The impacts of institutional implementation on provincial economic performance: the case of Vietnam (2007-2011)

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS THE IMPACTS OF INSTITUTIONAL FACTORS ON PROVINCIAL ECONOMIC PERFORMA

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

THE IMPACTS OF INSTITUTIONAL FACTORS ON PROVINCIAL ECONOMIC PERFORMANCE: THE CASE OF VIETNAM

(2007-2011)

By

PHAN THACH TRUC

MASTER OF ARTS IN DEVELOPMENT ECONOMICS

Ho Chi Minh City, December, 2013

UNIVERSITY OF ECONOMICS INSTITUTE OF SOCIAL STUDIES

VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS

THE IMPACTS OF INSTITUTIONAL

FACTORS ON PROVINCIAL ECONOMIC PERFORMANCE: THE CASE OF VIETNAM

(2007-2011)

A thesis submitted in partial fulfillment of the requirements for the degree of

Master of Arts in Development Economics

DECLARATION

I declare that: "The impacts of institutional factors on provincial economic

performance: The case of Vietnam (2007-2011)" is my own work; it has not been

submitted to any degree at other universities

I confirm that I have made all possible effort and applied all knowledge for finishing this thesis to the best of my ability

Ho Chi Minh City, December 2013 PHAN THACH TRUC

ACKNOWLEDGEMENTS

First and foremost I would like to extend my gratitude to my supervisor, Dr Truong Dang Thuy, for invaluable comments, guidance and engagement through the learning process of the thesis I am much obliged to Dr Pham Khanh Nam for introducing me to the topic as well as for the availability of the dataset In addition, I am thankful for Le Duc Anh for all your kind help during my time in class 17 Then I feel lucky to have time doing the thesis with Nguyen Thi Ngoc Linh and Nguyen Quang, from whom I have a lot of things to learn Lastly and obviously, I am deeply indebted to my family members for all the kind understanding and spiritual support Forever, I will be wholeheartedly grateful for your love

ABSTRACT

Given the intensive PCI score race among Vietnamese provinces in the past year so as

to improve the enabling environment for local economic development, this study is to justify the imperativeness of such a race Employing the PCI and its nine sub-indices as proxies for institutional quality, we explore the institutional impacts on GDP for the dataset of 58 Vietnamese provinces from 2007 to 2011 In realizing this objective, we make an attempt to incorporate into the estimation models of both Cobb-Douglas and

Translog functional form the de facto overstatement of provincial GDP While the

positive impacts of PCI on provincial GDP are statistically confirmed, the regression results obtained for the nine sub-indices are divergent across functional forms and outlier treatment approaches To the extent that the economic significance of different institutional factors is assessed, our research objective is realized and respective policy implications are then suggested

Key words: PCI, GDP overstatement, institution, Cobb-Douglas, Translog

TABLE OF CONTENTS

DECLARATION i

ACKNOWLEDGEMENT ii

ABSTRACT iii

TABLE OF CONTENTS iv

Chapter 1 INTRODUCTION 1

1.1 Problem statement 1

1.2 Research objective and research questions 3

1.3 Contribution of the study 4

1.4 Organization of the study 4

Chapter 2 LITERATURE REVIEW 5

2.1 Theoretical literature: The New Institutional Economics 5

2.1.1 A renewed interest – some distinct features 5

2.1.2 Institutions matter for economic performance 7

2.1.2.1 Theoretical approaches 7

2.1.2.2 Mechanism and channels of influence 8

2.2 Theoretical model: the aggregate production functions 11

2.2.1 History of the production functions 11

2.2.2 Properties of the production function 14

2.3 Empirical literature 15

Chapter 3 DATA AND RESEARCH METHODOLOGY 19

3.1 Data and variable measurement 19

3.1.1 Data collection 19

3.1.2 Variable measurement 21

3.1.2.1 Response variable 21

3.1.2.2 Explanatory variables 21

3.1.3 Descriptive statistics 29

3.2 Research methodology 33

3.2.1 Model specifications 33

3.2.2 Estimation methods 35

3.2.2.1 Data cleaning – OLS regression for panel data 35

3.2.2.2 Modeling the overstatement of growth data – The Stochastic Frontier Analysis (SFA) 37

3.2.3 Model validation 41

3.2.3.1 Testing the model overall significance 41

3.2.3.2 Statistical test of individual coefficient significance 41

3.2.4 Analytical framework 42

Chapter 4 EMPIRICAL RESULTS AND DISCUSSION 43

4.1 Overall explanation and hypothesis testing 43

4.1.1 Overall explanation of model regression 43

4.1.2 Hypothesis testing 44

4.1.2.1 Statistical test of model overall significance 44

4.1.2.2 Statistical test of individual significance of institution variables 45

4.2 Model estimation – Institutional impacts on GDP 46

4.2.1 Transaction cost institution: Entry cost, Time cost and Informal charges 46

4.2.2 Property right and Contract enforcement institution: Access to Land and

Legal Institution 53

4.2.3 Information problem institution: Access to information, Private sector

development and Pro-activity of provincial leaders 56

4.2.4 Labor policy and training 62

4.2.5 The PCI 64

Chapter 5 CONCLUSION AND PERTAINING ISSUES 70

5.1 Conclusion 67

5.2 Policy implications 68

5.3 Limitations and further studies 69

REFEERENCES 70

APPENDIX 74

LIST OF TABLES

Table 3.1 Variable summary 28

Table 3.2 Summary statistics of key variables 29

Table 3.3 Between and within variations of key variables 30

Table 3.4 Overstatement of provincial growth data 35

Table 4.1 Attempted regressions of this study 44

Table 4.2 Hypothesis testing of overall significance of multiple regression 44

Table 4.3 Hypothesis testing of individual significance for institution variables 45

Table 4.4 Regression result– Transaction cost institution 48

Table 4.5 Regression result - Property right and Contract enforcement institution 54

Table 4.6 Regression result – Information problem institution 56

Table 4.7 Regression result – Labor policy and training 62

Table 4.8 Regression result –the PCI 64

LIST OF FIGURES Figure 3.1 Institution and economic outcomes: mechanism and channels of influence 27

