Fiscal Decentralization and Macroeconomic Performance: the Case of Pakistan

Fiscal Decentralization and Macroeconomic Performance: The Case of Pakistan

By: Imran Hanif

Supervised By:

Professor. Dr. Imran Sharif Chaudhry School Of Economics, Bahauddin Zakariya University, Multan (Pakistan)

A dissertation submitted to the School of Economics, Bahauddin Zakariya University Multan, for the fulfilment of the degree of Doctor of Philosophy (PhD) in Economics (2015)

IN THE NAME OF ALLAH, THE MOST BENEFICENT AND THE MOST MERCIFUL

DECLARATION

I, Imran Hanif, a Ph. D Research Scholar in School of Economics, Bahauddin Zakariya University, Multan (Pakistan), do hereby solemnly affirm and declare that this research work has been accomplished by me under the supervision of Professor Dr. IMRAN SHARIF CHAUDHRY, for the purpose of submitting it as a requirement to grant the degree of “DOCTOR OF PHILOSOPHY IN ECONOMICS”. All the sources of information have been acknowledged in this thesis.

(IMRAN HANIF) Ph. D Research Scholar, School of Economics, Bahauddin Zakariya University, Multan (Pakistan)

(2015)

iii

CERTIFICATE

It is hereby confirmed that this thesis is based on the research accomplished by Imran Hanif. It is also verified that the research work has not been previously presented for a higher degree. Mr. Imran Hanif has done this research work under supervision of Professor Dr. Imran Sharif Chaudhry. He has fulfilled all the requirements and is qualified to submit the accompanying thesis for the degree of “Doctor of Philosophy in Economics”.

SUPERVISOR

Professor Dr. Imran Sharif Chaudhry, School of Economics, Bahauddin Zakariya University, Multan

EXTERNAL EXAMINER

DIRECTOR

Professor Dr. Imran Sharif Chaudhry School of Economics Bahauddin Zakariya University, Multan

Dedicated

MY PARENTS, BROTHER AND SISTERS

Who instruct me how to live diligently, and educate me to live with capability?

TABLE OF CONTENTS

TITLE PAGE Declaration iii Certificate iv Table of Contents vi Abbreviations/Acronyms xi List of Tables xiii List of Figures xv Acknowledgements xvi Abstract of the Study xvii

CHAPTER 1 INTRODUCTION 1.1 Background of the Study……………………………………... 1 1.2 Statement of the Problem…………………………………….. 5 1.3 Significance of the Study…………………………………...... 7 1.4 Objectives of the Study………...……………………………. 8 1.5 Organization of the Study……………………………………. 9 CHAPTER 2 REVIEW OF LITERATURE 2.1 Introduction…………………………………………………... 11 2.2 A Glance at Fiscal Decentralization Studies………………..... 12 2.3 A Review of Significant Fiscal Decentralization Studies from

Pakistan……………………………………………………….. 13 2.4 A Review of Significant Fiscal Decentralization Studies from

International Portfolio……………………….. ………………. 16 2.5 Conclusion……………………………………………………. 36 CHAPTER 3 FISCAL DECENTRALIZATION AND PROFILE OF PAKISTAN’S ECONOMY 3.1 Introduction………………………………………………….. 42 3.2 History of the Fiscal Autonomy Campaign in Pakistan ……... 44 3.3 Allocation of Resources in Pakistan………………………….. 45 3.4 Resource Allocation before the Independence of Pakistan in

1947…………………………………………………………… 45 3.5 Resource Allocation after the Independence of Pakistan…...... 46 3.6 Resource Allocation in the Period of One Unit………………. 46 3.6.1 The Prize of 1961………………………………………. 47 3.6.2 National Finance Commission Award of 1964………… 47 3.6.3 National Finance Committee 1970…………………...... 48

3.7 Resource Allocation in the Financial Arrangement of the

1973 Constitution…………………………………………...... 49 3.7.1 The First NFC Award 1974…………………………...... 49 3.7.2 The 2nd NFC Award 1979……………………………… 50 3.7.3 The 3rd NFC Award 1985……………………………… 50 3.7.4 The 4th NFC Award 1990……………………………… 51 3.7.5 The 5th NFC award 1996……………………………... .. 52 3.7.6 The 6th NFC Award 2000………………………………. 54 3.7.7 The 7th NFC Award 2006………………………………. 55 3.7.8 The 8th NFC Award 2009………………………………. 58 3.8 Economic Performance of Pakistan…………………………... 60 3.8.1 Long Run Growth Trends of Basic Economic Indicators

in Pakistan……………………………………………… 60 3.9 Development of the Fiscal Situation in Pakistan…………….. 62 3.9.1 Tax Structure in Pakistan………………………………. 65 3.10 Conclusion…………………………………………………... 68 CHAPTER 4 FISCAL DECENTRALIZATION: CONCEPTUAL AND THEORETICAL FRAME WORK 4.1 Introduction…………………………………………………... 69 4.2 Concepts of Decentralization…………………………………. 69 4.2.1 Political Decentralization………………………………. 70 4.2.2 Administrative Decentralization ………………………. 71 4.2.3 Financial or Fiscal Decentralization …………………... 72 4.3 Fiscal Federalism……………………………………………... 74 4.4 Economic Decentralization…………………………………… 74 4.5 On Going Debate Regarding Fiscal Decentralization………... 75 4.6 Conventional Fiscal Decentralization Theories………………. 79 4.6.1 Stabilization ……………………………………………. 79 4.6.2 Distribution …………………………………………..... 80 4.6.3 Allocation ……………………………………………… 81 4.7 Tiebout Hypothesis; “Voting with Your Feet”……………….. 86 4.7.1 Basic Assumptions of the Tiebout Mechanism………… 88 4.7.2 Brief Awareness of the Characteristics of All the

Communities…………………………………………… 88 4.7.3 Costless Mobility in Different Competitive

Jurisdictions……………………………………………. 89 4.7.4 Impure Public Good Provision…………………………. 89 4.7.5 Large Population Size………………………………...... 89 4.7.6 Dependent on Tax Related Benefits……………………. 89 4.7.7 Externalities Created due to Mobility………………...... 90

vii

4.7.8 Economies of Scale…………………………………...... 91 4.7.9 Benefit Spillover……………………………………….. 92 4.7.10 Non- Static Preferences……………………………..... 92 4.7.11 Production and Location……………………………… 92 4.8 The Tiebout Hypothesis: Empirical Tests……………………. 93 4.8.1 Focus on Capitalization………………………………… 93 4.8.2 Focus on ‘Fiscally Provoked Migration’……………..... 94 4.8.3 Focus on ‘Taxpayer Satisfaction’………………………. 96 4.8.4 Focus on ‘Sorting’……………………………………… 97 4.9 Decentralization and ‘Leviathan’…………………………….. 97 4.10 New-classical Growth Theory……………………………..... 98 4.11 Conclusion…………………………………………………... 99 CHAPTER 5 DATA AND METHODOLOGICAL ISSUES 5.1 Introduction…………………………………………………... 100 5.2 Data Sources …………………………………………………. 100 5.3 Selection of Variables………………………………………… 101 5.4 Application of the Econometric Model………………………. 105 5.4.1 Model for Time Series Data Analysis………………...... 105 CHAPTER 6 FISCAL DECENTRALIZATION AND MACROECONOMIC PERFORMANCE: AN ELEMENTARY DATA ANALYSIS 6.1 Introduction…………………………………………………... 111 6.2 Descriptive Statistics…………………………………………. 111 6.3 Results of Pair-wise Correlation…………………………….... 114 6.4 Short Run Causality Analysis……………………………….... 116 6.5 Conclusion……………………………………………………. 119 CHAPTER 7 IMPACT OF FISCAL DECENTRALIZATION ON GROWTH, INVESTMENT AND INFLATION: A TIME SERIES ANALYSIS 7.1 Introduction…………………………………………………... 120 7.2 Impact of Fiscal Decentralization on Economic Performance

of Pakistan: A Time Series Analysis………………………… 121 7.2.1 The Wald Test (F-statistics)……………………...... 122 7.2.2 Long-Run Impact of Fiscal Decentralization on

Economic Performance………………………………... 124 7.2.3 Estimation of Error Correction Model………………… 126 7.2.4 Diagnostic and Stability Test………………………….. 127 7.3 Impact of Fiscal Decentralization on Economic Growth in

Pakistan………………………………………………………. 130

viii

7.3.1 Fiscal Decentralization Expenditures and Productive

Efficiency………………………………………………. 130 7.3.2 Fiscal Decentralization Revenue and Productive

Efficiency………………………………………………. 132 7.3.3 Fiscal Decentralization and Economic Growth: A Time

Series Analysis…………………………………………. 133 7.4 Impact of Fiscal Decentralization on Public Investment in

Pakistan………………………………………………………. 140 7.4.1 Fiscal Decentralization Expenditures and Productive

Efficiency of Public Investment……………………….. 140 7.4.2 Fiscal Decentralization Revenue and Productivity of

Public Investment……………..……………………….. 142 7.4.3 Fiscal Decentralization and Public Investment: A Time

Series Analysis………………………………………… 143 7.5 Impact of Fiscal Decentralization on Inflation in Pakistan…... 150 7.5.1 Productive Efficiency of Fiscal Decentralization

(Revenue & Expenditure) and Price Stability…………... 150 7.5.2 Fiscal Decentralization and Price Stability: A Time

Series Analysis………………………………………… 151 7.6. Conclusion…………………………………………………... 158 CHAPTER 8 IMPACT OF FISCAL DECENTRALIZATION ON EMPLOYMENT AND SAVINGS: A TIME SERIES ANALYSIS 8.1 Introduction…………………………………………………... 159 8.2 Impact of Fiscal Decentralization on Employment in

Pakistan……………………………………………………...... 160 8.2.2 Productive Efficiency of Fiscal Decentralization (Revenue & Expenditures) and Employed Labour Force……………………………………………………. 160 8.2.3 Impact of Fiscal Decentralization on Employment in

Pakistan: A Time Series Analysis……………………… 161 8.3 Impact of Fiscal Decentralization on Private Savings in

Pakistan………………………………………………………. 167 8.3.1 Fiscal Decentralization (Revenue & Expenditures) and

Productive Efficiency of Private Savings ……………... 167 8.3.2 Impact of Fiscal Decentralization on Private Savings: A

Time Series Analysis…………………………………... 168 8.4 Conclusion……………………………………………………. 175 CHAPTER 9 SUMMARY, CONCLUSION AND RECOMMENDATION 9.1 Summary……………………………………………………… 177

9.2 Conclusion of the Study………………………...... 179 9.3 Suggestions……...…………………………………………..... 181 REFERENCES 184

Abbreviations/acronyms

ADF Augmented Dickey-Fuller AIC Akaike Information Criterion ARCH Autoregressive Conditional Heteroskedasticity ARDL Autoregressive Distributive Lags Model CDA Capital Development Authority CE Co-integration Equations CM Chief Minister CPI Consumer Price Index ECM Error Correction Model EJF Efficient Jurisdiction’s Frontier ELF Employed Labour Force EU European Union FBR Federal Board of Revenue FBS Federal Bureau of Statistics FD First Difference FDR Fiscal Decentralization Revenue FGLS Feasible Generalized Least Square GDP Gross Domestic Product GFS Government Finance Statistics GLS Generalized Least Square GMM Generalized Moving Method GST Gross Sales Tax IMF International Monitory Fund IV Instrumental Variable JB Jarque-Bera KPK Khyber Pakhtunkhawa (new name of NWFP) LR Likelihood Ratio LCU Local Currency Unit LSDV Least Square Dummy Variable MC Marginal Cost MPNS Municipal President’s Nation Survey NFC National Finance Commission NWFP North West Frontier Province

OECD Organization of Economic Cooperation and Development OLS Ordinary Least Square PAM Partial Adjustment Model PFC Provincial Finance Commission PSDP Public Sector Development Program SBC Schwartz Bayesian Criterion SBP State Bank of Pakistan SRO Statutory Regulatory Order UK United Kingdom USA United State of America VAR Vector Auto Regressive VECM Vector Error Correction Model WAPDA Water and Power Development Authority WDI World Development Indicators

xii

List of Tables Table No. Name of the Table Page# Table 2.1: Summary of Macroeconomic Variables and Their Relationships with 38 Fiscal Decentralization in Existing Empirical Studies Table 3.1: Participation in Provincial Award 1970 48 Table 3.2: Provincial Percentage According To the 1974 NFC Award 49 Table 3.3: Provincial Share Table, 1979 Award 50 Table 3.4: Provincial Share Table, 1990 Award 51 Table 3.5: Special Annual Grants to Provinces in 4th NFC, 1990 (Rs. Million) 52 Table 3.6: Matching Grants under the 1996 Award (Rs. Million) 53 Table 3.7: Percentage Provincial Share, 1996 Award 54 Table 3.8: Special Annual grants to The Provinces (Rs. Billion) 54 Table 3.9: Percentage Provincial Share, 2006 Award 56 Table 3.10: Special Grants to Provinces under the 2006 Award (Rs. Billion) 57 Table 3.11: Shares of Provinces According to 8th NFC Award 59 Table 3.12: Economic Performance of Major Economic Sectors in Each Decade 61 Table 3.13: Performance of Pakistan’s Fiscal Indicators (% of GDP) 64 Table 3.14: Tax Assignments of State and Provincial Governments 66 Table 3.15: Structure of Federal Direct and Indirect Tax Revenue (Rs. Billion) 67 Table 6.1: Descriptive Summary of the Variables (period covered from 1972 to 210) 112 Table 6.2: Results of Correlation Matrix 115 Table 6.3: VAR Granger Causality/Block Exogeneity Wald Tests 117 Table 7.1.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root. 122 Table 7.1.2: Results of Bound Test for Co-integration 123 Table 7.1.3: Long- Run Impact of Fiscal Decentralization on Economic 124 Performance Table 7.1.4: The Results of Error Correction Model 126 Table 7.1.5: Results of Diagnostic Tests 127 Table 7.2.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root. 134 Table 7.2.2: Result of Unrestricted Co-integration Rank Test (Trace) 135 Table 7.2.3: Normalize Co-integration Coefficients 136 Table 7.2.4: Result of Vector Error Correction Model 137 Table 7.3.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root. 144 Table 7.3.2: Result of Unrestricted Co-integration Rank Test (Trace) 145 Table 7.3.3: Normalize Co-integration Coefficients 146 Table 7.3.4: Result of Vector Error Correction Model 147 Table 7.4.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root. 152 Table 7.4.2: Result of Unrestricted Co-integration Rank Test (Trace) 153

xiii

Table No. Name of the Table Table 7.4.3: Normalize Co-integration Coefficients 154 Table 7.4.4: Result of Vector Error Correction Model 155 Table 8.1.1: Estimation Results for Short Run 162 Table 8.1.2: Estimation Results for Long Run 163 Table 8.2.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root 169 Table 8.2.2: Result of Unrestricted Co-integration Rank Test (Trace) 170 Table 8.2.3: Normalize Co-integration Coefficients 171 Table 8.2.4: Result of Vector Error Correction Model 172

xiv

List of Figures

Figure No. Name Page# Figure 3.1: The Performance of Major Economic Sectors in Pakistan 61 Figure 3.2: Trends in Total and Partial Revenue and Expenditures 63 Figure 3.2 (a): Trends in Total and Partial Revenue and Expenditures 63 Figure 3.2 (b): Trends in Total and Partial Revenue and Expenditures 63 Figure 3.2 (c): Trends in Total and Partial Revenue and Expenditures 63 Figure 4.1: Public Finance and Public Choice 83 Figure 4.2: Tiebout Adjustments 87 Figure 4.3: Migratory Incentives 94 Figure 4.4: Optimal Location 96 Figure 7.1 Stability Test CUSUM and CUSUMS 129 Figure 7.2: Trends of FDX and GDP_FDX in Pakistan 132 Figure 7.3: Trends of FDR and GDP_FDR in Pakistan 133 Figure 7.4: Trends of FDX and GTI_GDP in Pakistan 141 Figure 7.5: Trends of FDR and GTI_GDP in Pakistan 142 Figure 7.6: Trends of FDX_GDP, FDR_GDP and inflation in Pakistan 151 Figure 8.1: Trends of FDX_GDP, FDR_GDP and ELF in Pakistan 161 Figure 8.2: Trends of FDX , FDR and SAPVT_GDP in Pakistan 168

Acknowledgements

My heartiest praises and thanks to the Lord of Lords and Creator of the universe, whose unlimited power and glory accomplished all good things. Peace and blessing to the last Prophet (PBUH) of Almighty Allah who brought peace and blessing to all creatures. I feel great pleasure and happiness on the successful completion of my dissertation for the award of PhD degree in economics, which is nothing but all because of gracious and Almighty Allah and Great Messenger (PBUH) of Allah. It is matter of great pleasure for me to pay my profound gratitude and recognition to my highly knowledgeable and esteemed supervisor Dr. Imran Sharif Chaudhary who was always a torch of hope and inspiration for me during my thick hours. His cooperation, generous efforts, constant encouragement and supervision made this mile stone achievable. Without his precious support it would not be possible to conduct this research. My sincere thanks also goes to Dr. Sally Wallace and Dr. Musharraf Rasool Cyan who provided me an opportunity to join their team as research scholar, and who gave access to the advance research facilities. I am also benefited and steered from the constructive comments, criticism and suggestions of Dr. Paritam Singh, Dr. Maria Pilar Gago, Dr. Dennis Carlson, Dr. Hammyat Ullah Khan, Dr Shahnawaz Malik, Dr. Muhammd Zahir Faridi, Dr. Muhammad Omer Chaudhry and Dr. Muhammad Ramzan Sheikh. I am grateful to them for their generous encouragement, painstaking cooperation and guidance in the successful culmination of this research work. I would certainly take this valued opportunity to pay my deepest thanks and sincere gratitude to my sweet mother, my father M. Hanif (Late), elder brother Irfan Hainf, sisters, Robert N. and Nancy Leitch, who proved themselves as siblings of mine in alien land of USA, also stood by me and encouraged me when the light of hope seems to vanish. I extend special gratitude to my sincere friends Muhammad Zahid Ikram and Syed Asad Ahmad Shah, Ozlem Tuba and Tugba Kalafatoglu. They always stood behind me to encourage me cherish me.

IMRAN HNAIF

xvi

Abstract of the Study:

The study aimed at investigating the impact of fiscal decentralization on macroeconomic performance of Pakistan. For gauging macroeconomic performance of Pakistan key economic indicators like economic growth, public investment, inflation, employment and private savings have been used. Most of the work regarding fiscal decentralization is related to developed economies to examine the relationship of fiscal decentralization and economic performance or growth. The results of these studies are not established a clear cut relationship between decentralization and economic growth as the findings of most of studies are quite contradictory with each other for example; studies by Davoodi and Zou (1998), Woller and Phillips (1998), Zhang and Zou (1998), and Xie et al. (1999) found an inverse association between fiscal decentralization and economic growth. While other studies like Akia and Sakata (2002), Iimi (2005), Lin and Liu (2000), Stansel (2005) and Yilmaz (2000) found a positive relationship between fiscal decentralization and economic growth. Besides that very little work has been done in developing and transitional economies to explore the connection between fiscal decentralization and economic performance. This association demanded more attention and research work to provide explicit guidelines for policy makers to design and propose successful completion of fiscal decentralization in developing economies like Pakistan.

To investigate the impact of fiscal decentralization on macroeconomic performance in Pakistan study summarized the brief history of resource distribution among the provinces. It has been concluded that since 1991, the NFC awards have improved the process of resource allocation to the Provincial Governments in Pakistan. Moreover, due to these awards direct shifting of funds and grants have been enhanced to all provinces. Likewise, the incentive of matching grants forced the provinces to bring improvement in their efficiency, self-reliance and resource generation and in turn obtain financial independence.

To explore the impact of fiscal decentralization on macroeconomic performance in Pakistan the study used a time series data for the period from 1972 to 2010. The empirical results show that fiscal decentralization (revenue and expenditures) is an effective tool to enhance the economic stability, encourage the public investment and foster the economic growth in Pakistan. Fiscal decentralization is also an effective tool to promote the employment level and induce the rise in private savings. The increase in subnational share of total government expenditures due to fiscal decentralization generate more development projects and create more employment opportunities in Pakistan. Specifically, expenditure decentralization is a mean of economic development and can bring positive change in employment opportunities. Contrary to this, the rise in subnational revenue share in total government

xvii

revenue has negative association with employment. The major part of subnational revenue share is collected through taxes. It means that higher tax rates can have a discouraging impact on productive ability of local industries which may raise the level of unemployment in Pakistan. For empirical investigation study used aggregate and disaggregate macroeconomic performance indicators and suggested that fiscal decentralization has significant and positive influence on macroeconomic performance of Pakistan. These findings are matched with the majority of previous literature which stated the positive impact of fiscal decentralization on macroeconomic indicators (Naqvi and Khan, 1989; Gregorio, 1996; Frenkel and Mehrez, 1998; Burdekin et al. 2004; Lin and Liu, 2000; Zhang and Zou, 2001; Qaio et al. 2002; Ahmad and Mortaza, 2005 and Jin et al. 2005). However, these studies also indicate that in developing economies the numbers of constraints which can curb the effectiveness of fiscal decentralization are greater than in developed economies.

Finally, study verified that on one hand fiscal decentralization will promote macroeconomic stability by controlling the inflation and unemployment in Pakistan and on other it will enhance public investment, private savings and encourage economic growth in the long run. The findings of the study revealed that the expenditures and revenue autonomy to subnational governments has a positive impact on macroeconomic performance in. These findings are also endorsed by the empirical findings of Oates’ (1999) which suggested that the fiscal autonomy to provincial and local governments will promote economic growth and ensure the efficient provision of public good and services at local level.

xviii

Chapter 1 Introduction

1.1 Background of the Study

Fiscal decentralization or devolution of economic powers is considered an important reform process and a policy debate all over the globe. The wave of fiscal decentralization has gained momentum in South Asia, particularly in Pakistan, India and Malaysia as these have been adopted as federal systems. Fiscal decentralization currently has become a basic tool of rescuing the emerging economies. In spite of little empirical work carried out to examine the relationship between macroeconomic efficiency and fiscal decentralization, it is not what should have been done.

Decentralization, in more than one tier the devolution of powers from national government to sub-national government or autonomy of certain functions to the sub-national government is known as decentralization. While shifting fiscal responsibilities in terms of expenditure and income from national government to sub national government is known as fiscal decentralization. As sub-national governments have a closer interaction with local citizens, it is generally considered that sub-national departments can allocate different baskets of desirable goods and finance various projects according to the desire of local citizens. Therefore, fiscal decentralization is an effective means to provide a better division of public goods according to the demand (Musgrave and Musgrave, 1959 and Oates, 1972), to lower the signaling and mobility costs (Breton and Scott, 1978), to promote the horizontal competition (Tiebout, 1956) and to provide the optimal type and quantity of public goods (Breton, 1996). Fiscal decentralization on one hand is considered an effective tool to promote economic growth at

the local level due to greater participation of local citizens in development and economic activities while on the other it can enhance the coordination and administrative costs.

Oates (1993) clearly argued in favor of a positive relationship between fiscal decentralization and economic growth. Moreover, it is important to consider that it may improve the efficiency of allocation of resources and improve economic growth, though it can make it difficult to carry out the stabilizing policies. For instance, China’s empirical analysis has shown that, after rapid economic decentralization, it experienced both higher growth and instability simultaneously (Prudhomme, 1994 and Tanzi, 1996). All in all, Fiscal decentralization has become an important policy issue demanding more attention for understanding, recognizing that there is not a consensus on Oates’ presentation including mobility, accountability of government, public interest, among other issues. As a matter of fact, some scholars believe that decentralization is inherently not welfare enhancing.

According to Bird and Smart (2002) “The provision of receiving transfers needs a clear authority, sufficient resources and sufficient flexibility”. Similarly, Rondinelli and Dennis

(1981) explains decentralization as “the method through which the duties as well as means from federal to provincial governments are developed”.

In decentralization new tiers of government are established becoming responsible for the design and implementation of economic policies. The concept of fiscal decentralization is becoming popular due to its strength in political matters. Moreover it increases competition on economic grounds by raising the competition of the allocation of resources among the localities. Thus, in a fiscal decentralization process the powers of collecting taxes, covering

expenses and correcting imbalances in resource mobilization and obligations are transferred to the local units of government.

The term decentralization may refer to fiscal, political, administrative and economic decentralization. The term administrative decentralization is the allocation of administrative powers among different tiers of government. The term political decentralization is a process in which the power of making political decisions is shifted at the local level to citizens or their elected representatives. Similarly, fiscal decentralization measures how the power of generating revenue and making expenditures is shifted to local authorities, while in economic decentralization certain functions are transferred from the public to the private sector.

Phenomenon of fiscal decentralization is now well recognized. Given that most countries are practicing different forms of decentralization it is important to evaluate the impact of decentralization on economic welfare. It is a common belief that economic growth is enhanced through fiscal decentralization even though there could be implications for resource redistribution (Martinez-Vazquez and McNab, 2001). Due to the popularity of the phenomenon it is considered worthwhile in this study to evaluate its impact on economic growth.

The attainment of healthier economic performances is a serious debate in developing nations and a prerequisite for their endurance. Besides the fact that they are confronting problems of poor allocation of social services such as access to clean water, sanitation services, education and housing, which are far from acceptable levels as well as a deficiency of good governance. Raising the participation of poor people in economic decision making and in political affairs is the biggest achievement of decentralization. Thus, with the help of decentralization, barriers in social and institutional setup can be removed. Through

decentralization, service agencies come closer to local people and therefore they can better control the provision of services. One can say that decentralization is an effective tool for strengthening local people (World Bank, 2001).

No doubt the demand for decentralization is rising all over the world and it has become a debated subject among the policy experts. All have different opinions of the decentralization process but agree on the major principle which deals with the provision of public goods and services. This principle will be followed effectively when the liability for providing goods and services is assigned to local government that can expertly deliver the goods and services to the local citizens (Martinez-Vazquez and McNab, 2002 and Mello, 2000).

Number of studies have been examined the direct impact of fiscal decentralization on economic growth with major emphasize on developed economies but the nexus between fiscal decentralization and macroeconomic performance in developing economies was seldom explored. Therefore, the major emphasis of current work is to study the impact of fiscal decentralization on economic performance of a developing country and this work will focus on Pakistan. This study will consider the simultaneous effect of revenue and expenditure decentralization on key macroeconomic indicators of Pakistan. On one hand, it will examine the direct impact of fiscal autonomy on economic growth and on other it will explore the impact of inflation, public investment, employment and private savings on fiscal autonomy in

Pakistan. In fiscally decentralized systems the provincial governments play a vital role in the program financing, the provision of public services and the process of tax collection. In centralized systems, funds are exclusively generated by national government and purely used to finance the nationally preferred programs. However for a better provision of services at

lower levels, it is generally considered, fiscal decentralization would be less expensive and easier for the national government to assign powers to provincial and local level governments.

1.2 Statement of the Problem

Over the past two decades, the world has paid attention to raising economic performance and finding alternative ways to foster economic growth. Great deals of large and small economies have paid attention to fiscal decentralization and they have considered it as a successful way to improve economic performance. Instead of centralization, when central governments exercise control over state and municipal levels of government, fiscal decentralization is now taking place in developing countries. It is considered an effective tool to avoid poor governance, inefficient economic growth and poor economic performance.

Various researchers have proven a positive relationship between fiscal decentralization and economic growth (Bird and Wallich 1993; Rivlin, 1992; Sylla et al. 1995; Legler, 1990 and

Gramlich, 1993) but in spite of that, a number of studies have proven a reverse relationship

(Davoodi and Zou, 1998 and Zhang and Zou, 1998). So, the decision of national governments to give autonomy to the provincial or state governments to raise their efficiency and promote economic performance is still an unsolved question.

In this context, this study will deal with decentralization and different fiscal responsibilities in terms of income and expenditure and examine the impact on macroeconomic indicators in

Pakistan. No doubt the positive or negative influence of fiscal decentralization on macroeconomic performance basically depends on design and policy implementation. So this study will not only determine the relationship between fiscal decentralization and key

determinants of macroeconomic performance, it will also devise a suitable policy designed for the successful fiscal decentralization process.

Oates (1972) argued that mostly central governments remain inefficient in the provision of services, in making economic progress and in achieving an efficient distribution of income.

Tiebout’s hypothesis also states that if people remain unable to find their desired goods and services then they vote by their feet and move to the localities which contain their desired mix of services and taxes through which they can increase utility and welfare level.

Some experts argued that fiscal decentralization may increase economic growth and improve the provision of public goods and services at the local level. Local governments can better appreciate the needs of their local citizens and allocate goods and services according to their demand and desires. Moreover the higher degree of fiscal decentralization exerts pressure on local governments to explore ways to increase their performance which may raise economic performance at the gross root level. It also compels the local governments to find ways to accelerate economic growth by finding new investment opportunities and extending business at the local level.

Some experts deviated to former ideas and highlighted the problems concerning funds accumulation methods adopted by the local governments. They argued that the local governments increased the burden of taxes to accumulate funds to finance economic activities at the local level. While enduring a heavy tax burden, the purchasing power of the local people was be lower due to the heavy cut in disposable income and as a consequence they spent less resulting in the reduction of economic growth.

The execution of the decentralization process is also a very important aspect and needs special attention before embarking on the decentralization process. Due to different geographical and ecological characteristics, decentralization execution varies from region to region. The primary emphasis of the present work is to study the composition of fiscal decentralization in Pakistan and to determine its contribution in raising economic performance.

This study will highlight the importance of fiscal decentralization to raise the economic performance and its implementation in developing countries like Pakistan. Instead of using only a macroeconomic performance index this study will also examine the impact of fiscal decentralization on various macroeconomics performance indicators. We believe that the results will provide the mechanism of successful decentralization and the importance of various economic indicators which can improve the quality and effectiveness of fiscal decentralization for Pakistan. Moreover, this study will provide guidelines for our policy makers and help them to focus on and formulate such policies which can improve the quality of fiscal decentralization and enhance economic performance of Pakistan.

1.3 Significance of the Study

The present study will analyze and discover the impact of fiscal decentralization on macroeconomic performance of Pakistan. Schaltegger and Feld (2009) suggested that a decentralized government is more successful than centralized governments in developing countries. So, this study may be helpful to find out how fiscal decentralization can assist to improve the economic performance of a developing economy like Pakistan. In the light of this study, it may be possible to find out in depth which macroeconomic indicators are contributing

positively to economic performance and which are contributing less than the required level and need to be improved.

1.4 Objectives of the Study

The world has been organized in large free trade zones and many European countries have taken steps towards monetary and economic unions. Due to this, it is a natural phenomenon that decentralization has become the center of economic and political debates in developing countries. So this study will examine and suggest some policies regarding how the Pakistani government can embark on fiscal decentralization and may achieve its established targets.

i. To analyze the process of fiscal decentralization in Pakistan

ii. To explore the performance of various macroeconomic indicators in Pakistan

iii. To investigate the impact of fiscal decentralization on macroeconomic performance of

Pakistan

iv. To investigate the impact of fiscal decentralization on economic growth, inflation and

public investment in Pakistan

v. To determine the impact of fiscal decentralization on employment and private savings

in Pakistan

vi. To suggest an appropriate policy/recommendations regarding fiscal decentralization

and macroeconomic performance of Pakistan.

1.5 Organization of the Study

The study is being planned as follows.

Chapter one includes the introduction of fiscal decentralization and emphasizes the need of the subject area of the study. Importance of this study to investigate the impact of fiscal decentralization on macroeconomic performance have been discussed in the present chapter.

Finally, the statement of the problem and the key objectives of the study have been presented in present chapter.

A review of the literature will be presented in chapter two which may include national and international studies related to fiscal decentralization expenditure and revenue and its impact on various macroeconomic indicators. This chapter will emphasize on different variables, methodologies and procedures which have been adopted in the previous studies.

Chapter three will incorporate the execution process of fiscal decentralization with special reference to Pakistan since its emergence. It will also provide a comprehensive detail of various constitutional amendments and resource distribution methods with the passage of time. In addition, economic performance of major macroeconomic indicators of Pakistan has been discussed in this chapter.

Chapter four consists of the conceptualization of fiscal decentralization and discusses the basic theories of fiscal decentralization. This chapter will shed light on various types, dimensions and theories of decentralization to propose an effective methodology to examine the impact of fiscal decentralization on macroeconomic performance in Pakistan.

After the literature and the theoretical review, chapter five incorporates the data and methodological issues. This chapter will describe the data sources and types of data, econometric strategies and finally the description of variables which will be used to investigate the relationship between fiscal decentralization and macroeconomic performance.

Chapter six will present an elementary data analysis. In this chapter study will present the statistical summary and pairwise correlation by using the correlation matrix. Chapter seven will include empirical analysis to explore the impacts of fiscal decentralization expenditures and revenue on macroeconomic performance in Pakistan by using a time series data for the period from 1972 to 2010. For macroeconomic performance the key economic indicators like economic growth, Public investment and inflation rate will be used.

In chapter eight the study will explain the empirical analysis to examine the Impact of fiscal decentralization on employment and private savings in Pakistan.

Finally, the summary, conclusions and suggestions for the improvement of macroeconomic performance and to swift the process of fiscal decentralization’s execution in Pakistan will be presented in chapter nine.

Chapter 2

Review of Literature

2.1 Introduction

Money plays the basic role in the life of an economy whether it is centralized or decentralized. Every country in the world is confronting the issues of distribution and money collection as well as facing various issues such as tax imposition, tax collection and its redistribution. Nearly all countries have at least two or more tiers of government. The rising debates on federalism have highlighted the complexity of distribution of powers and responsibilities on each tier of government. So we can say that the critical and dynamic character of fiscal decentralization has made it the most studied and researched aspect of decentralization.

There are a number of studies which have examined the distribution of these responsibilities at different tiers of government and determined their links to economic growth.

Fiscal decentralization is a multidimensional phenomenon and always attracts the interest of international organizations, public authorities and researchers. Their aim has been to propose and design strategies for the execution of successful fiscal decentralization processes. On the basis of ancient studies of decentralization, in this chapter we will briefly explain the relationship of different determinants of fiscal federalism and economic growth. A comprehensive review of related studies is a prerequisite to design and incorporate the data and methodological issues. We will review country and cross-country level studies and studies selected for a review due to their significance with the work in process.

2.2 A Glance at Fiscal Decentralization Studies

Fiscal decentralization has proven to be a successful tool to promote economic growth in developed or industrialized countries but its consequences in developing economies are still debatable. Oates (1993) stated that the infrastructure investment could be better targeted and outcomes can be improved by the means of fiscal decentralization. Ultimately fiscal decentralization could enhance economic growth. Thus, Oates and his followers generally considered that fiscal decentralization is a successful way to increase the efficiency and productivity of public expenditure and revenue (Oates, 1972; Bhal and Linn, 1992; Yilmaz,

2000 and Stansel, 2005).

Later on Davoodi and Zuo (1998) used national and subnational goods as separate input in the Cobb-Douglas production function and stated that if the subnational spending share will match the Cobb-Douglas exponent then a certain degree of fiscal decentralization can enhance economic growth. So the majority of fiscal decentralization literature in late 1990s found an adverse impact of fiscal decentralization on economic growth for developing countries (Zhang and Zuo, 1998; Davoodi and Zuo, 1998 and Demello, 2000).

