Matched change of financial systems and economic
concentration in the European Union
Simon P. Muller
August 2007
Supervisor: dr. ir. Dirk J. Bezemer
Matched change of financial systems and economic concentration in the European Union
Abstract
Table of Contents ABSTRACT... 2 TABLE OF CONTENTS... 3 PREFACE... 5 INTRODUCTION... 6 Outline ... 7 LITERATURE REVIEW... 8
1.1 APPROACHES TO MEASUREMENT OF THE PERFORMANCE OF TRANSITION ECONOMIES... 8
1.1.1 Policies ... 8
1.1.2 Institutions ... 8
1.1.3 Business system ... 9
1.1.4 Pace of transition processes ... 9
1.1.5 Combination of policies, institutions, the business system and economic performance... 9
1.2VARIETIES OF CAPITALISM... 9
1.2.1 Topics in the varieties of capitalism literature...10
1.2.2 Initial varieties of capitalism literature, state institutions ...11
1.2.3 Neo-corporatism...11
1.2.4 Business system and institutional system...12
1.2.4.1 Coordinated market economies versus liberal market economies approach ... 13
1.2.4.2 The societal effects approach... 13
1.2.4.3 Comparative business systems approach ... 14
1.3 OPENING IN THE LITERATURE...14
1.4 CONCLUSION...15
COMPARATIVE BUSINESS SYSTEMS MODEL...16
2.1 WHAT MAKES IT WORTHWHILE TO INVESTIGATE THE MATCH OF CHANGE?...16
2.2 SELECTION OF BASIC MODEL FROM VARIETIES OF CAPITALISM LITERATURE...16
2.2.1 Requirements model ...17
2.2.2 Comparison models...17
2.3 COMPARATIVE BUSINESS SYSTEMS MODEL...18
2.3.1 Business system ...19
2.3.2 Institutional system ...19
2.3.3 Selection of measurable characteristics from the comparative business systems model... 20
2.3.3.1 Proxy for the business system characteristic ownership coordination,... 21
Hirschmann-Herfindahl Index... 21
2.3.3.2 Proxy for the institutional system characteristic financial system, structure-size measure... 22
2.4 CONCLUSION...23
METHODOLOGY...24
3.1 FRAMEWORK...24
3.1.1 Focus of the study and research question ...24
3.1.2 Hypotheses ...24 3.2METHODOLOGY...25 3.2.1 Estimation model...25 3.2.2 Dependent variable...26 3.2.3 Explanatory variable ...26 3.2.4 Control variables ...27 3.2.5 Interaction variables ...29 3.3 CONCLUSION...29
DATA COLLECTION AND ANALYSIS...30
4.2 DATA EXPLORATION...33
4.3DATA ANALYSIS...34
4.3.1 Collinearity ...34
4.3.2 Normality ...34
4.3.3 Analysis estimation model ...35
CONCLUSIONS, LIMITATIONS AND APPRAISAL...38
5.1CONCLUSIONS...38
5.2 LIMITATIONS...39
5.3 APPRAISAL...39
REFERENCES...41
Preface
My interest for political economics and more specific institutional economics started with a course in Rome about topics related to innovation economics. Innovation, the development and implementation of technology, is a phenomenon where the proper interplay between technology, businesses, government policies and public institutions is crucial to success. If that interplay works out well the effects can be immense for economic growth as well as for improving welfare in the form of anything from wheels to micro-processors that makes life easier and richer. This ‘whirl’ of interlinked actors and effects fascinates me since the day I started to read newspapers and watch the news.
Besides reading about economics and institutions I also had the luck to have some practical experiences from different sides and in different roles in the ‘whirl’. As a student, I experienced a little bit of the endless difficulties academics encounter when trying to approximate reality in models attempting to discover something that could be called ‘the objective truth’. As treasurer of the Economics Congres in Groningen I noticed what it is to deal with companies, both as sponsors and contractors, and I experienced what it is to take care of the ‘bottom line’. As the ‘guy-studying-economics’ in an engineering team developing a Human Powered Aircraft I became aware of what it is to be an entrepreneur and owner of a technology start-up, having to deal with public institutions, technology partners and sponsor partners. Besides experiencing how great and powerful it is to work with technology of which I learned a lot among my technology crazy engineering friends in Delft. As a consultant for the largest business consultancy firm in the Netherlands I worked on optimization processes for a large ministry and for a multinational energy company. Both are large players in the same complex economic and societal system but have totally different objectives, approaches, attitudes and problems. In my hopefully many professional years to come I hope to participate in this fascinating dynamic game in many more roles and from many different sides.
One of the places where the whirl of developments is most prominent and dynamic is in transition economies, for example in the new EU member states, which is the reason I choose it as the subject of my thesis.
I would like to thank professor Dirk Bezemer for his course Advanced International Economics & Business and especially for his advice and support my thesis writing process.
I would like to thank my family and friends for their endless care and energy.
Simon P. Muller
Introduction
Businesses are the core of the economy they organize the creation of products and services. Government policies and other public laws and guidelines form the rules of the game in which the businesses play. Institutions also have a strong influence in the economy. They facilitate businesses, helping them to boost growth and economic progress, but they can also hinder businesses, stop change and cause chaos. It is obvious that those actors are intertwined in action and effect. There are ties among themselves and with anything else in a society.
The role of the major spheres in an economy, the business system and institutional system, and their interrelatedness becomes even more profound in extreme situations and scenarios, such as in economic transition processes. Transition economies are in a highly complex situation where economic, political and institutional settings are all changed dramatically and change society with them. That complexity makes transition processes extremely vulnerable to errors and mistakes, while its success is crucial for the overall welfare of a society.