Figure 3.2 Time-series plots of real GDP against year and PCI against year 31

Figure 3.3 Overall scatter plot of real GDP against PCI 32

Figure 3.4 Scatter plots for within and between variations of real GDP against PCI 32

Figure 3.5 Illustration of growth data overstatement 38

Figure 3.6 Different distributions of u 39

Figure 4.1 Effects of Time cost scores on GDP 51

Figures 4.2 Effects of Informal charge scores on GDP 52

Figure 4.3 Effects of Information access scores on GDP 59

Figure 4.4 Effects of Private sector development scores on GDP 60

Figure 4.5 Effects of Pro-activity of provincial leader scores on GDP 61

Figure 4.6 Effects of Labor policy and training scores on GDP 63

Figure 4.7 Effects of the PCI scores on GDP 65

Chapter 1 INTRODUCTION

1.1 Problem statement

There has recently been a resurgence of interests and discussions on the relationship between institutional framework and economic performance among economists and policy-makers across countries The interest in investigating this issue has been motivated by the increasing consensus in considering institutions as a crucial determinant shaping economic performance From a methodological perspective, the study of this relationship has been initiated by the debut of the New Institutional Economics, the origin of which could be traced back to Coase (1937) with the well-known message emphasizing the importance of institutions in the presence of high transaction costs “When it costly to transacts, institution matters” According to Bates (1995), this strand of analysis contributes to explaining sources of economic growth by integrating institutions, which are defined as the formal and informal “rules of the games and their enforcement characteristics” (North, 1994, p.361) into economic theory The basic argument for the roles of institution in the growth literature is that in the absence of efficient institutions, it is challenging for standard factors of production

to deliver rapid growth (Eicher, Garcia-Penalosa & Teksoz, 2006), especially in transition economies Theoretical and empirical literature on the issue suggests that channels of institutional influence on economic outcomes may be direct via affecting relevant costs incurred or indirect via affecting the incentives for investment in factors

of production Despite a great number of studies in the empirical literature focusing on the institutions – growth nexus, institutional analysis is still in its development stage (Brousseau & Glachant, 2008; Chang, 2006) and further research needs to be conducted

“before the institutional perspective can be fully operationalised” (Pelikan, 2003; Rodrik, Subramanian & Trebbi, 2004)

Being a transition economy experiencing recent institutional reforms, Vietnam offers

an appropriate empirical context for addressing the above-mentioned issue A provincial level study for the country is of imperative concern for plenty of reasons Since the 1986 economic reform called “Doi moi”, Vietnam has witnessed a significant

growth of the non-state sector with its legal recognition in the early 1990s and remarkably, the promulgation and enactment of Enterprise Law in the year 2000 Despite this legal milestone, the sector still faces numerous constraints to its establishment and business operation, the majority of which is closely related to the implementation of central policies by local governments, which varies substantially across provinces Such institutional variations arise, for the most part, due to the complexity of law, which entails the issuance of quite a great number of sub-law gazettes for enforcement purposes This, in turns, makes the implementation of laws depend a great deal on the interpretation of local officials, who can always apply their own interpretation to central policies even when regulations are not vague (Nguyen, Pham, Bui & Dapice, 2004; Tenev, Carlier, Chaudry & Nguyen, 2003) Put differently, there exists a high level of discretion by Vietnam’s local authorities compared with other developing countries, which is a prominent feature of Vietnam’s institutional reforms (Fforde & Vylder, 1996; Tran, Grafton & Kompass, 2009)

On the other hand, Vietnam’s provinces differ substantially in their achievement of economic outcomes A preliminary analysis reveals, to some extent, high dispersion of income both in terms of growth and level across provinces for the 2007 – 2011 periods Notably, the provincial GDP 5 - year average growth rate ranges from a striking low 2.92% to an approximate ten-time greater figure of 29.52% Similar divergence is evident in provincial output levels as well In particular, Ba Ria- Vung Tau exhibited highest GDP of around VND122,000 billion and VND170,000 billion for the year 2007 and 2008, respectively Likewise, Ho Chi Minh City maintained its reign as top performer with impressive figures of VND131,000 billion, VND 151,000 billion and VND158,000 billion respectively in the 2009, 2010 and 2011 scores Standing at the other end were Bac Can for the year 2007-2009 and Lai Chau for the 2010-2011 period with GDP scores even below the VND3,000 billion level Against this background, a crucial question was posed on the contribution of institutions in explaining tremendous variations in economic performance among Vietnam’s provinces

As the concept of institution means different things to different scholars and varies from studies to studies (Nelson & Sampat, 2001), inquiries firstly fall on how to

measure such a multi-dimension notion Compiled from survey-based responses of randomly selected private enterprises and hard data from published sources, which is subsequently aggregated into a provincial level score, the Provincial Competitiveness Index (PCI) is an annual composite index representing the voice of the local business community in ranking Vietnam’s provinces according to their governance quality Specifically, nine aspects of governance are measured by nine corresponding sub-indices, namely Entry costs, Land access and security of tenure, Transparency and access to information, Time costs of regulatory compliance, Informal charges, Pro-activity of provincial leadership, Business Support Services, Labor and Training and Legal institutions Since its launch in 2005 by the collaboration of the Vietnamese Chamber of Commerce and Industry (VCCI) and the United States Agency for International Development-funded Vietnam Competitiveness Initiative (USAID/VNCI), PCI has been considered as a critical tool for analyzing and monitoring progress in the local regulatory environment from the perspective of the private sector Importantly, using PCI ranking to measure progress in provincial institutional reform, Schmitz, Dau, Pham and McCulloch provide evidence for the essential role of Vietnam’s private sector in driving the economic reform at provincial level, stating that “there was no formal public-private coalition but the dynamic was one of proactive government seeking the input from the private sector, and the latter lobbying for and contributing to responsive and effective government” (Schmitz et al.,

2012, p 3) On the other hand, facts and figures obtained from VCCI at the relevant website www.pcivietnam.org reveal that there has been an intensive race to improve PCI scores among provinces in the past years in an effort to create a better enabling environment for economic development This context raises the imperative question of whether higher PCI scores really result in better economic performance for provinces in Vietnam Employing PCI and its sub-indices as proxy variables for institution, this study investigates the impacts of institutional implementation on economic performance of Vietnam’s provinces for the 2007-2011 period

1.2 Research objective and research questions

This thesis aims to investigate the contribution of institutional factors to Vietnam’s provincial economic performance through answering the following research question:

- How much of the variation in cross-provincial GDP of Vietnam could be explained by institutional quality, both in a broad sense and as specific factors of economic governance?