On the basis of previous findings we can classify the fiscal decentralization literature into three groups. The first group consists of the studies which confirmed the Oates ideology that fiscal decentralization has positive impact on economic growth (Huther and Shaw, 1998; Jin et al. 1999; Yilmaz, 2000; Lin and Liu, 2000; Akai and Sakata, 2002; Behnich et al. 2002;

Qiao et al. 2002; Thiessen, 2003; Stansel, 2005 and Iimi, 2005). The second group consisted of those studies which remained indifferent and found no significance links between decentralization and economic growth (Woller and Philipps, 1998; Xie et al. 1999 and Bodman

and Ford, 2006). The third group of the literature disconfirmed the Oates findings and concluded the negative impact of fiscal decentralization on economic growth (Davoodi and

Zou, 1998; Zhang and Zou, 1998 and Woller and Phillips, 1998). Recently, numbers of studies have shifted from direct to indirect channels to explore the relationship between fiscal decentralization and macroeconomic performance so we will review some selective and relevant literature on fiscal decentralization.

2.3 A Review of Significant Fiscal Decentralization Studies from Pakistan

Malik et al. (2006) investigated fiscal decentralization’s impacts over economic development within Pakistan. Here in Pakistan, policy makers as well as economists view that fiscal decentralization of a nation could be an influential way while promoting economic development as well as accelerating the economic growth’ pace. The underlying study intends to investigate fiscal decentralization’s impacts over economic development within Pakistan and for the time duration of 1971 to 2005. Technique of cointegration has been employed in order to analyze the empirical data. Findings of the study have recommended that the underscored fiscal decentralization creates positive impacts over economic development. It was summarized that at development’s earlier stages, central government had got to be found in quite better position in order to deal with fiscal decentralization. Furthermore, it was stated that if revenues as well as expenditures’ shares of provincial government rise with the passage of time then there would be possibilities of slowing down the economic development’s pace.

Khattak et al. (2010) understood fiscal decentralization as very significant policy weapon for achieving the efficiency regarding economy as well as ensuring influential governance via financial autonomy which belongs to provincial governments. For the purpose of

accomplishing tasks of national level with more efficacy, fiscal decentralization could assist federation’s smaller level units in taking part country’s economic growth as well as endowing central government with lots of opportunities. It had also been taken as very important measure regarding the growth enhancement. Governments of lower level are empowered i.e. provincial etc. via the empowerment of administration as well as financial autonomy. Through this study, it was highlight that within Pakistan, distribution criteria of the resources has often been criticized by many of federal units over various grounds. This investigation also attempted to investigate sharing of fiscal resources apart from investigating the fiscal decentralization’s effects over Pakistani economic growth and using data of time series within 1980 to 2007.

Findings of the study revealed that resource pool has got to be expanded because of taxes’ inclusion within divisible pool as well as development while collecting the tax. The major population’s criterion in order to distribute the resources has become caused friction between different provinces therefore resulted in NFC awards’ deadlocks. In earlier times, Pakistani distribution mechanism of resources remained unsuccessful while influencing country’s prolong economic growth. Findings of the study recommended that type of criteria which had been used to address distribution of horizontal resources ought to broad via integrating the criteria regarding tax collection efficacy lag related to infrastructure as well as area. In the same way, administrative decentralization must be carried along passable fiscal decentralization as well as capacitating different provinces with useful taxation powers. This all can result to have positive competition between jurisdictions along with ensuring greater efficacy or human capital for promoting country’s economic development.

Faridi (2011) explained that Pakistan has got to be as one of underdeveloped economy which doesn’t have any significant development rate. Several sources regarding economic

growth could be there but this study remained focused over fiscal decentralization in terms of its key sources regarding economic growth. Fiscal decentralization could intend to improve public sector’s efficacy or empowering the economic growth. This study’s analysis was setting its base over annual data in the time duration of 1972 to 2009. Autoregressive model has also been used with relation to ordinary least square estimation. Findings revealed that both variables related to fiscal decentralization including revenue and expenditure autonomy create significant as well as positive impacts over economic development. It was summarized at the end that federal government ought to delegate kind of fiscal power towards both local as well as provincial governments in order to raise welfare as well as economic growth.

Faridi et al. (2012) inspected fiscal decentralization’s role at times of producing opportunities regarding the employment within Pakistan. The study had been used secondary data undertaking time duration of 1972-2009. Ordinary Least Square had been used to calculate the impact of fiscal decentralization and employment and in the end, it was concluded that type of expenditure decentralization manages to have significant and positive impact over employment generation however revenue decentralization was not found suitable in order to generate employment. Both inequality as well as poverty has been decreasing employment within Pakistani sphere. This is the reason; study recommended that fiscal autonomy has got to be very important in order to create opportunities regarding employment within Pakistan.

Faridi et al. (2012) put an emphasis that process of an effective fiscal decentralization is quite useful way for increasing citizens’ participation in both development of local level as well as economic activities. They revealed that fiscal decentralization creates positive impacts over generation of employment and economic development however the negative association with inflation. Improvement in relation with educational activities as well as economic

integration intends to promote activities of employment within state sphere. On the other hand, fiscally decentralized economies have large capacity for promoting the employment activities as compared to unitary governments. Perhaps it has been suggested by this study that in order to promote activities regarding employment fiscal decentralization has got to be as an important tool as it also get to discourage nominal wages’ rise which ensures the control over high inflation rate.

Iqbal et al. (2013) intended to investigate fiscal decentralization’s impacts over economic growth. They scrutinized kind of complementarity among democratic institutions to promote growth as well as fiscal decentralization. Modeling framework had been utilized in this work in terms of model of endogenous growth. Moreover, three of varied measure regarding the fiscal decentralization had been utilized for capturing the multidimensionality. The analysis as results revealed that decentralization in revenue gets to promote economic advancement however expenditure decentralization retards kind of economic development within Pakistan.

Composite decentralization effect economic growth positively via inferring an immediate decentralization strengthens one another for promoting economic advancement. Findings of the study also revealed that kind of democratic institutions come to play an important as well as significant role at times of realizing fiscal decentralization’s benefits.

2.4 A Review of Significant Fiscal Decentralization Studies from International Portfolio

Prudhomme (1995) stated that the demand for decentralization is very high in developed and developing economies. The benefits of decentralization are not up to the mark as theory of fiscal decentralization suggested. Therefore, to avoid the danger of decentralization, the

choice of certain functions, regions and sectors is a primary requirement to design the decentralization program for an economy. The provision of services is not a problem but the joint production of services is a serious problem for the central, regional or local government.

The fiscal decentralization acts like some potent drug that can be prescribed for a relevant illness in correct dose and for specific time, but in case of an overdose and wrong timing it can harm rather than heal. Finally the study highlighted the adverse effects of decentralization for better understanding of its dangers and wise application of decentralization programs.

Smoke and Lewis (1996) explored the relationship between fiscal decentralization and financing of local public services in Indonesia. The study highlighted the key obstacles to decentralization in developing countries and introduced strategies to minimize those obstacles.

The author stated that before embarking on fiscal decentralization state government should analyze which functions or assignments local governments can manage in a better way. The study indicates that in Indonesia, like other developing countries, that system of providing local public services is highly centralized. However, after knowing the weaknesses of a centralized system, Indonesia has been attempting to decentralize its public sector with heavy support from international donors. In spite of these efforts the local government’s performance still seems stagnant. Finally, the study concluded that fiscal decentralization can improve the standard of living in developing countries like Indonesia, only when local governments are capable of meeting the given tasks efficiently.

Zhang and Zou (1998) analyzed the relationship between fiscal decentralization and economic growth in developed and developing economies. The study was based on a panel dataset of 46 countries and considered the period from 1970 to 1989. The analysis used three government levels that were federal, state and local while the level of fiscal decentralization

was determined by measuring the spending of state and local levels of governments as a fraction of total government spending. It means, if spending by local and state governments was rising as compared to the spending by the federal government, then this represents a rising level of fiscal decentralization. The study deviated from Barro’s model and assumed three levels of public spending. The results of the study showed that the developed countries are more decentralized as compared to developing countries. Developed countries have a positive relationship between fiscal decentralization and economic growth while developing countries’ data showed a negative relationship.

Zhang and Zou (2001) examined the allocation of fiscal resources between central and subnational governments and their impact on economic growth in China since the reforms began in the late 1970’s. The study used the provincial panel data during the period 1978-1992 for

28 provinces. By using the Least Square Dummy Variable (LSDV) the results showed that labour growth has a positive insignificant effect on economic growth while central and provincial taxes have negative insignificant effects on growth. The conventional results would have suggested that fiscal decentralization has a positive relationship with economic growth.

So results suggested that the higher degree of fiscal decentralization had a negative relationship on provincial economic growth from 1978 to 1992 in China. The study highlighted the important point for other transition and developing economies; if local shares in fiscal expenditure and revenue were already higher then further decentralization may slower economic growth.

Davoodi and Zou (1999) stated that the provision of human capital and infrastructure were more sensitive to local or regional conditions and more effective to enhance economic

development than centrally determined policies. He explored the linkages between fiscal decentralization and economic growth in The United States’ economy by using a general analytical model. The study was based on an endogenous growth model and used two inputs: public spending and private capital, while empirical analysis was based on a time series dataset for the period of 1948 to 1994. For data analysis the Ordinary Least Square (OLS) method was used and it was concluded that the existing spending share for local and state government are consistent with growth maximization at the three levels of government. Finally the study suggested that if growth and efficiency are the key objectives, then additional fiscal decentralization in The United States can prove to be harmful for economic growth.

Demello (2000) determined the relationship between fiscal decentralization and intergovernmental fiscal relations. The study stated that the process of fiscal decentralization basically consists of moving the revenue resources and expenditure functions to lower government tiers. It is expected that fiscal decentralization is a process which brings government closer to the local people and a means to enhance transparency in the delivery of services, accountability, improvements in public sector efficiency and policy making. The econometric analysis of the study based on a sample of 30 countries and the results showed that the process of fiscal decentralization did not prove very fruitful in developing countries due to coordination failures in intergovernmental fiscal relations. Mostly developing countries were unable to meet the primary requirements for successful fiscal decentralization which leads to deficit bias in decentralization policy making. The study suggested that better design of institutions is prerequisite for successful decentralization which can minimize the coordination failures in intergovernmental fiscal relations by improving fiscal discipline at subnational levels of governments.

Jin and Zou (2000) examined the connection between expenditure and revenue decentralization at the province level and the provincial economic growth in China. They collected panel datasets for 30 provinces for the period 1979-1993 and 1994-1999 respectively.

They evaluated how the link between economic growth and fiscal decentralization in China is affected by the shifting of contracted revenue sharing (1980- 93) to a tax assignment system

(1994-99). The provincial GDP growth rate was considered as a dependent variable while the provincial tax rate, control tax rate, revenue decentralization, expenditure decentralization, investment rate, growth rate of labour, provincial inflation and openness were considered as explanatory variables. In both time periods revenue and expenditure decentralization will lead to the enhancement of provincial growth.

Ozcan (2000) explored the linkage between decentralization and local development in

Turkey. The study used primary data collected from medium sized cities called Kayseri,

Denizli and Gaziantep and showed that in the recent decade decentralization has enabled local groups and municipalities to realize and formulate development projects at local levels. The results showed that in Kayseri the local economic development through local initiatives has developed a competitive environment among the local political interests and civic groups. The study concluded that in Turkey, decentralization should be considered together with raising participatory democracy in government at the national and local levels and it can be achieved with a credible and strong state, and if a state is strong and credible then there is no need for a supranational body like The European Union (in which localities can interact and cooperate).

So, decentralization can support strong and credible states by promoting competition and economic development at the local level.

Akia and Sakata (2002) explored the relationship between fiscal decentralization and economic growth by using the states’ level data from 50 states of The United States. The study considered four decentralization indicators to reflect the necessary viewpoints of fiscal decentralization and growth linkages. By using the variables like share of revenue in the total budget, the share of expenditure in the total budget, combining the share of revenue and expenditure in the total budget and how public spending maintained at lower level governments, the study applied the Ordinary Least Square (OLS) method on cross sectional data of states. The result showed a positive relationship between fiscal decentralization and economic growth. Further studies suggested that the moves toward fiscal decentralization in developing countries may foster economic growth. However, to investigate the process of fiscal decentralization it is a prerequisite to construct the accurate decentralization indicators which may guide the institutional and political authorities about how to raise or collect taxes and undertake public expenditures.

Jin and Zuo (2002) analyzed the effect of fiscal decentralization on national, sub-national and aggregate government size. Instead of using ordinary regression the study analyzed the relationship between fiscal federalism and government size and used panel data of thirty two developing and industrial countries. Feasible Generalized Least Square (FGLS) results showed that the borrowing constraints have no significant effect on sub-national, national and aggregate government size. Similarly, revenue decentralization leads to reduced national government size and larger sub national governments, while expenditure decentralization leads to smaller national government and larger sub-national and aggregate governments. So revenue decentralization will lead to smaller aggregate government size because it reduces national government size more than it increases sub-national governments. Further results showed that

the vertical imbalance is the only variable which can raise the size of the public sector at the national, sub-national and aggregate government levels.

Martinez-Vazquez and McNab (2003) explained that the theoretical relationship between fiscal decentralization and economic growth is under developed. The empirical correlation between economic growth and fiscal decentralization was not very reliable due to several methodological issues. The results of theoretical analysis showed that much work remains to be done on preference differences among different levels of government, and to explore the direct influence of decentralization on economic growth. Meanwhile, a majority of the literature showed non-monotonic and unidirectional relationships between fiscal decentralization and economic performance. Whether, in developing and transitional economies either decentralization improved macroeconomic performance or not is still a question. Studies suggested that to examine the true sense of fiscal decentralization and its impact on economic performance theoretical and empirical caution must be used.

Jimoh (2003) explored the implementation of the strategies of fiscal federalism in Nigeria.

Results of the regression analysis showed that delegating power to local government, especially in the case of increased transfer of revenue to local tiers, will stimulate economic growth. This study also explored the extent to which decentralization of expenditure and revenue powers will affect the Nigerian economy. This effect was measured by some explanatory variables as the number of seats, number of local governments and the ratio of their expenditure, ratio of concentration of revenues, ratio of fiscal autonomy on Nigerian output and level of poverty prevalence in the state. The Linear functional form of the model

was evaluated by using Ordinary Least Square (OLS) technique and it was concluded that more decentralized government leads to accelerate growth in the Nigerian economy.

Meloche (2004) explored the importance of autonomy in fiscal matters while evaluating the effects of decentralization. They used two models for this analysis and took published data by OECD. The first model was about decentralization and economic growth used by Davoodi and Zou (1998) and the other about fiscal decentralization and public sector size incorporated by Oates (1985). The results regarding economic growth showed two directions. First, in

European transition countries the share of expenditure in sub national governments is not related to economic growth. Secondly, revenue decentralization has no positive relationship to economic growth. Empirical results about the second model also showed the same indications. For European transition countries there was no negative relation between public sector size and decentralization. There existed a negative link between public sector size and fiscal autonomy of sub national government. This relation vanished in a multiple regression analysis but showed that fiscal autonomy affected the size of public sector. This study concluded that the size of the public sector is affected by fiscal autonomy.

Thun (2004) examined the relationship between decentralization and industrial development in China. The author stated that decentralization improves the flexibility of policy making and it is often used to improve the economic policy. The study analyzed decentralization and its impact on auto industries in two states called Guangzhou and Shanghai where the auto sector served as a pillar industry. Analysis based on data from 1985 to 2000 showed that during the 1990’s in China the pressure on local governments was great in terms of balance creation between fiscal decentralization policies with fiscal responsibilities. As a

result, most localities were bound, to a great extent, to rely on their own resources to meet those responsibilities. Similarly in 1994 when the central government revised its allotment of revenue to provincial governments then the share of official budgetary revenue declined but expenditures were the same. Finally, the study suggested that an effective policy for heavy industry required strong states’ institutions. Moreover, the balance between fiscal revenue and fiscal expenditure is a very basic element of successful federalism.

Desai et al. (2005) examined the relationship between fiscal decentralization and regional growth in Russia. This study was based on fiscal data from 80 Russian regions for the period from 1996 to 1999. To analyze the data the study used the ordinary least square and two-stage least square estimation methods. Results showed that there should be certain limits of fiscal autonomy in resource dependent regions compared to resource rich regions. Results indicated that additional fiscal autonomy is justifiable for most Russian regions. However, to get control over expenditure policy, the restrictions by the central government on fiscal autonomy may be proven useful. So there is a need to introduce or formulate different polices for different regions, as in the case of bankrupt regions or fiscally depressed regions where the role of central government is to control their expenditure policy as necessary. Similarly, a higher degree of centralization is required in the case of wealthy regions. Their natural resource revenue should be used to finance public infrastructure projects in low income regions, instead of revenue distribution as budgetary transfer.

Feltenstein and Iwata (2005) explored the linkages between fiscal decentralization and economic performance in China. The study was based on macroeconomic time series data for the period of 1952 to 1996. For empirical analysis study used the Vector Autoregressive (VAR)

model with latent variables and analyzed the impact of fiscal decentralization on economic growth and the inflation rate in China. The results showed the positive association between economic decentralization and macroeconomic performance. Similarly real output growth also has a positive relationship with economic decentralization. While the study explored negative relationships (especially after the 1970s) between economic decentralization and the country’s inflation rate. So, the study suggested that economic decentralization seems to be good for economic growth while bad for price stability.

Arzaghi and Henderson (2005) stated a model and an empirical analysis of the factors which can promote the decentralization process and its positive impact on economy.

Decentralization evolved due to rising demands of hinterland regions for local autonomy, which have large populations and a respectable share in rising national income. The degree of decentralization can be measured by the share of the local government representation in the national government. This applies from the conception to devolution of power to local governments. To measure the degree of fiscal decentralization the study developed the federalism index of forty eight countries and decomposes the total period (from 1960 to 1995) in a five-year period and calculated the federalism index for each five years. For the empirical analysis the study used a Generalized Moving Method (GMM) and found that the degree of population concentration in large cities, land areas, per capita income and population have a large impact on the degree of federalism. Finally, the study suggested that national democratization leads to development of regional demands for greater fiscal autonomy and potential regional representation. Decentralization will be proven successful if the federal constitution encourages fiscal decentralization and federalism.

Jin et al. (2005) compared the relationship of fiscal decentralization to economic growth in China. The study was based on the panel dataset of 30 provinces in China and captured the correlation between fiscal decentralization and economic growth. This happened in two phases of fiscal decentralization where the fiscal contract system and tax assignment system were used. The first phase was concerned with the time period from 1979 to 1993 while the later phase considered the period from 1994 to 1999. To analyze the effects of fiscal decentralization on provincial economic growth during the two fiscal regimes the study used expenditure decentralization (provincial budgetary expenditure and provincial extra-budgetary expenditure as a share in total budgetary expenditure and total extra-budgetary expenditure) and revenue decentralization (provincial budgetary revenue as a share in total extra- budgetary revenue).

The study results showed that in both time periods, revenue and expenditures decentralization have diverged to benefit economic growth at the provincial level.

Iimi (2005) stated that the previous studies showed a mixed picture of relationships between fiscal decentralization and economic growth. Theoretically, it seemed perfect that fiscal decentralization leads to better provision of public goods, but this study proved it empirically by using the cross sectional data from 1997 to 2001. The results showed that fiscal decentralization has a positive significant impact on per-capita GDP growth. The study also highlighted the importance of economic reforms through decentralization of the public sector since the late 1990’s. It was concluded that fiscal decentralization will be a key challenge for developing countries and it will be important to reinvestigate the linkage between decentralization and economic growth in the twenty first century. The econometric analysis was based on 22 high income, ten lower middle income and seven low income countries, and found the positive significant relationship between per capita growth rate and fiscal

decentralization, which was measured by the local share of expenditure to total government expenditures.

Brueckner (2004) examined the relationship between fiscal federalism and economic growth by using endogenous growth models with overlapping generations. He explored that with the help of fiscal federalism, which can provide public goods and can be provided to different consumers according to their choice and these consumers may be young or old, and living in different localities. The analysis of this study is based on the earlier work of

Brueckner (1999) in which he used an overlapping generation model to explore that incentive to save as suggested by the federalism theory. In this work they replaced the common tax burden on provision of public goods with a head tax burden that varies between old and young customers. Thus, federalism affected the pattern of saving by affecting the income left after paying taxes. This work used a technique of Yakita (2003), where more investment is made in human capital when people are young so that their earnings may be enhanced when they get old. This analysis concluded that fiscal decentralization showed a positive impact for growth.

Guess (2005) examined the rising trends of adopting decentralization in Latin America and considered a sample of 17 countries for the period from 1985 to 1995. The study showed that countries which were suffering from crises of legitimacy were more engaged in decentralization. Moreover, in 1995 countries which had a unitary system, as Bolivia and

Colombia, were more decentralized than decentralized countries like Venezuela which showed that other factors must also explain decentralization. For empirical analysis the study used

Ordinary Least Square (OLS) with panel corrected standard errors. The results indicated that countries with powerful presidents have a higher degree of fiscal decentralization. While

legitimacy and democracy did not have a significant effect on fiscal decentralization, the results showed that federalism, structural adjustment, country size, level of development and presidential power exerted a positive impact on fiscal decentralization.

Stansel (2005) examined the link between economic growth and government structure. For this purpose he collected data from 314 Metropolitan areas of the US and the sample period examined was from 1960 to 1990. By using the Ordinary Least Square (OLS) method the results showed that fiscal decentralization encourages economic growth. The results also showed that in Los Angeles and New York metro areas from 1960 to 1990 each has grown by an extra 1.1 million dwellers and per capita income has also grown by extra $490 and $530 in

Los Angeles and New York respectively.

Martinez-Vazquez and McNab (2006) examined the direct and indirect relationships of fiscal decentralization with economic growth and macroeconomic stability. The theoretical analysis of the study was based on an augmented neoclassical model of economic growth and for the empirical analysis the study used a panel dataset of more than 180 countries (including

66 developing countries). The data ranged from 1972 to 2003 and used data from Government

Finance Statistics (GFS) of The International Monetary Fund (IMF). To measure the fiscal decentralization the study used the ratio of total sub-national government revenue to general government revenue and the ratio of total sub-national government expenditure to general government expenditure. To measure the macroeconomic stability the study used the

Consumer Price Index (CPI) and infant mortality was used as a proxy of evolution of human capital. The study found that decentralization promoted price stability in higher income countries while it diminished in developing economies. In the full sample of countries the

results highlighted that there exists no direct role of fiscal decentralization in promoting economic growth. While it directly reduced the growth rate in the sample of developed countries, this negative impact could be slightly offset due to the positive impact of fiscal decentralization on economic growth through inflation. However, in developing countries the sample results are not very clear, neither decentralization lower pace nor the higher pace of economic growth through inflation.

Zhang and Xiaobo (2006) stated that in China the fiscal system is in decentralization while the governance structure is strongly centralized. The study used time series data from 1993to

2000 which was based on the China County Public Finance Statistics Yearbook. The results showed that the transaction cost is a serious problem under fiscal decentralization. Similarly, a higher agriculture tax collection cost and an excessive government size make local government hardly able to provide necessary services and public goods. The outcome showed that firms in the rich regions enjoyed lower tax burdens and generous support while the farmers and firms in the poor regions paid heavy taxes and received less support. This shows that the fiscal decentralization is beneficial for rich localities while promoting the regional gap between poor and rich localities. Finally, the study suggested that the reduction of employees in those areas where agriculture is a major source of revenue is very essential, while fiscal decentralization alone is not sufficient to deal with distributional problems of services and public goods to fulfill the need of local residents.

Barankay and Lockwood (2007) analyzed fiscal decentralization and how it can enhance the capabilities of the government in terms of an increase in efficiency for the delivery of government services. The study was based on the time series data from Swiss cantons over the period from 1982 to 2000. For the empirical analysis, to analyze the level of fiscal

decentralization the study used spending on education within each canton. Instead of using total expenditure as an indicator of fiscal decentralization, the study considered only education expenditures to avoid aggregation bias (which appears only when we considered total expenditure as an indicator). The results showed a positive relationship between decentralization and educational attainment. This indicated that expenditure decentralization was also beneficial when central governments were not very competent. The study recommended that quality of data and correct model specification are essential tools to determine the effect of decentralization on the efficiency of public goods provision.

Facchini and Testa (2008) described the decentralization and regional inequality relationship in Brazil by using two-period models of public goods provision. On the basis of the two-period model study they determined the optimal level for the provision of public goods through borrowing by federal government. So this model developed an interaction between the central government and the state government. The study showed that the national public goods were provided to the state government by the federal government, while local public goods were provided by the state government to the local government. The state government will receive a share of revenue collected to finance the best provision of a local public good, and the local government could also borrow to finance the provision of public goods but this must be repaid in the following period. In cases of hard budget constraints, the state government will not be able to meet the best optimal level or provision of local public goods by the federal grants. However, the local government can maintain the provision of local public goods by the federal borrowing.

Thornton (2007) examined the relationship between fiscal decentralization and economic growth. The study examined the extent of independent taxing powers available to local

governments which were not incorporated clearly while determining the relationship between fiscal decentralization and growth. The empirical analysis based on 19 OECD member countries used cross sectional data from 1980 to 2000. By using an Ordinary Least Square

(OLS) method and considering real GDP per capita as a dependent variable the results showed insignificant relationship between economic growth and fiscal decentralization. The study highlighted the process of measuring decentralization by only considering sub-national governments’ expenditures and revenue shares not showing the actual degree of decentralization because it did not consider the autonomy in taxing powers available to subnational governments. So, sub-national governments, in spite of having full autonomy over revenue and expenditure shares (meaning a high degree of fiscal decentralization) fiscal decentralization has no significant impact on economic growth in OECD economies.

Kappeler and Valila (2008) analyzed the impact of fiscal decentralization on the determinant of productive public investment in Europe. For the empirical analysis study he used a panel of ten European Union countries from the period 1990 to 2005. Analysis excluded those countries whose public investment was significantly influenced by the receipt of EU support. The study used the General Method of Movement (GMM) to estimate the dynamic panel data and concluded that fiscal decentralization increased economic productivity of public investment. But, there is no significant relationship between public investment in consumption oriented local public goods or in the redistribution side. The study also concluded that decentralization is an effective tool to enhance the level of investment particularly in infrastructure as well as schools and hospitals which provide a public good. So the empirical examination showed that fiscal decentralization not only contributed to productive public

investment but it also raised investment in public spillover goods like schools, hospital and infrastructure.

Trillo and Rabling (2008) analyzed the weak effectiveness of federal transfers and poor absorption ability of local governments in poor areas. The study used the Municipal President’s

Nation Survey (MPNS) which contains information about 2429 municipalities in Mexico. By using the Generalized Moving Method (GMM), he examined the empirical evidence of decentralization of funds in the presence of weak institutions and legal framework. The results showed that the distribution of conditional transfers is mostly politically based and mostly the distribution of poverty reduction funds, which were made by community based development authorities, did not reach to the poor population. So the study highlighted the evidence that in the presence of elite capture local may not be beautiful. The study suggested that there was a need to explore the mechanisms through which local capture works.

Brueckner (2009) examined the rising trends of partial fiscal decentralization and explored the magnitude of the emergence of partial fiscal decentralization. In his study stated that economies mostly, instead of giving full autonomy to the local governments which are funded by central transfer, offer partial decentralization. This seems to be the better option, and yet existing literature provides no empirical and economic analysis of a partial decentralization in a Tiebout style model. The study highlighted that the provision of public goods depends on government officials and different levels of public good, which can be provided by different degrees of effort. Analysis also highlighted that the variety of a public good under full decentralization would be greater than under partial decentralization. The local officials may lean towards over spending under full decentralization. So, partial fiscal decentralization would

prove to be an appropriate move to overcome this problem. Yet while the economy has full decentralization capacity it may be recognized that partial decentralization could be the best option.

Frienkman and Plekhanov (2009) examined the behaviour of entire regions in the presence of fiscal decentralization. Panel data from 77 Russian states from the period 1996 to 2001 was considered for economic analysis. Results showed that fiscal decentralization can strengthen the sub-national governments’ abilities to improve the economic growth, promote reforms, enhance development outcomes and improve business climate at the local level. However, fiscal decentralization can distort the economic conditions of those territories that heavily drive a major share of revenue from natural resources, rents or federal grants. Results indicated that in entire regions fiscal decentralization or autonomy to local governments can produce inferior development or worsen the economic conditions. The study suggested that to improve the decentralization process, it should be implemented gradually and incorporated with the monitoring of delivery of services and utilization of funds at municipal levels. So the implementation of advance intraregional fiscal decentralization in an entire region can support the local governments to improve development and reduce the policy distortion.

Herath (2009) compared the pre and post fiscal decentralization process in Sri Lanka. The study examined the relationship of fiscal decentralization with poverty and per-capita income.

The results showed that the real degree of fiscal decentralization to promote per-capita income is not significantly conducive in Sri Lanka. In lower tiers of government the public official were not very competent to get the benefits of fiscal decentralization. Moreover, in developing countries there existed a huge imbalance in regional development, which hid the benefits of

fiscal decentralization. The study suggested that for a successful execution of fiscal decentralization there is a need to redefine the functions of national and sub-national governments. Mostly the overlapping in national and sub-national government’s functions limited the effectiveness of fiscal decentralization in an economy.

Jutting (2009) examined the relationship between fiscal decentralization and health outcomes in China. The study was based on a panel data set from 1995 to 2001 with nationwide country level fiscal data. The study used two indicators, namely the local share of total fiscal expenditure to capture the local expenditure responsibility in the public sector and the degree of local fiscal autonomy of local government. The study considered the share of local expenditure covered by local government’s own revenue. For the empirical analysis the study used a fixed-effect model and showed that fiscal decentralization is a powerful tool to enhance the provision and quality of public goods such as healthcare. Results showed that fiscal decentralization can help to reduce the infant mortality rate in China. Indeed, it is very crucial result for achieving pro-poor decentralization in the health sector and the better provision of public goods in terms of good healthcare facilities.

Gemmell et al. (2009) explored whether any efficiency gains obtained through decentralization have a positive effect in decentralized economies. For this purpose they collected data from 23 OECD countries for the period of 1972 to 2005. To analyze the panel dataset they applied more extensive and flexible econometric techniques i.e. the technique of instrumental variables and a pooled mean group. In this study they constructed three measures for revenue decentralization and two measures for expenditure decentralization which were based on the study of Stegarescu’s et al. (2005). While for econometric analysis they followed

approaches that were used by Xie, et al. (1999) and Davoodi and Zou (1998). The study concluded that expenditure decentralization has an adverse effect on economic growth of

OECD countries but revenue decentralization leads to enhance economic growth. But the

OECD countries are practicing more spending decentralization than revenue decentralization.

Therefore, this study argued that in order to take the efficiency gains of decentralization there must be a closer match between spending and revenue decentralization in these countries.

Neyapti (2010) investigated the macroeconomic effects of fiscal decentralization and used a panel of sixteen countries from the period 1980 to 1998. To present sub-national government activity the study used the sum of provincial and local government. Revenue decentralization was measured as a share of total government revenue with provincial and local government revenue. Expenditure decentralization was also measured in the same manner. To examine the relationship of fiscal decentralization (revenue and expenditure) with fiscal deficit the study used Ordinary Least Squares (OLS) with robust standard errors. The results found the significant negative relationship between fiscal decentralization and budget deficit although it may have been largely influenced by and dependent upon institutional features and characteristics of a country. This study also highlighted the effectiveness of fiscal decentralization in reducing budget deficit in case of a large population although population size and budget deficit typically have a positive association.

Baskaran (2011) analyzed the process of fiscal decentralization and its impact on the public sector size. The study explained the relationship on theoretical grounds and also verified this relationship on empirical grounds. A panel dataset from 18 OECD countries over the period of

1980 to 2000 was used. In an empirical analysis study he explored the correlation between the

public sector size and fiscal decentralization and determined that government ideology has a major influence on both aspects. However, the empirical results were not fully matched with theoretical hypothesis due to some shortcomings in the former analysis. The study suggested that there is a need to develop the strong assumptions about fiscal constitutions and a generalized model should be adopted to understand the relationship between fiscal decentralization and the public sector size. The study indicated that the ideology of the central government for the effect of fiscal decentralization on the public sector size and decentralization leads to a smaller increase in the size of the public sector.

Kyriacou and Roca-Sagalés (2011) examined the impact of fiscal decentralization and government quality in 27 OECD economies over the period from 1996 to 2005. The study used sub-national expenditure and revenue as two indicators of fiscal decentralization. Sub-national government expenditure was taken as a proportion of consolidated general government spending and adjusted by subtracting from state and local expenditure and transfer paid to national government. Similarly sub-national revenue was calculated by subtracting the local revenues from state revenue and grants received from the state government. The regression results of the study showed that the fiscal decentralization has statistical significance and a positive relationship with the quality of the government. In the presence of regional elections, the regression results showed the positive contribution of fiscal decentralization to the performance of the government which may vanish under bicameralism and federalism.

2.5 Conclusion

The literature showed that most of the work on decentralization is accomplished to examine the relationship of fiscal decentralization and economic performance or growth rate of

developed economies. The results of these studies do not present a very clear relationship between decentralization and economic growth as Davoodi and Zou (1998), Woller and

Phillips (1998), Zhang and Zou (1998), and Xie et al. (1999) found an inverse association between fiscal decentralization and economic growth. While other studies like Akia and Sakata

(2002), Iimi (2005), Lin and Liu (2000), Stansel (2005) and Yilmaz (1999) found a positive relationship between fiscal decentralization and economic growth. More importantly, number of studies performed in Pakistan were limited only to investigate the direct impact of fiscal decentralization on economic growth (Malik et al. 2006; Khattak et al. 2010 and Faridi 2011).

So, this study will be an additional opinion of various aspects of decentralization and could be an effort to clear the association of decentralization with the macroeconomic performance in a developing economy like Pakistan. The review also made it clear that very little work has been done in developing and transitional economies to explore the connection between fiscal decentralization and macroeconomic performance. This association demanded more attention and research work to provide clear guidelines for policy makers to design and propose successful execution of fiscal decentralization in developing economies. Moreover, present work will be a continuity of fiscal decentralization literature on developing economies and also highlight the important macroeconomic indicators which may positively contribute to enhance the macroeconomic performance in Pakistan. After a comprehensive review of the literature on fiscal decentralization, the next chapter will shed light on basic concepts and important theories of fiscal decentralization.

