12, 231 – 254.
Johansen, S., (1991) “Estimation of Hypothesis Testing of Cointegration Vectors in Gaussian Vector
Autoregressive Models” Econometrica, 59, 1551 – 1580.
Johansen, S., (1992) “Determination of Cointegration Rank in the presence of a Linear Trend” Oxford
Bulletin of Economics and Statistics, 54, 383 – 389.
Granger, C. W. J., and Newbold, P., (1974) “Some properties of Time Series Data and their use in
Econometric Model Specification” Journal of Econometrics, 16, 121 – 130.
Sobel, Russell S. and Randall G. Holcome (1996) “ Measuring the Growth and variability of Tax Bases
over the Business Cycle” National Tax Journal, vol. 49.
Bilquees, Faiz (2004) “ Elasticity and Buoyancy of the Tax System in Pakistan”
Pakistan Development Review. Vol 43.
Milwood, T.T (2012) Elasticity and Buoyancy of the Jamaican Tax System.
Fiscal and Economic Programme Monitoring Development Bank of Jamaica,
Jamaica.
lxxi
APPENDIX I: UNIT ROOT TEST RESULTS
Dickey fuller test
. dfuller DD.totaltax, lag(4)
Augmented Dickey-Fuller test for unit root Number of obs =
25
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -3.031 -3.750 -3.000 -
2.630
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0321
.
. dfuller DD.govexp, trend lag(8)
Augmented Dickey-Fuller test for unit root Number of obs =
21
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
lxxii Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.017 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0083
.
. dfuller DD.realgdp, lag(0)
Dickey-Fuller test for unit root Number of obs =
29
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -7.666 -3.723 -2.989 -
2.625
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
.
. dfuller oilrevenue, trend lag(8)
lxxiii Augmented Dickey-Fuller test for unit root Number of obs =
23
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -3.721 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0210
.
. dfuller D.companyincometax, trend lag(2)
Augmented Dickey-Fuller test for unit root Number of obs =
28
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.405 -4.352 -3.588 -
3.233
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0022
lxxiv
. dfuller D.petroleumprofittax, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -10.538 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller D.educationtax, lag(0)
Dickey-Fuller test for unit root Number of obs =
10
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -3.161 -3.750 -3.000 -
2.630
lxxv -----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0224
. dfuller D.customandexciseduties, trend lag(1)
Augmented Dickey-Fuller test for unit root Number of obs =
29
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.384 -4.343 -3.584 -
3.230
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0023
. dfuller DD.exratereal, lag(4)
Augmented Dickey-Fuller test for unit root Number of obs =
25
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
lxxvi Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -3.610 -3.750 -3.000 -
2.630
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0056
. dfuller DD.import, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
29
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -10.356 -4.343 -3.584 -
3.230
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller D.tot, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
lxxvii
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -5.200 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0001
. dfuller D.vat, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
16
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.593 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0011
lxxviii
. dfuller D.primelendrate, lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -8.790 -3.716 -2.986 -
2.624
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller extgrantnaira, lag(0)
Dickey-Fuller test for unit root Number of obs =
31
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
lxxix Z(t) -3.018 -3.709 -2.983 -
2.623
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0332
. dfuller D.vat, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
16
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.593 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0011
. dfuller D.lag1totaltax, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
29
lxxx ---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -6.124 -4.343 -3.584 -
3.230
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller D.Lag1companyincometax, trend lag(2)
Augmented Dickey-Fuller test for unit root Number of obs =
27
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.600 -4.362 -3.592 -
3.235
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0010
lxxxi
. dfuller D.lag1petroleumprofittax, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
29
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -10.424 -4.343 -3.584 -
3.230
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller D.lag1customandexciseduties, lag(0)
Dickey-Fuller test for unit root Number of obs =
29
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
lxxxii Z(t) -7.349 -3.723 -2.989 -
2.625
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0000
. dfuller D.lag1vat, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
15
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -4.438 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.0019
. dfuller etottax, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
lxxxiii Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -0.531 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.9822
. dfuller ecit, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -0.925 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.9534
lxxxiv . dfuller evat, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
16
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -2.074 -4.380 -3.600 -
3.240
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.5609
.