In an attempt to explain the economic differentials across Vietnam’s provinces, the thesis explores the contribution of institutional factors, particularly the PCI and its sub-indices, including Entry Costs, Time Costs and Regulatory Compliance, Informal Charges, Land Access and Tenure Security, Legal Institution, Transparency and Access

to Information, Labor Policy and Training, Business Support Services and Pro-activity

of Provincial Leadership

1.3 Contribution of the study

This study contributes to the existing literature in the following two ways First, since there have been few studies on the issue for the case of Vietnam so far, this thesis is an attempt to provide the first empirical evidence for the impact of institutional factors on the country’s provincial economic outcomes Specifically, using PCI as the proxy for institution quality, a limited number of previous researches examine the issue for Vietnam at firm-level in a particular year, that is they offer a cross-section analysis of how institutional implementation impact on firm’s productivity This study, on the other hand, explores the issue at provincial-level over a five-year time span Second, in

realizing the above-mentioned objective, efforts are made to incorporate the de facto

overstatement of growth data into estimation models Treating such a sensitive dataset

in this way helps, to some extent, to deliver a more reliable regression results for analysis

1.4 Organization of the study

The thesis is structured in the following way The second chapter reviews the theoretical literature underlying the institution – economic growth nexus along with the empirical dimension of this relationship Chapter 3 describes the data, research methodology and empirical models for investigating the impacts of provincial institutional quality on economic performance Chapter 4 reports and analyses the regression results The last chapter concludes the study with policy implications and discusses the limitations and directions for further research

Chapter 2

LITERATURE REVIEW

This chapter is to review the theoretical and empirical literature on the central role of institution quality on economic performance The chapter is divided into three sections The first section reviews the prominent economic theory underlying the relationship and discusses the theoretical models to be applied The second part reviews empirical studies on the issue The final section concludes the chapter with summary of empirical evidence and conceptual framework

2.1 Theoretical literature: The New Institutional Economics

There has been a resurgence of interest in institution literature during the last few decades From a methodological viewpoint, the incorporation of institutional perspective into economic analysis was initiated by a particular strand of analysis: the New Institutional Economics (NIE) Crucially, NIE covers a wide range of institutional issues with a growing proliferation of literature devoted to study the impacts of institution quality on economic performance

2.1.1 A renewed interest – some distinct features

Despite an early interest in the institution literature dated far back to the time of Adam Smith’s work, institutions did not regain its prominence in mainstream economics until the 1990s in response to the failures of neoclassical economics in interpreting economic growth and development issues (Nye, 2008) Specifically, experience of the ex-Soviet regimes in their transition period of worsening economic performance lends significant support to the essentiality of including “institutional factors in the corpus of mainstream economics” (North, 1994, p.367) In addition, given the technology diffusion and capital mobility as the expected result of globalization, the widening gap of inter-country inequality in the late twenties has justified for the World Bank’s acknowledgement that “the main challenge for development policy at the turn of the

twenty-first century was the supply of effective market supporting institutions, and the creation of demand for such institutions” (World Bank, 2002, p.4)

NIE draws on the backbone analysis of Williamson (1985) and North (1990), with their definitions of institutions as “governing structures” and “rules of the game”, respectively As such, NIE assumes that individuals have incomplete information and limited rationality, which accounts for the arising of uncertainty and transaction costs (Menard & Shirley, 2008) In this light, humans create institutions, both in written forms (the formal institution) as well as norms of conduct and beliefs (the informal ones) In addition, they develop organizational arrangements to support transactions and cooperation in production and exchange According to Brousseau and Glachant (2008), substantial efforts have been made in studying how institutions interact with these organizational arrangements as well as “how these arrangements act in turn to change the rules of the game” (Menard & Shirley, 2008, p.12)

NIE differs from its former variant, the Old or Traditional Institutional Economics, whose name and core elements were owed to Hamilton (1919) in his article “The institutional approach to economic theory”, in two significant ways Firstly, NIE is built

on the fundamental assumption of neo-classical assumption of scarcity and competition Secondly, while the former being shortage of “a systematic theoretical foundations as well as supportive empirical analysis” (Joskow, 2003, p.6), NIE is, according to Menard and Mary (2008), “built around a backbone of some fundamental and original contributions” and has so far been characterized by extensive empirical studies

Remarkably, NIE is not an integrated theory; but a “combination of bricks coming from different traditions” (Brousseau & Glachant, 2008, p.2) Being the outcome of an evolutionary process, NIE is open to academic contributions of scholars from a wide range of social and scientific backgrounds This high level of openness, to a certain extent, helps to explain for NIE’s evolutionary perspective as well as the heterogeneity

of academic contributions in NIE

Overall, in exploring the contribution of institutional factors to variations in economic performance, NIE does not reject the basic progress and analytical tools that has been

developed in the neoclassical tradition (Joskow, 2003); rather it acknowledges the latter’s strengths and drawbacks as well as employs existing tools in its own way to address a broader set of relevant issues

2.1.2 Institutions matter for economic performance

According to Smith (1976), institutional factors can explain not only the cross-country economic differences in growth rates, but also regional disparities within a country Institutions being “the underlying determinants of the long-run performance of economies” (North, 1990), a significant body of institution literature has recently been revolving around the inquiry: “How do institutions matter?”