Table 2.1: Summary of Macroeconomic Variables and Their Relationships with Fiscal Decentralization in Existing Empirical Studies

Main Findings Author’s Period of Econometric Dependent Countries Fiscal decentralization Fiscal Decentralization Other Supporting Name Study Model Variables revenue Expenditure Variables Prudhomme Theoretical ------Negative Relationship Negative Relationship (1995) Base Smoke Theoretical A general discussion on fiscal decentralization process in Indonesia and other developing -- -- (1996) Base economies. Positive Relationship Positive Relationship Heng-fu- Panel of 46 Panel Economic with Developed with Developed 1970-1989 Zou (1998) countries Regression Growth Negative with Negative with Developing Developing Positive Least Square Zhang Economic Relationship 1978-1992 China dummy Negative Relationship Negative Relationship (1998) Growth with Labour Variable Growth Davoodi Economic 1948-1994 USA OLS Negative Relationship Negative Relationship (1999) Growth Inter Demello, Panel of 30 Panel government Negative Relationship in Negative Relationship in 1970-1995 J.R. (2000) countries Regression Fiscal developing countries developing countries Relationship Positive relationship with Jin and Zou Panel Economic 1979-1999 China Positive Relationship Positive Relationship Investment, (2000) Regression Growth Openness and Labour growth Ozcan Panel Local 1998-1999 Turkey Positive Relationship Positive Relationship (2000) Regression Development Akia and Panel of 50 Economic Sakata 1988-1992 OLS Positive Relationship Positive Relationship US states Growth (2002) Panel of 32 Generalized Government Zhuo (2002) 1987-1997 Positive Relationship Positive Relationship countries Least Square Size

Table 2.1: Summary of Macroeconomic Variables and Their Relationships with Fiscal Decentralization in Existing Empirical Studies

Main Findings Author’s Period of Econometric Dependent Fiscal Fiscal Countries Name Study Model Variables decentralization Decentralization Other Supporting Variables revenue Expenditure McNab Theoretical Economic Positive Positive -- -- (2003) Base Growth Relationship Relationship Jimoh Regression Economic Positive Positive Nigeria (2003) Analysis Growth Relationship Relationship Meloche, No Significant No Significant Regression Economic Negative Relationship with et al. OECD Positive Positive Analysis Growth Public Sector Size (2004) Relationship Relationship Thun Regression Industrial Positive Positive 1985-2000 China (2004) Analysis Development Relationship Relationship OLS and Two Regional Desai Positive Positive 1996-1999 Russia stage least Economic (2005) Relationship Relationship Square Growth Feltenstein Economic Positive Positive Inflation have Negative 1952-1996 China VAR (2005) performance Relationship Relationship relation with FD Henderson Panel of 48 Economic Positive Positive 1960-1995 GMM (2005) Countries Growth Relationship Relationship Panel of 30 1979-1993 Panel Economic Positive Positive Jin (2005) provinces 1994-1999 Regression Growth Relationship Relationship from China Limi Panel of 32 Panel IV Per Capita Positive Positive 1997-2001 (2005) countries Regression Growth Relationship Relationship Brueckner Theoretical Economic Positive Positive -- -- (2005) Study Growth Relationship Relationship

Table 2.1: Summary of Macroeconomic Variables and Their Relationships with Fiscal Decentralization in Existing Empirical Studies Main Findings Author’s Name Period of Econometric Dependent Fiscal Fiscal Countries Other Supporting Study Model Variables decentralization Decentralization Variables revenue Expenditure Country OLS with Panel 1985- Panel of 17 Size, Guess (2005) Corrected Positive Relationship Positive Relationship 1995 countries Presidential Standard Error Power 314 1960- Economic Stansel (2005) Metropolitan OLS Positive Relationship Positive Relationship 1990 Growth Areas USA Martinez- 1972- Panel of 180 Panel IV Economic No direct No direct FD promote Price Vazquez and 2003 countries Regression Growth Relationship Relationship Stability McNab (2006) 1993- Cost of Zhang (2006) China OLS Positive Relationship Positive Relationship 2000 Transaction Barankay 1982- Swiss Panel Government Positive Relationship Positive Relationship (2007) 2000 cantons Regression Efficiency 1995- Two period Public Goods Testa (2007) Brazil Positive Relationship Positive Relationship 2001 Model Provision Thornton 1980- GDP per No Significant No Significant OECD OLS (2007) 2000 capita Positive Relationship Positive Relationship Kappelor 1995- Panel of 10 Public GMM Positive Relationship Positive Relationship (2008) 2005 EU Countries Investment Mexico Weak No significant No significant Trillo (2008 2002 GMM Survey based Institutions Positive Relationship Positive Relationship Brueckner Theoretical Provision of Partial FD Positive Partial FD Positive -- -- (2009) Based Public Goods Relationship Relationship Panel of 77 Frienkman 1996- Panel Economic Russian Positive Relationship Positive Relationship (2009) 2001 Regression Growth States Per-capita No significant No significant Herath (2009) Siri Lanka OLS Income Positive Relationship Positive Relationship

Table 2.1: Summary of Macroeconomic Variables and Their Relationships with Fiscal Decentralization in Existing Empirical Studies

Main Findings Period Author’s Name Econometric Dependent Fiscal Fiscal of Countries Other Supporting Model Variables decentralization Decentralization Study Variables revenue Expenditure 1995- Fixes Effect Jutting (2009) China Health care Positive Relationship Positive Relationship 2991 Model Panel of 23 1972- Pooled Mean Economic Negative Negative Gemmell (2009) OECD 2005 Group Growth Relationship Relationship countries Significant Significant 1980- Panel of 16 Fiscal +ive relation with Neyapti (2010) OLS Negative Negative 1998 Countries Deficit Large Population Relationship Relationship Not significant Not significant 1980- Panel of 18 Fixed Effect Public Baskaran(2011) Positive Positive 2000 OECD Regression Sector Size Relationship Relationship 1996- Panel of 27 Govt. Positive Positive Kyriacon (2011) Regression 2005 OECD Quality Relationship Relationship 1971- Time Series ARDL Economic Positive Positive Malik et al. (2006) 2005 Pakistan Cointegration Growth Relationship Relationship Khattak et al. 1974- Time Series Johnson Per Capita Negative +ive relation with -- (2010) 2007 Pakistan Cointegration GDP Relationship Trade Openness 1972- Time Series Auto Economic Positive Positive -ive relation with Faridi (2011) 2009 Pakistan Regressive Growth Relationship Relationship Inflation 1972- Time Series Negative Positive -ive relation with Faridi et al. (2012) OLS Employment 2009 Pakistan Relationship Relationship Inflation 1972- Time Series Economic Positive Negative +ive relation with Iqbal et al. (2013) GMM 2010 Pakistan Growth Relationship Relationship Tax to GDP ratio

Chapter 3

Fiscal Decentralization and Profile of Pakistan’s Economy

3.1 Introduction

Pakistan emerged as an independent country on the world map in August 1947 consisting of two parts, which were known as Western Pakistan and Eastern Pakistan. Due to political and economic reasons the eastern part separated (which at present is known as Bangladesh) from the western part in 1971. After the partition of the western part, which was known as the

Islamic Republic of Pakistan, it became a centralized country consisting of four provinces and a federal area (the so-called central administration area).

Although there were several social, political and economic reasons which caused the separation of Eastern Pakistan, most economic experts consider that in a different scenario with less inequality in the distribution of resources between the two territories, Eastern

Pakistan would had never split off. However, that was just one of the many reasons for the division. Actually, though most of the production industry was placed in Eastern Pakistan,

Western Pakistan was getting the lion share of the revenues while Eastern Pakistan was getting much less of its expected share. That fact is considered to be one of the main reasons for

Bengalis to having felt uncomfortable and having decided to fight for their separation from

Western Pakistan.

After losing the eastern part a commission was established to make sure the fair distribution of resources among the rest of provinces with Western Pakistan. The National Finance

Commission (NFC) is the organization that seeks/ implements the Fiscal decentralization as a

best resource distributor. Its main aim is the distribution of annual revenue between the federation, as well as distribution among the provinces.

Nowadays, fiscal decentralization is considered a significant tool to raise the economic performance in developing countries. Pakistan is a developing country and currently is confronting various socio-economic problems such as regional discrepancy, illiteracy, overpopulation and unequal distribution of wealth, among others. We can assess the evolution of Pakistan's economic performance by comparing some basic macroeconomic indicators since its independence. Today there are many indicators to measure a country's economic performance and the shooting speed of development. Some economists believe that the GDP is the best indicator to measure economic performance of a country while others believe structural changes in a society may capture the true picture of its economic performance. In this work, the concern is not just to capture the exact measurement of the economic performance but also there is a need to explore the linkages between economic performance and fiscal decentralization. Because the aim is to study how it may foster economic performance, the study will analyze the process of fiscal decentralization in Pakistan. In the present chapter our focus/ objective is to provide a brief summary on the process of fiscal decentralization in Pakistan and on the performance of its macroeconomic variables. The composition of present chapter can be described as follows.

The first section will describe a brief history of resources distribution in Pakistan. In second section the economic performance of major economic sectors will be described. Third section will explain the performance of key macroeconomic indicators in Pakistan and some concluding remarks will be presented at the end.

3.2 History of the Fiscal Autonomy Campaign in Pakistan

Decentralization is the course through which tasks and assets from national to sub-national governments are transferred (Rondinelli and Dennis, 1981). Thus, through decentralization, the central government delegates powers to sub-national governments, to help create a better asset utilization, higher values of public life and at the same time to divide the burden of effort

(Gordin, 2004). Pakistan is a state with a strong central government. Currently there are three performance levels of the Pakistani government: the federal, the provincial and the local

(district) level governments.

Due to competition and distributional aspects, the allocation method has remained under discussion. According to Jaffery and Sadaqat (2006), organized resource transfers take place in four stages. In the first stage, the National Finance Commission (NFC) allocates resources between the federal and the provincial governments. During the second stage, the Provincial

Finance Commission (PFC) forwards these resources to the local provincial levels. In the third stage, transfers are made from the federal to the local jurisdictions and, finally, in the fourth stage, the distribution of vertical resources locally produced, i.e. to the District Governments and Tehsil Municipal Administrations are accomplished. While in contrast, accidental transfers such as grants, and special optional funds for executives, parliamentary development funds were also decided in this step.

Study in the next sections will summarize all awards presented over the years after independence.

3.3 Allocation of Resources in Pakistan

The cornerstone of the Pakistani economy stands on its federal structure and it further divides into provincial and local governments. The federal structure depends on the collection of taxes which afterwards channel to the federating units. This distribution is carried out on the basis of income generation, population, degree of development and the location of natural resources. The federal government is the only authority liable to collect the revenue and distribute it to the federal units. The population is the only stakeholder in the process that generates the revenue.

Speaking about history of allocation of resource disbursement there are four different time periods which can be explained as follows:

i. Resource Allocation in the region before independence

ii. Resource allocation after independence

iii. Resource distribution in the period of one unit

iv. Resource distribution in term of award according to 1973 Constitution of Pakistan.

3.4 Resource Allocation before the Independence of Pakistan in 1947.

The Niemeyer Award under the government of India Act, 1935, was announced for the allocation of resources between federal and provincial governments of British India. Under this award 50% of the total revenue collection was advised to be distributed among provincial and local governments. The government of Pakistan after its inception followed the same award until March1952. The new empire had no funds nor grants to disburse but when the funds came to Sindh initially were used to settle federal debt. In addition, the annual grant of

Rs. 10.5 Million to Sindh and a grant of Rs. 10 Million for NWFP were disbursed. The same planning was continued until March 1952.

3.5 Resource Allocation after the Independence of Pakistan

After the independence of Pakistan, Sir Jeremy Raisman made the first disbursement for the feasible revenue sharing between the federation and the federal units of the country. The formula of revenue sharing was presented after December 1947 and afterward adopted on April

1, 1952 (Pakistan, 1991 and Jaffery and Sadaqat, 2006).

The Raisman award was particularly planned to face the poor financial situation of the federal government and 50% of the sales tax collection was distributed to the federal government. According to the formula to fight against fiscal crises taking place at the time of the partition, the federal government was given 50% of the ad-hoc sales tax. From the national divisible pool (comprised of proceeds of 50% income Tax) Eastern Pakistan received 45% of the share while western Pakistan received the remainder. According to the formula of the

Raisman Award the revenue share was distributed as follows: The Punjab province was granted

27% Sindh 12%, NWFP8%, Bahawalpur 4%, Khairpur 0.6%, and finally, residual and

Baluchistan got 2.8% and 0.6 %, respectively (Pakistan, 1991). Moreover the NWFP was supported with an additional amount of rupees 12.5 million (Ahmad et al. 1997 and lftikhar et al. 2007).

3.6 Resource Allocation in the Period of One Unit

In the earlier period, the four provinces (Punjab, Sindh, Balochistan and NWFP) were considered as separate identities while in the period of 1955 for the first time they were

considered as a single unit of Western Pakistan. Later in 1961 and 1964, two awards were announced by considering two identities called East and West Pakistan. Therefore, in 1955 arrangements were first made to allocate resources between only two units i.e. East and West

Pakistan. In 1970 a finance committee was organized to allocate fund between two units. The detail of resources distribution can be explained as.

3.6.1 The Prize of 1961

According to the 1961 Prize from the divisible pool, consisting of 70% share of sales and other taxes, Western Pakistan and Eastern Pakistan each had shares of 46 and 54% respectively, (Ahmad et al. 1997 and Iftikhar et al. 2007). The remaining 30% of the divisible pool was allocated among the provinces on the basis of collection of sales and other taxes from their own areas. Similarly the other taxes on agricultural land, the tax on the value of capital and property taxes were distributed among the provinces as per their collection.

3.6.2 National Finance Commission Award of 1964

In 1964 the National Finance Commission was established according to the Article 144 of the Constitution of 1962. Under this award, 30% of the sales tax from the divisible pool was distributed among the provinces and the distribution was based on the collection of revenues

(which consisted of the collection of excise duties, export taxes, sales tax and income tax) from each province. The total of the central and the provincial shares of the divisible pool were 35 and 65% respectively (Ahmad et al. 1997 and Iftikhar et al. 2007). While the distribution ration of Eastern and Western Pakistan remained unchanged, their respective shares were 54% and

46% of the divisible pool. Western Pakistan was split up into four provinces, Punjab, Sindh,

Balochistan, and NWFP on July 1, 1970. Its total share of the divisible pool was 46% which

was disaggregated into four provinces as; Punjab got 56.5% of the total divisible pool of

Western Pakistan (Pakistan, 1991) while Sindh, Blochistan and NWFP got a share of 23.5%,

15.5 and 4.5% respectively from.

3.6.3 National Finance Committee 1970

In Apri1970 a committee rather than a commission was established by the finance minister to recommend the intergovernmental resource allocation with mutual consent of the provinces.

The shares of the federal and the provincial government in the divisible pool were considered

20% and 80% respectively. There was no change in the divisible pool, even the share of each unit remained unchanged. Western Pakistan received 46 percent of the total divisible pool while East Pakistan received a 54 percent share. The allocation or share of each province according to the 1970 award can be observed on table 3.1.

Table 3.1: Participation in Provincial Award 1970 Punjab Sindh NWFP Baluchistan

56.50% 23.50% 15.50% 4.50%

Source: Pakistan (1991)

After the separation of Western Pakistan from the Eastern part, Western Pakistan’s provinces (Punjab, Sindh, NWFP and Balochistan) were awarded the same proportion in the total divisible pool of West Pakistan. However, due to the separation the size of the total divisible pool for Western Pakistan changed completely.

3.7 Resource Allocation in the Financial Arrangement of the 1973

Constitution

Immediately after the separation of Eastern Pakistan from Western Pakistan, it was inevitable for Western Pakistan to compose the National Finance Commission every five years.

All these arrangements were implemented under the constitution of 1973 and the finance commission was chosen to evaluate and advise the resource allocation in Pakistan. The main objective of this planning was to make the resource distribution method acceptable and satisfactory for each province. The next section will briefly summarize the events subsequent the 1973 constitution (Ahmad et al. 1997 and Iftikhar et al. 2007).

3.7.1 The First NFC Award 1974

The first NFC award 1974 allocation of resources was made by adopting the population and taxes as a distribution criterion. Income tax, sales tax, and export duties were considered in the distribution pool. The divisible pool distributed between the federal and the provincial governments with the ratio of 20% and 80% respectively. According to the first NFC award the Punjab share was improved from 56.50 % (1970 award) to 60.25% due to its population as it was the only measure for resource sharing. However, the other three provinces suffered and the Sindh suffered the most (Ahmad et al. 1997 and Iftikhar et al. 2007). The allocation of resources according to the NFC Award 1974 among the provinces is given in table 3.2.

3.2: Provincial Percentage According To the 1974 NFC Award Punjab Sindh NWFP Baluchistan

60.25% 22.50% 13.39% 3.86%

Source: Pakistan (1991)

According to the NFC award 1974, in order to offset the weak financial condition it was decided to allocate Rs.50 million along with Rs.100 million annual grants to Baluchistan and

NWFP respectively.

3.7.2 The 2nd NFC Award 1979

In February 1979, under the chairmanship of Federal Minister Mr. Ghulam Ishaq, the second NFC Award was constituted by President General Zia-ul-Haq without any plan. Thus, the allocation of resources was the same as the one proposed by the award of 1974.

Notwithstanding, after the census in 1981, when there was evidence that the proportion of population had changed, the resources were also reallocated. Thanks to following the population criterion, the situation in Sindh and Baluchistan improved while the participation of NWFP remained unchanged (Ahmad et al. 1997 and Iftikhar et al. 2007). Table 3.3 shows the distribution of resources according to 1979 Award.

Table: 3.3 Provincial Share Table, 1979 Award Punjab Sindh NWFP Baluchistan

57.97% 23.34% 13.39% 5.30%

Source: Pakistan (2006)

3.7.3 The 3rd NFC Award 1985

During Zia-ul-Haq’s regime in 1985 the 3rd NFC was constituted under the chairmanship of Federal Minister of Finance Dr. Mahbub-ul-Haq. At that time Pakistan was suffering from serious political problems due to external and central political instability. Even after nine sessions, the commission was unable to finalize its recommendations. The allocation of resources from the divisible pool remained the same from 1974 to 1990 (Ahmad et al. 1997

and Iftikhar et al. 2007). So in the distribution process this NFC award was unable to offer any improvement. Once again the allocation of resources continued by following the distribution principles (with modified provincial population) of the 1974 NFC award.

3.7.4 The 4th NFC Award 1990

In April 1990, with few positive suggestions of the democratic government of Prime

Minister Nawaz Sharif, the 4th NFC Award was proposed after 16 years of interruption. The enlargement of the divisible pool under this award was an important progression. For the very first time it was decided to include excise duties in the divisible pool. In the 4th NFC award the divisible pool had a number of duties and taxes like excise duty, duty on tobacco and sugar, export duties, sales tax and income tax. According to the 4th NFC award in the 1990s customs duties were not included in the divisible pool and remained with the central government.

Although all provinces demanded some change in the distribution formula, in spite of that the committee could not reach a consensus for adopting any changes and the same formula (based on a population criterion) for distribution of resources was maintained. Once again the federal and provincial governments adopted the same population based criterion of revenue distribution from the divisible pool. The federal and provincial governments continued to share the divisible pool at the rate of 20 and 80 percent respectively (Ghaus et al. 1994). The distribution of the divisible pool can be observed in table 3.4.

3.4 Provincial Share Table, 1990 Award Punjab Sindh NWFP Baluchistan

57.88 % 23.28% 13.54% 5.30%

Source: Pakistan (1991)

There was no census since 1981, and the population remained the only criterion for distribution of the divisible pool, so the ratio of the horizontal distribution remains the same.

However the transfers of funds from federal and regional levels were improved given the larger size of the divisible pool. Federal tax revenues increased from 28 to 45 percent and as a result the horizontal funds share of the provinces improved by 17% as compared to the 1974 award

(Ghaus et al. 1994). Thanks to the addition of the excise duty on sugar and tobacco the financial share of the provinces increased and they got more fiscal autonomy. Moreover, in terms of straight transfers, the provinces were given a right to excise duties on crude oil, development surcharge on gas and net hydel profit. Therefore, this award was a positive step towards fiscal decentralization. More fiscal autonomy was given to the provinces in terms of finance but motivation of generating their own revenue was still a question mark. Extra funding to the provinces that was provided to meet the need of development is shown in table 3.5.

Table 3.5: Special Annual Grants to Provinces in 4th NFC, 1990 (Rs. Million) Amount / Years Punjab Sindh NWFP Baluchistan

Amount 1000 700 200 100

Next Years 3 5 3 3

Source: Pakistan (1991)

3.7.5 The 5th NFC award 1996

The caretaker Prime Minister Mr. Malik Meraj Khalid constituted the 5th NFC award in

December 1996.The Finance Minister and Chairman of the commission, Mr Shahid Javed

Burki, announced it in February, 1997. According to the award all taxes such as income tax, wealth tax, sales tax, capital value tax, export duties, excise duties (excluding excise duty on gas charged at the wellhead), custom duties and any other collected by the Federal Government

were included in the divisible pool. Furthermore, all provinces were given royalties on net development surcharges on natural gas, royalties on crude oil and matching grants were introduced if they exceeded their revenue growth target of 14.2% (Pakistan, 1996). Table 3.6 shows the maximum limits of matching grants given to each province (Ahmad et al. 1997 and

Iftikhar et al. 2007).

Table 3.6: Matching Grants under the 1996 Award (Rs Million) Punjab Sindh NWFP Baluchistan 500 500 100 100 Source: Pakistan (1996)

According to the NFC 1996 award overhead expenses were classified into non-priority and priority expenses. Defense, debt service, social sector and development expenditures were considered as priority expenditures while law and order, community services and general administration expenditures were described as non-priority expenditures (Sabir and Khan,

2001). This bifurcation of expenditure was made to cope with emerging challenges and other financial issues on a priority basis and to improve the development level. To explain the divisible pool, in this award there were drastic changes in the shares to the provinces. Indeed, the ratio of federal and provincial shares in the 1985 NFC award changed significantly to

62.5% instead of 20% for the federal government and 37.5% instead of 80% of the provincial level of governments. The allocation of resources from the divisible pool was revised on the basis of the population census of 1998 and decided to be implemented as of July 1, 2002. The drastic changes in provincial shares were based on optimistic projections of desirable inflation and a higher GDP growth rate. Unfortunately, when the award was exercised the economy did not respond accordingly and due to internal and external shocks all the provinces were

adversely affected during that time period. At that time it was realized that the implementation of the previous award of 1990 will be beneficial in improving the financial position of the provinces (Sabir, 2001). Without any change, population still being the sole criterion of resource allocation, the percentage of the provincial share remained the same as with the minimum 5.3 percent and maximum 57.88 percent of Baluchistan and Punjab respectively.

The percentage share of each province is shown in table 3.7.

Table 3.7: Percentage Provincial Share, 1996 Award Punjab Sindh NWFP Baluchistan

57.88% 23.28% 13.54% 5.30%

Source: Pakistan (1996)

To deal with the poor situation it was decided to give special grants of rupees 4 billion and

3.3 billion to the Baluchistan and NWFP. These funds were granted on a five year basis to cope with the rising inflation rate. The special annual grants are presented in the table 3.8 as follows.

Table 3.8: Special Annual grants to Provinces (Rs Billion) Amount / Years Punjab Sindh NWFP Baluchistan

Amount - - 3.31* 4.08*

Next Years - - 5 5

Source: Pakistan (1996)

3.7.6 The 6th NFC Award 2000

The previous distribution awards of 1991 and 1997 were considered as respectable achievements of resource allocation among the provinces although their execution stages faced many problems. Under the chairmanship of Federal Finance Minister Mr. Shaukat Aziz, the

6th NFC award was constituted by President of Pakistan General Pervaiz Musharraf on July

22nd 2000. Even after 11 meetings this award recommendation could not be finalized because of a lack of consensus among its participants. In this award the major focus of the provinces was a greater share of resource allocation in the divisible pool, and a change in the distribution formula was also demanded. The provinces were demanding a 50% share while the central government was demanding a 45% share from the divisible pool (Ahmad et al. 1997 and

Iftikhar et al. 2007).

3.7.7 The 7th NFC Award 2006

After the fruitless ending of the 6th award, the 7th NFC award was formed on July 1st

2005. Unfortunately the results were similar and the commission faced difficulties due to lack of mutual acceptance and absence of consensus among the members through which an amicable resource allocation mechanism could be implemented. Finally according to the

Article No. 160 (6) of the 1973 constitution of the Islamic democracy of Pakistan the entire

CM's of the provinces gave the authority to the president to announce the formula for distribution of fiscal resources. This authority to the president stemmed from a new amendment in the “Distribution of Revenues and Grants-Aid Order, 1997 by the president under Article

160(6) of the constitution of the Islamic Republic of Pakistan, through ordinance No. 1 of

2006”.

The new NFC award based on the proposed formula of the president was declared to take effect from July 1, 2006 (Pakistan, 2006a). According to the proposed formula, for the first financial year the improved share of provinces was decided up to 45% (Grants plus a Share in the total divisible Pool) and would reach up to 50% with an annual increment of 1% per annum.

Like previous awards the divisible pool is comprised of these items; capital value tax, income tax, wealth tax, sales tax, custom duties, export duties, excise duties (excluding excise duty on gas, charged at the well head) and any other tax collected by the federal government (Pakistan,

2006a). At that time due to higher economic growth the federal government decided to allocate enough resources to the provinces. Instead of Rs.418 billion in financial year 2006-07, the total amount of Rs.497 billion was projected to be given in financial year 2007-08. This allocation of funds consisted of 46% from the divisible pool that was 1% more than the previous year’s allocation. This was a respectable enlargement in revenue transfer and fiscal autonomy to the provinces (Federal Budget Speech, 2007-08). Moreover, the provinces were granted net proceeds equal to 1/6th of the sales tax which was further transferred to the cantonment boards and at district level governments. The distributed shares to the provinces were as; Balochistan

5.22%, NWFP 9.93%, Shindh 34.85% and Punjab50.00% (Pakistan, 2006a). According to the

2006 award the Provincial share from the divisible pool is shown in table 3.9.

Table 3.9: Percentage Provincial Share, 2006 Award Punjab Sindh NWFP Baluchistan

57.36% 23.71% 13.82% 5.11%

Source: Pakistan (1996)

Due to the higher annual growth rate the total grants for the provinces were raised from Rs.

8.7 billion to Rs. 27.75 billion. In this award it was decided to give a Rs. 3.05 billion grant to

Punjab province and Sindh was endowed with Rs. 5.83 billion. In terms of grants they did not receive any grants in the previous awards. The estimated net addition to the resource transfer in term of grants and the share from central governments to provincial governments were

almost Rs. 51 billion (Pakistan, 2006b). Table 3.10 shows the amount of special grants to each province.

Table 3.10: Special Grants to Provinces under the 2006 Award (Rs. Billion) Name of Provinces 2005-06 (BE**) 2006-07 Estimates)

Punjab - 3.05

Sindh - 5.83

NWFP 3.9* 9.71

Baluchistan 4.8* 9.16

Total 8.7 27.25

Source: Pakistan (1996b)

A rise in direct transfers of royalties on crude oil, gas, gas development surcharge and excise duty on gas was another significant advancement considered in that award. Additionally at a capped level Rs. 6 billion was given to the NWFP government on a yearly basis from

WAPDA’s net hydel profit (Pakistan, 2006b). Over time there was a respectable annual addition in the grants and Sindh was the greatest beneficiary through the straight transfers.

According to this NFC award all the provinces were endowed with greater autonomy in monetary terms. Through that fiscal autonomy it was expected all provinces would have greater opportunities to generate their own revenue and improved economic development at the local level. The federal government under the Public Sector Development Program (PSDP) financed the assorted development plans in all provinces such as straight transfers, with full funds and with 50:50 shares. Moreover to handle the victims of uncontrolled situations and natural calamity at the provincial level, non-development funds were also transferred to the provinces.

3.7.8 The 8th NFC Award 2009

With the passage of time different awards were introduced and until the 6th NFC award announced by Pervaiz Musharraf provinces were forbidden to accept grants by the government and the provinces were unable to build a consensus over different issues. For achieving harmonious economic development and prosperity in all regions the system of NFC awards has played a significant role in promoting and allocating the resources among provinces.

Although there were some shortcomings in previous awards and various reasons for dissatisfaction of systems of collection and distribution among provinces, we can classify them as follows:

i. Population was the only criterion for resource distribution.

ii. Inappropriate technical skills in the whole process of NFC award. iii. Instability in continuation in the process of NFC award.

On December, 11 2009 Shukat Tareen addressed the sources of disagreement amongst provinces, and after building a consensus among them, the 8th NFC award was announced.

This award was signed at Gwadar in Sindh, “all the states have shown nobility and flexibility to provide accommodation to one another” (Finance Minister Shaukat Tareen's Press

Conference after the declaration of NFC, December 11, 2009). On demand of smaller provinces (according to population size), since 1980's for horizontal distribution process instead of the population criterion a new formula was introduced for resources distribution. On the basis of new criteria, Punjab and all other provinces shared funds from the federal tax pool.

Due to a fear of a sharp cut in their share from the federal transfers, Punjab was against the

demand of small provinces of a change in horizontal distribution criterion. The selected criteria for resource allocation in 8th NFC award can be stated as follows.

i. 82 percent weight was given to population.

ii. 10.3 percent weight was considered for backwardness and poverty.

iii. 5 percent weight was considered for revenue collection and generation.

iv. Finally 2.7 percent weight was given to Inverse population density.

Provincial shares are shown in the table 3.11.

Table 3.11: Shares of Provinces According to 8th NFC Award

Punjab Sindh NWFP Baluchistan

51.74% 24.55% 14.62% 9.09%

Source: Pakistan (1996)

In a nutshell, based on the experience of the NFC awards the resource allocation in Pakistan has proved futile. However, this process of resource allocation has both aspects of disgruntlement as well as constructive approaches. If we considered the constructive dimension, NFC has proved a premium strategy to ensure a friendly resource allocation among provinces and the federal areas. In addition, with the passage of time more financial autonomy was assigned to the provinces and it proved to be a means of better understanding of fiscal delegations, particularly in the previous NFC awards of 1997 and 2006.

It is also necessary to take a close look at key economic indicators and their performance in Pakistan. Next we will take a brief look at the performance of major economic indicators and their contribution in economic growth.

3.8 Economic Performance of Pakistan

The economic performance of Pakistan’s major economic sectors and their contribution in

GDP growth can be described as follows:

3.8.1 Long Run Growth Trends of Basic Economic Indicators in Pakistan

By casting a glance at the growth rate of major economic indicators one can observe that Pakistan experienced healthier GDP growth rates in the 1960’s and 1980’s while in the

1950’s and 1990’s it remained quite low. The growth rate in the manufacturing sector seemed dominant till the 1980’s but it performed poorly in the 1990’s and again flourished during the

2000’s. The services sector grew quite moderately and consistently over the period of the

1950’s to the 2000’s and shifted from 3.61 percent to 5.31 percent respectively. It remained highest during the 1960s and 1980s when it grew at the rate of 6.70 percent as presented in table 3.12.

Pakistan is classified as an agriculture country according to the world classification criterion and agriculture output has had a major influence over the country’s economic growth.

As figure 3.1 shows, the GDP growth rate is consistent with the agriculture sector growth rate.

Except for the 1990’s, over all the entire period the growth rate in the agriculture sector remained much lower than other sectors like manufacturing and services. The GDP growth rate in Pakistan followed the trend in the agriculture growth rate except in the last decade when the manufacturing and the services sectors acted as compensatory sectors and the trend in the

GDP growth rate followed the trend in manufacturing and services. The growth rate of the above discussed sectors can be observed from table 3.12 while trends in growth can be judged from figure 3.1.

Figure: 3.1

The Performance of Major Economic Sectors in Pakistan

0 1950s 1960s 1970s 1980s 1990s 2000s

Agriculture Services Manufacturing GDP

Table 3.12: Economic Performance of Major Economic Sectors in Each Decade Real Growth Rates

Years Agriculture Services Manufacturing GDP

1950s 1.76 3.61 7.73 3.11

1960s 5.10 6.70 9.90 6.80

1970s 2.40 6.30 5.50 4.80

1980s 5.40 6.70 8.20 6.50

1990s 4.40 4.60 3.90 4.60

2000s 2.83 5.31 5.67 5.41

Source: Economic Survey Government of Pakistan (various issues).

3.9 Development of the Fiscal Situation in Pakistan

Pakistan’s fiscal situation is grave and can easily lead to economic instability in terms of monetary and fiscal crises. The continuous underperformance may lead to losses in output, short fall in investment and insurmountable debt from financial institutions like The IMF and

The World Bank. However, control over these problems can be achieved by controlling fiscal deficits. Every year we propose budgets at state and provincial levels to promote economic activity and to improve the provision of public goods to the people. But the existence of the off-budget items, which cannot be measured and monitored properly, creates serious problems and makes budget allocations more complicated. So the factor of uncertainty with budget allocations makes it difficult to design a meaningful fiscal policy, and obviously the impact of fiscal policy cannot be measured or examined in isolation.

In the previous decade, Pakistan introduced various fiscal strategies to promote economic stability despite fiscal expenditures remaining dominant over fiscal revenue (generating fiscal deficit). The gap between fiscal revenue and expenditure was observed at a minimum level in

2003 and 2004 (minimum level of fiscal deficit) but rose consistently and reached the maximum bound in 2008, but was reduced by 2011. The rising trend in government expenditure emerged due to a disastrous earthquake in 2005. The fiscal deficit was at a minimum level in 2004 but a continuous rising trend was observed from 2005 to 2008. The trends in fiscal revenue and expenditures can be observed in figure 3.2(c). Similarly, trends in total and partial revenue and expenditures can be observed in figure 3.2(a) and 3.2(b) respectively while the growth rate of overall fiscal deficit, total expenditures and revenue as a percentage of GDP from 2000 onward can be judged from table 3.13.

Figure: 3.2

Trends in Total and Partial Revenue and Expenditures Figure 3.2 (a) Figure 3.2 (b)

16 25

14 20 12

10 15

6 10

4 5 2

0 0

Total Revenue Tax Non-Tax Total Exp Current Development

Figure 3.2 (c)

Total Expnditure Total Revenue

Table 3.13: Performance of Pakistan’s Fiscal Indicators (% of GDP) Expenditure Revenue Real Overall Financial GDP Fiscal Year Total Total Non- Growth Deficit Current Development Tax Expenditure Revenue Tax

2000 3.9 5.4 18.9 16.4 2.5 13.4 10.6 2.8

2001 3.9 4.3 17.4 15.3 2.1 13.1 10.5 2.6

2002 3.9 4.3 18.5 15.7 2.8 14 10.7 3.3

2003 3.9 3.7 18.8 16.2 2.6 14.8 11.4 3.4

2004 3.9 2.3 16.5 13.7 2.8 14.2 11 3.2

2005 3.9 3.3 16.8 13.3 3.5 13.8 10.1 3.7

2006 3.9 4.3 18.4 13.6 4.8 14.1 10.5 3.6

2007 3.9 4.3 20.8 15.8 5.0 14.9 10.2 4.7

2008 3.9 7.6 22.2 18.1 4.4 14.6 10.6 4.4

2009 3.9 5.3 19.9 16 3.8 14.5 9.5 5.1

2010 3.9 6.3 20.5 16.9 3.5 14.2 10.2 3.9

2011(B) 3.9 4 18 14.6 3.4 14.3 10.5 3.8

Source: Economic Survey of Pakistan (various issues); Statistical discrepancy (both positive and negative) has been adjusted in arriving at overall fiscal deficit numbers.

Table 3.13 shows that the total revenue grew at a constant average rate of 14.5 percent, but a sudden rise in total expenditure (due to the rise in current and development expenditure) lead to additional fiscal deficit onward after 2005. A severe earthquake in northern Pakistan added greatly to this deficit. Despite local and foreign donations the rise in government development and current expenditures was inevitable for Pakistan to foster the rehabilitation process.

3.9.1 Tax Structure in Pakistan

Taxes play a central role in modern public finance and are the most important source of revenue but they can also have repercussions if they become confiscatory or at least they lead to heavy tax burdens. The major objective behind taxes is to generate revenue to meet the financial requirements at the national and provincial levels. The tool of taxation can also be influential to achieve certain social objectives like lessen the inequality and improve the means of wealth distribution and control inflation by drawing away money from discretionary consumption.

Taxes are levied on individuals and firms or industries by the state authorities or equivalent of state-like revolutionary or secessionist movements and tribes and classified into two major segments called direct and indirect taxes. They also have different name like customs, duty, excise, toll, supply, gabel, tribute, aid, impost, tallage and subsidies etc. Some well-known taxes and their mode of application can be stated as;

Sales Tax: Imposed on the sale of goods and services and transaction of a business

Stamp Duty: Imposed on the documents verification and validation

Excise Tax: Imposed on the sale and production of certain goods

Road Tax: Imposed on a vehicle and known as Vehicle licensing Fee (Brazil), Registration fee

(USA), Vehicle excise duty (UK), Regco (in Australia) and toll tax (in Pakistan).

Income Tax: Imposed on families and individuals’ income

Since independence Pakistan is facing problems regarding tax imposition and collection methods. The tax system in Pakistan is not much supportive to foster economic growth. The heavy tax burden on low income sectors and nominal taxes on high income sectors is

unjustifiable. Mostly in underdeveloped countries a number of sectors are exempted from taxes. Unfortunately Pakistan lies in the list of those countries which have huge numbers of sectors exempted from taxes.