. dfuller eppt, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
lxxxv Z(t) -1.280 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.8928
.
. dfuller eced, trend lag(0)
Dickey-Fuller test for unit root Number of obs =
30
---------- Interpolated Dickey-Fuller --------
-
Test 1% Critical 5% Critical 10%
Critical
Statistic Value Value Value
-----------------------------------------------------------------------------
-
Z(t) -2.296 -4.334 -3.580 -
3.228
-----------------------------------------------------------------------------
-
MacKinnon approximate p-value for Z(t) = 0.4361
lxxxvi
APPENDIX II: REGRESSION RESULTS
. reg totaltax lag1totaltax realgdp govexp oilrevenue extgrantnaira inflation
taxrefdummy
Source | SS df MS Number of obs =
31
-------------+------------------------------ F( 7, 23) =
246.43
Model | 2.2596e+14 7 3.2280e+13 Prob > F =
0.0000
Residual | 3.0129e+12 23 1.3099e+11 R-squared =
0.9868
-------------+------------------------------ Adj R-squared =
0.9828
Total | 2.2898e+14 30 7.6325e+12 Root MSE =
3.6e+05
-----------------------------------------------------------------------------
-
totaltax | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+---------------------------------------------------------------
-
lag1totaltax | .2953151 .1023608 2.89 0.008 .0835657
.5070646
realgdp | 1.124253 2.979693 0.38 0.709 -5.039711
7.288217
lxxxvii govexp | .0745441 .3293506 0.23 0.823 -.6067695
.7558576
oilrevenue | .4474055 .2003626 2.23 0.035 .034751
.86006
extgrantna~a | -53520.98 118504.6 -0.45 0.656 -298666.5
191624.6 I inflation | -1858.53 6291.394 -
0.30 0.770 -14873.27 11156.21
taxrefdummy | 3219970 467163.8 6.89 0.000 2253569
4186372
_cons | -257965.9 638203.8 -0.40 0.690 -1578191
1062259
-----------------------------------------------------------------------------
-
.
. estat dwatson
Durbin-Watson d-statistic( 8, 31) = 2.683756
. reg vat tot primelendrate taxrefdummy realgdp
Source | SS df MS Number of obs =
18
-------------+------------------------------ F( 4, 13) =
79.61
Model | 2.7668e+11 4 6.9170e+10 Prob > F =
0.0000
Residual | 1.1295e+10 13 868839018 R-squared =
0.9608
-------------+------------------------------ Adj R-squared =
0.9487
Total | 2.8798e+11 17 1.6940e+10 Root MSE =
29476
lxxxviii
-----------------------------------------------------------------------------
-
vat | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+---------------------------------------------------------------
-
tot | .0218214 .0153818 1.42 0.180 -.011409
.0550518
primelendr~e | -4610.64 3599.67 -1.28 0.223 -12387.25
3165.973
taxrefdummy | -104694.1 52198.67 -2.01 0.066 -217462.5
8074.283
realgdp | .7962885 .1229545 6.48 0.000 .5306614
1.061916
_cons | -130461.4 68804.37 -1.90 0.080 -279104.2
18181.39
-----------------------------------------------------------------------------
-
. estat dwatson
Durbin-Watson d-statistic( 5, 18) = 1.993072
.