2.1.2.1 Theoretical approaches

From a theoretical perspective, the research of institutional issues proposed within NIE falls into three broad categories of theoretical approaches, namely The Historical Perspective Approach, The Comparative Institutional Analysis and the Imperfect Information Theory

Pioneered by North (1990), “The Historical Perspective Approach” analyzes institutions from a historical perspective As institutions are historically specific, it is of vital importance that historical context be taken into consideration when coping with institutional change (Alston, 1996) In an attempt to accomplish this purpose, North (1990), in his seminal book “Institutions, Institutional Change and Economic Performance” combines a theory of Human behavior with a theory of Transaction cost The main message delivered is that institutions matter for economic performance by eliminating uncertainty and transaction costs incurred in social transaction, thereby enhancing economic growth

The second strand of theoretical framework, The “Comparative Institutional Analysis”, owes its most systematic elaboration to Aoki (2001) Attempting to include both Evolutionary and Repeated game theory into analysis, Aoki’s definition of institution reflects the presence of these two sub-strand of approaches, whereby institution is considered as “self-sustaining system of shared beliefs” (Aoki, 2001) that may change

the incentive structures of games, thus influencing the strategic choices of and interaction between agents

Another influential strand of NIE is the one associated with the “Imperfect Information Theory”, according to which, institutions are explained “in terms of strategic behaviors under asymmetric information among the different parties involved” (Bardhan, 2000) This analytical framework emphasizes that information and enforcement costs drive some markets to be non-existent and others uncompetitive Against this background, institutions may be relied on as a response to missing markets or may help to eliminate information problems in other ones (Hoff, Braverman, Stiglitz, & Arnott 1993)

In brief, the basic difference among these three approaches lies in their analytical tools

to conduct investigation While there is an attempt to integrate economic theory and economic history in the Historical Perspective approach, the second one of Comparative Institutional approach makes good use of game theory in addition to historical information Finally, the Imperfect information framework is the most mathematically-oriented out of the three lines of study (Nabli & Nugent, 1989)

2.1.2.2 Mechanism and channels of influence

The above-mentioned theoretical approaches lay fundamental analytical frameworks for analyzing the “institutions matter” issue, specifically the institutional mechanism and channels of influence on economic performance

Among the institutions that are significant determinants of economic performance are those involving property rights and contract enforcement According to Coase (1937), one way in which institutions can help resolving the problems of prisoners’ dilemma as well as overcome the failures of their collective action is to provide for a well-defined property right Being norms and rules that confer the control of the returns to invested assets and/or produced values as well as stipulate the procedures for exchange of executive, institutions provide the essential level of internal security and trust By eliminating the potential risks of expropriation, institutions affect the economic agents’ decisions and extent of saving and investing in physical and human capital Put another

way, the lower the risk of expropriation and the higher the level of security, the higher

is the level of investment and growth, all other factors remain unchanged

Moreover, institutions that provide for secure property rights lend great support to the process of patent registration, ideas dissemination, to list some, across creative individuals; all factors that stimulate research and development activities In this sense, good institutions contribute to boosting an economy’s rate of innovation (Tebaldi & Elmslie, 2008), thereby promoting economic growth

With reference to contract enforcement, institutions stipulate rules and regulations that can prevent or mitigate market failures and agency problems Measured by the extent of regulatory burden, tax evasion, corruption and the like, institutions affect economic performance by enhancing the efficiency of public policy and by mitigating the risk of anti-competitive behaviors and rent-seeking by corporate actors Only with the right institution in place that the collective action problem is resolved and the scope for rent-seeking is minimized

In the same line of study on collective action problems, Olson (1982) argues that facing these problem prevent large groups of “self-interested economic agents” from cooperation Subsequently, Axelrod (1997), in his comprehensive analysis of co-operation, raises the importance of internalized norms that encourage cooperative behavior within such groups In the context that cooperation is essential for individuals

to capture trade gains in societies, the evolution of institutions helps facilitate a more favorable environment for cooperative solutions, thereby stimulating economic growth

Besides, institutions that set ground for a well-functioning human capital market affect the rate of returns to education, which, in turn, affects the productivity and allocation of labor In this sense, institutions stimulate human capital accumulation According to Wolf (1955), the “right institutions” are conducive to the formation of productive capital as well as the adoption of new technology across regions and nations

In general, in terms of property rights and contract enforcement, institutions create an incentive structure that leads to the reallocation of efforts to engage in productive legal market activities Put differently, it facilitates the conduct of appropriate and

constructive property transactions as well as other contracting of mutual interests among economic agents; thereby promotes economic outcomes

Another predominant source of institutional factors contributing to economic growth is those closely related to transaction costs According to North (1990), this point justifies for the existence of institutions” In the sense that institutions, in combination with other factors of production, determine the structure of exchange that influences production and distribution costs, they affect the feasibility and profitability of economic activities It is through this mechanism that institutions are considered the

“underlying determinant of long-run economic performance”, according to North (1990)

In addition, transaction costs being on the path upward together with the increasing level of development (Wallis and North, 1986), institutions emerge as a significant factor that offer a two-fold contribution to economic growth: one is efficiency gained implied by a lower measurement of costs per transaction and the other is a growing number of activities of mutual economic interest signaled by a higher volume of transactions Institutions are, therefore, conducive to better economic performance in the long run

The relevance of the underlying institution economic theory of information problems to that of transaction costs lies in the fact that the former represents a significant source of the latter (Stiglitz, 1985) In this line of study, institutions are considered to be “well-understood rules of the game” that bring about a certain degree of predictability to individual or corporate behaviors With the right institutions in place, uncertainty is reduced and contracting activities of mutual interests between economic agents will be encouraged to a certain extent, which results in better economic performance

The mechanisms of how institutions contribute to economic performance illustrated by previous paragraphs reveal relevant channels of influence First, institutional factors may enter a growth regression in a linearly additive term As such, they may display a direct, linear effect on growth, which is the direct channel of influence Second, in interacting with standard factors of production, institutions may have indirect effects on growth by altering the marginal effects of these production factors, which is the

interactive effect Another possibility that institutions indirectly matter to economic performance is by affecting the relationship between economic growth and its determinants through the latter’s parameter heterogeneity, which is the threshold effect (Minier, 2007) The basic argument for such an effect is that how growth fundamentals determine a country’s growth may, to a certain extent, depend on relevant institutions being adequately strong In other words, there exist some possible threshold levels of institutional quality that must be reached before say, physical and human capital as well

as technological advances could exert their influences on growth There are theoretical reasons to expect such indirect effects According to Wolf (1955), “growth-promoting institutions” stimulate, rather than impede the formation of physical and human capital

as well as allow for the adoption of technological advances; and thereby promote economic growth In a nutshell, the literature review on the contribution of institutions

to economic performance from the perspective of NIE shows that institutional economists have “won the war” in the sense that much insight has been increasingly provided on the proposition that “institutions matter for growth” (Joskow,2003) Despite substantial endeavors as well as progress having been made in the field over the last three decades, pertaining inquiries about the institution-growth nexus, for instance the potential endogeneity of institutional variables, the causality of the relationship, to name some, are still open for further research And the hard work is still ahead