In Pakistan, both state and provincial governments have different assignments regarding tax (direct and indirect) collection, which can be examined in table 3.14.

Table 3.14: Tax Assignments of State and Provincial Governments Federal Taxes Provincial/ District Tax Direct Taxes Direct Taxes Corporate (profit) Tax Capital Gains Tax Personal Income Tax Tax on Professions, Trade and Calling Indirect Taxes Tax on Transfer of Property Custom Duties Land Revenue Sales Tax/ General Sales Tax on Goods Agriculture Income Tax Excise Duty Urban Immoveable Property Tax Indirect Taxes Stamp Duties Provincial Duties Sales Tax or General Sales Tax on Services Motor Vehicle Tax Entertainment Tax Royalties on Natural Resources* Miscellaneous Duties *Royalties on hydal power and natural gas generation commonly not constituted under tax section but considered major source of revenue for Khyber Pakhtunkhawa (KPK), Sind and Balochistan.

During the 2000’s, the ratio of direct and indirect (Custom, Excise and Sales) tax collection and share of tax revenue as a percentage of GDP can be examined in table 3.15. During the last ten years total tax revenue increased gradually with a significant rise in indirect taxes. Sales taxes remained a major contributor to promote indirect taxes. The contribution of tax revenue

remained lower at the level of 8.9 percent in 2005 and 2010. Moreover with the passage of time Pakistan has introduced various reforms to improve the share of revenue collection in terms of taxes.

Table 3.15: Structure of Federal Direct and Indirect Tax Revenue (Rs. Billion) Tax Rev Indirect Taxes Total Direct Year as % of (FBR) Taxes Custom Sales Excise Total GDP 124.6 65 153.6 49.1 267.7 2000-01 392.3 9.4 [31.8] [24.3] [57.4] [18.3] [68.2] 142.5 47.8 166.6 47.2 261.6 2001-02 403.9 9.2 [35.3] [18.3] [63.7] [18j [64.7] 148.5 59 205.7 47.5 312.2 2002-03 460.6 9.6 [32.3] [18.9] [65.9] [15.2] [67.8] 165.3 89.9 219.1 44.6 353.6 2003-04 518.8 9.2 [31.9] [25.4] [62] [12.6] [68.1] 176.9 117 235.5 58.7 411.4 2004-05 588.4 8.9 [30.1] [28.5] [57.2] [14.3] [68.9] 224.6 138 294.6 55 487.9 2005-06 713.4 9.4 [31.5] [28.3] [60.4] [11.3] [68.5] 333.7 132.3 309.4 71.8 513.5 2006-07 847.2 9.7 [39.4] [25.8] [60.3] [13.9] [60.6] 387.5 150.6 376.9 92.2 619.7 2007-08 1007.2 9.8 [39.6] [24.7] [60.3] [14.6] [60.4] 444.0 148.0 452.0 117.4 717.4 2008-09 1161.1 9.1 [38.2] [20.6] [63.0] [16.4] [61.8] 529.0 162.0 516.0 121 799.0 2009-10 1327 8.9 [39.9] [20.3] [64.6] [15.1] [60.2] 657.7 180.8 674.9 153.6 1009.3 2010-11B 1667 9.2 [39.4] [17.9] [66.9] [15.2] [60.5] Source: Federal Board of Revenue *as % of total taxes ^ as % of Indirect Taxes

3.10 Conclusion

Since 1991, the NFC awards have improved the process of resource distribution to the provincial level of government in Pakistan. The share of financial resources was also enhanced either thanks to the senior provincial share besides that of federal funds or because of the addition of taxes in the divisible pool. Moreover, due to these awards direct shifting of funds and increased grants have been channeled to all provinces. Likewise, the incentive of matching grants forced the provinces, attracting them to increase efficiency, to have their own resource production and obtain financial independence (Ahmed et al. 1997). However, with the passage of time provinces have insisted mostly on adding up various criteria in the resource distribution formula. For example, NWFP is famed for its backwardness and anxiety about the division of revenue claiming that their backwardness itself should demand for a greater percentage of the funding. Sindh is in favor of the revenue creation criterion, while Baluchistan raised the notion that geographic area criterion takes precedence. Besides, according to Punjab, agriculture manufacturing should be the supreme distribution criterion of resources at the provincial level.

Thus, due to these contradictory resource distribution approaches and the lack of agreement among all provinces, they are inclined to not follow a single measure, which would have been the best possible solution. It is also concluded that tax reforms have increased the revenue share through taxes but still there is a need to promote the transparent tax mechanism and fair collection and redistribution of the resources to lower tiers of government. While the poor progress of various economic sectors has decreased the economic prosperity in Pakistan. To deal with challenges of fiscal decentralization and poor economic performance in next chapter we will take a comprehensive review of some ancient studies to find out affective determinants of fiscal decentralization and appropriate research methodology for empirical analysis.

Chapter 4

Fiscal Decentralization: Conceptual and Theoretical Frame Work

4.1 Introduction

Decentralization can affect the performance and the role of the government in a state which activates decentralization. Its impact on the performance of its institutions and the type of improvement to be experienced in a country is still a debatable question in many countries, particularly in developing economies.

No doubt fiscal decentralization is a very effective tool to promote economic growth and a number of studies have proved that statement (showing a positive effect of decentralization) but some studies showed different results (negative effects of decentralization). However, to understand the decentralization and its execution process in an economy, it is primarily demanding a basic understanding of the subject.

The objective of this chapter is to provide a comprehensive knowledge about different types of decentralization, its conceptualization and its theoretical framework. Before embarking on the literature review and proposing the effective methodologies to capture the true impact of fiscal decentralization on macroeconomic performance of a country, it is a prerequisite to get knowledge about various types, dimensions and theories of decentralization.

4.2 Concepts of Decentralization

Decentralization is a process of shifting the power from a national or a central government to a regional or local government. Moreover it is a system in which more autonomy is allotted to the people to deal with different issues at the local or municipal level. More local autonomy

represents a higher degree of decentralization. Decentralization is the shifting of powers and autonomy to the local government to make various decisions according to their own priorities

(Akia and Sakata, 2002 and Thiessen, 2003) defined decentralization in the following words.

“A transfer or shifting of duties related to duties of sub-national governments”.

In simple words it is the autonomy of the provincial or state government to collect taxes and spend them according to their preferences (Thiessen, 2003). Decentralization is a process of transferring the control of accountability and community functions from national level to lower tiers of the government. Due to different types, the term ‘decentralization’ can be divided into political, administrative, fiscal and market decentralization. The familiarity with all these terms can be helpful to properly understand the various useful aspects of decentralization.

4.2.1 Political Decentralization

The political decentralization promotes the elected government to facilitate the local communities at their door step in various decision making activities. It is frequently connected with many political affairs and representative government, but it can also maintain democratization by giving the nation, or their legislative body, more control in the formulation and execution of policies. Individuals in favor of political decentralization believe that the policies or decisions made by political administration are not as much supported or appreciated as compared to those decisions which are made by large grants, known locally as “donations”.

So we can say that the political decentralization shows that the elected leaders are well connected with local citizens and know more about the needs and wishes of their respective constituencies. Moreover, political decentralization allows the local citizens to participate in decision-making activities.

4.2.2 Administrative Decentralization

The dispersal of fiscal funds and responsibilities among different government tiers to lend services to the local communities can be achieved by administrative decentralization. But by itself it is the transfer of responsibilities in terms of financial resources, rendition of bills and project execution responsibilities from the central authorities to the provincial, district or to local municipal government or to the semi-autonomous public authorities. In short, the shifting of management, financing and planning responsibilities to the lower tiers of a government is known as administrative decentralization. Deconcentration, delegation and devolution are three major types of administrative decentralization which have different characteristics, as they are explained below. i. De-concentration

De-concentration, which is considered a type of decentralization, is generally known as the weakest type and mostly used in centralized states. It concerns the shifting of responsibilities from national government officials to the officials at provinces, districts and regional levels to facilitate the central government operations. Simply in de-concentration the tasks of central government are achieved through provincial, district and local governments’ management. The isolation in central government tasks ultimately reduces the autonomy of lower tiers of government and makes this type of decentralization weaker. ii. Delegation

Delegation is considered a more extensive type of decentralization. It is a process in which the national government transfers the responsibility to provincial, district or semi-autonomous

organizations to follow the operation of the national government in addition to decision making powers but they remain accountable to the state government. However, the intervention of national government in the delegation system is less than in the deconcentration stage. As the national government transfers powers to lower level government it establishes, for example, local transportation authorities to generate transportation facilities and special project units to execute projects at the local level, among others. These organizations have discretion in decision making and they are exempted from various constraints. iii. Devolution

There is a third type of administrative decentralization and it deals at the managerial level.

In the devolution system the central government transfers financial, management and decision- making functions to local government or quasi-autonomous units (like municipalities and union councils) and it gives them autonomy to collect revenues and make their own investment decisions. In this system each local territory or union council has its own legally defined geographical boundary over which they can affect their public functions. Mostly devolution is considered in the sphere of political decentralization.

4.2.3. Financial or Fiscal Decentralization

Financial decentralization is a process which deals with the transfer of taxing and spending powers by the central government to a sub-national government. Mostly the industrialized or developed countries of the world are more decentralized as compared to developing countries.

The unsatisfactory performance of centralized economic systems has diverted the attention of centralized economies toward decentralization. It has broken the control of the central

government and lifted up the participation rate in democratic governance. In the last two decades in many developing countries the fiscal decentralization has become a crucial theme of governance. The devolution of fiscal responsibilities is the key component of decentralization. It is generally considered that if lower tiers of government carry out decentralization effectively, they must have adequate revenue (collected locally or transferred by central administrations) as well as authority to make decisions about expenditures. Financial decentralization can appear in different types which can explain as:

i. Co-production or co-financing arrangement through which the consumers can

participate in different services and infrastructure with the help of labour or monetary

assistance.

ii. Cost recovery or self-support through the consumer charges; iii. Adding up revenues at local level by means of indirect charges and sales or property

tax. iv. Permission of municipal borrowing and recruitment of local or national resources

through loan guarantees.

v. Inter-government shifting which can transfer general revenues from taxes charged by

the federal government to a lower level of government for specific or general uses.

Mostly in developing economies local governments are endorsed to impose taxes but this source of revenue is not much healthier in meeting the current expenditures, and dependence

of local government on central government subsidies is inevitable just as the subsidies are dependent on their processing.

4.3 Fiscal Federalism

In an officially declared federation and in non-federation states (consisting of different tiers of governments) fiscal federalism is not connected with fiscal decentralization, both have different concepts. So, fiscal federalism is limited up to the principles which a federation or a non-federation state applied to execute fiscal decentralization. In other words, fiscal federalism is generally considered a structure to assign functions at different governments’ level and provision of fiscal tools for execution.

Fiscal federalism consists of guiding concepts and principles which facilitate the formulation of a financial relation between the central and the state or provincial governments while fiscal decentralization is a process of applying the principle of fiscal federalism. On the basis of federal or non-federal states the implementation of these principles will be different due to differences in their legislative and political framework.

4.4. Economic Decentralization

The privatization and deregulation are also considered kinds of decentralization. They deal with economic liberalization and expansion of the markets but not all the times. They shift powers from the public sector to the private sector and known as another form of decentralization. The terms ‘privatization’ and ‘deregulation’ can be explained as follows:

i. Privatization

Privatization is a process in which the responsibilities for services provision in transferred to private partners, subject to market discipline or a regulator. Moreover privatization cannot be considered equivalent to decentralization. It can include the following points:

i. Give autonomy to the private sector to perform a function which was previously

monopolized by the federal government.

ii. By making a contract with a commercial company to make it possible the provision

of public services or goods.

iii. Allowing the private sector to participate in public sector programs through capital

market. ii. Deregulation

Deregulation abolished the legal constraints on private participation in services provision.

It allows the private sector to make the provision of those services which were previously provided by the federal or monopolistic government. In developing countries privatization and deregulation are considered a more efficient substitute to the government and local governments are, also privatizing by contracting out service provision.

4.5 On Going Debate Regarding Fiscal Decentralization

Decentralization can increase the performance of public and private sectors and foster economic development due to an improvement in financial allocation (Oates, 1972). According to Oates, decentralization has enhanced economic growth because local governments have a

closer interaction with local citizens. Due to such a close interaction local governments have financed various projects or plans according to the desire of the local citizens (Brennan and

Buchannan, 1980). It has also increased the efficiency of local governments by developing a competitive environment and improved the quality of services provision at lower governments’ tiers.

Accountability is another important aspect which acts as a catalyst to improve the efficiency of decentralization, particularly fiscal decentralization. The roots of fiscal decentralization cannot be firm without it. Accountability enhanced the efficiency by improving the quality of production, ensured the better provision of services and improved the delivery of goods at the local level (Martinez- Vasquez and McNab, 2003).

Fiscal decentralization is considered helpful in order to improve reflection in local political participation. This can enhance democratic values and local political effectiveness. It provides an opportunity to discuss local preferences and may enable the development of future political leaders. It also contributes to the creation of the special laboratory tests, which in the end, may provide a model being implemented by the central government or other lower level governments.

Oates (1999) stated that the decentralization process changed and specified the government levels that increased the provision of services and enhanced local production, which ultimately improved the welfare level in a society. The biggest difference in the preferences of regions causes greater public demands which can be fulfilled only from the payment of decentralized taxes.

There is a considerable disagreement regarding the effects of fiscal decentralization. First of all, financial decentralization may affect the participation across the country where the movement of families and businessmen badly affects and limits the redistribution of income.

Fiscal decentralization is expected to re-arrange policies to support the poor which have become the victim of discrimination. (Persson and Tabellini, 1994). Financial decentralization may affect the growth by making the organization more difficult. In addition, interest benefits and its consequences on the several geological sites may hamper per capita growth because of disparities in the region, communications, healthcare, education and other public services

(Thiessen, 2003). In this case, central governmental sectors can be allocated to generate funds that would lead to more influence over participation. Other economic financial decentralization dissenting opinion include the potential damage of macroeconomic and financial policies through the department (Tanzi, 1996); its breakdown to neglect economies of scale and scope

(Prudhomme, 1995); and inter-jurisdictional ‘leakages’ linked with local expenditures (Oates,

1972).

Due to the increase in the degree of fiscal decentralization, the local governments or municipalities went to give ‘the race toward the abyss” on the judgment of primary issues animation, and then in the provision of fruitful government expenditures to solve these issues on priority basis (Brueckner, 2004). It is said that local officials are closer to the people and can know more about the needs of the concerned local groups (Prudhomme, 1995; Tanzi,

1996).

The summary of criticism of Tanzi (1996) which collected a number of cases especially in emerging countries where the application of direct financial decentralization is less than an optimal result can be stated as:

i. The taxpayer’s information is incomplete or does not have the political power to

pressurize the policy makers to make local decisions about effective resources.

ii. Technology opportunity and mobility weakened the services that are totally local. iii. The local government often has shortage of expenditures on public goods and lack of

budget choice. iv. Fiscal decentralization can also hamper the ability of the central government towards

structural balances.

v. The authority and effectiveness of national bureaucrats are more than local bureaucrats. vi. Local politicians are frequently less competent and can spoil more resources as

compared to national politicians or find themselves at least in less optimistic

circumstances.

Prudhome (1995) finds some disadvantages in the fiscal decentralization theory. He says that economic competency requires local fiscal capacities and this situation should not be offered in underdeveloped countries. Moreover, he argued that fiscal inequities also rise with the process of decentralization and the mode of action of revenue decentralization is quite different from expenditure decentralization. The provision of goods and services through local government channels is not always the issue, but the key issue is how we can categorize the joint production of the services. So, we can say that local governments are acting like agents of the central government and they are obligated to implement the policies at local or regional levels.

4.6 Conventional Fiscal Decentralization Theories

Conservative economists usually define a local role for major industrialized countries using financial and fiscal decentralization theory known as federalism. This ideology is based on the

“Musgrave model of public accounting sector for distribution and allocation”. The theories and the model of decentralization not only provide the assistance in the execution process but it also facilitate the definition of the role of national and sub-national or local governments in terms of shifting and sharing different functions or tasks at different levels.

4.6.1 Stabilization

The primary liability for the stabilization function is predictably allocated to the central or the federal government, in part due to the problem that would result if sub-national jurisdictions had autonomy over funds. Moreover, local economies were believed to be very open, and that can improve the performance of the local sector. It can also enhance the provision of locally produced goods and services to meet the requirements of local citizens. Finally, the lack of policies in support of the local level was not considered particularly interesting as it is of real concern that the reimbursement could result in substantial income transfers to the devices of debtor/creditor jurisdictions. Although the revenue side of fiscal decentralization seems more appropriate for local level governments, the autonomy in revenue collection and distribution at the local level may also hamper the ability of local governments to follow a suitable fiscal policy in terms of national interests.

Due to severe economic depressions which were based on macroeconomic instability the tools to achieve the stability in macroeconomic context are the key issue in global economic forums. Most of the stabilized (developed) countries are following a decentralized government

system. So it is commonly considered that decentralized governments can better meet the challenges on economic stabilization. So with no doubt stabilization is a primary issue in the emerging economies. In addition, the formulation of fruitful stabilizing policies is also a big challenge to promote a healthier economic environment in developing economies. First of all, macroeconomic trends may be mainly hard in those economies which are based on factors that are linked with natural climate (as agriculture based economies) or unpredictable changes (like political instability). Therefore, to face such types of challenges or circumstances the services of experienced and submissive people is a prerequisite to design stabilization policies.

Secondly, in developing countries the role of local governments in the economy is minimal as compared to developed or industrialized countries. In addition, local governments are bound to depend heavily on resources endowed by the federal government and only follow the federal government operations. As a result, such types of economies are categorized as low decentralized regimes and this attribute restricted the impact of fiscal policies at the grass root levels. Thirdly, the impact of fiscal behaviour is restricted in underdeveloped countries because their local government’s demand for local revenue is greater than in developed economies. In nearly all developing economies the demand of revenue is greater than local revenue

(depending on local taxes) and thus the local government is bound to depend on the central government.

4.6.2 Distribution

Fiscal decentralization theories give stress on the distribution of powers and resources between the national and the sub-national governments. It means that decentralization limited the role and the intervention of the central government in states or local level issues. In

centralized economies, only the central government is considered the supreme authority to reallocate the resources among the different localities.

Fiscal federalism itself gives the prime responsibility for its distribution with the federal government. First of all, only the national government is in a position to transfer the resources of the jurisdictions from the richer to the poorer. Secondly, the programs of redistribution would be likely to create problems if the factors of production are mobile. Affluent residents and businesses can move outside of the control with the redistribution, while the poor people who have the right to receive benefits would approach, thus limiting the base of redistribution.

Thirdly, local governments were persuaded to have contact with fiscal resources that are not easily perceived as progressive with regard to the income.

But there were also some problems with the traditional suggestion that the distribution has been centralized and it is still generally accepted that the administrations are generally more limited compared to the central government regarding the distribution of income. They can play a role as an intra-jurisdictional redistribution, but it is clear what happens between the jurisdictions. Limits on intra-jurisdictional redistribution are not as much problematic in the developing countries with compared to the advanced countries. The mobility of the rich can be less important in the developing countries because it only affects a few cities with a high level of public services and other facilities.

4.6.3 Allocation

The default act of fiscal decentralization is important for the government in the distribution functions because the demand for a lot of public services can only be met with decentralization.

We can improve welfare benefits through decentralization because inhabitants of various

jurisdictions could choose the combination of public goods and the taxes that are consistent with their preferences. In the event that there is no market and competitive pricing of public services the demand at the community level is expressed through collective decision-making and casting. In this context, the decentralization could be regarded not only on the basis of pointing out uneven distribution of resources, but on the basis that the expense layouts are deeply committed to the actual cost of the resources in smaller jurisdictions. On the other hand, on account of decentralized government increases, it is obvious that there will be more testing and progress in stipulation of the available goods in that locality and potentially leading to improvements in overall progress of resource availability.

The limits of the existing political jurisdictions, which provide social benefits lead to the creation of new levels of government with effective borders to provide the most public good with a given territory. This must be weighed against the local institutional benefits and the cost for the large number of jurisdictions. It seems realistic to pursue the alternative selected policies that show the similar whole effect as the configuration of cooperation, understanding between those represented, decentralized opinionated jurisdictions. On the other hand, there are four fundamental instructions that are recommended for explaining and evaluating the structure of revenue (Oates, 1972). Additionally, there should be a balance between local taxes with high achievability and a positive impact on the local economy. All the advantages and expenses of local taxes in the provision of services should be clearly understood by the people.

Further, the sketch or outline of the taxes collected from the locality should adhere to the criteria of equality. Complex taxes should be avoided for better management and minimum costs.

Figure 4.1: Public Finance and Public Choice

The Welfare Loss of Centralization.

Figure 4.1 shows the demand for a local public good ‘G’ of two individuals in two different localities ‘A’ and ‘B’. The DA shows the demand of an individual in first locality ‘A’ while DB shows the demand of a person in second locality ‘B’. it is assumed that the marginal cost MC for the provision of public good ‘G’ is constant as presented by the line PMC. This also shows that marginal cost will be distributed equally between the individuals.

Figure 4.1 also shows that the uniform level of good is provided by the centralized economy. The demand of individuals in each locality ‘A’ and ‘B’ is compromised at level ‘q’ which is less than that quantity demanded by an individual in locality ‘A’ while greater than demand of an individual in locality ‘B’. As a result welfare loss will be experienced by both individuals in their respective localities. So the individual in locality ‘A’ will face the loss equal to triangle 1-2-3 because this locality is consuming less than that which they demanded (if there is no compromise). Therefore to achieve the additional good qqa the individual in locality

A would gladly pay the cost equal to q23qa but actually it will be available at cost equal to

q13qa . Similarly the Individual in locality ‘B’ will generate welfare losses equal to triangle 1-

4-5 due to excessive use of good and he should pay qb51q for additional qbq but he is actually

paying qb54q. However this deadweight loss can only be avoided if the distribution authority

will allocate the good exactly according to the demand of each locality. In decentralized

systems the provision of goods can easily be made according to the demand of each locality so

decentralization permits each locality to avoid such losses. This idea described the

‘decentralization theorem’ which was elucidated by Oates as follows:

Oates (1972) stated that the consumption of a public good can be described by uniting the

geographical subsets of the total population. Although for central and local governments the

allocation cost for each level of output is the same in each jurisdiction. But despite that, it will

be very difficult for the central government to make available uniform levels of goods across

all the jurisdictions while it will be more convenient for local government to provide the pareto

efficient level of output for relevant jurisdictions. Some other points may also add up which

can be explained as: i. The areas of the shaded triangles depend on the slope of the demand curve meaning that

the more inelastic the demand curve the greater the increase in the areas of the triangles

which means that the losses from the centralized distribution of the good can be determined

by the size of the triangles. By using multiplicative demand functions, Bradford and Oates

(1971) made various attempts to determine the welfare losses. It is also helpful to determine

the loss in consumer surplus that would appear from an effort to achieve uniform levels of

expenditure. ii. Heterogeneity in choice by localities will increase the welfare losses on the centrally

distributed good. If each locality is different in preference then welfare loss will be greater

which means that the q will provide a close approximation if qa and qb are close to each

other. iii. Decentralization is an effective tool to avoid welfare losses, but despite that it is considered

that the centralized government can apply macroeconomic stabilization policies more

effectively as compared to other levels of government. Government expenditure

multipliers will be lower due to the openness of small local economies. Alternatively it will

lower the impact of stabilizing fiscal policies at the local level. iv. Moreover, in the case of local economies, when individuals are mobile between the local

authorities then the redistributive policies will not be proven to be very effective. Suppose

if local governments want to increase taxes on upper income residents to generate revenue

that would simply induce them to move to another locality with a lower tax rate. Thus

centralized government can apply redistributive policies more effectively because

individuals are less mobile internationally.

v. The optimum size of the locality will be influenced if there are economies of scale in the

production of goods. If other things remain the same, due to lower average cost, the larger

communities will be in a better position to take advantage as compared to small

communities.

Oates (1999) stated that decentralization is required because the political pressure may

influence the distributive authorities of national governments in the provision of certain

government services. Local government, due to a closer interaction with local peoples, can

make such provision of goods matching best local demands. Moreover, for empirical estimates

Oates found the inverse relationship between fiscal shares of national governments with a

sectionalism index in sampled countries. It means the governments with more decentralized public sectors gain more welfare as compared to centralized governments.

4.7 Tiebout Hypothesis; “Voting with Your Feet”

The previous section discussed the provision of public goods according to the choices of individuals in different localities and the best possible size of societies to which they belong.

In addition, in 1956 Tiebout discussed the preferences of individuals and argued that individuals will choose the community according to tax levels and provision of public goods and services. Ultimately, they will choose those localities which will best compliment their choices or preferences. While contrarily Samuelsson in 1954 suggested that in the case of public goods individuals would not reveal their preferences. So Tiebout’s arguments were a direct response to Samuelson’s ideology. He explained that in a local community context if mobility is relatively less costly, the individuals can move from one locality to another to meet their preferences and tax benefits. So in case of public goods people can reveal their preferences. Hugehes et al. (1987) also supposed that individuals will shift to those societies

(or local governments) whose principles (in terms of provision of public goods and tax benefits) match or best suit their own desire. Samuelson himself highlighted the problems for allocation of public goods or absence of revealed preference in case of large communities.

Tiebout’s hypothesis proved to be the best remedy for Samuelson’s findings and promoted the idea that individuals will reveal their choices by ‘voting with their feet’. It means that the individuals who reside with ‘other’ will always pursue their own priorities or preferences and join their respective local alliances accordingly. For instance, the people who like expenditure on art and libraries can live among ‘other’ peoples but indulge their own urges. In other words if a local community tenders diverse types of public goods then each person would like to

select that group or community which makes best provision of output to meet their preference.

In short, the main objective of the Tiebout decentralization theorem is to improve the welfare gain and lower the deadweight losses through decentralization.

Figure 4.2 Tiebout Adjustments

The above diagram shows the Tiebout equilibrium while Holcombe (1983) also used the similar diagram. The (b) part of figure 4.2 shows the total demand curve ‘D’ for a local public good, ‘G’. Here the demand curve ‘D’ is vertical summation of individual ‘D’ demand curves i.e. D1, D2 and D3.

At equal tax share O-T shows the marginal cost for the provision of public good ‘G’ and equilibrium quantity is O-q2. At this outcome the individual with demand curve Di will face the welfare loss equals to triangle 1-4-5 due to over provision of the public good. While individual 3 will face welfare losses (equal to triangle 1-2-3) due to under provision of the

public good, it is considered that the individuals ‘3’ and ‘1’ are living in their preferred communities and they are in competing jurisdictions. If there are three identical individuals then each individual will be in equilibrium as presented in figure (a), (b) and (c) and there will be no welfare losses. So a situation of ideal outcome will appear when each individual enjoys public good according to their demand O-P. If each locality provided different levels of public goods i.e. Oq1, Oq2 and Oq3, then fiscal migration can adjust the losses which appear due to unequal provision of public goods in different competitive jurisdictions. As individual ‘3’ will move towards that locality which offers greater provision of local public good means by voting with their feet both individual ‘3’ and ‘1’ will avoid welfare loss which appeared due to less or excess use of public goods respectively. No doubt the Tiebout hypothesis is very significant in theoretical prospects and comprehensively discussed the difficulties of public goods provision. But the question is how will it act to achieve the equilibrium state? To answer this question we will discuss limitations and assumptions of Tiebout’s model and observe the empirical work.

4.7.1 Basic Assumptions of the Tiebout Mechanism:

The basic assumptions (which are by no means exhaustive) add a note of skepticism.

Following postulates can play a vital role to enhance the efficiency of Tiebout’s mechanism:

4.7.2 Brief Awareness of the Characteristics of All the Communities:

It means that all individuals in different localities must have comprehensive information about local expenditures and tax burdens.

4.7.3 Costless Mobility in Different Competitive Jurisdictions:

It means that households can shift from one locality to another locality without getting worried about transportation costs and alternative earning opportunities. However, Hindriks and Myles (2006) explained that it is possible only when individuals obtain their income from capital (on shares, property and land ownership) and then there will be no issue in which locality they prefer to live.

4.7.4 Impure Public Good Provision:

It means that the pure public good can be provided only in the case of a single large and non- competitive community. It implies that the public goods are impure. Otherwise the optimal size of the community will be infinitely large in the case of pure public good provision.

4.7.5 Large Population Size:

In a population there should be sufficient and variety of communities representing different levels of tax burdens and provision of public goods which will satisfy the diversity in choice of public goods. Therefore an appropriate large size population can permit the optimal formation of various communities.

4.7.6 Dependent on Tax Related Benefits:

Usually the community or local laws are not involved with the issues of free riding. The free rider problem mostly emerges in local jurisdictions where they have financed various public goods and services (like parks, schools and roads) by means of tax revenue generated by the property tax. The individuals who do not own property will act as free riders and enjoy

the services without bearing any cost unless you also count residents who rent from a local taxpayer.

In case of a public good, Tiebout said that people would reveal their preferences. Moreover the total cost for the provision of public goods will be distributed equally among the individuals. Then the individuals with a higher public good demand will have to forego the consumer surplus if they shift from the locality of a higher tax burden to a locality with a lower tax burden. The losses in the form of consumer surplus will emerge due to relatively less provision of public goods in that locality with the lower tax burden. On the other hand, the individual with less demand for public goods will face losses (cost will be greater than marginal benefits) if they will reside in the jurisdiction with the higher public goods provision. So the

Tiebout model will offer another mechanism, the ‘voting with your feet’ if tax burden is decided on the basis of benefits.

4.7.7 Externalities Created due to Mobility:

The externalities (in the shape of extra overcrowding) could also come forward due to the migration of individuals from one locality to another. The Tiebout mechanism can maximize welfare at equilibrium when movement of individuals from one locality to another locality will increase their utility level. Contrarily the Tiebout model can create the situation where even costless mobility would not be approved for fruitful gains in the presence of certain externalities. As people migrate from one community to another, the externalities will be experienced by those people who already reside in that community.

Boadway (1979) has discussed the studies of Flatters et al. (1974) Buchanan and Goetz

(1972) and Buchanan and Wagner (1970). The migration process will continue until the

benefits of migration will be equal if people will move from one locality to another within a district. If an individual derives total utility TBx while residing in locality ‘X’ and total utility

TBy when residing in ‘Y’ locality then at equilibrium the total utility can be write as;

TBx= TBy → (a)

Is this equilibrium condition with maximum welfare level? As we previously discussed the extra overcrowding could also emerge due to migration of individuals from one locality ‘X’ to another ‘Y’. If MC denotes the cost of overcrowding in each locality then welfare will be maximized as;

TBy- MCy = TBx- MCx → (b)

It shows that an individual migrant will make the migration decision on the basis of the difference in TBx and TBy without considering MCx or MCy because it is concerned with those individuals who are already living in that locality. So an individual migrant will not migrate only when TBx will equal to TBy, however the equilibrium will be decided on the basis of the settlement of the equation (b). This means every individual has a right to join that group or community which best reflects his preferences.

4.7.8 Economies of Scale:

The gain from economies of scale is a prerequisite for better performance. In case of a large number of local organizations the preferences for public goods will be diverse and equilibrium will be achieved. This implies that there will be a lot of small local communities who are missing the potential gain which emerges due to economies of scale in the production of Gross

Domestic Product.

4.7.9 Benefit Spillover:

The above analysis showed that all gain will be experienced by the individuals who reside in different localities. However, the benefits will spill over from one locality to another and through direct bargaining, localities can centralize the spillover which shows the role for central government.

4.7.10 Non- Static Preferences:

The Tiebout mechanism will lose the efficiency if preferences for local public services changed over the life-cycle. As in an early age period people will prefer the educational facilities but in a later period their preference will divert towards facilities of old-age pensioners. It means that communities of individuals or groups of households all required changes simultaneously.

4.7.11 Production and Location:

When we take into account the ratio of workers required in different localities, the Tiebout hypothesis loses validity. If in locality ‘A’, 40 engineers and 20 doctors are required to provide the preferred local public good and locality ‘B’ required 25 engineers and 10 security officers then efficiency will only appear if the level of community preferences will meet the level of preferences for engineers, doctors and security officers in each locality (Hendricks and Myles,

2006). So if the location of professions is incompatible then efficient provision of local public services will not be possible and put a question mark on the Tiebout hypothesis.

4.8 The Tiebout Hypothesis: Empirical Tests

To judge the efficiency of Tiebout Mechanism and to examine the hypothesis empirically the following example can be given;

4.8.1 Focus on Capitalization

Oates (1969) considered education expenditure per student to examine the quality of public services provision in New Jersey and data was collected from 53 municipalities located in the northern region. He found that people will shift towards that locality which has better provision of services and the property price in that area will be increased if the property level is considered constant. The result shows an inverse relationship between the tax burden and the property prices in a locality. Positive correlation was determined between property values and education expenditure per student which means that the individual will prefer to move to that locality with additional fiscal spending. Oates’ finding goes in favor of Tiebout’s hypothesis which means that individuals will prefer to reside in those localities which provide high quality services, and they will be ready to pay more toward these services.

Oates also explained the instructive strength of the Tiebout’s hypothesis by building the relationship between local expenditure (incurred by taxes) and capitalization. However, some studies deviate from the Oates ideology and argue that there will be no capitalization of local expenditure “crop up” if Tiebout’s mechanism worked (Edel and Sclar, 1974 and Hamilton,

1976). It can be explained as follows; if an individual can move to different localities then each locality will consist of those communities who have common preferences of public goods. In other words, the convergence towards the Tiebout equilibrium is not feasible when local

expenditure (incurred by taxes) will perfectly reflect the preferences of each community. It means individuals are already in their preferred local communities, so it will restrict the further mobility of individuals from one locality to another (means the Tiebout mechanism is not working). Some studies also argued that if capitalization decreases over time that does not mean the Tiebout equilibrium would be achieved (Pauly, 1976 and Chaudry, 1988).

Figure 4.3 Migratory Incentives

4.8.2 Focus on ‘Fiscally Provoked Migration’

Aronson and Schwartz (1973) introduced the fiscally provoked migration between the localities. Figure 4.3 indicating the migration incentive with the help of four locations. For proper understanding of the figure we can explain its various points as.

Along X-axis = (Bad) Effective Tax Rate

Y-axis = (Good) Per-capita Local Government Expenditure

While point ‘1’ indicates the jurisdiction in which an individual is currently living and point 2, 3 and ‘4’ show different localities. Locality 2, 3 and ‘4’ with reference to location one can be explained as follows;

Locality 2 = Less of a Bad (situated in West)

Locality 3 = More of a Good (situated in North)

Locality 4 = Both Good or Bad (situated in North-west)

Localities 2, 3 and ‘4’ will gain population because of more public services as compared to locality ‘1’. However, the migration process will be repeated from locality ‘2’ to 4, 3 and from locality 2, 4 to locality ‘3’. Local governments in The United States and The United

Kingdom provide the best empirical example of such a mechanism.

We can also explain this mechanism by considering the frontier of efficient jurisdictions

(Aronson, 1974). Figure 4.4 shows the frontier of efficient jurisdiction which, against effective tax rate, provided the higher per capita expenditure. Alternatively, it offers per-capita expenditure against the lowest possible tax rate as compared to other local communities.

Here,

Along X-axis= Per-capita Expenditure ; Along Y-axis= Effective Tax Rate

EJF= Efficient Jurisdiction’s Frontier ; I = Utility of 1st Individual

H= Jurisdiction

When the Indifference curve of the ‘1st’ person is applied it shows the utility level ‘I’ in jurisdiction ‘H’. It means that ‘H’ will be an ideal jurisdiction for the ‘1st’ person where he gets maximum utility (benefits) from public goods provision against effective taxes. However, few studies concentrate on migration in neighboring territories analyzed that the provision of public

good improved due to exogenous factors. For instance, the changes in environment can raise the demand for land in those territories with improved environment. As Banzhaf and Walsh

(2008) have found in empirical evidence, the rise in demand of land in neighboring localities with better environment follows.