. reg companyincometax Lag1companyincometax realgdp primelendrate
taxrefdummy
Source | SS df MS Number of obs =
31
-------------+------------------------------ F( 4, 26) =
187.09
lxxxix Model | 5.0661e+11 4 1.2665e+11 Prob > F =
0.0000
Residual | 1.7601e+10 26 676958273 R-squared =
0.9664
-------------+------------------------------ Adj R-squared =
0.9613
Total | 5.2421e+11 30 1.7474e+10 Root MSE =
26018
-----------------------------------------------------------------------------
-
companyinc~x | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+---------------------------------------------------------------
-
Lag1compan~x | .5885913 .1232995 4.77 0.000 .3351456
.8420369
realgdp | .2912466 .1097986 2.65 0.013 .0655524
.5169407
primelendr~e | -1534.227 1098.549 -1.40 0.174 -3792.328
723.8733
taxrefdummy | 19755.97 28500.09 0.69 0.494 -38826.81
78338.74
_cons | -39863.13 25087.2 -1.59 0.124 -91430.61
11704.35
-----------------------------------------------------------------------------
-
. estat dwatson
Durbin-Watson d-statistic( 5, 31) = 2.127995
xc . reg petroleumprofittax lag1petroleumprofittax realgdp exratereal oilrevenue
taxrefdummy
Source | SS df MS Number of obs =
31
-------------+------------------------------ F( 5, 25) =
53.77
Model | 3.7336e+12 5 7.4671e+11 Prob > F =
0.0000
Residual | 3.4716e+11 25 1.3886e+10 R-squared =
0.9149
-------------+------------------------------ Adj R-squared =
0.8979
Total | 4.0807e+12 30 1.3602e+11 Root MSE =
1.2e+05
-----------------------------------------------------------------------------
-
petroleump~x | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+---------------------------------------------------------------
-
lag1petrol~x | -.1447168 .2180011 -0.66 0.513 -.5936986
.3042649
realgdp | 2.159399 .5593121 3.86 0.001 1.007474
3.311323
exratereal | 1181.898 603.8829 1.96 0.062 -61.82179
2425.619
oilrevenue | .0524146 .0255328 2.05 0.051 -.0001711
.1050003
taxrefdummy | -335894 135507.1 -2.48 0.020 -614976.1 -
56811.82
_cons | -518434.8 146974.5 -3.53 0.002 -821134.5 -
215735.2
-----------------------------------------------------------------------------
-
xci
. estat dwatson
Durbin-Watson d-statistic( 6, 31) = 1.479792
.
. reg customandexciseduties lag1customandexciseduties realgdp exratereal
import taxrefdummy
Source | SS df MS Number of obs =
31
-------------+------------------------------ F( 5, 25) =
84.66
Model | 2.4655e+11 5 4.9310e+10 Prob > F =
0.0000
Residual | 1.4562e+10 25 582468511 R-squared =
0.9442
-------------+------------------------------ Adj R-squared =
0.9331
Total | 2.6111e+11 30 8.7037e+09 Root MSE =
24134
-----------------------------------------------------------------------------
-
customande~s | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+---------------------------------------------------------------
-
lag1custom~s | .4246799 .2649663 1.60 0.122 -.1210285
.9703882
realgdp | .4474055 .2003626 2.23 0.035 .034751
.86006
exratereal | 146.4382 134.0966 1.09 0.285 -129.739
422.6154
xcii import | -.0017859 .0045904 -0.39 0.701 -.0112401
.0076682
taxrefdummy | -69161.37 31117.52 -2.22 0.036 -133249.1 -
5073.637
_cons | -97773.28 48087.91 -2.03 0.053 -196812.2
1265.62
-----------------------------------------------------------------------------
-
. estat dwatson
Durbin-Watson d-statistic( 6, 31) = 2.308298
. reg totaltax lag1totaltax realgdp govexp lextgrantnaira inflation
oilboomdummy
Source | SS df MS Number of obs =
31
-------------+------------------------------ F( 6, 24) =
123.76
Model | 2.2181e+14 6 3.6968e+13 Prob > F =
0.0000
Residual | 7.1692e+12 24 2.9872e+11 R-squared =
0.9687
-------------+------------------------------ Adj R-squared =
0.9609
Total | 2.2898e+14 30 7.6325e+12 Root MSE =
5.5e+05
-----------------------------------------------------------------------------
-
totaltax | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
xciii -------------+---------------------------------------------------------------
-
lag1totaltax | -.4804423 .2491356 -1.93 0.066 -.9946329
.0337484
realgdp | 16.96742 5.347215 3.17 0.004 5.931307
28.00352
govexp | .0480777 .3739461 0.13 0.899 -.723709
.8198644
lextgrantn~a | -407403.5 199374.9 -2.04 0.052 -818892.9
4086.013
inflation | -15186.37 9490.314 -1.60 0.123 -34773.42
4400.671
oilboomdummy | -6071377 1307770 -4.64 0.000 -8770481 -
3372272
_cons | 1.18e+07 4847121 2.44 0.022 1827305
2.18e+07
-----------------------------------------------------------------------------
-
estat dwatson
Durbin-Watson d-statistic( 8, 31) = 2.333181