In modeling institutional effects, a large number of empirical works on growth make use of the production function Section 2.2 provides a brief introduction

institutions-to the development of the production function The section also discusses the properties

of the production function that are relevant for the application of institutions-growth analysis

2.2 Theoretical model: the aggregate production functions

2.2.1 History of the production functions

There is widespread consensus among scholars in considering production function as a critical tool of analysis in the neoclassical economics According to Schumpeter (1954), the history of the production function dates back to the year 1767 when Turgot, in his

work of Observations on a Paper by Saint-Peravy, discusses how the variations in the

proportion of standard factors of production account for marginal productivities, whereby the production function was implicitly formulated Later on, Philip Wicksteed (1894) was generally believed to be the first economist to algebraically formulate the input-output relationship as P f x x 1, 2, ,x m On the other hand, Humphrey (1997) provides some evidences for the fact that the first formulation of production function was owed to Johann von Then in the 1840’s

During the 1950’s-1970’s period, production function attracted the attention and discussion of a large number of economists, which results in various propositions and analysis of the specifications for input-output relationship as well as their corresponding conclusions (Mishra, n.d) In particular, Stigler (1952) noted that the logarithmic production function was introduced by Malthus while (Humphrey, 1997) demonstrated how the Cobb-Douglas function was firstly presented in “disguised form”

by Von Then in The Isolated State, vol – II In this line of investigation, it is noted by

Velupillai (1973) that Wicksell in his work dated 1900-1901 formulated his production function identical to that of Cobb-Douglas Specifically, in Wicksell’s 1923 review of Gustaf Akerman’s doctoral dissertation, his production function was formulated as

P AL C , with  1 Similarly, the well-known Leontief production function was evidenced to be jointly formulated by Jevons, Menger and Leon Walras Overall, certain academic achievements associated with individual scholars are not the sole accomplishment of such individual; rather they are the “culmination of the research efforts of an entire epoch” (Russel, 1984)

Arrow, Chenery, Minhas, & Solow (1961) formulated the Constant Elasticity of Substitution (CES) production function allowing for the elasticity of substitution between capital and labor to be valued between zero and infinity, which stays fixed along and across the isoquant regardless the size of output or inputs used in the production process In this sense, the Cobb - Douglas, the Leontief and the linear production functions are only its special cases In the classical form, these production functions assume that the technological progress is Hicks-neutral (Hicks, 1932), whereby the marginal rate of substitution between any two factors of production is independent of technical progress or level of output Generalization of the CES

production function was initialized by Brown and Cani (1963) to allow for constant returns to scale Fundamental contributions were made by Uzawa (1962) and Kazuo Sato (1967) concerning the incorporation of more than two factors of production Likewise, Nervole (1963) generalized the Cobb-Douglas production function with variable returns to scale, which was further studied by Ringstad (1967) Diewert (1971) proposed a functional form for the production process that permits variable elasticities of substitution Generalization of the Cobb-Douglas and CES production functions in their classical form was almost complete by mid 1970’s

non-The prominently restless area of research in the economics of production is the empirical estimate of the aggregate production function, an essential tool of analysis in macroeconomics According to Felipe and Fisher (2003), there are two strands of literatures questioning the notion of an aggregate production function: the so-called Cambridge (UK)- Cambridge (USA) capital controversy and the aggregation literature The former raises questions concerning the unit of measurement for capital goods as a factor of production in the aggregate production function while the latter explores the conditions for neoclassical micro production functions to be aggregated into a neoclassical aggregate one As there are considerable mismatches between characteristics possessed by firms and that of industries, or moving upwards, the economy as a whole, when attempts were made in identifying the recipe for the production functions at macroeconomic levels, they have to be handled with significant caution

As advocates of the aggregate production functions, Knut Wicksell (1923) justified the application of aggregate linearly homogeneous functions by making substantial contributions to conclude that non-homogeneous production functions for firms are perfectly compatible with a linear homogeneous function for the entire industry (Humphrey, 1997) This line of investigation progressed rapidly with activity analysis

of Koopmans (1979), aggregate linear production function of Georgescu-Roegen (1951), separation theorems and generalization of Von Neumann’s model (1937) as well as well-grounded proofs provided by Nikaido (1968), all of which strengthened the foundation of using the aggregate production function in economic analysis

According to Felipe and Adams (2005), the Cobb-Douglas production function is still today “the most ubiquitous form in theoretical and empirical analyses of growth and productivity” Its origin officially dates back to the seminal work of Cobb and Douglas

in 1928, whereby an aggregate production function was, for the first time, econometrically estimated with the employment of 1899-1922 data for the United States manufacturing sector and the results were accordingly presented to the contemporary economists (Cobb & Douglas, 1928)

2.2.2 Properties of the production function

The production function reflects the physical relationship between inputs and outputs, which can be mathematically defined as

The production function has the following properties

Firstly, there is no “free-lunch”, that is (0) 0f  Secondly, the production function is monotonic As is well-known with respect to a (single output) production function, monotonicity requires positive marginal products with respect to all inputs, that is

  max | , 

2.3 Empirical literature

The last decade has witnessed a vast majority of empirical studies employing the production function of some forms to address the institution-growth issue across countries In particular, the production function is mostly applied under the so-called