Figure 4.4 Optimal Location

4.8.3 Focus on ‘Taxpayer Satisfaction’

Tiebout (1956) hypothesizes that the taxpayer’s level of satisfaction will be greater in case of a large number of competitive jurisdictions. Ostrom and Whitaker (1973) used primary data to analyze the relationship between the provisions of public services controlled by large versus small jurisdictions. They found that the smaller jurisdictions are more targeted and efficient in delivering public services. Lowery and Lyons (1989) developed a comparison by collecting data from consolidated (autonomous in collection and redistribution services) and fragmented

(concentrated on allocation of services only) jurisdictions. They did not find a strong relationship between the level of satisfaction and jurisdictional structure. Later on Parks and

Oakerson (1993), they evaluated the education, street, fire and police services and determined positive results while they found opposite results when they examined provision of social services, air pollution and water supply.

4.8.4 Focus on ‘Sorting’

Tiebout (1956) also hypothesizes that a larger number of competitive jurisdictions will lead to greater homogeneity among them. When there will be a large number of competitive jurisdictions, it will be more convenient for individuals to move to their preferred localities.

This phenomenon can ultimately promote homogeneity in preferences among those localities

(Gramlich and Rubinfeld, 1982). It indicates greater homogeneity in demand and exists within peripheries near many other communities. While in contrast, peripheries near very few communities were more heterogeneous. Sorting concerned with those variables which determined the preferences, like income, occupation and education. Ottensmann (1982) concentrated on sorting and found heterogeneity in terms of home ownership.

4.9 Decentralization and ‘Leviathan’

Brannan and Buchanan (1980) argued that the large size of a central government can make it incompetent. To maximize revenue through taxes, such types of government will mature like a ‘Leviathan’. They also suggested that to allocate services and finance projects at the local level, a single-tier government will act as a monopolist. In Brennan and Buchanan’s words it can be written as follows;

“The potential for fiscal exploitation varies inversely with the number of competing governmental units in the inclusive territory”. (Brennan and Buchanan, 1980, p.185).

A highly segregated system allows the individuals to ‘vote with their feet’ and if mobility is compulsory, decentralization can amplify political accountability and information.

Moreover, ‘voting with their feet’ restricts the abilities of policy makers to act independently according to the individuals’ preferences. Oates (1972 and 1985) examined the Tiebout’s statement that fiscal decentralization with mobile individuals is limited to concentrating on public expenditure. For the empirical examination, national and state level data were used. He also controlled the effects of intergovernmental grants, population urbanization and income.

Despite that, the results did not show a significant relationship between public spending and government size.

4.10 New-classical Growth Theory

In an economy, capital and labour are the key driver of economic growth. Therefore, present work will be based on Neo – classical growth theory and used the transposed form of

Cob-Douglas production function to examine the impact of fiscal decentralization on macroeconomic performance. In an economy with two factors of production the traditional production function can be written as follows;

푌푡 = 푓(퐾푡 , 퐿푡)

Where

푌푡 = level output in an economy in the current period,

퐾푡 = stock of capital in an economy in the current period,

퐿푡 = supply of labor in an economy in the current period.

Now we will transpose the Neo-classical growth model by introducing the fiscal decentralization variables and other growth related variables. The new functional form will be written as follows:

푌푡 = 푓(퐾푡 , 퐿푡, 퐷푋푖 , 푍푖)

Here 푌푡 is the nominal GDP in local currency unit which represents the total output in an economy and further in our extended model, we have peroxide capital stock with total fixed investment and employed labour force is used for labor stock. We also introduced Fiscal decentralization variables which are represented here by 퐷푋푖 and 푍푖 indicates some other control variables like inflation and openness.

4.11 Conclusion

This chapter provides a comprehensive knowledge about various concepts, terminologies and types of decentralization, and arguments of various researchers in favor of and against fiscal decentralization. Finally, it discussed the basic model, theories and hypothesis of fiscal decentralization. After establishing basic terminology and concepts of decentralization, the brief discussion on theories, models and arguments of various researchers we are now in a position to take a comprehensive review of decentralization activities in Pakistan. At the international forum Pakistan is considered a front line decentralized country but still the benefits of decentralization are not very visible in Pakistan. The impact of fiscal decentralization on macroeconomic performance is still a debatable question in various policy agendas. So, in the next chapter we will propose a suitable methodology to analyze the available data sets.

Chapter 5

Data and Methodological Issues

5.1 Introduction

Data and Methodology play a noteworthy role in research studies. Particularly, for empirical investigation it is very essential to use reliable data sources along with correct estimation techniques. After the literature review on fiscal decentralization studies and theories showed that a number of theoretical and empirical studies have been carried out to investigate the relationship of fiscal decentralization, with various economic indicators, particularly with macroeconomic performance.

The present chapter will briefly describe the data and methodologies to examine the inter linkages between fiscal decentralization and macroeconomic variables. Moreover, this chapter will briefly explain the sources of data collection and eventual use of econometric strategies for the current study.

In short, the present chapter will describe the data sources and types of data, econometric strategies and finally the description of the variables which we will use to investigate the relationship between fiscal decentralization and key macroeconomic indicators with special emphases on Pakistan.

5.2 Data Sources

To explore the association between fiscal decentralization and various macroeconomic indicators (like economic growth, Public Investment, Inflation, employment and private

100

savings) of Pakistan, the present study used the Economic survey of Pakistan (various issues), the annual reports of the State Bank of Pakistan and the handbook of statistics compiled by the

Federal Bureau of Statistics (FBS) to retrieve data from the time period 1972 to 2010.

Bearing in mind the weaknesses of the secondary sources of data, we have applied the maximum data mining tools to examine the real impact of fiscal decentralization on macroeconomic performance in Pakistan. Finally, the descriptions of the dependent as well as the explanatory variables are given in table 5.1.

5.3. Selection of Variables

According to the nature and the economic classification we have segregated the variables into different sections. The detail description of the selected variables is as follows:

Description of Variables: Previously, federalism had been measured in terms of revenue

(ratio of sub-national or provincial government revenues to total national government revenues) and expenditures (ratio of sub-national government expenditures to total government expenditures). Some studies suggest that the usage of these two ratios is not reliable but it is still a significant tool to capture the true intensity of federalism (Phillips and Woller, 1997).

Particularly, in the case of Pakistan, the revenue side may generate misleading results because all types of taxes are levied by the state government. So this study will follow the two adjustments to capture the true picture of fiscal decentralization on the revenue and expenditure side that was prescribed by Wasylenko in 1987. We can define the variables as follows:

101

Gross Domestic Product (GDP): Annual nominal GDP (in Millions Rupees) is also used in fiscal decentralization literature to measure the direct relationship between fiscal decentralization and economic growth.

Public Investment (INV): The sum of total investment made by the government during a fiscal year in Millions Rupees. Hypothetically, fiscal decentralization enhances the level of public investment which leads to more economic growth.

Inflation Rate (INFR): The inflation rate is measured by the overall consumer price index at the national level. It is hypothesized that fiscal decentralization is a mean of price stability in an economy.

Employed Labour Force (ELF): Employed labour force (in Million) indicated the proportion of total labour force which consisted of the aggregate of salaried employed, self-employed and casual employed. It can be hypothesized that fiscal decentralization has positive impact on employment.

Private Savings (PSAV): National saving (in Millions Rupees) is the aggregate of public saving (comprised of general government savings) and private saving (consisting of household saving plus corporate savings). According to our model we will limit our interest to private saving. Hypothetically, fiscal decentralization is considered a mean to enhance the level of private savings in an economy.

Capital Stock (CAP): To examine the yearly output level in an economy capital Stock is an important factor therefore present study will use the gross fixed capital formation (in Local currency units in Millions) to measure the available capital stock.

102

Decentralization with Expenditures (FDX): To measure the spending side of fiscal decentralization we have constructed a simple variable which represents the ratio of subnational government expenditures to total government expenditure. A rise in the share of federal expenditure lowers the level of fiscal decentralization. On the other hand a rise in subnational government expenditures results in rise of fiscal decentralization.

Adjusted Decentralized Expenditure (AFDX): As far as the analysis of expenditure side of fiscal decentralization is concerned, the expenditures which are pure responsibility of central or sub-national government are taken into account. This variable is composed of the ratio of sub-national government expenditure to national government expenditure after deducting the expenditure on defense and debt servicing.

Decentralization with Revenue (FDR): To estimate the revenue side of fiscal decentralization we will take the ratio of sub-national government revenues to total national government revenues.

Adjusted Decentralized Revenue (AFDR): Federal spending does not include federal grants awarded to the local governments. State spending comprised of the total amount of grants a state receives minus the total amount of transfer by the state to the local governments.

Therefore, in order to measure the revenue side of the fiscal decentralization an adjusted variable is used and the ratio of sub-national government’s revenue deducting the grants in aid is taken into account.

Economic Performance Index (EPI): For economics performance we have taken the absolute value of Inflation rate minus budget deficit as a percentage of GDP, minus unemployment rate,

103

plus percentage change in rea GDP and finally subtracted from 100 percent (Khramov and Lee,

2013).

Trade Openness (OPEN): Trade openness is the net foreign trade to GDP ratio and net foreign trade is the sum total of imports and exports. Trade openness directly affects growth and employment. The positive influence of trade openness on growth and employment is caused by benefits of resource allocation raised from external competition.

Per Capita Income (PCI): Nominal per capita income is the ratio of annual nominal GDP (in

Millions) to total annual population (in Million).

Population Growth Rate (POP): Annual population in Millions is also considered as an explanatory variable.

Literacy Rate (LITR): Annual literacy rate is used as an explanatory variable to examine the impact of literacy rate on employed labour force in Pakistan.

FDX_GDP: Ratio of decentralization expenditures to GDP is used to examine the productivity of decentralization expenditures to accumulate GDP.

FDR_GDP: Ratio of decentralization revenue to GDP is used to examine the productivity of decentralization revenue to accumulate GDP.

GTI_GDP: Ratio of public sector investment to GDP is used to examine the productivity of public sector investment in GDP accumulation.

PSAV_GDP: Ratio of private savings to GDP is used to determine the productivity of private savings in GDP accumulation.

104

Log of Gross Fixed Investment (LGFI): Log of annual gross fixed investment (in Millions

Rupees) is used as an explanatory variable to examine the impact of gross fixed investment on employment in the observed time period.

5.4. Application of the Econometric Model

This study is concerned with fiscal decentralization and its impact on economic performance. From the literature review, we came to know that several studies have been analyzed including the empirical trends between fiscal decentralization and the variety of economic variables. Sometimes those studies found a positive and sometimes showed a negative relationship. However, the present study will explore the impact of fiscal decentralization on macroeconomic indicators by following a different approach. This study will be an addition of new literature in fiscal decentralization studies may divert the attention of researchers in new directions. Likewise, the present study is using latest available data for

Pakistan and for World’s Federations. In Pakistan, Previously we believe there is no comprehensive work performed to analyze the macroeconomic performance by means of fiscal decentralization.

That is also first attempt in which we only targeted those developed and developing counties which are categorized as federations and working under federalism. Therefore the panel used in the study is unique form the previous work.

5.4.1 Model for Time Series Data Analysis

To examine the short run and long run dynamic the study used a time series econometrics, such as the Co-integration Analysis, Error Correction Model and Granger Causality Analysis.

In time series analysis, testing stationary data is very important for reliable results because in

105

OLS estimation non-stationary variable produce spurious results (Granger and Newbold,

1974).

Stationarity of Data

To test the non-stationary of the variable, Dickey and Fuller (1979 and 1981) developed a procedure which is known as the Augmented Dickey-Fuller (ADF) test. So, to examine the stationary of the variables present study follows the ADF test as the error term is unlikely to be white noise so Dickey and Fuller extended their test that included extra lag terms of the dependent variable in order to eliminate the problem of autocorrelation.

It is very essential to determine an appropriate lag length to whiten the residuals. Akaike

Information Criterion (AIC) or Schwartz Bayesian Criterion (SBC) is used to determine the lag length on extra terms. The ADF test can be performed with intercept, trends and intercept and none of them. It is very important to consider that if all the data series in an observed model are stationary at the level with the same integration order, the application of co-integration will produce spurious results.

Johansen (1988 and 1991) introduced an approach to examine the co-integration in multiple equations; if the variable or series have a co-integration relationship then the linear relationship of these series would be stationary and yield long run relationships among the variables. Further, to generate the short run relationship, the Error Correction Model (ECM) is used and Vector Error Correction Model (VECM) is a conventional model which determines the correction from disequilibrium of the previous period.

106

Keeping in view these variables, we plan the methodology. The adaptive expectation model is one way of rationalizing the Koyck Model. Another rationalization is provided by Marc

Nerlove’s Partial Adjustment Model (PAM). The partial adjustment model can be explained as follows:

Partial Adjustment Model

The partial adjustment model has two parts that is static part and dynamic part. The static part represents determination of the desired amount and dynamic part is used to describe the dynamic partial adjustment process as follows:

z *     x  u t 0 1 t t * zt  zt1  (zt  zt1 )

If z* is the desired level of z we can obtain the following expression by substituting the value of z* into the other equation:

zt   0  (1 )zt1  1xt  ut

The general Auto Regressive Distributive Lag model equation can be written as:

zt  0  1zt1  2 xt  3 xt1  vt

If partial adjustment occurs the following restriction will be imposed:

3  0

The estimates of the parameters can be obtained which contain the desire level of z and the adjustment parameter λ. In the above case:

1  (1 )    (1 1)

2  1

0  0

107

The value of adjustment parameter λ lies between 1 and 0 and used to measure the speed of adjustment. It further shows that the adjustment parameter closer to 1 represents high speed of adjustment.

The Error Correction Model

Mostly in economic and financial time series for short-run dynamic Error Correction

Model is used. Error Correction Model without error correction term can be represented in first differences as follows:

zt  zt  zt1

Error Correction Model (ECM) is based on Auto Regressive Distributive Lag Model

(ARDL) and can be derived as follows:

yt  0  1zt1  2 xt  3 xt1  vt

If we subtract zt-1 from ARDL equation on both sides we will get the following equation.

zt  zt1  zt  0  1zt1  2 xt  3 xt1  zt1  vt

If we add and subtract 2 xt1 in the right hand side we obtained the following expression:

zt  0  1zt1   2 xt   2 xt1  3 xt1  zt1   2 xt1  vt

More properly we can write the above equation as follows:

zt  0  2xt  (1 1)zt1  (2  3 )xt1  vt

For ECM the coefficient of zt-1 should be equal to minus the coefficient of xt-1 that can be written as follows:

108

 1  (   ) 1 2 3 1  2  3  1

To apply ECM, the sum of coefficient will be equal to one in ARDL if we exclude the constant terms. In literature short term ECM is usually presented with τ as the coefficient of error correction term as given below.

z     x  (z  x )  u t 0 2 t t1 t1 t   (1 1)  ( 2  3 )

If the difference term is growing at a constant rate N then long-run relationship between z and x can be written as:

N   0   2 N  (z * x*)

 (z * x*)   0  ( 2 1)N   ( 1)N z*  0 2  x * 

If the original model without and with log is

z *t  kx *t

By taking log on both sides

log z *t  log k  log x *t

Then long run expression will be the anti-log of the above equation as follows.

  ( 1)N k  exp[ 0 2 ] 

109

Here k represents the long-run relationship between z and x. After examining the long run relationship we used the Error Correction Model (ECM) framework to determine the short run relationship. The ECM exhibits the introduction of past dis-equilibrium as an explanatory variable in the dynamic behaviour of an existing variable. It shows both short run as well as long run relationships among the variables. The private saving and its determinant can be expressed as:

푛

∆푁푡 = ∑ 휏1∆푁푡−1 + 휏2∆푁푡−1 + (푌푋푡) + 푢푡 푖=1

Here 푋푡 is a vector of exogenous variables and 휏1, 휏2, 푌 are vectors of parameters while 푢푡 denotes random disturbance term. Finally, the Granger Causality test has been applied to examine the causality of the selected variables.

110

Chapter 6

Fiscal Decentralization and Macroeconomic Performance: An Elementary Data Analysis

6.1 Introduction

In this chapter elementary data analysis will be performed. Before time series analysis it is necessary to take an overview some basic properties of the dependent and independent variables. The summary statistics provides and important information of data and it shows

Mean, Median, Maximum and Minimum values, skewness and kurtosis etc. The results of correlation matrix will also be reported in this chapter which is used to measure the strength between pair of variables and widely used by the analysts to identify the problem of

Multicollinearity in the estimated data series. Therefore, in present chapter we will discuss various data properties before performing the time series analysis and rest of the chapter will be organized as follows;

In section two we will discuss the summary statistics of the dependent and independent variables and section three will report the results of correlation matrix to avoid the problem of

Multicollinearity and finally section four will conclude the chapter.

6.2 Descriptive Statistics

To examine the impact of fiscal decentralization on economic growth the descriptive summary will be presented as follows.

111

Table 6.1: Descriptive summary of the variables (Period covered from 1972 to 2010)

Mean Median Maximum Minimum Std. Dev. Skewness Kurtosis GDP (Million) 4430925 834112 91921584 49784 14825664 5.620 3.366 INV (Million) 466797.9 177761 2414749 6521.000 640426.1 1.830 5.412

EPI 84.077 85.725 91.457 67.169 6.544 -0.965 2.972

INFR 9.600 8.650 30.000 3.100 5.759 1.907 7.163 ELF (Million) 27.8621 28.000 29.920 25.980 1.068 -0.023 2.146 PSAV (Million) 388862.8 146324. 18244239 6310.00 504703.1 1.430 3.388

FDX 0.255 0.248 0.368 0.192 0.039 0.559 2.956

AFDX 0.463 0.442 0.850 0.287 0.134 0.591 3.019

FDR 0.326 0.338 0.478 0.063 0.099 -1.538 5.072

AFDR 0.290 0.294 0.436 0.058 0.092 -1.038 3.997

OPEN 0.340 0.333 0.480 0.259 0.045 0.752 3.879 CAP (Million) 16.613 17.001 20.956 11.435 2.108 -0.323 3.366

LITR 36.913 35.450 58.000 19.600 12.518 0.228 1.675 GFI (Million) 392421.2 162861 1833861. 6547.000 500829.9 1.616 4.661

POP 111.45 112.07 163.07 63.34 30.687 0.175 1.771

Source: Authors own calculation by using E-view 5

112

The descriptive statistical summary of selected variables is given in table 6.1. The first column shows that the average GDP is Rs.4430925 million and Investment (INV) is

Rs.466797.9 million while average inflation rate (INFR) is 9.6 for our period of analysis. The variables for non-adjusted decentralized revenue (FDR) and non-adjusted decentralized expenditures (FDX) are 0.33 and 0.26 respectively on the average in Pakistan. According to the statistical analysis, the average values for trade openness (OPEN), adjusted variables of decentralization revenue (AFDR) and decentralization expenditures (AFDX) in Pakistan are

0.34, 0.29 and 0.46 respectively. The difference in mean value of adjusted fiscal decentralization and non-adjusted fiscal decentralization variables represent the adjustment effects in fiscal decentralization variables for Pakistan.

The skewness values of the variables are given in second last column of table 6.1 which indicated that almost all variables are a little bit skewed. The variables like, GDP, Public

Investment (INV), Inflation rate (INFR), private savings (PSAV), openness (OPEN), adjusted decentralized expenditures (AFDX), non-adjusted decentralized expenditures (FDX), gross fixed investment (GFI) and literacy rate (LITR) are positively skewed. It also shows that the variables for employed labour force (ELF), adjusted decentralized revenue (AFDR), non- adjusted decentralized revenue (FDR) and capital formation (CAP) are negatively skewed.

Last column of table 6.1 shows the kurtosis values. In statistical summary Kurtosis is used to measure the peakedness or flatness of the data relative to normal distribution. Normal distribution has Kurtosis value equal to 3 and if the value is great than 3 this indicate the probability distribution is high peaked and known as Leptokurtic and if the value of Kurtosis is less than 3 it is known as Platykurtic.

113

In our observed sample the variables like GDP, public investment (INV), inflation (INFR), private savings (PSAV) openness (OPEN), capital formation (CAP), non-adjusted decentralized revenue (FDR), adjusted decentralized revenue (AFDR) and gross fixed

Investment have Leptokurtic or high peaked probability distributions. It also shows that the variables like non-adjusted decentralized expenditure (FDX) and adjusted decentralized expenditures (AFDX) have an approximately normal distribution while employed labour force

(ELF) and Literacy rate (LITR) have relatively flatter probability distributions i.e. Platykurtic.

6.3 Results of Pair-wise Correlation

It is necessary to examine the correlation between the dependent and the independent variables. Usually, the pair-wise coefficient of correlation is used to identify the problem of multicollinearity between the variables. The high coefficient of correlation shows severe multicollinearity among the variables.

Table 6.2 shows a correlation matrix of all the variables included in present study. The correlation matrix shows the pair-wise correlation between the variables. It shows that the GDP is correlated with public investment (INV) and inflation rate (INFR) about 0.59 and 0.67 points in Pakistan and the strength of their correlation is bearable and less than 0.80 (which shows sever Multicollinearity problem). Table 6.2 also shows that GDP is correlated with employed labour force (ELF) and private savings (PSAV) but the strength of correlation is less than the critical value. Results also shows that the adjusted fiscal decentralization (AFDX, AFDR) and non-adjusted fiscal decentralization (FDX, FDR) variables are also correlated with GDP but there is a moderate correlation and there is no issue of Multicollinearity.

114

Table 6.2: Results of Correlation Matrix GDP INV INFR EPI ELF PSAV FDX AFDX FDR AFDR OPEN CAP LITR GFI POP GDP 1.00 INV 0.59 1.00 (0.00) INFR -0.67 -0.71 1.00 (0.00) (0.00) EPI 0.56 0.27 0.36 1.00 (0.00) (0.11) (0.02) ELF 0.36 0.55 -0.44 0.19 1.00 (0.08) (0.00) (0.06) (0.26) PSAV 0.73 0.37 -0.45 0.23 0.69 1.00 (0.00) (0.03) (0.01) (0.18) (0.00) FDX 0.15 0.14 -0.20 0.26 0.17 0.13 1.00 (0.34) (0.38) (0.21) (0.13) (0.30) (0.42) AFDX 0.57 0.77 -0.72 0.46 0.78 0.34 0.09 1.00 (0.00) (0.00) (0.00) (0.00) (0.00) (0.08) (0.56) FDR 0.68 0.66 -0.65 0.37 -0.63 0.74 0.01 0.62 1.00 (0.00) (0.00) (0.00) (0.02) (0.00) (0.00) (0.99) (0.00) AFDR 0.54 0.77 -0.78 0.34 -0.80 0.73 0.09 0.65 0.70 1.00 (0.00) (0.00) (0.00) (0.03) (0.00) (0.00) (0.57) (0.00) (0.00) OPEN 0.23 0.24 -0.10 0.22 0.03 0.26 0.11 0.25 0.35 0.22 1.00 (0.15) (0.13) (0.54) (0.14) (0.82) (0.11) (0.49) (0.11) (0.02) (0.18) CAP 0.43 0.32 0.63 0.15 0.51 0.16 0.18 0.74 0.59 0.76 0.08 1.00 (0.04) (0.04) (0.00) (0.44) (0.00) (0.23) (0.26) (0.00) (0.00) (0.00) (0.61) LITR 0.54 0.13 0.74 0.23 0.48 0.21 0.16 0.75 0.64 0.78 0.02 0.49 1.00 (0.00) (0.44) (0.00) (0.15) (0.01) (0.14) (0.33) (0.00) (0.00) (0.00) (0.90) (0.00) GFI 0.40 0.52 0.64 0.12 0.69 0.71 0.14 0.77 0.66 0.48 0.24 0.80 0.76 1.00 (0.03) (0.00) (0.00) (0.49) (0.00) (0.00) (0.38) (0.00) (0.00) (0.04) (0.14) (0.00) (0.00) 0.52 0.17 0.70 -0.33 0.46 0.22 0.18 0.70 0.41 0.71 0.53 0.59 0.29 0.71 1.00 POP (0.00) (0.44) (0.00) (0.05) (0.01) (0.13) (0.30) (0.00) (0.02) (0.00) (0.00) (0.00) (0.13) (0.00) Note: The Probability values are given into the parenthesis

115

6.4 Short run Causality Analysis

Granger (1969) introduced a simple test to examine the causality among the variables and the results of the Granger causality test are reported in table 6.3. According to the Pairwise

Granger causality test, a variable 퐴푡is said to be the Granger Cause of 퐵푡, if 퐴푡 can be predicted with greater accuracy by using the past value of 퐵푡 rather than not using such past values if all other terms remain constant. If two variables are uncorrelated the Granger Causality Test for two stationary variables 퐴푡 and 퐵푡 can be performed by selecting the optimal lag length (k=1) through AIC which can be found by following the Vector Auto Regressive (VAR) model.

However, the Vector Error Correction Mechanism is very significant method which provides the useful information about long run as well as short causality in the variables. To determine the short run causality between the variables Granger causality or block exogeneity Wald test which is based on Vector Error Correction model is performed to find the short run causality.

Therefore the results of the Granger Causality/ Block Exogeneity Wald test which is based on vector error correction are presented in table 6.3.

The results of Block Exogeneity Wald test show that there is bidirectional causality between adjusted decentralized revenue and inflation in Pakistan.

There is unidirectional causality between adjusted decentralized revenue and employed labour force and causality is running from adjusted decentralized revenue to employed labour force. Adjusted decentralized expenditures also have unidirectional causal relationship with employed labour force and again causality is running form adjusted expenditures decentralization to employed labour force. Therefore it could be concluded that fiscal decentralization my help to promote the level of employment in Pakistan.

116

Table 6.3: VAR Granger Causality/Block Exogeneity Wald Tests AFDR AFDX INFR ELF EPI FDR FDX GDP GTI LITR OPNS CAP SAPVT TGFI AFDR 0.41 11.42*** 8.68*** 6.97** 0.05 1.88 93.13*** 17.88*** 0.79 0.55 83.52*** 9.63*** 16.65*** -- (0.81) (0.00) (0.01) (0.03) (0.97) (0.53) (0.00) (0.00) (0.67) (0.75) (0.00) (0.00) (0.00) AFDX 10.66*** 10.72*** 7.86*** 6.09** 0.47 0.78 137.81*** 51.04*** 1.70 2.31 97.56*** 4.78* 37.18*** -- (0.00) (0.00) (0.01) (0.05) (0.79) (0.67) (0.00) (0.00) (0.42) (0.31) (0.00) (0.09) (0.00) INFR 5.83** 8.11*** 12.11*** 3.57 2.12 7.97*** 10.34*** 0.19 4.39 0.84 21.68*** 3.71 3.54 -- (0.05) (0.01) (0.00) (0.16) (0.34) (0.01) (0.00) (0.90) (0.11) (0.65) (0.00) (0.15) (0.17) ELF 4.32 1.20 9.93*** 9.31*** 3.64 8.24*** 67.44*** 0.85 0.69 3.09 10.72*** 8.88*** 1.44 -- (0.11) (0.54) (0.00) (0.00) (0.16) (0.01) (0.00) (0.65) (0.70) (0.21) (0.00) (0.01) (0.48) EPI 10.11** 3.31 5.13* 1.20 4.16 9.97*** 38.56*** 0.30 1.06 1.24 21.19*** 1.58 0.77 * -- (0.19) (0.07) (0.54) (0.12) (0.00) (0.00) (0.86) (0.58) (0.53) (0.00) (0.45) (0.67) (0.00) FDR 2.65 2.17 6.01** 2.02 6.82** 19.42*** 114.79*** 1.02 0.68 2.06 13.84*** 14.68*** 5.49* -- (0.26) (0.33) (0.04) (0.36) (0.04) (0.00) (0.00) (0.59) (0.71) (0.35) (0.00) (0.00) (0.06) FDX 1.57 3.86 0.80 0.72 9.41*** 2.94 9.31*** 0.90 0.86 2.23 54.82*** 0.44 0.57 -- (0.45) (0.14) (0.66) (0.69) (0.00) (0.22) (0.00) (0.63) (0.64) (0.32) (0.00) (0.80) (0.74) GDP 6.51** 9.25** 1.32 4.91* 3.59 5.29* 12.12*** 11.97*** 2.54 3.06 29.56*** 0.48 3.52 -- (0.03) (0.00) (0.51) (0.08) (0.16) (0.07) (0.00) (0.00) (0.28) (0.21) (0.00) (0.78) (0.17) GTI 50.48*** 3.46 21.29*** 1.71 2.30 0.17 11.36*** 64.92*** 0.97 0.95 9.93*** 6.17** 0.69* -- (0.00) (0.17) (0.00) (0.42) (0.31) (0.91) (0.00) (0.00) (0.61) (0.61) (0.00) (0.04) (0.70) LITR 1.66 0.87 8.32*** 2.03 6.89** 10.73*** 3.88 80.19*** 3.16 3.58 1.19 12.35*** 7.47** -- (0.43) (0.64) (0.01) (0.36) (0.03) (0.00) (0.14) (0.00) (0.20) (0.16) (0.55) (0.00) (0.02) OPNS 0.30 2.06 3.05 6.01** 0.30 0.20 5.26* 1.82 0.82 1.27 86.79*** 1.90 0.39 -- (0.85) (0.35) (0.21) (0.04) (0.85) (0.90) (0.07) (0.40) (0.66) (0.52) (0.00) (0.38) (0.81) CAP 3.37 3.61 17.22*** 3.01 0.43 0.93 0.76 41.29*** 0.56 0.57 2.95 5.50* 1.77 -- (0.18) (0.16) (0.00) (0.22) (0.80) (0.62) (0.68) (0.00) (0.75) (0.75) (0.22) (0.06) (0.41) SAPV 1.85 1.29 6.76** 6.53** 4.60* 1.88 1.50 32.19*** 0.45 1.16 0.25 35.45*** 2.02 -- T (0.39) (0.52) (0.03) (0.03) (0.10) (0.39) (0.47) (0.00) (0.79) (0.55) (0.87) (0.00) (0.36) TGFI 52.16*** 3.59 19.82*** 0.80 2.41 0.22 10.65*** 72.02*** 8.54*** 0.86 0.89 5.13* 6.19** -- (0.00) (0.16) (0.00) (0.66) (0.29) (0.89) (0.00) (0.00) (0.01) (0.64) (0.63) (0.07) (0.04) 1606.20** 331.98* 627.91** 78.70** 148.08** 271.40** 1528.68** 3829.87** 21.66** 33.15** 211.45** 245.79** 1843.78** 94.47*** ∑ 푿ퟐ * ** * * * * * * * * * * * (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.70) (0.15) (0.00) (0.00) (0.00)

117

The results also reported the unidirectional causality between decentralized revenue and economic performance index for Pakistan and direction of causality is running from decentralized revenue to economic performance. In other words revenue decentralization is helpful to raise the level of economic performance in Pakistan. Similarly adjusted expenditure decentralization also have unidirectional causal relationship with economic performance and direction of causality is running from expenditures decentralization to economic performance.

The results suggested that fiscal decentralization could be affective mean to boost the economic performance of Pakistan.

There is no causal relationship between adjusted and non-adjusted decentralized revenue and expenditure variables reported in the short run.

Bidirectional causality is also reported in adjusted decentralized revenue and economic growth variables. Economic growth and adjusted decentralized expenditures also have bidirectional causality.

Public investment have bidirectional causality with adjusted decentralized revenue while unidirectional causality from decentralized expenditures to public sector investment is observed in the short run. It means that fiscal decentralization is helpful to accumulate the account of public sector investment in Pakistan.

The results did not reported short run causal relationship between decentralization expenditures and trade openness. There is no short run causal relationship existed between adjusted fiscal decentralization variables and Literacy rate in Pakistan.

118

Decentralized revenue and stock of capital have unidirectional causality and causality running form decentralized revenue to capital formation which highlighted that revenue decentralization my help to generate capital in the economy. There is also unidirectional causality between decentralized expenditures and capital stock and directional of causality is running from decentralized expenditures to capital stock. Therefore it can be concluded that fiscal decentralization is helpful to raise the level of capital stock in Pakistan.

There is strong unidirectional causal relationship between decentralized revenue and private savings and causality is running for decentralized revenue to private savings while a weak causal relationship is reported between expenditure decentralization and private savings.

It could be suggested that fiscal decentralization (revenue and expenditure) may be helpful to raise the level of private savings. Finally, the results show that there exists bidirectional causal relationship between gross fixed investment and adjusted decentralized revenue while unidirectional causality from adjusted decentralization expenditures to gross fixed investment is running.

6.5 Conclusion

The chapter shows the time series to estimate the cointegration relationship between fiscal decentralization and macroeconomic indicators are normally distributed and also have short run causal relationship. The results of short run causality confirm the important implication of

Granger Causality test which suggested that if two time series are co-integrated then at least there exists a unidirectional causation. Therefore the results confirmed the presence of cointegration (stable long run relationship) between macroeconomic indicators and fiscal decentralization in Pakistan.

119

Chapter 7 Impact of Fiscal Decentralization on Growth, Investment and Inflation: A Time Series Analysis

7.1. Introduction

Fiscal decentralization is a successful tool to promote economic growth in developed or industrialized countries but its consequences in developing economies are still debatable. It is a general perception that fiscal decentralization is a successful way to increase the efficiency of public expenditure and revenue (Oates, 1972; Bahl and Linn, 1992; Yilmaz, 2000 and

Stansel, 2005). Fiscal decentralization is also a way to shift the fiscal authority to the lower tiers of the government and limited the control of central governments. So, mostly economies were not only inspired from the benefits of fiscal decentralization but they also wanted to get rid of the large centralized bureaucracies.

Over the past two decades mostly developing and transitional economies have taken steps towards fiscal decentralization. More importantly the notion of fiscal decentralization is not limited to the transitional or developing economies; it has also become a primary policy issue in the agenda of most OECD countries (Martinz-Vazquez and McNab, 2002).

Pakistan is also in the front line of developing countries which have taken various steps towards fiscal decentralization. The constitution of Pakistan gives various revenue instruments and expenditure responsibilities to the subnational governments. Provincial governments are free in their choices of functions and assignments to the local governments, however, the local governments are treated as a subject of provincial governments.

120

Therefore, the current chapter will examine the impact of fiscal decentralization on macroeconomic economic growth, investment and inflation in Pakistan. Firstly, section 7.2 will examine the impact of fiscal decentralization on economic performance of Pakistan.

Secondly, section 7.3 will examine the relationship between fiscal decentralization and economic growth. Thirdly, section 7.4 will develop a relationship between fiscal decentralization and investment and fourthly, section 7.5 will analyze the impact of fiscal decentralization on inflation. Finally, section 7.6 will draw the main conclusions of chapter 7.

7.2 Impact of Fiscal Decentralization on Economic Performance of Pakistan:

A Time Series Analysis

The following model will examine the impact of adjusted fiscal decentralization variables on macroeconomic performance of Pakistan. For macroeconomic performance they study will use economic performance index (EPI) as a dependent variable. Here inflation rate (LITR), employed labour force (ELF) and Population (POP) are used as controlled variables. The linear functional from of the model can be written as;

퐸푃퐼 =∝0+∝1 퐴퐹퐷푋 +∝2 퐴퐹퐷푅 +∝3 퐿퐼푇푅 +∝4 퐸퐿퐹 +∝5 푃푂푃 + 휇푖 → (7.1)

Estimation and Results:

To avoid the spurious regression results and to select the proper econometric methodology the Augmented Dickey Fuller (ADF) test for unit root will be applied. The result of ADF test is reported in table 7.1.1

121

Table 7.1.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root.