“extended production function” specification that “combines institutional proxies with traditional growth factors” (Glaeser, La Porta, Lopez-de-Silanes & Shleifer, 2004) There is also a wide range of other “extended production function” specification in which other growth determinants say, trade openness, geography, macroeconomic policy, to list some, are further included as explanatory variables in modeling the institutional effects on economic performance (Jacob & Osang, 2007) In addition to independent variables, other source of heterogeneities among model specifications arises from the choice of dependent ones, which may be grouped into two broad categories: output-level related variables and output-growth related ones According to Eicher and Leukert (2006), empirical evidences of output levels show more robust effects than those of output growth In the same line of investigation, main findings report that the choice of institutional proxies as well as the (non)treatment of its potential endogeneity is also a primary source of heterogeneity Particularly, institutional factors have been proxied mainly by standardized aggregated indices and/or survey-based indicators constructed at national level According to Aidis, Estrin and Mickiewicz (2009) and Shirley (2008), the multi-dimensionality together with the disparity among institutional proxies, to a certain extent, poses skepticism on the validity of an ideal variable highly representative of the concept of institution In general, empirical institutional research is characterized by such substantial heterogeneities that it would be challenging to infer a representative empirical impact magnitude for the whole literature (Efendic, Pugh & Adnett, 2011)

Dias and Tebaldi (2012) address the importance of interactions between institutions and human capital accumulation for economic growth process The paper employs panel OLS and GMM dynamic panel estimation technique for the sample of 61 countries over the 1965- 2005 interval Regarding institution variables, democracy and autocracy are used as proxies for political institution while share of educated labor in the economy represents structural institution These institutional variables enter the formal growth model together with the human capital variable generated based on Hall and

Jones’(1999) piecewise function and the capital stock calculated following Perpetual Inventory Method proposed by Easterly and Levine (2001) Empirical estimates lend support to the proposition that structural institutions are conducive to productivity and long-term economic performance

Siddiqui and Ahmed (2013) investigate how institutional factors shape economic performance in 84 countries from 2002 to 2006 In order to capture the multi-dimensionality of institution, the aggregate index of Institutionalized Social Technologies (IIST) is used together with its sub-indices of Institutional and Policy Rents, Political Rent and Risk Reducing Technologies Other than institution, human capital, savings and trade openness variables are used to specify a formal extended growth model with the application of OLS and GMM methods Notably, interactive variables are introduced into the model to measure the possible complementarities among institutional variables Estimation results suggest a high degree of positive complementarities between institutions that protect property rights and political institution The authors conclude that institutions positively affect economic outcomes

Atul and Sal (2012) employs Solow’s growth accounting model with World Bank regulation indicators as proxy variable for institutional quality to study the impacts of regulation framework on economic performance Fixed and random effects estimation

is applied to the data of 23 OECD countries over the 2002-2008 period Findings confirm that regulation quality has a positive effect on growth through its impact on total factor productivity

The paper by Gagliardi (2008) reviews the major theoretical and empirical literature exploring the channels of influence of institutional framework on economic performance Theoretical approaches within the New Institutional Economics are proposed, comprising of the historical perspective approach, the institutional comparative approach and the imperfect information theory Within the institutional approach, the concept of institutional complementarities is elaborated and has important implication from the policy perspective Despite difficulties in institutional measurements, empirical evidences lend support to the proposition that a country’s institutional framework is critically conducive to its economic development

Using state-level cross-sectional data of the U.S and OLS regression, Sobel (2008) examines the relationship between institutional quality and entrepreneurial productivity, with the latter proxied by economic freedom index and net entrepreneurial productivity index In addition to institutional variables, the paper includes variables of geographic and demographic controls such as median age of state population, population density, and percentage of male population along with the percentage of the population with a college degree The paper provides the first empirical evidence supporting Baumol’s theory that good institutions help to channel entrepreneurial efforts toward productive entrepreneurship further concludes that entrepreneurship productivity helps to explain differences in economic performance across states

Hasan, Wachtel, and Zhou (2007) review the link between China’s institutional development and economic growth with emphasis on the three facets: financial deepening, development of legal institutions and the degree of political pluralism Size

of private sector, Rule of law and Awareness of property rights are used as proxy variables, which are measured by ratio of private sector total fixed investment to overall fixed investment, number of lawyers per 10,000 people and ratio of number of domestic trademark applications to number of firms, respectively The model specification based on Barro (1991)’s growth equation is applied to the annual dataset

of 31 provinces for 1986-2002 period with the employment of OLS, annual system GMM and three-year average system GMM estimation method Conclusion confirms that institutional development contributes to explain for provincial economic growth differentials

Jalilian, Kirkpatrick and Parker (2007) investigate the role of regulatory framework on economic growth in developing countries Data set covers 117 countries for the 1980 –

1999 interval and 96 countries for the 1980 – 2000 period corresponding to two methods of estimation: the cross-section and panel regression, respectively Represented by the two indicators of Regulatory quality and government effectiveness, the impact of regulation on GDP per capita growth is estimated after controlling for initial GDP per capita, gross capital formation, schooling, trade, inflation and government expenditure Findings are supportivefor the theoretical literature that there exists a strong association between regulatory framework and economic growth

Using Vietnam’s firm-level data in 2005 and the Provincial Competitiveness Index

2006 as proxy variable for local governance practice, Tran, Grafton and Kompas (2009) analyze Vietnam’s institutional reforms in 2000 and their impacts on economic performance of the non-state sector Other than institution variable, the model is specified with the inclusion of control variables that accounts for provincial initial endowments affecting firm’s performance, such as human capital and market size In addition, dummy variables are also introduced into the estimated model to control for firm’s specific characteristics, including firm’s size, age, capital intensity and ownership type Findings confirm that improved institutional quality concerning market information, securing land tenure and labor training promotes firms’ labor productivity

There are some studies that do not support the proposition that institution causes growth Glaeser et al (2004) revisits the proposition that institutions cause growth using cross country data from 1960 to 2000 Their results do not support the proposition They also find that the literature analyzing the relationship between institutions and growth is prone to the poor measurement of institutional variables and irrelevant instrumental variables for institutions

Unlike most of empirical studies on the subject assuming parameter- invariant, Jenny Minier (2007) uses panel data of 70 countries (1960-2000) to analyze the indirect effects of institutions on economic growth through parameter heterogeneity Other than executive constraint, which is the proxy variable for institution, control variables of growth fundamentals in the estimation model comprise of physical capital accumulation, education, geography and initial income Applying endogenous sample splitting technique proposed by Hansen (2000), the author attempts to measure the varied parameters of growth determinants in recognition of such a threshold Findings report that there is little evidence supportive for the threshold effect of institution on factors of production through parameter heterogenity; rather institutions influence the marginal effect of policy variables on growth