Results with intercept Results with trends and intercept

Variable Level 1st difference Conclusion Level 1st difference Conclusion

EPI -3.773 -- I(0) -3.825 -- I(0)

AFDX -0.028 -5.521 I(1) -1.723 -5.323 I(1)

AFDR -1.325 -8.963 I(1) -3.087 -9.040 I(1)

ELF -0.907 -5.954 I(1) -2.567 -6.057 I(1)

LITR -3.025 -- I(0) -4.707 -- I(0)

POP -2.961 -3.537 I(I) -2.80 -4.39 I(I)

The results of the ADF test are reported in table 7.1.1, which shows that the time series are not stationary at the similar order like EPI and LITR stationary at level with order I(0) while

AFDX, AFDR and POP are difference stationary with order I(I) . According to the ADF test we can easily concluded that the time series are not stationary at the same order.

According to the results of ADF test, the Auto Regressive Distributive Lag procedure will be followed to determine the co-integration relationship in equation 7.1.1. To test the presence of co-integration in selected time series Wald test will be applied.

7.2.1 The Wald Test (F-statistics)

First on lagged level variables used in equation 7.1 we have applied the Joint significance

F-test or Wald coefficient test and computed the F-statistics. In next step we compared the computed and tabulated F-statistics. Here it is important to describe that the chart for tabulated

F-statistics was developed by Pesaran et al. (2001) and they devised two critical bound called upper bound and the lower bound.

122

It had been proposed that if the computed F-statistics will likely to be excess of upper bound then there will be probably exists the long run relationship or the existence of cointegration in the variables. However, if the calculated F-statistics are less than the specified lower bound, it indicates the absence of long run relationship while the estimated F-statistics within the upper and lower bound will suggest inconclusive results.

The results of Wald Coefficient test on all lagged explanatory variables are reported in table 7.1.2. Here the null hypothesis is that there exists no long run relationship among the variables.

Table 7.1.2: Results of Bound Test for Co-integration Upper Bound Equation (7.1) F-statistics Conclusion Critical Value

Model-Equation 4.002 3.79 Co-integration EPI/AFDR,AFDX, LITR,ELF,POP [0.0006] (95%) exists

Note: Computed F-statistic is 4.002 which is Significant at 1 percent and shows the rejection of null hypothesis i.e. no long run relationship exists. Here, critical values at k=6-1=5 is cited by Pesaran et al. (1999). The values in parenthesis shows the probabilities of F-statistic.

The results reported in table 7.1.2 showed that the calculated F-statistics are higher than the upper bound value, as suggested before if the calculated F-statistics is higher than upper bound value confirmed the existence of long run relationship among the estimated model. Here we reject the null hypothesis and the results showed there exists long run relationship in our estimated model. Moreover the results reported in table 7.2.1 confirmed the variables in the estimated model are co-integrated and the estimation for long run relationship among the variables will not yield sporious results.

123

This next section will explore the Long run impact of expenditure decentralization and revenue decentralization on economic performance of Pakistan. Results are reported in Table

7.1.3.

7.2.2 Long- Run Impact of Fiscal Decentralization on Economic Performance

The following long run relationship indicates that fiscal decentralization (adjusted revenue decentralization and adjusted expenditure decentralization) has statistically significant impact on economic performance of Pakistan and these findings are reporting the actual insight.

Table 7.1.3: Long- Run Impact of Fiscal Decentralization on Economic Performance Dependent variable is EPI

Regressors Coefficient Standard Error T-Ratio

AFDR 0.97*** 0.31 3.19

AFDX 0.76*** 0.29 2.66

LITR 3.16** 1.69 1.86

ELF 1.33 1.45 0.91

POP -1.04** 0.56 -1.87

Note: (*) shows significance level at 10 percent, ** shows significance level at 5 percent and *** shows significance at 1 percent.

Table 7.1.3 shows the long run relationship between fiscal decentralization and economic performance of Pakistan. In this model we examined the impact of adjusted decentralized revenue (AFDR) and adjusted decentralized expenditures (AFDX) on economic performance

(EPI). While Literacy rate (LITR), employed labour force (ELF) and Population have been used as a control variables.

124

The results show that adjusted decentralized revenue has positive and statistically significant impact on economic performance of Pakistan in the observed time period. In simple words we can say that a unit increase in adjusted decentralized revenue will increase the economic performance 0.97 units by considering all other factors constant. So the findings support that the additional revenue autonomy at local level will help to accumulate the revenue at gross root level affectively and ensure the financing for new projects which may ultimately put an economy to the road to progress (Jin et al. 2005; Limi, 2005 and Gil-serrate and Lopez-

Laborda, 2006).

The results also reported the positive impact of expenditures decentralization on economic performance of Pakistan and shows that a unit increase in expenditures decentralization will help to stimulate economic performance about 0.76 units. The expenditure autonomy to lower tiers of the government permits the local governments to spend on those goods and services which are demanding by the local citizens. In other words the expenditure autonomy at local level will be supportive for the local governments to tailor the public good and services according to the demand of the local citizens. So the results confirmed that fiscal decentralization in term of revenue and expenditures will be supportive to raise the economic performance of Pakistan.

Apart from the variables of fiscal decentralization, other control variables also yielded very good results. According to the results, the value of population (POP) coefficient is -1.04 and it has posed significant but negative impact on economic performance of Pakistan in the long run.

125

7.2.3 Estimation of Error Correction Model

The results of error correction model has been given in table 7.1.4

Table 7.1.4: The Results of Error Correction Model ARDL(0,1,1,0,1,1) selected based on Schwarz Bayesian Criterion

Dependent variable is dEPI

Regressor Coefficient St. Error T-Ratio

dAFDR 0.83 0.89 0.09

dAFDX 0.49 0.13 3.76

dLITR 2.05 0.85 2.42

dELF 0.86 1.03 0.86

dPOP -0.68 0.32 -2.12

dC 40.64 12.05 3.37

ecm(-1) -0.65 0.14 -4.53

R-Squared 0.68 R-Bar-Squared 0.59

S.E. of Regression 4.72 F-stat. F(6, 29) 9.78 [0.00]

Akaike Info. Criterion -110.42 Equation Log-likelihood -102.42

DW-statistic 2.03 Schwarz Bayesian Criterion -116.75

The error correction model in ARDL frame work exhibits the short run and long run results of the estimated model. More importantly the term ecm (-1) represents the speed of adjustment of the estimated mode which is statistically significant and with negative sign. In other words the coefficient of ecm (-1) shows that 65 percent error will corrected from short run to long run equilibrium per year.

126

Overall the results of estimated model depicted the existence of long run as well as short run relationship between fiscal decentralization and economic performance of Pakistan. The results showed that over the long run, revenue decentralization and expenditure decentralization both have positive and statistically significant impact on macroeconomic performance of Pakistan. However in short run study did not find the statistically significant impact of revenue decentralization on economic performance of Pakistan. The results showed that fiscal decentralization is helpful to uplift the macroeconomic performance of Pakistan.

7.2.4 Diagnostic and Stability Tests

We have applied the synergy of diagnostic tests and also performed the stability test to examine the stability of the estimated model. The results of various diagnostic tests are reported in table 7.1.5

Table 7.1.5: Results of Diagnostic Tests

Test Statistics LM Version F Version Serial Correlation CHSQ (1) = 0.72 [0.39] F(1, 26)= 0.56 [0.46]

Functional Form CHSQ (1) = 0.29 [0.58] F(1, 26)= 0.22 [0.64]

Normality CHSQ (2) = 4.16 [0.12] --

Heteroscedasticity CHSQ (1) = 0.24 [0.62] F(1, 37)= 0.23 [0.63]

To examine the reliability of estimated results number of diagnostic tests have been applied and the results of these testes reported in table 7.1.5. First of all the results of Breusch-Godfrey

LM test are reported. This test is commonly used to identify the problem of serial correlation and the results showed that the probability value (0.39) is higher than 5 percent level of significance and therefore we reject the null hypothesis that there is problem of serial

127

correlation. So the rejection of null hypothesis confirmed that there is no serial correlation in the estimated model.

The model misspecification was also tested by applying the Ramsey reset test. The results of Ramsey reset test reported that the probability value is higher than 5 percent level of significance therefore we rejected the null hypothesis i.e. the model is miss specified. So it can concluded that the estimated model have no misspecification issue.

According to the assumptions of Classical Linear Regression Model the residuals should be distributed normally with constant variance. Jarque- Bera normality test was applied to examine the residuals normality. The results reported that the probability value are greater than

95 percent confidence interval and showed that the residuals are normally distributed in the estimated model.

Finally the Autoregressive Conditional Heteroskedasticity (ARCH) test was applied to examine the autocorrelation in the variance of error term. The probability value is higher than

5 percent level expressed that there exists no ARCH effect in the estimated model. The results of all the diagnostic tests confirmed the validity of the results of estimated model to examine the impact of fiscal decentralization on economic performance of Pakistan.

Stability Test

We have plotted the cumulative sum of recursive residuals CUSUM and cumulative sum of recursive residual square to examine the stability of the estimated coefficients. Figure 7.1 shows that graphs of CUSUM and CUSUMS are lies between the upper and lower critical bounds which ensure that the stability of the estimated model.

128

Figure 7.1 Stability Test CUSUM and CUSUMS

Plot of Cumulative Sum of Recursive Residuals

20 15 10 5 0 -5 -10 -15 -20 1976 1981 1986 1991 1996 2001 2006 2010 The critical bounds are at 5% significance level

Plot of Cumulative Sum of Squares of Recursive Residuals

1.5

1.0

0.5

0.0

-0.5 1976 1981 1986 1991 1996 2001 2006 2010

The critical bounds are at 5% significance level

129

7.3 Impact of Fiscal Decentralization on Economic Growth in Pakistan

In this section we will examine the impact of fiscal decentralization on economic growth in Pakistan by using a time series data from 1972 to 2010.

7.3.1 Fiscal Decentralization Expenditures and Productive Efficiency

Before presenting the empirical investigation we will analyze the basic hypothesis for implementing a successful fiscal decentralization process in Pakistan.

It is generally considered that the allocation of budgetary resources to non-productive lower tiers of governments will indicate that the local governments are inefficient to provide the public goods at local level. If local governments are less productive than central government then the idea of fiscal decentralization would be harmful for an economy. To analyze the efficiency study we will examine a trend between subnational expenditures to GDP ratio and the share of subnational governments’ expenditures in total expenditures (fiscal decentralization expenditures). The relationship between fiscal decentralization expenditure and efficiency of subnational governments’ expenditure can be examined by figure 7.2.

In the case of Pakistan, subnational expenditures to GDP ratio (FDX_GDP) and decentralized expenditures (FDX) both ratios have had an increasing trend for the period from

1972 to 1986 and from 1996 to 2010.When FDX_GDP decreased the FDX also decreased as it was the case for the period from 1987 to 1996. It also shows that the rise in decentralized expenditures has had a positive association with subnational expenditures to GDP ratio and vice versa. Here it is important to consider that if ratio of subnational expenditure to GDP is equal to 1, means that provincial governments are fully efficient. However in case of total

130

government expenditures to GDP, the ratio may not be equal to 1 because of some non- productive expenditures like debt servicing and defence expenditures. As sub-national expenditures are enumerator in both decentralization expenditures and subnational expenditures to GDP ratios therefore, the addition in sub-national government expenditures on one hand represents expenditure autonomy (rise in subnational expenditures share in total expenditures) and on other hand productivity of expenditures if we considered all other factors constant. The addition in these two ratios will enhance GDP if the expenditures will be productive. Thus, the addition in decentralization expenditures (by shifting budgetary resources from central government to productive subnational governments will positively influence GDP in Pakistan.

Figure 7.2 shows that subnational expenditures to GDP ratio (FDX_GDP) and the share of subnational government in total expenditure (FDX) have had an increasing trend for the period from 1972 to 1986 and from 1996 to 2010. When FDX decreased the FDX_GDP also decreased as it was the case for the period from 1987 to 1996. It also shows that the rise in the share of subnational government expenditures in total expenditures has had a positive association with subnational expenditures to GDP ratio and vice versa. It means that the addition in sub-national government expenditures will influence GDP accumulation positively.

Thus, the addition in fiscal decentralization expenditures (by shifting budgetary resources from central government to the subnational governments) will positively influence GDP in Pakistan.

131

Figure 7.2 Trends of FDX and GDP_FDX in Pakistan 0.35

0.3

0.25

0.2 FDX 0.15 FDX_GDP 0.1

0.05

7.3.2 Fiscal Decentralization Revenue and Productive Efficiency

For a successful revenue decentralization process, it is generally hypothesized that the increase in the share of subnational governments’ revenue in total revenue (decentralize revenue) will have a positive impact on GDP. However, the fiscal decentralization revenue contributes positively to GDP accumulation when the potential of revenue collected by lower tiers of the governments contribute for GDP.

Figure 7.3 shows that revenue decentralization (FDR) and the ratio of total tax revenue to

GDP (FDR_GDP) have increasing trends from the period of 1972 to 2010 but the curve of

FDR is much steeper than FDR_GDP’s curve which means that the marginal impact of revenue decentralization is less than unity. However, a continuous increase in curve of FDR_GDP means that the addition in revenue autonomy to subnational governments will contribute

132

positively to GDP. It also shows that revenue decentralization will prove helpful to raise the account of GDP in Pakistan.

Figure 7.3

Trends of FDR and FDR_GDP in Pakistan

0.5 0.45 0.4 0.35 0.3 0.25 FDR 0.2 FDR_GDP 0.15 0.1 0.05

2000 2006 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2002 2004 2008 1972

7.3.3 Impact of Fiscal Decentralization on Economic Growth in Pakistan: A Time Series Analysis

The following model will examine the impact of adjusted and non-adjusted fiscal decentralization variables on economic growth. For economic growth rate the study will use the log of annual GDP (LGDP) as a dependent variable. Here inflation rate (INFR), employed

(ELF) and capital formation (CAP) are controlled variables. The linear functional from of the model can be written as;

퐿퐺퐷푃 =∝0+∝1 퐴퐹퐷푋 +∝2 퐴퐹퐷푅 +∝3 퐹퐷푋 +∝4 퐹퐷푅 +∝5 퐼푁퐹푅 +∝6 퐸퐿퐹 +∝7 퐶퐴푃

+ 휇푖 → (7.2)

133

Estimation and Results:

To avoid the spurious regression results and to select the proper econometric methodology, first I will examine the stationarity level of the data series. The results for unit root test have been reported in table 7.2.1.

Table 7.2.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root.

Results with intercept Results with trends and intercept

Variable Level 1st difference Conclusion Level 1st difference Conclusion

LGDP -0.763 -4.983 I(1) -3.126 -4.865 I(1)

AFDR -1.325 -8.963 I(1) -3.087 -9.040 I(1)

AFDX 0.028 -5.521 I(1) -1.723 -5.323 I(1)

FDX -2.232 -4.948 I(1) -2.185 -5.012 I(1)

FDR -2.839 -5.807 I(1) -3.560 -5.788 I(1)

INFR -2.415 -6.261 I(1) -2.415 -6.172 I(1)

ELF 0.907 -5.954 I(1) -2.567 -6.057 I(1)

CAP -2.741 -5.807 I(1) -2.754 -5.719 I(1) Source: Authors own calculation by using E-view 5

The results of the ADF test are reported in table 7.2.1. The results show the presence of unit root at level while the null hypothesis of non-stationary can be rejected at first difference which confirms the difference stationarity in the time series. Thus the results of unit root test suggested the adoption of Johansen test procedure to examine the long run relationship

134

between dependent and independent variables. However it is necessary to identify the co- integration in the time series to determine the long run association between dependent and independent variables therefore the result of trace statistics to identify the cointegration relationship are reported in table 7.2.2.

Table 7.2.2: Result of Unrestricted Co-integration Rank Test (Trace)

Eigen value Trace Statistic 5%Critical Value Prob.** Hypothesized No. of CE(s)

0.926 238.788 * 159.529 0.000 None *

0.749 147.444 * 125.615 0.001 At most 1 *

0.605 99.039 * 95.753 0.029 At most 2 *

0.528 66.469 69.818 0.089 At most 3

0.376 40.140 47.856 0.217 At most 4

0.326 23.623 29.797 0.216 At most 5

0.154 9.797 15.494 0.296 At most 6

0.105 3.917 * 3.841 0.047 At most 7 * Note: * ( * ) indicates rejection of the hypothesis and indicates four co-integration equation at 5% significance level.

The results in table 7.2.2 show that the trace statistics are higher than 5% critical value which means the rejection of the null hypothesis. Finally one can conclude that there are no deterministic trends in the data and there exists cointegration in the observed time series and fortitude of the long run association of fiscal decentralization with economic growth will not yield sporious results.

135

All variables have significant relationships with the dependent variable i.e. LGDP. The results show that the adjusted and non-adjusted fiscal decentralization (revenue and expenditure) variables have a positive impact on economic growth in Pakistan. The growth rate also has significant and positive relationships with capital formation (CAP) and employed labour force (ELF) while negative relationship with Inflation rate (INFR).

Table 7.2.3: Normalize Co-integration Coefficients

Dependent Variable = LGDP

Independent Variables Coefficient St. Error t-statistics

AFDX 2.244*** 0.190 17.042

AFDR 4.897*** 1.179 6.695

FDX 5.559*** 0.490 11.339

FDR 5.5137*** 1.548 3.560

INFR -0.065*** 0.006 -10.796

ELF 0.0081*** 0.0012 6.494

CAP 7.148*** 0.707 10.106 Note: Here ( * ) shows 10% level of significance, ( ** ) shows 5% level of significance and ( *** ) shows 1% level of significance.

After examining the long run relationship between growth rate and fiscal decentralization, the study will apply the Vector Error Correction Model (VECM) and results can be expressed as:

136

Table 7.2.4: Result of Vector Error Correction Model

Dependent Variable = ΔLDGP

Error Correction: D(LGDP) St. Error t-statistics

C -0.044 0.121 -0.369

D(LGDP(-1)) 0.197 0.187 1.050

D(AFDX(-1)) 2.257 0.752 2.999

D(AFDR(-1)) 3.847 2.560 1.502

D(FDX(-1)) 1.392 2.194 0.634

D(FDR(-1)) 6.357 3.488 1.822

D(INFR(-1)) -0.006 0.016 -0.383

D(ELF(-1)) -0.014 0.040 -0.367

D(CAP(-1)) 0.156 2.494 0.062

ECt-1 -0.514 0.193 -3.685

Table 7.2.4 gives the result of VECM and reported that the statistically significant ECt-1 coefficient. More importantly, the coefficient of ECt-1 is reporting the speed of adjustment and negative sign showing the convergence of the estimated model from short run to long run equilibrium. The results show that 51% disequilibrium could be corrected from short run to long run each year and shows the long run causality between fiscal decentralization and economic growth in Pakistan.

137

Results and Discussions

The study provides empirical evidence for the long run relationship between economic growth and fiscal decentralization in Pakistan based on time series data for the period of 1972 to 2010.

Results indicate that fiscal decentralization is helping to improve the economic growth in

Pakistan. It means more autonomy for provincial or local governments on both the revenue and the expenditure side will promote economic growth. While the empirical results of present study differ from various studies that have shown a negative impact of fiscal decentralization on economic growth in developing economies (Davoodi and Zuo, 1998; Zhang and Zuo, 1998;

Phillips and Woller, 1997; Thode, 2001; Marinez-Vazquez and McNab, 2002; Naumets, 2003 and Jin and Zou, 2005).

In simple words, the results of our empirical investigation in table 7.2.3 shows that 1 unit increase in adjusted fiscal decentralization expenditure will induce to increase .0224 percent growth rate in Pakistan. The positive influence of decentralized expenditures on economic growth in Pakistan is harmonious with previous literature (Jin at al. 1999; Yilmaz, 2000; Qiao et al. 2002; Akia and Sakata, 2002; and Behnich et al. 2002). Results in table 7.2.3 also show that one unit increase in adjusted fiscal decentralization revenue will induce an increase of

0.048 percent growth rate in Pakistan (Desai et al. 2003; Feld and Kirchgassner, 2004;

Schaltegger, 2005; Jin et al. 2005; Limi, 2005 and Gil-serrate and Lopez-Laborda, 2006).

According to these findings the fiscal autonomy at local level is helping to raise the productivity of public sector and fostering economic growth Pakistan as Oates (1972) suggested.

138

The impacts of non-adjusted fiscal decentralization revenue and expenditure on economic growth are stronger than adjusted fiscal decentralization variables. The study shows that the non-adjusted fiscal decentralization variables overstated the impact of fiscal decentralization in Pakistan. It means that in the revenue side the presence of grants, federal transfers and shares of non-tax revenue is depicting a strong influence of revenue autonomy on economic growth, while in expenditure side the inclusion of payments of interest on debt and defense expenditures also shows stronger impact of expenditures autonomy on economic growth

(Mello and Barenstein, 2001).

Notwithstanding, we would caution that the impact of fiscal decentralization on economic growth could be declining with further adjustments in revenue and expenditures.

To capture the precise impact of fiscal decentralization on economic growth, the employed force (ELF), inflation rate (INFR) and capital formation (CAP) are controlled for GDP growth rate. The results show that inflation rate (INFR) has a negative association with GDP growth in Pakistan, while employed labor (ELF) and capital formation (CAP) have a strong positive association with GDP growth rate (Naqvi and Khan 1989; Gregorio, 1996; Frenkel and

Mehrez, 1998; Burdekin et al. 2000 and Ahmad and Mortaza, 2005).

Although the present model has been determined the direct positive influence of fiscal autonomy on economic growth, despite that the fiscal autonomy from provincial government to district government is still not perceptible. On one hand, the tax system in Pakistan is still centralize and the local governments have authorities of tax collections on behalf of central government. While on other hand the phenomena of shifting the fiscal resources from the central authorities to local authorities still seems challenging and complicated in Pakistan

139

Finally, it can be observed from the estimated results, the impact of decentralized revenue on economic growth is almost twice as compared to the decentralized expenditures and these findings advocate to further fiscal autonomy in revenue side. Different studies on fiscal decentralization highlighted that a balance between two sides of fiscal decentralization

(revenue and expenditure) could increase the welfare of the people and ensure the long run growth process in an economy. The fiscal autonomy on revenue and expenditure to the lower tiers of government could generate a competitive environment among the local governments

(Brennan and Buchanan, 1980). This competitive decentralization could take local governments towards the novelty in revenue generation and a better service provision. Thus, a balanced and competitive fiscal decentralization is a prerequisite to enhance the availability of public good according to the desire of local citizens and stimulus revenue generation by creating financial competition at local level and it would later account for higher GDP growth.

7.4 Impact of Fiscal Decentralization on Public Investment in Pakistan

In present section we will examine the impact of fiscal autonomy on public investment in

Pakistan by using a time series data from 1972 to 2010.

7.4.1 Fiscal Decentralization Expenditures and Productivity of Public

Investment

It is generally hypothesized that fiscal decentralization will promote the productivity of public investment by shifting the budgetary resources from central government to lower tiers of the governments. However, this ideology will be true only if the lower tiers of governments are efficient. If local governments are less productive than central government then the idea of fiscal decentralization would harm the outcomes of public sector investment. To analyze the

140

productivity of public sector investment, the study will examine a trend between public sector investment to GDP ratio and decentralized expenditures. The relationship between expenditure autonomy and public sector investment to GDP ratio can be examined by figure 7.4.

Figure 7.4

Trends of FDX and GTI_GDP in Pakistan

0.35

0.3

0.25

0.2 FDX 0.15 GTI_GDP

0.1

0.05

1980 1996 1974 1976 1978 1982 1984 1986 1988 1990 1992 1994 1998 2000 2002 2004 2006 2008 1972

Figure 7.4 shows that public sector investment to GDP ratio (GTI_GDP) and the share of subnational government in total expenditure (FDX) have an increasing trend for the period from 1972 to 1986 and from 2001 to 2010. When FDX decreased the GTI_GDP remain almost stagnant as it was the case for the period from 1985 to 1996. It shows that the rise in the level of expenditure autonomy has approximately positive association with public sector investment to GDP ratio. It means that the addition in sub-national government expenditures will raise

GTI_GDP and indirectly promote the account of GDP. Thus, the shifting of budgetary

141

resources from central government to the subnational governments shows an addition in public sector investment’s productivity and ultimately positively influence on GDP in Pakistan.

7.4.2 Fiscal Decentralization Revenue and Productivity of Public Investment

For revenue decentralization it is generally hypothesized that the increase in the share of subnational governments’ revenue in total revenue (fiscal decentralization revenue) may be helpful to enhance public sector investment. However, the public sector investment will contribute positively to GDP accumulation when the revenue collected by the governments would be invested in productive projects. In the case of Pakistan there is a continuous increase in fiscal decentralization revenue but the contribution of public sector investment to GDP is not that much responsive.

Figure 7.5

Trends of FDR and GTI_GDP in Pakistan

0.5 0.45 0.4 0.35 0.3

0.25 FDR 0.2 INV_GDP 0.15 0.1 0.05

1998 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 2000 2002 2004 2006 2008 1972

142

Figure 7.5 shows that the revenue decentralization (FDR) and the ratio of public sector investment to GDP (INV_GDP) have increasing trends but the curve of FDR is much steeper than INV_GDP’s curve which means that the marginal impact of public sector investment is minimal to accumulate GDP. However, an overall addition in INV_GDP means that the addition in revenue autonomy to subnational governments will increase public sector investment and promote GDP growth. However a clear picture of fiscal decentralization revenue (FDR) and public sector investment (INV) in Pakistan can be developed by analyzing a time series data as presented in the next section.

7.4.3 Fiscal Decentralization and Public Investment: A Time Series Analysis

The following model will inspect the effect of revenue and expenditure decentralization, on public Investment (LINV) in the presence of different controlled variables like employed labour force (ELF), Inflation rate (INFR) and capital formation (CAP), and the linear functional form can be written as:

퐿퐼푁푉 =∝0+∝1 퐴퐹퐷푋 +∝2 퐴퐹퐷푅 +∝3 퐹퐷푋 +∝4 퐹퐷푅 +∝5 퐿 퐸푃퐿 +∝6 퐼푁퐹푅 +∝7 퐶퐴푃

+ 휇푖 → (7.3)

Estimation and Results:

The ADF test procedure will be adopted to examine the presence of unit root in the time series the results are reported below in table 7.3.1.

143

Table 7.3.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root.

Results with intercept Results with trends and intercept Variable Level 1st difference Conclusion Level 1st difference Conclusion

LINV -2.277 -4.455 I(1) -3.040 -4.620 I(1)

AFDR -1.325 -8.963 I(1) -3.087 -9.040 I(1)

AFDX 0.028 -5.521 I(1) -1.723 -5.323 I(1)

FDX -2.236 -4.948 I(1) -2.185 -5.012 I(1)

FDR -2.839 -5.807 I(1) -3.560 -5.788 I(1)

ELF 0.907 -5.954 I(1) -2.567 -6.057 I(1)

INFR -2.415 -6.261 I(1) -2.415 -6.172 I(1)

CAP -2.741 -5.807 I(1) -2.754 -5.719 I(1)

To examine the stationarity the unit root test was performed at level and the results reported the existence of unit root in the time series. However when unit root test was performed at first difference then the results of ADF test reported that all the series are difference stationary as shown in above table. Therefore Johansen test procedure will be applied to diagnose the cointegration in the model 7.3.

144

Table 7.3.2: Result of Unrestricted Co-integration Rank Test (Trace)

Eigen value Trace Statistic 5%Critical Value Prob.** Hypothesized No. of CE(s)

0.904 237.397 * 159.529 0.000 None *

0.799 155.216 * 125.615 0.000 At most 1 *

0.624 98.998 * 95.753 0.029 At most 2 *

0.554 64.706 69.818 0.119 At most 3

0.395 36.395 47.856 0.376 At most 4

0.271 18.763 29.797 0.510 At most 5

0.170 7.683 15.494 0.499 At most 6

0.031 1.137 3.841 0.286 At most 7 Note: (*) indicates rejection of the null hypothesis and indicates three co-integration equations at the 5% significance level.

The results of trace statistical rejected the null hypothesis in first three rows where the value of trace statistics is greater than the critical value at 5 % which confirmed the existence of three cointegration relationships. So we can estimate the long run relationship between dependent and independent variables.

To examine the long run relationship of fiscal decentralization the co-integration vector has been normalized on the public investment and the results are reported in table 7.3.3. The results showed that all variables have significant relationships with public investment (LINV). The adjusted and non-adjusted decentralization revenues (AFDR and FDR) have a significant and positive relationship with public investment (LINV). In the expenditure side the adjusted and non-adjusted decentralization expenditures (AFDX and FDX) also have a positive relationship

145

with the dependent variable (LINV). In control variables the employed labour force (ELF) and

Capital formation (CAP) also have a significant and positive relationship with public investment (LINV) while inflation rate (INFR) has a negative relation with public investment

(LINV).

Table 7.3.3: Normalize Co-integration Coefficients

Dependent Variable = LINV

Independent Variables Coefficient St. Error t-statistics

AFDX 2.012*** (0.206) 9.749

AFDR 8.315*** (1.300) 7.554

FDX 2.885*** (0.447) 6.031

FDR 8.690*** (1.489) 5.833

ELF 7.371*** (0.680) 10.832

INFR -0.372*** (0.050) -7.424

CAP 6.698*** (0.904) 7.406 Note: Here ( *** ) shows 1% level of significance.

After examining the long run relationship the vector error correction mechanism for public investment and fiscal decentralization can be expressed as:

146

Table 7.3.4: Result of Vector Error Correction Model Dependent Variable = ΔLINV

Error Correction: D(LINV) St. Error t-statistics

C 0.125 0.038 0.038

D(LINV(-1)) 0.141 0.231 0.610

D(AFDX(-1)) 0.064 0.255 0.251

D(AFDR(-1)) 0.676 0.931 0.726

D(FDX(-1)) 0.371 0.749 0.495

D(FDR(-1)) 1.553 1.252 1.240

D(ELF(-1)) 0.042 0.073 0.574

D(INFR(-1)) -0.009 0.006 -0.145

D(CAP(-1)) -0.361 0.928 -0.391

ECt-1 -0.23 0.064 -3.593

Table 7.3.4 gives the result of Vector Error Correction Mechanism and showed that the

ECt-1 coefficient is statistically significant and also has the negative sign. The coefficient of

ECt-1 which is also known as speed of adjustment reported that 23% disequilibrium or error

147

will be corrected from short run to long run per year. It also confirms the long run causality between fiscal decentralization and public investment in the estimated model.

Results and Discussions

This model provides empirical evidence for the relationship between public investments, fiscal decentralization, employed labour force, inflation rate and capital stock. The empirical results show that public investment has a significant positive relationship with adjusted and non-adjusted fiscal revenue. It means that more autonomy for provincial or local governments on the revenue side will raise the public sector investment level. Results in table 7.3.3 show that a 1 unit increase in adjusted decentralized expenditure and revenue will lead to a 2 percent and 8.3 percent increase in public investment respectively. The correlation between non- adjusted fiscal decentralization (expenditure and revenue) and public investment is much stronger than adjusted fiscal decentralization variables. In the model we controlled for the employed labour force and expected the rise in employed labour force would lead to increase in public investment. Like previous studies, capital stock (CAP) and the consumer price index

(INFR) are also used as control variables (Rogoff, 2003 and Romer, 1993).

Although our model shows a positive correlation between fiscal decentralization and public investment some previous studies found a negative relationship (Sturm, 1999 and Valila,

2005). However, the former idea seems right, only when fiscal goals achieved by shifting public investment at local levels instead of increasing the taxes or reducing the government spending. In Pakistan the positive relationship between fiscal decentralization and public investment is quite straight forward because a big proportion of public investment is financed by the tax revenue. In Pakistan the positive association of fiscal decentralization with public

148

investment also shows the better provision of public goods and services and the improvement in infrastructure. Similarly the experience of Spain shows that fiscal autonomy at local level has proved helpful to enhance public investment, which has substantial improvement in the infrastructure (Esteller and Sole, 2005) or in Colombia and Bolivia, where the public investment has been an effective tool to improve the local needs (Faguet, 2005).

More importantly the coefficients of decentralized revenue (adjusted and non-adjusted) have a much healthier influence on public investment as compared to the expenditures decentralization. The visible difference between revenue and expenditure decentralization directs our attention toward public expenditures in Pakistan. The balance between fiscal decentralization revenue and expenditure (with no fiscal gap) is the prerequisite of successful decentralization. Therefore, to raise the level of public investment and to regenerate revenue through taxes, by following the distribution formulas, do not seem like an adequate tool. It could be demoralizing to the regions or states that have significance contribution to the national revenue because through the distribution formulas they would be able to capture a specified proportion of revenue. So, adopting the marginal retention rate will increase the marginal benefits of the states or regions in Pakistan. It means that the increase in revenue decentralization would increase the percentage of revenue retained by the regions and promote the marginal benefits of public investment (Cereaga and Weingast, 2003). Despite that, some studies suggested that the grants can be used to remove the revenue bias between the high and low revenue generating territories (Hindriks et al. 2008 and Ferreira et al. 2005). While in practice, the grants can diminish the marginal benefits of extra revenue generated by the subnational government who has to share that extra revenue with the rest of the country. In short, the marginal benefits of revenue generation can create a competitive environment among the

149

sub-national governments and it can promote public investment by means of fiscal decentralization. Fiscal decentralization also considered a successful tool to achieve the primary goal of fiscal policy like macroeconomic stability. Means high economic growth with lower inflation rate. Therefore, in the next section the study will explore a relationship between fiscal decentralization and inflation rate in Pakistan.

7.5. Impact of Fiscal Decentralization on Inflation in Pakistan

To analyze the impact of fiscal autonomy on price stability or inflation, a descriptive summary of selected variables, an efficiency analysis and a time series analysis will be presented as follows.

7.5.1 Productive Efficiency of Fiscal Decentralization (Revenue and

Expenditures) and Price Stability

The economies with high decentralization have lower inflation rate. It is generally hypothesized that fiscal decentralization will contribute to achieve price stability in an economy. In Pakistan, the price stability cure is explosive which indicates that the price stability is a challenging matter for her economy. Figure 7.6 shows there almost exists a negative association between productivity (contribution of fiscal decentralization expenditure in GDP accumulation) of fiscal decentralization expenditure and inflation rate but the relationship between fiscal decentralization revenue and inflation has no clear association.

However to develop a precise association between inflation rate and fiscal decentralization

(expenditures and revenue) the study will analyze the time series data of Pakistan for the period of 1972 to 2010.

150

Figure 7.6 Trends of FDX_GDP, FDR_GDP and Inflation in Pakistan

0.35

0.3

0.25

0.2 INFR 0.15 FDX_GDP FDR_GDP 0.1

0.05

1972 1990 2008 1976 1978 1980 1982 1984 1986 1988 1992 1994 1996 1998 2000 2002 2004 2006 1974

7.5.2 Fiscal Decentralization and Price Stability: A Time Series Analysis

Literature shows that number of studies that have tried to investigate the association fiscal autonomy and price stability concluded a positive association. According to these studies fiscal autonomy at local level could be helpful to achieve macroeconomic stability however these finding were limited to the developed economies (Rodden and Wibbels, 2002; Tanzi, 2000;

Qian and Rolland, 1998; Swell, 1996 and McLure, 1995). However, few of them suggested that fiscal decentralization has no significant relationship with price stability (Tresiman, 2000;

Roden and Wibbles, 2002 and Thornton, 2007). While some of them opposed the idea and presented fiscal decentralization as an obstacle to achieve economic stability (Prudhomme,

1995). From the previous debate it could be concluded that the efforts to explore the impact of fiscal autonomy to achieve macroeconomic stability are limited to developed economies.