Chapter 3 DATA AND RESEARCH METHODOLOGY

This chapter is to present the data and research methodology applied in modeling institutional effects on economic outcomes The chapter is divided into two broad sections The first section displays data analysis along with variable measurement and generation The second section of research methodology discusses model specifications, estimation methods and concludes the chapter with analytical framework as summary of estimation strategy employed in the study

3.1 Data and variable measurement

3.1.1 Data collection

In this study, we use secondary data from published sources, which vary across employed variables In particular, investment and GDP data are obtained from the Provincial Statistics Year Book of two sets: 2005 - 2009 and 2010 – 2011 Data for Labor force variable derive from the national General Statistics Office (GSO) at www.gso.gov.vn Regarding PCI and its sub-indices, data scores are obtained via the PCI Annual Reports from 2007 to 2011 and their corresponding website

The PCI is the result of an intensive and ongoing cooperation between the Vietnam Chamber

of Commerce and Industry (VCCI) and the United States Agency International Development/ Vietnam Competitiveness Initiatives (USAID/VNCI) managed by Development Alternatives, Inc (DAI) Notably, the creation of the PCI includes three steps: (1) collection of survey-based data and “hard” data; (2) construction of sub-indices and finally, (3) calibration of sub-indices to ensure that their relative weights reflect their contribution to the private sector development This means the PCI is generated from two general types of data: the perception indicators drawn from mail-out surveys and the hard indicators (“hard” data) deriving from published sources While perception data may deliver the specific information of interest, they are often prone to perception biases Comprising of more objective built-in indicators,

hard data, on the other hand, often limit researchers to a particular set of information Combining perception or “soft” data with hard data, the PCI research team strives to seek for

an optimum approach that could minimize the drawbacks as well as maximize the advantages

of both data sources (PCI Annual Report 2007)

Regarding the collection and processing of “soft” data, stratified sampling method is applied

at provincial level in terms of business age, sector and legal form to ensure that the sample is representative of the local business community with a sampling error of 3% Besides, the sample size ranges from approximately 7,000 to 10,000 private domestic firms as there are considerable fluctuations in rate of response over years Remarkably, the mail-out survey respondent component has been expanded to include more than 1,000 foreign invested firms

into the sample since 2010 Finally, the survey questions, which are separately designed for

domestic and foreign-invested firms, undergo a certain number of alterations and supplementations over years in order to capture major updates in the provincial business and regulatory environment

It is worth noticing that alterations were implemented to PCI methodology in 2009 Complying with the principle set up from the outset that construction measures could be

“tweaked at the margins” (PCI Annual Report, 2009) while fundamental revisions should be considered with great caution so as to ensure the over-time comparisons of PCI, the 2009 methodological changes fall into the following three categories First, against the background

of massive equitization of local state-owned enterprises (LSOES) around that time, the SOE bias sub-index was dropped Second, there were minor changes in indicators within each sub-index Finally, a new approach was adopted in calculating sub-index weights so as to eliminate the possibility that weightings were driven by accidental correlations (PCI Annual

analysis and estimation model building so as to capture any significant changes in the institutional impacts on provincial economic performance due to such methodological alterations

The Data for this study is annual data covering Vietnam’s 58 provinces for the 2007-2011 interval Data for PCI index is available from 2005 to 2011, but the year 2005 and 2006 are removed from the sample because provincial statistics of these two years does not display sufficient figures for the majority of selected variables of this study, such as provincial GDP

and investment In addition, 6 provinces including Ha Tay, Cao Bang, Hung Yen, Thanh Hoa, Long An and Ninh Binh are dropped due to insufficient data for the variables of interest

3.1.2 Variable measurement

3.1.2.1 Response variable

RGDP stands for Real Gross Domestic Product Being a general indicator of economic

outcomes that reflects the final results of production and business activities of the province in

a year, this variable is calculated in constant price by deflating corresponding nominal values (GDP calculated in current price) with provincial annual CPIs, with 2005 as the base year RGDP is measured in VND million

As annual provincial CPIs are not available for 5 out of 58 provinces, including Dac Nong, Hau Giang, Hoa Binh, Lang Son, Phu Yen, corresponding national CPIs are used as substitution in calculating real GDP and investment for these provinces

3.1.2.2 Explanatory variables

LABFO stands for labor force, by which name it denotes the work force aged from 15 and

upwards who actively participate in the province’s business and production activities This variable’s measurement unit is thousands of people

CAP denotes capital input variable Applying the Unified approach of the Perpetual

Inventory Method (PIM) proposed by Berlemann and Wesselhöft (2012), we generate the capital input variable with the employment of provincial nominal investment values According to GSO’s definition, investment is a measurement of expenses to increase and maintain material assets in a given period, usually a year Investment can be calculated through investment projects and national programs aimed to increase fixed and liquid assets

In this study, real values of investment are obtained by deflating nominal values using provincial annual CPIs, with 2005 as the base year

The net capital stock at the beginning of period t, let say K , could be mathematically t

expressed as

I is gross investment in the current period

For the sake of simplicity, depreciation is assumed to be at a constant rate δ of 5% and then

the capital stock could be rewritten as

GDP

I K

Concerning the initial value of investmentI t1, we apply the alternative procedure suggested

by Nehru and Dhareshwar (1993) in their later application of the Steady State Approach According to this procedure, the time-series of log investments is regressed on time using OLS

The advantage of this procedure as opposed to the traditional stead-state approach is that the initial value of the investment time series can be generated without the sole reliance on the investment figure of any single year, thereby a more reliable initial value could be obtained

For all the institution variables, relevant information is obtained via PCI Annual Report from