151

Therefore in present section, to determine the effectiveness of fiscal decentralization to achieve macroeconomic stability will be tested empirically by using time series data of Pakistan.

For time series analysis annual change in the Consumer Price Index (CPI) which is generally known as inflation rate (INFR) will be considered as an independent variables.

Mostly, in economics literature the changes in prices level within a country or inflation rate is used as a measure of economic stability. The following model will examine the impact of fiscal decentralization, growth rate and trade openness on economic stability by using the equation

7.4. The linear functional from may be written as:

퐿퐼푁퐹푅 =∝0+∝1 퐴퐹퐷푋 +∝2 퐴퐹퐷푅 +∝3 퐹퐷푋 +∝4 퐹퐷푅 +∝5 (퐿 퐺퐷푃/푃푂푃) +∝6 푂푃퐸푁

+ 휇푖 → (7.4)

Estimation and Results:

Table 7.4.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root.

Results with intercept Results with trends and intercept

Variable Level 1st difference Conclusion Level 1st difference Conclusion

LINFR -2.415 -6.261 I(1) -2.415 -6.172 I(1)

AFDR -1.325 -8.963 I(1) -3.087 -9.040 I(1)

AFDX 0.028 -5.521 I(1) -1.723 -5.323 I(1)

FDX -2.236 -4.948 I(1) -2.185 -5.012 I(1)

FDR -2.839 -5.807 I(1) -3.560 -5.788 I(1)

LPCI -1.909 -5.995 I(1) -2.781 -5.019 I(1)

OPEN -2.741 -5.807 I(1) -2.754 -5.719 I(1)

152

The results in table 7.4.1 shows that the null hypothesis of non-stationary of data series can be easily rejected at 1% significance level when the ADF test was conducted at the first difference. So, we concluded that all the data series included in equation 7.4 are stationary at the first difference. Therefore we applied the Johansen co-integration procedure to examine the presence of co-integration in the estimated equation 7.4 and the results of trace statistics are presented in table 7.4.2.

Table 7.4.2: Result of Unrestricted Co-integration Rank Test (Trace)

Eigen value Trace Statistic 5%Critical Value Prob.** Hypothesized No. of CE(s)

0.847 174.951* 125.615 0.000 None *

0.686 109.150* 95.753 0.004 At most 1 *

0.590 68.502 69.818 0.063 At most 2

0.460 37.266 47.856 0.335 At most 3

0.242 15.661 29.797 0.735 At most 4

0.127 5.941 15.494 0.702 At most 5

0.033 1.184 3.841 0.276 At most 6 Note: * indicates rejection of the null hypothesis at the 5% level of significance.

The results of trace statistics reported the existence of cointegration because the value of trace statistics are higher than the critical value at 5% in the first two rows of table 7.4.2.

Results given in table 7.4.3 showed that all the variables have statistically significant long run relationships with inflation rate (LINFR). The adjusted and non-adjusted fiscal decentralization (expenditure and revenue) variables have a negative relationship with the

153

dependent variable i.e. inflation rate (LINFR). The per capita income also has statistically significant and negative relationships with the inflation rate. While the results show that the inflation rate has a positive and significant relation with trade openness.

Table 7.4.3: Normalize Co-integration Coefficients Dependent Variable = LINFR

Independent Variables Coefficient St. Error t-statistics

AFDX -1.070*** 0.118 -9.572

AFDR -1.527*** 0.294 -5.192

FDX -2.086*** 0.249 -8.362

FDR -2.569*** 0.487 -5.267

LPCI -1.079* 0.648 -1.665

OPEN 11.634** 0.717 5.064

Note: Here ( * ) shows 10% level of significance, ( ** ) shows 5% level of significance and ( *** ) shows 1% level of significance.

To results of the Error Correction Model (ECM) have been reported in table 7.4.4. The

ECM exhibits the negative association between the inflation rate and fiscal decentralization

(revenue and expenditure) variables.

Most importantly, the ECt-1 coefficient has the negative sign. It confirms the long run causality and also shows that the 5.9% disequilibrium from the short run to long run will be corrected each year.

154

Table 7.4.4: Result of Vector Error Correction Model

Dependent Variable = ΔINFR

Error Correction: D(INFR) St. Error t-statistics

C 0.068 0.484 0.140

D(LINFR(-1)) 0.203 0.163 1.244

D(AFDX(-1)) -3.916 7.180 -0.545

D(AFDR(-1)) -8.081 2.728 -2.961

D(FDX(-1)) -0.084 2.187 -0.038

D(FDR(-1)) -1.229 0.375 -3.276

D(LPCI(-1)) -2.736 1.928 -1.418

D(OPEN(-1)) 3.989 2.514 1.586

ECt-1 -0.059 0.011 -5.003 Source: Authors own calculation by using E-view 5

Results of the Model

To examine the relationship between economic stability and fiscal decentralization, the

Consumer Price Index (CPI), which is commonly known as the inflation rate (LINFR), has been used as a dependent variable. Like previous studies, we used the log of inflation rate to minimize the effect of big changes in the inflation rate (Rodden, 2002).

Results show that the inflation rate has a significant negative relationship with fiscal decentralization (adjusted and non-adjusted revenues and expenditures) in Pakistan. It is generally considered that there are a lot of indirect channels which may affect the economic stability, therefore the change in the degree of fiscal decentralization (revenue and expenditure) influence slightly to lessen the inflation rate. Table 7.4.3 shows that, the 1 unit change in

155

adjusted decentralized revenue (AFDR) and expenditure (AFDX) may lower the prevailing inflation rate (LINFR) 1.53% and 1.08% respectively. Table 7.4.3 showed that for non- adjusted fiscal decentralization variables, one unit increases in decentralized revenue (FDR) and expenditures (FDX) could induce to lower the prevailing inflation rate at the rate of 2.6% and 2% percent respectively (which is almost twice the adjusted fiscal decentralization variables).

These results could be different for different controlled variables. Moreover, these findings are limited to the time period of our sample. It is also clear from the results that the non- adjusted variables of fiscal decentralization (revenue and expenditure) show stronger influence to induce a decrease in the prevailing inflation rate in the economy of Pakistan. Here we will caution that these results are limited to the adjustments which we made with fiscal decentralization variables and the effect of fiscal decentralization could be minimized or even reversed with different adjustments and other controlled variables.

The negative association between fiscal decentralization and inflation rate also support the idea of other studies that countries with higher fiscal decentralization have relatively lower inflation rates. The significant results of our model shed some light on findings of previous studies, where it is stated that fiscal decentralization has no significant relationship with macroeconomic stability ((Tresiman, 2000; Roden and Wibbles, 2002 and Thornton, 2007).

Our results also oppose the ideas that fiscal decentralization is an obstacle to achieve economic stability (Prudhomme, 1995) or the positive association of fiscal decentralization with the inflation rate (Rodden and Wibbels, 2002; Tanzi, 2000; Qian and Rolland, 1998; Swell, 1996;

McLure, 1995; Barro and Gordon, 1983 and Kydland and Presscott, 1977). The empirical findings of our study highlighted that fiscal decentralization could be an effective tool for

156

Pakistan and for other developing economies to achieve macroeconomic stability (Martinz-

Vazquez and McNab, 2005; Neyapti, 2003 and King and Ma, 2001).

The model also includes the per capita GDP (PCI) and Trade Openness (the percentage change of total trade to GDP) to control the impact of per capita GDP and Trade Openness on the inflation rate like previous studies (Rogoff, 2003 and Romer, 1993). The classical views about this issue are that the macroeconomic policies should be the responsibility of central government rather than sub-national governments. But our findings support the idea that devolving of macroeconomic policies to subnational governments is not an obstacle in macroeconomic stability. However, this answer could be ineffective in the presence of poor institutional and political structure in the country of question. Generally, fiscal decentralization implies more responsibility and fiscal autonomy to the sub-national governments which could lower the control of central government. Due to this the macroeconomic conditions of the country could be less or more volatile. Our finding supports the idea that fiscal decentralization is an effective tool to describe a clear revenue sharing mechanism between federal and state governments. It means that there would be less competition for fiscal resources among the federal and state governments. So, fiscal decentralization will exert positive impact on economic stability and act as a catalyst to trickledown the effects of fiscal policy and promote the pro-cyclical fiscal policy.

The results for control variables show that there exists a negative relationship between the per capita GDP and the inflation rate. Table 7.6.4 reported that ‘1’ percent increase in per capita

GDP will reduce the prevailing inflation rate by 1.6%. These findings are similar to the previous studies (Thornton, 2007; Martinz-Vazques and McNab, 2005 and Neyapti, 2003).

157

7.6. Conclusion

The empirical results of aggregate analysis shows that fiscal decentralization is helpful to promote economic performance in Pakistan. Further the results of disaggregate analysis also confirmed that fiscal decentralization (revenue and expenditure) is an effective tool to enhance the economic stability, encourage public investment and foster economic growth in Pakistan.

In other words, fiscal decentralization has a significant and a positive influence on macroeconomic performance of Pakistan. These findings are matched with the majority of previous literature which stated the positive impact of fiscal decentralization on macroeconomic indicators (Naqvi and Khan, 1989; Gregorio, 1996; Frenkel and Mehrez,

1998; Burdekin et al. 2000; Lin and Liu, 2000; Zhang and Zou, 2001; Qaio et al. 2002; Ahmad and Mortaza, 2005 and Jin and Weingast, 2005). Although the empirical investigation shows that fiscal decentralization is beneficial for Pakistan, despite that, the results show the mismatch in expenditure and revenue assignments (Boadway and Habson, 1993). In a nutshell, the process of fiscal decentralization is favorable for encouraging economic growth in

Pakistan. Although in developing economies the numbers of constraints which curbed the effectiveness of fiscal decentralization were greater than in developed economies. Finally, to embark on a successful fiscal decentralization process, particularly in Pakistan, the fiscal balance in revenue and expenditure assignments is a must; otherwise its consequences may harm its economic performance. In Pakistan, lower levels of governments (provincial and local) still face the problem of centralized tax revenue. In addition the revenue is also distributed through the set formulas which snub the competitive environment among the local governments and also minimize the impact of fiscal decentralization in Pakistan.

158

Chapter 8 Impact of Fiscal Decentralization on Employment and Savings: A Time Series Analysis

8.1. Introduction

Fiscal decentralization is a potent enhancer of economic growth in developed countries but its impact in developing economies is still a question mark. In addition, fiscal decentralization is considered a way to check the central government from unnecessary defection from its normal path of financial management. In the words of Bird and Smart (2002), “for services to be effectively provided, those receiving transfers need a clear mandate, adequate resources and sufficient flexibility to make decisions”. Besides the efficient delivery of public goods, the consensus view is that fiscal decentralization may be a powerful tool to raise the level of employment, to support the public and private savings by improving the level of education and to achieve price stability. Some of those aims could be attained given that fiscal decentralization has some leeway to improve economic integration and educational activities which could positive influence the employment activities. In the present chapter, our study will explore the direct influence of fiscal decentralization on employment and private savings. The rest of the chapter will be organized as follows:

The first section will examine the relationship between fiscal decentralization and employment in Pakistan by using a time series data for the period from 1972 to 2010. The second section deals with the relationship between fiscal decentralization and private savings.

159

8.2 Impact of Fiscal Decentralization on Employment in Pakistan

We have carried out a descriptive summary of the key variables used in the model, a trend analysis and a time series analysis in order to explore the impact of fiscal decentralization on employment.

8.2.2 Productive Efficiency of Fiscal Decentralization (Revenue &

Expenditures) and Employed Labor Force

Fiscal decentralization can improve the level of employment by providing better education facilities to the local citizen. However, the direct impact of fiscal decentralization on employment is still undermined. Before carrying out the time series analysis, our study will explore a link between fiscal decentralization expenditures and revenue’s productivity with the employed labour force In Pakistan.

To examine the productivity of fiscal decentralization revenue, the ratio of fiscal decentralization revenue to GDP has been used. Figure 8.1 shows that there exist continuous increasing trends in fiscal decentralization revenue’s productivity (FDR_GDP) curve and employed labour force (ELF) curve for the time period from 1972 to 2010. According to this finding, it could be stated that in Pakistan, the rise in fiscal decentralization revenue’s productivity has shown a positive association with the employed labour force over the past three decades.

The employed labour force ELF has a positive association with the ratio of fiscal decentralization expenditure to GDP for the period from 1972 to 1986 and also from 2000 to

2010. But there exists a negative association between FDEX_GDP and ELF from 1987 to 1999

160

as figure 8.1 indicated. However, a clear picture of fiscal decentralization revenue and expenditures impact on employed labour force in Pakistan can be developed by analyzing a time series data for the period of 1972 to 2010.

Figure 8.1

Trends of FDX_GDP, FDR_GDP and ELF in Pakistan

0.5 0.45 0.4 0.35 0.3 FDX_GDP 0.25 FDR_GDP 0.2 0.15 %Elf 0.1 0.05

1992 1974 1976 1978 1980 1982 1984 1986 1988 1990 1994 1996 1998 2000 2002 2004 2006 2008 1972

8.2.3. Impact of Fiscal Decentralization on Employment in Pakistan: A Time

Series Analysis

For the empirical investigation, our study tried to analyze the effect of non-adjusted and adjusted fiscal decentralization variables on the employed labour force in Pakistan. Thus, the linear form of our models can be written as follows:

퐿퐸퐿퐹 =∝0+∝1 퐼푁퐹푅 +∝2 푂푃퐸푁 +∝3 퐿퐼푇푅 +∝4 퐿퐺퐹퐼 +∝5 퐹퐷푅 +∝6 퐹퐷푋 +

∝7 퐿퐸퐿퐹푡−1 + 휇푖 ……….. (8.1)

161

Instead of using the simple ratio, to capture the effect of fiscal decentralization on employment the linear functional form of our model with adjusted fiscal decentralization revenue (AFDR) and adjusted fiscal decentralization expenditures (AFDX) can be written as follows.

퐿퐸퐿퐹 = 훽0 + 훽1퐼푁퐹푅 + 훽2푂푃퐸푁 + 훽3퐿퐼푇푅 + 훽4푇퐺퐹퐼 + 훽5퐴퐹퐷푅 + 훽6퐴퐹퐷푋 +

훽7퐿퐸퐿퐹푡−1 + 휇푖 …………. (8.2)

The empirical findings of equations 8.1 and 8.2 are reported in table 8.1.1.

Table 8.1.1: Estimation Results for Short Run Dependent Variable: LELF VARIABLE OLS 8.1 OLS 8.2 INFR -0.003** 0.002** (-1.98) (-1.92) OPEN 0.285 0.338** (1.45) (1.83) LGFI 0.047* 0.059** (1.68) (2.03) LITR 0.004** 0.008*** (1.92) (3.33) FDR -0.280*** -- (-3.65) FDX 0.334* -- (1.73) AFDR -- -0.297*** (-4.67) AFDX -- 0.091** (1.94) LELF(-1) 0.447*** 0.24* (2.89) (1.58) Observations 38 38 R-squared 0.89 0.88 F-statistic 558.83 663.45 Durbin-Watson 2.16 2.27 Source: Authors own calculation by using E-view 5 Notes: Dependent variable Employed Labour Force (ELF) is the proportion of total labour force that fall in any category of employment such as self-employed, casual employed, salaried employed etc. It is taken in millions of people who are included in the labour force. Here t-statistics are in parenthesis and the coefficient is shown with stars. *** 1% Level of Significance; ** 5% Level of Significance; *10% Level of Significance.

162

Table 8.1.2: Estimation Results for Long Run

Dependent Variable: LELF

Variables OLS 8.1 OLS 8.2

INFR -0.006 -0.003

OPEN 0.552 0.447

LGFI 0.085 0.078

LITR 0.009 0.010

FDR -0.508 -

FDX 0.604 -

AFDR - -0.393

AFDX - 0.121 Source: Authors own calculation by using E-view 5 Note: Estimated value of coefficients of adjacent variables in the long run.

Results of the Model

Table 8.1.2 shows the regression results of the equations (8.1) and (8.2). The empirical investigation exhibits that the association between dependent and independent variables endorsed the proposed hypothesis i.e. fiscal decentralization expenditures have positive impact on employed labour force in Pakistan (Faridi, 2012 and Khan, 1989). In simple words a unit increase in fiscal decentralization expenditures will raise 0.344 percent in employment. The results also indicate that fiscally decentralized revenue has a negative impact on employment i.e. A unit increase in fiscal decentralization of revenue collection will decrease the employment at the rate of 0.28 percent. No doubt, in Pakistan taxes are the major source of

163

revenue therefore when the ratio of revenue will increase only by increasing tax rates, it may exert high tax burden on the producers which may decrease the industrial production.

Ultimately, lower production will crumble the employment creation. Nevertheless, when the degree of expenditure decentralization increases it will bring an increase in developmental projects. The increase in development projects will generate new employment opportunities and raise the level of employed labour force. Therefore, the first equation shows a positive relationship between decentralized expenditure and employment although this relationship is not very strong. The results also show the negative relationship between revenue decentralization and employment since the rise in revenue decentralization may reduce the level of employment. When given the option people will tend to prefer leisure over labour. The results of first equation show that FDR has a negative relationship with employment while

FDX indicated a statistically significant and positive relationship with employment. However to measure the intensity of fiscal decentralization by taking the ratios of provincial expenditures to total expenditures and provincial revenue to total revenue are not adequate

(Phillips and Woller, 1997). For instance, in Pakistan federal government collected all types of taxes that can create misunderstandings to specify variable of fiscal decentralization revenue.

Similarly to construct the variable of the fiscal decentralization expenditures we will consider only those expenditures which are the key responsibility of provincial government and federal government. Therefore, fiscal decentralization variables used in this study incorporate the possible adjustments to reflect the real impact of fiscal decentralization on employment.

Equation (8.2) exhibits the impact of adjusted fiscal decentralization’s variables on employed labour force in Pakistan. The results articulate a statistically significant (at ‘1’ percent) relationship between the adjusted variables of fiscal decentralization and employed

164

labour force. Empirically, a unit increase in adjusted fiscal decentralization expenditures

(AFDX) will stimulate an increase of 0.091 percent in employed labour force. It is clear that if government increases its expenditures after cutting defense expenditures will generate new employment opportunities and raise the level of employed labour force.

These findings also confirm the Oates’ (1972) arguments that fiscal decentralization is an effective way to enhance the long-term economic growth and development. In short, fiscal decentralization not only promotes the provision of public goods it is also a successful mode to deliver the public utilities like health, education and recreation facilities at a lower cost. In addition, the local or provincial governments usually have more checks and balances on the institution which tend to lead to a more producer efficiency. Creating these external and internal economies would increase production and economic growth and ultimately they would provide better opportunities for employment.

In the context of the previous literature on Leviathan and regarding the size of the public sector, our main result supports Oates (1985) views that the public sector employment tends to be larger with more fiscal decentralization. The results of this study further favor the argument

(Brennan and Buchanan, 1980) that fiscal decentralization promotes the competitive atmosphere among various levels of governments. Our results also support Lin and Liu (2000),

Akai and Sakata (2002), Thiessen (2003), Ebel and Yilmaz (2004), Limi (2005), and Gemmell et al. (2009) findings that fiscal decentralization positively influences economic growth and employment.

In order to control the robustness in the model we introduced some controlled variables like inflation rate (INFR) and the trade openness (OPEN). The study finds that the inflation

165

rate (INFR) has a significant negative relationship with employed labour force which is similar to the Philips Curve (which represents the trade-off between unemployment and inflation rate).

This finding shows that the policies adopted by the government to reduce or control the inflation rate bear a cost in terms of higher unemployment. In other words, the policies which hold the unemployment below the natural rate will induce an increase in the inflation rate. This implies that the addition in aggregate expenditures can increase the prices and wages and throw the markets out of equilibrium. On one hand when the firms will realize that their wage costs have been increased, they will increase the prices of their production while on other will hand, the workers will realize the purchasing power of their higher wages has been reduced, they would persistently demand higher wages. Inflationary expectations will take hold as a result wages and prices will push each other higher and higher. Only high unemployment for an extended period of time could be capable of bringing the inflation rate back to an acceptable level (Faridi, 2012; Ahmad and Mortaza, 2005; Burdekin et al. 2000; Frenkel and Mehrez,

1998; Naqvi and Khan, 1989 and De-Gregorio, 1996).

The results showed a positive and statistically significant association between trade openness (OPEN) and employment in our observed model. This relationship is quite simple and shows that the open economy has more employment opportunities as compare to the closed economy (Faridi et al. 2012). Other controlled variables such as literacy rate (LITR), total fixed investment (GFI) have statistically significant and positive impact on employment opportunities in Pakistan.

166

8.3 Impact of Fiscal Decentralization on Private Savings in Pakistan

To examine the direct impact of fiscal decentralization on private savings in Pakistan the descriptive summary of variables used in the model and a time series analysis can be presented as follows:

8.3.1 Fiscal Decentralization (Revenue & Expenditures) and Productive

Efficiency of Private Savings

There exists a positive relationship between the ratio of private savings to GDP and the share of subnational governments’ revenue in total revenue for the period from 1972 to 2010.

But figure 8.2 shows that the addition in the share of subnational governments in total revenue

(AFDR) is greater than the addition in the ratio of private savings to GDP (PSAV_GDP). It indicates that the fiscal decentralization revenue has a positive association with private savings but the share of private savings (which turned into investment in next fold) in GDP accumulation is minimal. Figure 8.2 shows no clear association between fiscal decentralization expenditures (FDEX) and the ratio of private savings to GDP (PSAV_GDP). However to develop a clear association between fiscal decentralization expenditure, fiscal decentralization revenue and private savings in Pakistan, a time series analysis would be presented in next section.

167

Figure 8.2

Trends of FDX , FDR and PSAV_GDP in Pakistan

0.5 0.45 0.4 0.35 0.3 FDX 0.25 FDR 0.2 PSAV_GDP 0.15 0.1 0.05

1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 1972

8.3.2. Impact of Fiscal Decentralization on Private Savings: A Time Series

Analysis

The following model will examine the impact of fiscal decentralization, growth rate and inflation rate on private saving. The natural log on both sides gives us a linear estimable function which may be written as:

퐿푃푆퐴푉 =∝0+∝1 퐿퐴퐹퐷푅 +∝2 퐿퐴퐹퐷푋 +∝3 퐿퐺퐷푃 +∝4 퐿퐼푁퐹푅 + 휇푖 → (8.3)

Estimation and Results:

To examine the short run and long run results of the model, the study will adopt these tests but to precede further the ADF test for unit roots is a pre-requisite for a time series analysis.

168

Table 8.2.1 reflects the results of ADF test which indicated that the null hypothesis of non- stationary cannot be rejected at levels. It means that the time series are not stationary at level however the logarithmic transformations of observed time series are stationary at first difference. So we concluded that the data became stationary at the first difference where ADF statistics are greater than 99% critical value as shown in table 8.2.1.

Table 8.2.1: Results of Augmented Dickey-Fuller Test (ADF) for Unit Root.

Results of unit root test with intercept Results of unit root test with trends and intercept

Variable Level 1st difference Conclusion Level 1st difference Conclusion

LPSAV -1.873 -7.798 I(1) -2.819 -8.072 I(1)

LAFDR -1.325 -8.963 I(1) -3.087 -9.040 I(1)

LAFDX 0.028 -5.521 I(1) -1.723 -5.323 I(1)

LGDP -0.763 -4.983 I(1) -3.126 -4.865 I(1)

LINFR -2.415 -6.261 I(1) -2.415 -6.172 I(1)

Note: the null hypothesis is that the series is non-stationary, or contains a unit root. The rejection of null hypothesis for the ADF test is based on the Mackinnon critical values 1%.

We also determined the optimal value of the Schwartz Bayesian Criterion (SBC) and the

Akaike Information Criterion (AIC). For this result we started with a large number of lags and then reduced down, and by doing this we found that the optimal lag length was lag-1. Then we applied the maximum likelihood based Johansen test procedure and determined the presence of the co-integration equation in a set of time series data. Moreover, to test the null hypothesis of the Z co-integration vector we used trace statistics as reported in table 8.2.2.

169

Table 8.2.2: Result of Unrestricted Co-integration Rank Test (Trace)

Hypothesized No. of Eigen value Trace Statistic 5%Critical Value Prob.** CE(s)

0.630 78.719* 69.818 0.008 None *

0.423 42.911 47.856 0.134 At most 1

0.403 23.095 29.797 0.241 At most 2

0.114 4.493 15.494 0.860 At most 3

0.002 0.107 3.841 0.742 At most 4 Source: Authors own calculation by using E-view 5 Note: * ( * ) indicates rejection of the hypothesis at the 5% significance level, L.R. test indicates one co-integration equation at the 5% significance level.

Here we observed the trace statistics, either it accepted or rejected the null hypothesis. As in the first model trace statistics are higher than the critical value at 5%. So, we reject the null hypothesis and proceed further in the second model and examine that the trace statistics are smaller than the critical value i.e. at 5%. Finally one can conclude that our model is most appropriate with restricted intercept and no deterministic trend in the data as shown in table

8.2.2.

To examine the magnitude and sign of a long run relationship and elasticity, the co- integration vector has been normalized on Private Saving (PSAV). It is concluded that there exists one co-integration relationship. The normalized co-integration coefficient can be observed in table 8.2.3. All variables have significant and elastic relationships with Private

Saving (PSAV). The adjusted revenue side (LAFDR) of fiscal decentralization has a positive relationship with private saving while the expenditure side has a negative relationship. The growth rate also has significant and positive relationships with private saving while the inflation rate has a negative relation with private saving.

170

Table 8.2.3: Normalize Co-integration Coefficients

Dependent Variable = LPSAV

Independent Variables Coefficient St. Error t-statistics

LAFDX 0.920*** 0.084 10.922

LAFDR -0.344*** 0.066 -5.182

LGDP 0.521*** 0.109 4.754

LINFR -0.639*** 0.047 -13.477 Note: Here ( * ) shows 10% level of significance, ( ** ) shows 5% level of significance and ( *** ) shows 1% level of significance.

After examining the long run relationship we used the Vector Error Correction Model

(VECM) framework to determine the long run causality. The VECM exhibits the introduction of past dis-equilibrium as an explanatory variable in the dynamic behaviour of an existing variable. It shows both short run as well as long run relationships among the variables. The private saving and its determinant can be expressed as:

푛 ∆푁푡 = ∑푖=1 휏1∆푁푡−1 + 휏2∆푁푡−1 + (푌푋푡) + 푢푡 (8.4)

Here 푋푡 is a vector of exogenous variables and휏1,휏2, 푌 are vectors of parameters while 푢푡 denotes random disturbance term. Table 8.2.4 gives the result of VECM which relate the error terms in the lagged period and change in LPSAV to changes in other variables.

The ECt-1 coefficient has the correct negative sign and the value of the coefficient shows the speed of adjustment. In our case it is observed at the rate of 46% which is quite respectable.

In simple words the 46% disequilibrium in private savings towards its long run level is corrected each year in Pakistan.

171

Table 8.2.4: Result of Vector Error Correction Model

Dependent Variable = ΔLPSAV

Error Correction: D(LPSAV) St. Error t-statistics

C 0.245 0.110 2.224

D(LPSAV(-1)) -0.039 0.284 -0.139

D(LAFDR(-1)) -0.233 0.298 -0.783

D(LAFDX(-1)) -0.069 0.149 -0.461

D(LGDP(-1)) -0.212 0.666 -0.319

D(LINFR(-1)) 0.080 0.164 0.490

ECt-1 -0.462 0.149 -3.100 Source: Authors own calculation by using E-view 5

Results of the Model

Table 8.2.4 provides empirical evidence for the relationship between private saving, fiscal federalism, economic growth and inflation rate. The empirical analysis of the study is based on Johansen’s co-integration and Error Correction Model for Pakistan’s time series data for the period of 1972 to 2010. Private saving has a positive relationship with adjusted fiscal revenue. It means more autonomy for provincial or local governments on the revenue side which will contribute more in private saving. In an economy taxes, both direct and indirect, are the main source of revenue. In Pakistan, taxes have a major effect on private saving but the positive effect emerges only when funds (collected in the form of taxes) are distributed by the state government at a grass roots level.

The results in table 8.2.4 show, if the degree of fiscal decentralization would be higher, the revenue collection process of the economy will be more efficient than the centralized system, because in Pakistan the local governments have autonomy only in the tax collection process:

172

they have no powers to specify tax rates and also to distribute the revenue independently.

Alternatively, the higher degree of fiscal decentralization will give more autonomy to the provincial or local governments to reinvest the collected revenue. This autonomy will encourage economic activities in the private sector and enhance private saving. As a result, in

Pakistan the promotion of fiscal decentralization will lead to significant contribution to private savings. Private saving has a negative relationship with adjusted fiscal expenditures. It suggests that the autonomy of local governments in expenditures will injure the private savings balance.

Mostly in developing countries the transfer received by the local governments from the central government to meet the expenditures of local citizens needs a clear mandate and the local governments must be accountable for results which is only possible when expenditure made by local governments must be properly designed and accountable (Bird and Smart, 2002). A negative relationship in expenditure autonomy to local governments and private saving emerges when central government gives power to local governments to meet the required expenditures of local citizens without proper accountability. Mostly in developing countries, due to poor quality administration, corruption and poor quality of projects, the outcome of funds invested by the local government is not up to the mark. As a result local governments demand more funds from the federal government, and an additional part of the expenditure that emerges due to fiscal autonomy comes from the public sector saving or from external savings. As a result the overall impact of expenditure autonomy will reduce the level of private saving. The negative relation of fiscal decentralized expenditure with private saving also highlights another important issue: if in an economy the intensity of non-productive expenditure is already higher, the further autonomy in expenditures for the local government will be more harmful to economic prosperity.

173

As results of the study show, private saving and growth are highly correlated over the long term. In different regions and stages of development saving is associated with a higher growth rate (Schmidt-Hebbel, 1996). As private saving in one fold will become the investment which generates economic activities and in the next fold encourages economic growth. Feldstein and

Horioka (1980) stated that the increase in the domestic savings rate induced approximately an equal increase in the domestic rate of investment and contributed to economic growth.

However, to understand this logic we must explain the sources of saving from which aggregate investment can be financed. As the three major segments of national saving consist of private sector saving which emerge due to households’ savings and apart from business enterprises, while the second part comes from public sector savings that reflect the difference between tax revenue and government expenditure, and the third part of saving is the foreign sectors saving

(it emerges only in an open economy) which indicates the net balance of trade. By the logic of our international economic relationships these three parts of national saving will contribute to overall economic growth of a country when they generate future investment. Although, in the case of developing economies the last two parts are not very supportive in enhancing the investment level (due to deficit financing by the government and a deficit balance of payments or mostly near to close economies) in the economy and may be one of the important reasons to lower economic growth.

Results show that the private saving has a negative relationship with the inflation rate in an economy. All of the income generated in any period is attributable to the household sector, since the owners of firms (whose income takes the form of profits) are also a part of that sector.

More importantly the income received by households may be classified into three major parts; a part of income spent on consumption goods, the part they required to pay taxes on and the

174

part that they save. So the third segment of household income will be directly influenced by the inflation rate and will be lowered or disappear when a higher inflation rate prevails in an economy. Mostly in developing countries the role of public sector saving in national saving is nominal while a massive part of national saving and its behaviour is reflected in private saving.

Therefore in an economy the private saving account will develop more swiftly when the inflation rate will be lowered. In short, the raise in inflationary pressures in an economy condenses the purchasing power of both public and private sectors, which compel both sectors to increase their spending level and lower their saving to acquire the same level of services and real goods. Inflation is thus a depressant to private saving.

8.4 Conclusion

The results show that fiscal decentralization (expenditure and revenue) is an effective tool to promote the employment level, to induce the rise in private savings and to encourage public sector investment. An increase in the subnational share of the total government expenditure would generate more development projects and it could create more employment opportunities in Pakistan. Specifically the fiscal decentralization expenditure is a means of development and it can bring a positive change in employment opportunities. However, the raise in the subnational revenue share in total government revenue has a negative association with employment. Mostly in Pakistan the local government has a mandate to collect revenues or taxes which are defined by central authorities. Most of the subnational revenue share is collected through taxes. It means that higher tax rates can lower the productive ability of local industries and it can increase the production cost which may end raising the level of unemployment in Pakistan.

175

It is also concluded from the results that the rise in the subnational share in total expenditure it can reduce the private sectors’ saving. On the revenue side, fiscal decentralization has had a positive impact on the level of private savings in Pakistan in the observed time period. These findings are encouraging to execute a fiscal decentralization process in Pakistan. So far, the findings of the study remain limited to investigating the impact of fiscal decentralization on various macroeconomic indicators in Pakistan and various adjustments were considered in a time series data (1972 to 2010) to develop that association.

176

Chapter 9

Summary, Conclusion and Recommendation

9.1 Summary

Zhang and Zou (1998), and Xie et al. (1999) found an inverse association between fiscal decentralization and economic growth. While other studies like Akia and Sakata (2002), Iimi

(2005), Lin and Liu (2000), Stansel (2005) and Yilmaz (1999) found a positive relationship between fiscal decentralization and economic growth. Besides that very little work has been done in developing and transitional economies to explore the connection between fiscal decentralization and economic performance. This association demanded more attention and research work to provide explicit guidelines for policy makers to design and propose successful completion of fiscal decentralization in developing economies like Pakistan.

To investigate the impact of fiscal decentralization on macroeconomic performance in

Pakistan study summarized the brief history of resource distribution among the provinces. It has been concluded that since 1991, the NFC awards have improved the process of resource

177

allocation to the Provincial Governments in Pakistan. Moreover, due to these awards direct shifting of funds and grants have been enhanced to all provinces. Likewise, the incentive of matching grants forced the provinces to bring improvement in their efficiency, self-reliance and resource generation and in turn obtain financial independence.

The empirical results show that fiscal decentralization (revenue and expenditures) is an effective tool to enhance the economic stability, encourage the public investment and foster the economic growth in Pakistan. Fiscal decentralization is also an effective tool to promote the employment level and induce the rise in private savings. The increase in subnational share of total government expenditures due to fiscal decentralization generate more development projects and create more employment opportunities in Pakistan. Specifically the fiscal decentralization expenditures are means of development and can bring positive change in employment opportunities. Contrary to this, the rise in subnational revenue share in total government revenue has negative association with employment. The major part of subnational revenue share is collected through taxes. It means that higher tax rates can have a discouraging impact on productive ability of local industries which may raise the level of unemployment in

Pakistan. In other words, fiscal decentralization has significant and positive influence on macroeconomic performance of Pakistan. These findings are matched with the majority of previous literature which stated the positive impact of fiscal decentralization on macroeconomic indicators (Naqvi and Khan, 1989; Gregorio, 1996; Frenkel and Mehrez,

1998; Burdekin et al. 2000; Lin and Liu, 2000; Zhang and Zou, 2001; Qaio et al. 2002; Ahmad and Mortaza, 2005 and Jin and Weingast, 2005). However, these studies also indicate that in developing economies the numbers of constraints which can curb the effectiveness of fiscal decentralization are greater than in developed economies.

178

The findings of the study revealed that the expenditures and tax revenue autonomy to subnational and local governments has a positive impact on macroeconomic performance of

Pakistan. This finding is also endorsed by the empirical findings of Oates’ (1999) which suggested that the provincial and local governments are efficient in both ways (expenditures and revenue side) to meet the efficiency property of decentralization.

9.2 Conclusion of the Study

Zhang and Zou (1998), and Xie et al. (1999) found an inverse association between fiscal decentralization and economic growth. While other studies like Akia and Sakata (2002), Iimi

179

To investigate the impact of fiscal decentralization on macroeconomic performance in

Pakistan study summarized the brief history of resource distribution among the provinces. It has been concluded that since 1991, the NFC awards have improved the process of resource allocation to the Provincial Governments in Pakistan. Moreover, due to these awards direct shifting of funds and grants have been enhanced to all provinces. Likewise, the incentive of matching grants forced the provinces to bring improvement in their efficiency, self-reliance and resource generation and in turn obtain financial independence.

Study also highlighted the shortcomings in NFC awards due to the lack of consensus among the provinces resulted into deadlocks most of the time. In Pakistan, where each province has a conflict of priorities the process of resource allocation has created the situation of tug of war among them. With the passage of time provinces have insisted mostly on adding up various criteria in the resource distribution formula. For instance, KPK (former NWFP) is famed for its backwardness and anxiety about the division of revenue claiming that their backwardness itself should demand for a greater share in funding. Sindh is in favor of the revenue creation criterion, while Baluchistan raised the notion that geographic area criterion takes precedence.

While, according to Punjab, agricultural production should be the prime criterion for distribution of resources to the provinces. Thus, due to these contradictory resource distribution approaches and the lack of agreement among all provinces they are inclined to not follow a single yardstick, which could have been the best possible solution. As a result, short term agreements have produced short term grants and awards, which proved to be beneficial for bigger provinces but not fruitful for smaller ones.

The empirical results show that fiscal decentralization (revenue and expenditures) is an effective tool to enhance the economic stability, encourage the public investment and foster

180

the economic growth in Pakistan. Fiscal decentralization is also an effective tool to promote the employment level and induce the rise in private savings. The increase in subnational share of total government expenditures due to fiscal decentralization generate more development projects and create more employment opportunities in Pakistan. Specifically the fiscal decentralization expenditures are means of development and can bring positive change in employment opportunities. Contrary to this, the rise in subnational revenue share in total government revenue has negative association with employment. Mostly in Pakistan the local government has a mandate to collect revenue or taxes which are assigned to them by central authorities. The major part of subnational revenue share is collected through taxes. It means that higher tax rates can have a discouraging impact on productive ability of local industries which may raise the level of unemployment in Pakistan. In other words, fiscal decentralization has significant and positive influence on macroeconomic performance of Pakistan.

9.3 Suggestions

For successful fiscal decentralization to strike a balance between revenue and expenditures is a prerequisite. In Pakistan, the rising gap between fiscal revenues and expenditures is also acting as a barrier to reap the fruit of fiscal decentralization. The government is heavily relying on indirect taxes to meet the financial requirements. Tax reforms of 2010-11 increased the revenue share through taxes but still there is a need to promote the transparent tax mechanism and fair collection and redistribution of the resources to lower tiers of government. To deal with ever rising national and international debts are still a daunting challenge for Pakistan’s economy, so there is an urgent need to control the fiscal deficit because without that it is unavoidable to depend upon debts servicing and to lower the fiscal deficits. On the one hand,

181

government is required to improve the administration and collection methods of revenue

through taxes. In Pakistan, lower levels of governments (provincial and local) still face the

problem of centralized tax revenue on one hand while the revenue is also distributed through

the set formulas which snub the competitive environment among the local governments and

also minimize the impact of fiscal decentralization in Pakistan. The main suggestions for policy

implication are elaborated as under.

i. First of all, federal government should evolve a cogent and practical formula for

distribution of divisible pool among provinces. This should be accomplished after

consultation with all stakeholders, especially provincial governments. Chambers of

Commerce and Industry, Major entrepreneurs (both public and private), prominent scholars

and economists, regional political leaders, leading tax payers (individual and institutions)

should also be involved in the process of consultation.

ii. Provincial as well as local governments should be made autonomous. For this purpose,

they may be empowered to generate resources and achieve the goal of self-reliance. iii. There is no harm in incorporating the lower tiers of governments in collection of taxes like

income tax, capital gain tax, sales tax etc. iv. Transparency should be given prime importance in fiscal transactions from top to bottom

and from bottom to top.

v. Lower tiers of governments should be given full powers in utilization of funds allocated to

them. Intervention by the central government should be minimized. This would help in

making lower tiers more responsible and more confident. Only checks and balances may

182

be introduced to discourage misappropriation, misuse and leakage of funds. At the same

time, this will definitely improve the trust level and bring harmony among all tiers of

governments. vi. All subjects promised in 18th amendment of the constitution must be immediately devolved

to provincial governments. vii. An efficient and smooth mechanism should be urgently evolved for transfer of all types of

royalties, grants and funds lying pending with the central government to provincial

governments.

183

References

Ahmad, E., Hewitt, D., and Ruggiero, E. (1997). Assigning Expenditure Responsibilities. In Fiscal Federalism Theory and Practice, ed. T. Ter. Minassian, 25-48. Ahmed, S., and Mortaza, G. (2005). Inflation and Growth in Bangladesh: 1981-2005. Working Paper No. 0604. www.bangladeshbank.org.bd/research/pau.html Akai, N., and Sakata, M. (2002). Fiscal Decentralization Contributes to Economic Growth: Evidence from State-Level Cross-Section Data for the United States. Journal of Urban Economics, 52, 93-108. Arghazi, M., and Henderson, J. V. (2005). Why Country Fiscally Decentralizing. Journal of Public Economics, 89, 1157-1189. Aronson, J. R. (1974). Financing Public Goods and the Distribution of Population in Metropolitan Areas: An Analysis of Fiscal Migration in the US and England. Economic Policies and Social Goals, London: Martin Robertson. 313-41. Aronson, J. R., and Schwartz, E. (1973). Financing Public Goods and the Distribution of Population in a System of Local Government. National Tax Journal, 26(2), 137-60. Bahl, R. W., and Linn, F. J. (1992). Urban Public Finance in Developing Countries. Oxford: Oxford University Press. Bahl, R., and Wallace, S. (2003). Fiscal decentralization: The provincial–local dimension. Public Finance in Developing and Transition Countries: Essays in Honor of Richard Bird, 5-34. Barankay, I., and Lockwood, B. (2007). Decentralization and the productive efficiency of government: Evidence from Swiss cantons. Journal of Public Economics, 91, 1197-1218. Barenstein, M., and de Mello, L. R. (2001). Fiscal decentralization and governance: A cross- country analysis. International Monetary Fund. Barro R. J., and Gordon, D. B. (1983). A Positive Theory of Monetary Policy in a Natural- Rate Model. NBER Working Papers 807. Barro, R. (1990). Government Spending in a Simple Model of Endogenous Growth. Journal of Political Economy, 98 (1), 103-117. Baskaran, T. (2011). On the Link between Fiscal Decentralization and Public Debt in OECD Countries. Public Choice, 145, 351–78.

184

Behnisch, A. Büttner, T. and Stegarescu, D. (2002). Public Sector Centralization and Productivity Growth: Reviewing the German Experience. ZEW Discussion Paper No. 02- 03, Mannheim. Berthold, N., Drews, S., and Thode, E. (2001). Die föderale Ordnung in Deutschland-Motor oder Bremse des wirtschaftlichen Wachstums?. Univ. Würzburg, Wirtschaftswiss. Fakultät. Bird, R. M. and Smart, M. (2002). Intergovernmental Fiscal Transfers: International Lesson for Developing Countries. World Development, 30 (6), 899-912. Bird, R. M., and Wallich, C. I. (1993). Fiscal Decentralization and Intergovernmental Relations in Transition Economies: Towards a Systematic Framework of Analysis. Policy Research Working Paper WPS 1122, World Bank, Washington, D.C. Bloom, E. D. (2004).The Effect of Health on Economic Growth: A Production Function Approach. World Development, 32 (1), 1-13. Boadway, R. W. (1979). Public Sector Economics, Cambridge, Mass,; Winthrop Publishers. Boadway, R.W., and Habson, P. (1993). Intergovernmental Finance in Canada, Toronto: Canadian Tax Foundation. Bodman, P. and Ford, K. (2006). Fiscal Decentralization and Economic Growth in the OECD, Unpublished Manuscript, University of Queensland, Brisbane. Bradford, D. F., and Oates, W. E. (1971). The Analysis of Revenue Sharing in a New Approach to Collective Fiscal Decisions. The Quarterly Journal of Economics, 85(3), 416-439. Brandt, L., and Zhu, X. (2000). Redistribution in a Decentralized Economy: Growth and Inflation in China under Reform. Journal of Political Economy, 108, 422-439. Brennan, G., and Buchanan, J. (1980). The Power to Tax: Analytical Foundation of a Fiscal Constitution, (New York: Cambridge University Press). Breton A., and Scott, A. (1978). The Economic Constitution of Federal States, (Toronto: University of Toronto Press). Breton, A. (1996). Competitive Governments. An Economic Theory of Politics and Public Finance, (New York: Cambridge University Press) Brueckner, J. K. (1999). Fiscal Federalism and Capitan Accumulation, Journal of Public Economic Theory, 1, 205-224.

185

Brueckner, J. K. (2004). Fiscal Decentralization with Distortionary Taxation: Tiebout vs. Tax competition, International Tax and Public Finance, 11, 133-153. Brueckner, J. K. (2006). Fiscal Federalism and Economic Growth, Journal of Public Economics, 90, 2107-2120. Brueckner, J. K. (2009). Partial Fiscal Decentralization, Regional Science and Urban Economics, 39, pp. 23-32. Buchanan, J. M., and Goetz, C. J. (1972). Efficiency Limits of Fiscal Mobility: An Assessment of the Tiebout Model, Journal of Public Economics, 1, pp.25-43 Buchanan, J. M., and Wagner, R. (1970). An Efficiency Basis for Federal Fiscal Equalization, in J. Margolis (ed.), The analysis of Public Output, New York: Columbia University Press. Burdekin, R. C. K., Denzau, T., Keil, M. W., Sillihyot, T., and Willett, T. (2004). When Does Inflation Hurt Economic Growth? Different Nonlinearities for Different Economies. Journal of Macroeconomics, 26(3), 519-532. Caldwell, J. (1976). “Toward a Restatement of Demographic Transition Theory.” Population and Development Review, 2(3/4). Careaga M., and Weingast B. R. (2003) Fiscal Federalism, Good Governance, and Economic Growth in Mexico Dani Rodrik (Ed.), Search of Prosperity: Analytic Narratives on Economic Growth, Princeton University Press. Chaudhry, K. A. (1989). The price of wealth: Business and state in labour remittance and oil economies. International Organization 43(1), 101-145. Chaudry-Shah, A. (1988). Capitalization and the Theory of Public Finance: An Interpretive Essay. Journal of Economic Surveys, 2(3), 209-45. Davoodi, H., and Zou, H. (1998). Fiscal Decentralization and Economic Growth: A Cross- Country Study. Journal of Urban Economics, 43, 224-257. Demello, J.R. (2000). Fiscal Decentralization and Intergovernmental Fiscal Relations: A Cross-Country Analysis. World Development, 28 (2), 365-380. Desai, R.M., Freinkman L.M., and Goldberg I. (2005). Fiscal Federalism in Rentier Regions: Evidence from Russia. Journal of Comparative Economics, 33, 814–34. Desai, R.M., Freinkman L.M., and Goldberg, I. (2003). Fiscal Federalism and Regional Growth: Evidence from the Russian Federation in the 1990s, World Bank Policy Research Working Paper 3138, Washington, D.C.

186

Dickey D. A., and Fuller, W. A. (1979). Distribution of the Estimators for Autoregressive Time Series with Unit Root. Journal of American Association, 74, 427-431. Ebel, R. D., and Yilmaz, S. (2002). On the Measurement and Impact of Fiscal Decentralization. World Bank Study 2809. Policy Research Working Paper. Edel, M., and Sclar, E. (1974). Taxes, Spending and Property Values: Supply Adjustment in a Tiebout- Oates Model. Journal of Political Economy, 82(5), 941-54. Eller, M. (2004). The Determinants of Fiscal Decentralization and its Determinants on Economic Growth: Empirical Evidence from a Panel of OECD Countries, Diploma Thesis. Esteller, A., and Solé, A., (2005). Does Decentralization Improve the Efficiency in the Allocation of Public Investment? Evidence from Spain. Working Papers 2005/5. Institut d'Economia de Barcelona (IEB) Facchini, G., and Testa, C. (2008). Fiscal Decentralization, Regional Inequality and Bail-outs: Lessons from Brazil's Debt Crisis. The Quarterly Review of Economics and Finance, 48(2), 333-344. Faguet, J.P. (2005). Governance From Below A Theory of Local Government With Two Empirical Tests, STICERD-Political Economy and Public Policy Paper Series 12, Suntory and Toyota International Centers for Economics and Related Disciplines, LSE. Faridi, M. Z. (2011). Contribution of Fiscal Decentralization to Economic Growth: Evidence from Pakistan. Pakistan Journal of Social Sciences (PJSS), 31(1), 1-13. Faridi, M. Z., Chaudhry, I. S., Hanif, I., & Ansari, F. N. (2012). Fiscal Decentralization and Employment in Pakistan. International Business Research, 5(11), 54-64. Faridi, M. Z., Chaudhry, M. A., & Ansari, F. N. (2012). The Impact of Fiscal Decentralization, Inequality and Poverty on Employment: Evidence from Pakistan. Pakistan Journal of Social Sciences (PJSS), 32(2), 357-369. Feld, L. P., Kirchgassner, G., and Schaltegger, C.A. (2003). Decentralized Taxation and the Size of Government, Evidence from Swiss Sub-Federal Governments, CE Sifo paper No.1087, Munich. Feld, L.P., Kirchgässner, G., and Schaltegger, C.A. (2005). Fiscal Federalism and Economic Performance: Evidence from Swiss Cantons. Review of Regional Research, 25, 3-25. Feldstein, M., and Horioka, C. (1990). Domestic Savings and International Capital Flows. The Economic Journal, 90, 314-329.

187

Feltenstein, A and Iwata, S. 2005: Decentralization and macroeconomic performance in China, regional autonomy has its costs. Journal of Development Economics 76(2), 481–501. Ferreira, S.G., Varsano, R., and Afonso, J.R., (2005). Inter-jurisdictional Fiscal Competition: a Review of the Literature and Policy Recommendations. Revista de Economia Política, 25, 295–313. Finance Division Government of Pakistan. Federal Budget Speech (2007-08) Flatters, F., Henderson, V., and Mieszkowski, P. (1974). Public Goods, Efficiency, and Regional Fiscal Equalization. Journal of Public Economics, 3(2), 99-112. Freinkman, L., and Plekhanov, A. (2009). Fiscal Decentralization in Rentier Regions: Evidence from Russia. World Development 37(2), 503-512. Frenkel, M., and Mehrez, G. (1998). Inflation and Endogenous Technological Growth. Europa-Inst., Abt. VWL. Frenkel, M., and Mehrez, G. (2000). Inflation and the Misallocation of Resources. Economic Inquiry, 38(4), 616-628. Frenkel, Michael, and Mehrez, G. (1997). The Misallocation of Resources of Anticipated Inflation, Macroeconomics, EconWPA, Fukasaku, K., and De Mello, L. R. (1999). Fiscal Decentralization and Emerging Economic Governance Issues. Paris: Inter-American Bank/ OECD Development Centre. Galor, O. and Weil, D. (1996). The Gender Gap, Fertility, and Growth. American Economic Review, 86, 374-387. Galor, O. and Weil, D. (2000). Population, Technology and Growth: From Malthusian Stagnation to the Demographic Transition. American Economic Review, 110, 806-828. Gemmell, N., Kneller, R., and Sanz, I. (2009). Fiscal Decentralization and Economic Growth in OECD Countries: Matching Spending with Revenue Decentralization. P. T. N. 06/09. Ghaus, A., Pasha, A., and Hafiz, A. (1994). Dynamic Budgetary Consequences of the 1991 NFC Award. The Pakistan Development Review, 33(4), 627-645. Gil-Serrate, R., and López-Laborda J. (2006). Revenue Decentralization and Economic Growth in the Spanish Autonomous Communities. Unpublished Manuscript, University of Zaragoza. Gordin, J. P. (2004). Testing Riker's Party-Based Theory of Federalism: The Argentine Case. Publius: The Journal of Federalism, 34(1), 21-34.

188

Govt. of Pakistan (2005). Hand Book of Statistics on Pakistan Economy 2005. State Bank of Pakistan. Govt. of Pakistan (Various Issues). Pakistan Economic Survey 1991. Federal Bureau of Statistics, Islamabad. Gramlich, E. M. (1993). A Policymaker’s Guide to Fiscal Decentralization. National Tax Journal, 46 (2), 229-235. Gramlich, E. M., and Rubinfeld, D. L. (1982). Micro Estimates of Public Spending Demand Functions and Tests of the Tiebout and Median Voter Hypotheses. Journal of Political Economy, 90 (3), 536-560 Granger, C.W.J. (1969). Investigation Casual Relationship by Econometric Model and Cross- spectral Methods. Econometrica, 35, 199-211. Granger, C.W.J., and Newbold, P. (1974). Spurious Regression in Econometrics. Journal of Economics, 2, 111-120. Gregorio, D. (1996). Borrowing Constraints, Human Capital Accumulation, and Growth. Journal of Monetary Economics, 37, 49-72. Guess, G. M. (2005). Comparative Decentralization Lessons from Pakistan, Indonesia, and the Philippines. Public Administration Review, 65(2), 217-230. Hamilton, B.W. (1976). The Effect of Property Taxes and Local Public Spending on Property Values: A Theoretical Comment. Journal of Political Economy, 86 (3). 647-650. Hausman, J. A. (1978). Specification Tests in Econometrics. Econometrica, 46 (6), 1251-1271. Hausman, J. A. and Taylor, W. E. (1981). Panel Data and Unobservable Individual Effects. Econometrica, 49 (6), 1377-1398. Hayek, F.V. (1939). The Economic Condition of Interstate Federalism, in Friedrich von Hayek, Individualism and Economic Order (Chicago: University of Chicago Press). Hendricks, J., and Myles, G. (2006). Intermediate Public Economics, Cambridge Mass: MIT Press. Herath, T. N. (2009). Decentralization of Governance and Economic Development: The Sri Lankan Experience after Establishment of Provincial Councils. South Asia Economic Journal, 10(1), 157-185. Hernández-Trillo, F., and Jarillo-Rabling, B. (2008). Is Local Beautiful? Fiscal Decentralization in Mexico. World Development, 36(9), 1547-1558.

189

Hindriks, J., and Myles, G. D. (2006). Intermediate Public Economics (p. 16). Cambridge: MIT press. Hindriks, J., Peralta, S., and Weber, S. (2008). Competing in Taxes and Investment under Fiscal Equalization. Journal of Public Economics, 92(12), 2392-2402. Holcombe, R. G. (1989). The Median Voter Model in Public Choice Theory. Public Choice, 61(2), 115-125. Hughes, G., McCormick, B., and McCormick, B. (1987). Housing Markets, Unemployment and Labour Market Flexibility in the UK. European Economic Review, 31(3), 615-641. Huther, J., and Shaw, A. (1998). Applying a Simple Measure of Good Governance to the Debate on Fiscal Decentralization. World Bank Policy Research Working Paper, No. 1894 Iftikhar, A., Mustafa, U., and Khalid, M. (2007). National Finance Commission Awards in Pakistan, A Historical Perspective. Pakistan Institute of Development Economics, Islamabad. (PIDE Working Paper Series No. 33). Iimi, A. (2005). Decentralization and Economic Growth Revisited: An Empirical Note. Journal of Urban Economics, 57, 449-461. Iqbal, N., Musleh-ud-din, & Ghani, E. (2013). Fiscal Decentralization and Economic Growth: Role of Democratic Institutions. Pakistan Institute of Development Economics, Working Paper No. 2013:89. Jaffery, B. N., and Sadaqat, M. (2006). NFC Awards Commentary and Agenda. Pakistan Economic and Social Review, 44(2), 209–234 Jimoh, A. (2003). Fiscal policy and Growth in Africa: Fiscal Federalism, Decentralization and the Incidence of Taxation. Economic Commission for Africa Jin, H., Qian Y., and Weingast, B.R. (2005). Regional Decentralization and Fiscal Incentives: Federalism, Chinese Style. Journal of Public Economics, 89, 1719-1742. Jin, H., Qian, Y., and Weingast, B. R. (1999). Regional Decentralization and Fiscal Incentives: Federalism Chines Style. Department of Economics, University of Maryland. Jin, J. and Zuo, H. (2002). How Does Fiscal Decentralization Affect Aggregate, National, and Subnational Government Size?. Journal of Urban Economics, 52, 270-293. Jin, J., and Zou, H. (2005). Fiscal Decentralization, Revenue and Expenditure Assignments, and Growth in China. Journal of Asian Economics,16, 1047-1064.

190

Jing, j., and Zou, H. (2002). How Does Fiscal Decentralization Affect Aggregate, National, and Subnational Government Size?. Journal of Urban Economics, 52(2), 270-293. Johansen, S. (1988). Statistical Analysis of Co-integration Vectors. Journal of Economic Dynamic Control, 12, 231-254. Johansen, S. (1991). Estimation and Hypothesis Testing of Co-integration Vectors in Gaussian Vector Autoregressive Models. Econometrica, 59, 1551–1580. Johansen, S., and Juselius, K. (1990). Maximum Likelihood Estimation and Inference on Co- integration with Application for the Demand for Money. Oxford Bulletin of Economics Statistics, 52, 169-210 Jutting J.P. (2009). Fiscal Decentralization, Chinese Style: Good for Health Outcomes?. World Development, 37(12), 1926-1934. Kappeler, A., and Välilä, T. (2008). Fiscal Federalism and the Composition of Public Investment in Europe. European Journal of Political Economy, 24, 562–570. Khan, M. A. (1999). The Mordukhovich Normal Cone and the Foundations of Welfare Economics. Journal of Public Economic Theory, 1(3), 309-338. Khan, M. H., and Sundaram, J. K. (Eds.). (2000). Rents, Rent-seeking and Economic Development: Theory and Evidence in Asia. Cambridge University Press. Khattak, N. U. R., Ahmad, I., & Khan, J. (2010). Fiscal Decentralization in Pakistan. The Pakistan Development Review, 49(4-II), pp-419. Khramov, V., and Lee, J. R. (2012). The Economic Performance Index: An intuitive indicator for assessing a country’s economic performance dynamics in an historical perspective. Working Paper WP/13/214, International Monetary Fund. King, D., and Ma, Y. (2001). Fiscal Decentralization, Central Bank Independence and Inflation. Economics Letters, 72(1), 95-98. Kitunzi, A.n.d.(1998). Fiscal Decentralization in Developing Countries: An Overview. WorldBank,Washington,DC.(wbln0018.worldbank.org/network/prem/premdoclib.nsf/58 292ab451257bb9852566b4006ea0c8/8aa17d08e94022da852567e10009805c/) Kydland, F. E., and Prescott, E. C. (1977). Rules Rather Than Discretion: The Inconsistency of Optimal Plans. The Journal of Political Economy, 473-491.

191

Kyriacou, A., and Roca-Sagalés, O. (2011). Fiscal and political decentralization and government quality. Environment and Planning C: Government and Policy, 29 (2), 204- 223. Lardy and Nicholas, R. (1998). China’s Unfinished Economic Revolution. Brooking Institution Press, Washington, DC. Legler, J. B. (1990). Fiscal Federalism: Quantitative Studies. Edited by Harvey S. Rosen. Chicago: University of Chicago Press, 1988. Pp. ix, 262. $39.95. The Journal of Economic History, 50(01), 261-262. Lin, J. Y., and Liu, Z. (2000). Fiscal Decentralization and Economic Growth in China. Economic Development and Cultural Change, 1, 1-22. Lowery, D., and Lyons, W.E. (1989). The Impact of Jurisdictional Boundaries: An Individula Level Test of the Tiebout Model. Journal of Politics, 51, 73-97. Lucas, E. Jr. (1973). Some International Evidence on Output-Inflation Tradeoffs. American Economic Review, 63(3), 326-334. Malik, S., Mahmood-ul-Hassan, & Hussain, S. (2006). Fiscal Decentralization and Economic Growth in Pakistan. The Pakistan Development Review, 4(45), 845-854. Marlow, M. L. (1995). Fiscal Decentralization and Government Size. Public Choice, 56, 259- 269. Martinez-Vazques, J., and Mcnab R. M. (2001). Fiscal Decentralization and Economic Growth, International Studies Program Working Paper No.01-01, Atlanta: Andrew Young School of Policy Studies, Georgia State University. Martinez-Vazquez, J., and Jameson Boex, L. F. (2001). Russia’s Transition to a New Federalism. Washington, DC: World Bank, Institute Learning Resources Series. Martinez-Vazquez, J., and McNab R. M. (2006). Fiscal Decentralization, Macro Stability and Growth. Review of Public Economics, 179 (4), 25-49. Martinez-Vazquez, J., and McNab, R. M. (2002). Cross-Country Evidence on the Relationship between Fiscal Decentralization, Inflation, and Growth, in: National Tax Association (ed.), Proceedings of the 94thAnnual Conference on Taxation 2001, Washington, D.C, 42-47. Martinez-Vazquez, J., and McNab, R. M. (2003). Fiscal Decentralization and Economic Growth. World Development, 31(9), 1597-1616.

192

Martinez-Vazquez, J., and McNab, R. M. (2003b). Fiscal decentralization, macrostability and growth (No. 11-03 Classification-JEL: E62, H77, O20. O40.). Instituto de Estudios Fiscales. McLure, C. E., Jr. (1995). The Dangers of Decentralization: Comment. World Bank Research Ob•server, 10 (2), 221-226. Mello, L. (2000). Fiscal Decentralization and Intergovernmental Fiscal Relations: A Cross Country Analysis. World Development, 28 (2), 365-380. Meloche, J.P. (2004). Decentralization or Fiscal Autonomy? What Does Really Matter?. Policy Research Working Paper 3254, The World Bank. Musgrave, R. A., and Musgrave, P.B. (1959). Public Finance in Theory and Practice. New York: Mcgraw Hill. Naqvi, S. N. H., and Khan, A. H. (1989). Inflation and Growth: An Analysis of Recent Trends in Pakistan. Pakistan Institute of Development Economics. Naughton, B. (1995). Growing Out of the Plan: Chines Economic Reform1978-1993. Cambridge University Press, New York. Naumets, I. (2003). Fiscal Decentralization and Local Public Sector Efficiency. Unpublished Dissertation, National University of Kiew. Neyapti, B. (2003). Budget Deficits and Inflation: the Roles of Central Bank Independence and Financial Market Development. Contemporary Economic Policy, 21(4), 458-475. Neyapti, B. (2004). Fiscal Decentralization, Central Bank Independence and Inflation: A Panel Investigation. Economics Letters, 82(2), 227-230. Neyapti, B. (2005). Equalization via Fiscal Decentralization. Working Papers 2005/11, Turkish Economic Association. Neyapti, B. (2010). Fiscal Decentralization and Deficits: International Evidence. European Journal of Political Economy, 26, 155-166. Oates, W. (1972). Fiscal Federalism. Hamshire, England: Harcourt Brace Jovanovich 1972. Oates, W. (1985). Searching for Leviathan: An Empirical Study. American Economic Review, 75, 748-49. Oates, W. (1993). Fiscal Decentralization and Economic Development. National Tax Journal, 46, 237-243.

193

Oates, W. E. (1999). An Essay on Fiscal Federalism. Journal of Economic Literature, 37 (3), 1120-1149. Oates, W.E. (1969). The Effects of Property Taxes and Local Public Spending on Property Values: An Empirical Study of Tax Capitalization and the Tiebout Hypothesis. Journal of Political Economy, 77(6), 957-971. Ostrom, E., and Whitaker, G.P. (1973). Does Local Community Control of Police Make a Difference? Some Preliminary Findings. American Journal of Political Science, 17, 48-76. Ostrom, E., Schroeder, L., and Wynne, S. (1993). Institutional Incentives and Sustainable Development: Infrastructure Policies and Perspective. Boulder, Co West View Press. Ottensmann, J. R. (1982). Neighborhood Heterogeneity within an Urban Area. Urban Studies, 19, 391-395 Özcan, G. B. (2000). Local Economic Development, Decentralization and Consensus Building in Turkey. Progress in Planning, 54(4), 199-278. Panizza, U. (1999). On the Determinants of Fiscal Centralization: Theory and Evidence. Journal of Public Economics, 74, 97-139. Parks, R. B., and Oakerson, R. J. (1993). Comparative Metropolitan Organization-Services Production and Government Structures in St Louis (MO) and Allegheny-County(PA). Pblius-The Journal of Federalism, 6(3), 231-242. Pauly, M. V. (1976). A Model of Local Government Expenditure and Tax Capitalization. Journal of Public Economics, 6 (3), pp.231-242 Persson, T., and Tabellini, G. (1994). Is Inequality Harmful for Growth?, American Economic Review, 84(3), 600-621. Pesaran, M. H., and Smith, R. P. (1998). Structural analysis of cointegrating VARs. Journal of Economic Surveys, 12, 471-505. Pesaran, M. H., Shin, Y., and Smith, R. J. (2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics, 16, 289–326. Phillips, K. L., and Woller, G. (1997). Does Fiscal Decentralization Lead to Economic Growth?. Department of Economics, Brigham Young University. Prudhomme, R. (1995). Conflict and Dilemmas of Decentralization. Paper Presented at The Annual Bank Conference on Development Economics, World Bank, Washington, DC.

194

Qian, Y., and Weingast, B. R. (1997). Federalism as a Commitment to Reserving Market Incentives. Journal of Economic Perspectives, 11 (4): 83-92. Qiao, B. Martinez-Vazquez, J., and Xu Y. (2002). Growth and Equity Tradeoff in Decentralization Policy: China‘s Experience. Georgia State University, International Studies Program, Working Paper 02-16, Georgia. Rivlin A. (1992). Reviving the American Dream, The Economy, the States, and the Federal Government. Brookings Institution, Washington, D.C (1992) Rodden, J. (2002). The Dilemma of Fiscal Federalism: Grants and Fiscal Performance Around The World. American Journal of Political Science, 46, 670-687. Rodden, J., and Wibbels, E. (2002). Beyond the Fiction of Federalism: Macroeconomic Management in Multitier Systems. World Politics, 54 (4): 494-531. Rogoff, K. (2003). Globalization and Global Disinflation. Paper Presented at the Federal Reserve Bank of Kansas City conference on Monetary Policy and Uncertainty: Adapting to a Changing Economy, Jackson Hole, Wyoming. Romer, D. (1993). Advanced Macroeconomics, McGraw–Hill, New York, Rondinelli, and Dennis, A. (1981). Government Decentralization in Comparative Perspective: Theory and Practice in Developing Countries. IRAS, 2, 133-145. Sabir, M., and Khan, M. A. (2001). Dynamic Consequences of the 1997 NFC Award: Provincial Social Sector Expenditures [with Comments]. The Pakistan Development Review, 967-984. Samuelsson, P.A. (1954). The pure Theory of Public Expenditure. Review of Economics and Statistics, 36(4), 387-409 Schaltegger, C. A. and Feld, L. P. (2009). Are Fiscal Adjustments Less Successful in Decentralized Governments?. European Journal of Political Economy, 25, 115-123. Schmidt-Hebbel, K. (1996). Fiscal Adjustment and Growth: In and Out of Africa. Journal of African Economies, 5 (3), 7-59. Schmidt-Hebbel, K., Servén, L., and Solimano, A. (1996). Saving and Investment: Paradigms, Puzzles, Policies. The World Bank Research Observer, 11(1), 87-117. Sewell, D. O. (1996). The Dangers of Decentralization According to Prudhomme: Some Further As•pects. World Bank Research Observer, 11(1), 143-150.

195

Shah, A. (2004). Fiscal Decentralization in Developing and Transition Economics: Progress, Problems, and the Promise. World Bank Policy Research, working Paper No. 3282. Smith, A. (1776). An Inquiry into the Nature and Causes of the Wealth of Nations. Oxford World's Classics. Smoke, P., and Lewis, B. D. (1996). Fiscal Decentralization in Indonesia: A new Approach to an Old Idea. World Development, 24(8), 1281-1299. Stansel, D. (2005). Local Decentralization and Local Economic Growth: A Cross-Sectional Examination of US Metropolitan Areas. Journal of Urban Economics, 57, 55-72. Stegarescu, Dan, Büttner, Thiess, and Behnisch, Alexej, (2002). Public Sector Centralization and Productivity Growth: Reviewing the German Experience, ZEW Discussion Papers 02- 03, ZEW - Zentrum für Europäische Wirtschaftsforschung / Center for European Economic Research. Sturm, J. E., Jacobs, J., and Groote, P. (1999). Output Effects of Infrastructure Investment in the Netherlands. Journal of Macroeconomics, 21(2), 355-380. Sylla, R., Wallis, J., and Legler, J. (1995). Historical Economics: US State and Local Government. NBER Reporter, 14–16. Tanzi, V. (1996). Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomic Aspects, Annual World Bank Conference on Development Economics, The World Bank, Washington, DC. Tanzi, V. (2001). Pitfalls on the Road to Fiscal Decentralization. Carnegie Endowment for International Peace. Thiessen, U. (2003). Fiscal Decentralization and Economic Growth in High-income OECD Countries. Fiscal Studies, 24 (3), 237-274. Thornton, J. (2007). Fiscal Decentralization and Economic Growth Reconsidered. Journal of Urban Economics, 61(1), 64–70. Thun, E. (2004). Keeping up With the Jones: Decentralization, Policy Imitation, and Industrial Development in China. World Development, 32(8), 1289-1308. Tiebout, C. (1956). A Pure Theory of Local Government Expenditures. Journal of Political Economy, 416-424. Treisman, D. (2000). Decentralization and inflation: Commitment, Collective Action, or Continuity. American Political Science Review, 94 (4), 837-857.

196

Trillo, F., and Rabling, B. (2008). Is Local Beautiful? Fiscal Decentralization in Mexico. World Development, 36 (9), 1547-1558. Uchimura, H., and Jütting, J. P. (2009). Fiscal Decentralization, Chinese Style: Good for Health Outcomes?. World Development, 37(12), 1926-1934. Välilä, T., Kozluk, T., and Mehrotra, A. (2005). Roads on a Downhill: Trends in EU Infrastructure Investment. EIB Papers, 10, 18-40. Wasylenko, M. (1987). Fiscal Decentralization and Economic Development, Public Budgeting & Finance, 7 (4), 57-71. Woller, G.M., and Phillips K. (1998). Fiscal Decentralization and LDC Economic Growth: An Empirical Investigation. Journal of Development Studies, 34, 139-148. Wooldridge, J. M. (2002). Econometric Analysis of Cross-Section and Panel Data, MIT Press, Cambridge. World Bank (2000). China: Managing Public Expenditures for Better Results Washington DC. World Bank (2001). World Bank, 2001. China: Provincial Expenditure Review Washington D.C. Xie, D., Zou, H., and Davoodi, H. (1999). Fiscal Decentralization and Economic Growth in the United States. Journal of Urban Economics, 45, 228-239. Yakita, A. (2003). Taxation and Growth with Overlapping Generations. Journal of Public Economics, 87, 467-487. Yilmaz, S. (2000). The Impact of Fiscal Decentralization on Macroeconomic Performance, in: National Tax Association (ed.), Proceedings of the 92nd Annual Conference on Taxation 1999, Washington, D.C, 251-260. Yusuf, S. (1994). China’s Macro-economic Performance and Management during Transition. Journal of Economic Perspectives, 8, 71-92. Zhang and Xiaobo (2006). Fiscal Decentralization and Political Centralization in China: Implications for Growth and Inequality, Journal of Comparative Economics,34(4),713-726. Zhang, T., and Zou, H. (1998). Fiscal Decentralization, Public Spending, and Economic Growth. Journal of Public Economics, 67, 221-240. Zhang, T., and Zou, H. (2001). The Growth Impact of Inter-sectoral and Intergovernmental Allocation of Public Expenditure: With Applications to China and India. China Economic Review, 12, 58-81.

197