2007 to 2011

PCI: a 100-point scaled composite index measuring and ranking Vietnam’s provinces based

on their overall economic governance quality It is created through the “Three Cs” procedure, namely Collection, Construction and Calibration In the first step of collection, survey data are adjusted with hard data from published sources to correct for possible perception bias At the end of this step, indicators are available to be standardized into the ten-point scale sub-index construction in step 2 In the last step, sub-indices are calibrated into three weight classes of high, medium and low 15%-20%, 10% and 5%, respectively, such a way that the relative weight of each sub-index demonstrates its contribution to the sector’s growth, investment and profitability For instance, PCI report 2009 reveals that Transparency &

access to information sub-index together with Labor policy sub-index receives the highest weight of 20 percent as they are both considered crucial to the business outcomes of the sector Meanwhile, the three sub-indices of Land access, Legal institution and Private sector development receive the lowest weight of only 5 percent since there are few variations across provinces in their scores, making them not strongly correlated to the private sector development outcomes across the sample entities Correspondingly, medium weight class of 10% and 15% are assigned for sub-indices with average correlation across the three above-mentioned outcome variables Time cost of regulatory compliance sub – index stands out from this group with 15% weight, leaving 10% weight for the remaining sub-indices

For PCI as well as all of its sub-indices, a higher score represents a better governance quality For example, a higher value of labor policy sub-index indicates better evaluation of enterprises on labor policies of the province, and a higher value of informal charge sub-index indicates lower (better) unofficial fees required to “grease” business operations It is well - noted that major changes in PCI methodology has been implemented since the fifth iteration PCI 2009, consisting of changes in the number of sub-indices, changes within sub-indices and changes in sub-index weights in order to capture the dynamism of the provincial economic environment In particular, the index of State-owned enterprise (SOE) bias, which measures the biased perception of provincial authorities towards SOEs was dropped The key reason was that the local SOEs were equitized on a large scale As a result, the obstacle to private sector development previously posed by the so-called SOE bias no longer prominently existed

The below section offers an initial analysis of the nine sub-indices based on PCI report 2009 for the mere purpose of illustration Similar to PCI, each composite sub-index is constructed from a whole range of indicators, the scores of which are standardized from a combination of hard data and perception ones derived from survey responses

ENCOST stands for Entry costs, a measurement of costs in terms of money and time spent on

problematic procedures required for starting business Indicators included are Length of business registration in days, Number of licenses and permits necessary to start operation, Wait for land use right certificate, Percentage of firms waiting for more than one month to complete all steps necessary to start operation, Percentage of firms waiting for more than three month to complete all steps necessary to start operation, Percentage of firms having

difficulties obtaining all licenses and permits necessary to do business The final sub-index covers both original indicators of the 2005 survey and new indicators firstly used in the 2006 survey, which are equally weighted 50 percent each

TIMECO stands for Time costs of regulatory compliance, a measure of time firms have to

spend on bureaucratic compliance This sub-index, therefore, reflects the costliness of business-related administrative procedures and paperwork It is worth noticing that in building this sub-index, new indicators have been supplemented, reflecting improvement in local bureaucratic compliance, such as cuts in formal administrative fees as well as waiting time for administrative procedures

INFOCHARGE stands for Informal charges, indicating extra fees and other unofficial

payments firms have to pay so as to facilitate their business operations

LANDACC represents Land access and Security of tenure, that is firm’s accessibility to land

as well as tenure security upon land acquirement This two-dimension sub-index measures the costs to obtain business premises and the risk of property expropriation, respectively Each dimension accounts for 50% of the sub-index

LEINSTI represents Legal Institution, reflecting firms’ confidence in and reliance on

provincial legal institutions in settling business-related disputes, the actual costliness of judicial proceeding as well as firm’s actual use of local courts

INFOACC is a variable of Transparency and Access to information, measuring firms’

accessibility to legal documents as well as the availability of information essential to run a business This four-dimension sub-index covering Access to legal and planning documents, Equity and consistency, Participation of private enterprises in policy negotiation and Openness of web page The first three dimensions are computed from survey data and account for 60% of the final sub-index The fourth dimension with 40% weight reflects the business utility of provincial web page

PRIDEVELOP is coded for Business support services, a measurement of the availability of

business support services in the province This index is compiled of three equally weighted dimensions, including Use of business support services, Private business support and Quality

These dimensions cover access to business-related services, participation of private providers and the repeated usage of these services by private enterprises, respectively

PROACT is a variable of Pro-activity of provincial leadership, measuring the dynamism and

initiatives of provincial officials in working with regulatory framework in order to support private sector development

LABPOL stands for Labor and training, a measure of provincial officials’ efforts and

assistance in enhancing vocational training and local labor placement This sub-index is a measure of provincial government’s efforts to enhance vocational training for local enterprises It is compiled of three equally weighted indicators, namely labor recruitment, labor training facilities and labor quality

Economic outcomes

3 Information problems

Sub-index No 6,7,8

4 Functioning of human capital market

Sub-index No 9

Technical change Human capital

4 Joilson Dias & Edinaldo Tebaldi, 2012

Channel #1: Direct effect Channel #2: Indirect effect

6 Transparency and Access to Information

7 Business Support Services

8 Pro-activity of Provincial Leadership

9 Labor policy and Training

It is worth-noticing that as all sub-indices are constituted into the final composite index PCI, they are all scaled and scored in such a consistent way that any scoring incremental implies a corresponding improvement in institutional quality Consequently, all sub-indices bear a positive expected signs when entering the model estimation For more insights into these sub-indices, how they are constructed from numerous indicators as well as their alterations over years, see PCI annual reports from 2007 to 2011 which are available on the website: www.pcivietnam.org

Table 3.1: Variable summary

unit

Expected signs

Labfo Labor force of age 15 and upwards participating in

provincial economic activities

Thousands of people

(+)

PCI Provincial Competitiveness Index Percentage (+)

Landacc Land access and Security tenure 10-point scale (+) Infoacc Transparency and Access to information 10-point scale (+) Timeco Time costs of Regulatory compliance 10-point scale (+)

Proact Pro-activity of provincial leadership 10-point scale (+) Pridevelop Private sector development 10-point scale (+)

Labpol Labor policy and Training 10-point scale (+)

3.1.3 Descriptive statistics

Table 3.2 provides basic summary statistics of key variables used in the study A quick glance at the figures reveals that this is an unbalanced panel data as there are missing data in the variables of real GDP and capital Besides, there are substantial variations in real GDP variable, with its standard deviation of 29,600 which is almost 1.50 times of the mean of 20,500 Min and max values of PCI and its sub-indices also indicate that these variables exhibit a wide range of figures

Table 3.2: Summary statistics of key variables

Table 3.3: Between and within variations of key variables