A long list of historical examples shows that executing the optimal set of policies, both by governments and businesses, is practically unachievable. If a transition policy happens to work out well this is often the result of mere luck, although the responsible policy makers will naturally explain it as their achievement. What determines the success of transition economies?
The wide array of possible scenarios, extreme difficulty in the selection of the optimal course of action and the large potential gains make transition processes an interesting field of study. A lot of elements can be investigated which might yield insights into transition processes and economies in general. Still, transition processes share only very few common characteristics due to their unique circumstances. Transition processes are a period of strong change with the intention to switch from a backward system to a newer model influencing society as a whole.
This study combines the three elements introduced above in its approach: the interplay of institutions, companies and government; the complexity of transition processes; and change as a common characteristic. More concrete this study investigates the relationship of the two main spheres of the economy, the business system and institutional system, in transition economies by comparing their changes relative to each other. The research question that follows is:
What effect does a match of the change of the configuration of the financial system and the change of economic concentration have on the economic growth of a transition economy?
As will be explained in the literature overview the configuration of the financial system is a central characteristic in the institutional system and economic concentration is a central characteristic of the business system.
are tested using of empirical data and quantitative analytical techniques. Other non-transition economies are also taken into the analysis for comparison.
Outline
Chapter 1 Literature review
Chapter 1 first introduces the several approaches to the measurement of the success of transition economies. The second part outlines the varieties of capitalism literature in which several models are introduced that can form the basis for further quantitative analysis. Section 1.3 explains what is missing in the literature introduced in the first two sections. This omission is the starting point of chapter 2 which specifies the theoretical background for the methodology explained in chapter 3.
1.1 Approaches to measurement of the performance of transition economies
This study is related to literature that attempts to measure and compare the performance of economies in general and more specifically transition economies. For research that measures economic performance transition economies are a willing subject for the identification of influential elements on economic development and performance. Hare (2001) explains the generally assumed causal effect among the groups of actors that influence economic performance. Economic performance is most directly influenced by the behavior of companies, which together form the business system. The business system in turn is influenced by institutional settings and by government policies. The latter influences the business system directly and indirectly through the institutions.
1.1.1 Policies
The first element in the debate over the assessment of the success of economic transition processes is the configuration of government policies. Examples of studies that point to changes in government policies to explain differences in economic performance among countries are two articles of Sachs and Warner (1995 and 1997), which investigate what macroeconomic reforms indirectly improve economic development. The studies emphasize the importance of open trading policies to increase the openness of economies and the importance of the protection of private property rights, as well as the creation of institutions that secure the state’s income.
1.1.2 Institutions
Secondly, several studies focus on the influence of institutional settings on economic performance. An interesting part of the literature taking the institutional approach evolves around the ‘complementarity of institutions hypothesis’. This hypothesis states that the performance of institutions depends on its relations with other institutions and the context in which it operates (Boyer, 2005). Amable et al (2005), for example, explain the influence of the strength of labor unions on the strength of firm management. Nicita et al (2004) explain corporate diversity by pointing at the different combinations of complementarity among governance and financial institutions that influence the business system. In a similar way Pollin et al (2005) explain the European diversity in governance systems.
multiplicative interaction effects between variables proxying for institutions operating in the spheres of corporate governance and labor relations.
1.1.3 Business system
Third, the most direct influence on economic performance is by the behavior of companies that together form the business system. This relation is investigated by, for example, Hare (2001) who shows that three types of changes in the business system exist; first, the entry of new companies; second, the restructuring of existing companies; third, the exit of failing companies. Carlin et al. (2001) adds that the second effect, the restructuring of companies, is the result of incentives by policies or by institutions, while the first and third effects are the results of a selection process that is influenced indirectly by policies and institutions that help the market to signal opportunities and changes.
The role of policies and institutions on the business system and the related economic performance is the topic of a substantial line of research. Djankov and Murrell (2000) reviewed 125 studies in search for general lessons and implications for restructuring processes of companies. They highlight the importance of privatization, enterprise budget constraints and a moderate degree of product market competition.
1.1.4 Pace of transition processes
Finally, a fourth element in the debate over the assessment of the success of economic transition processes is the pace of changes in relationship to economic performance, gradual change versus a big bang. Kozminski (2002) compares the shock treatment in the Polish transition process with the gradual transition in Hungary. He finds that the whole debate around the pace of transition does not make sense because in reality shock reform is no option since the institutional system has not enough power and resources to direct the business system. He shows that any transition in post-communist countries will be lengthy and gradual no matter what, because of ‘the inertia of mental sets, institutions, and structures, the inevitable scarcity of material resources and managerial skills, and the resistance of the post-communist political formations’.
1.1.5 Combination of policies, institutions, the business system and economic performance
Two elements in the above described approaches are useful for the theoretical underpinning of this study. First, the influence which policies, economic institutions and the business system have on each other and on economic performance. Policies help developing institutions and business, institutions facilitate businesses and the performance of the business system directly determines the aggregate economic performance.
Second, the literature emphasizes the importance of the pace of transition processes. The scope of the discussion in the literature suggests a larger diversity in the pace of transition processes than can be observed in reality. Especially high pace shock reforms are only feasible in relatively short periods of time and to a very limited extent.
1.2 Varieties of capitalism
early developments of capitalism in eighteenth-century Britain. This section first provides an overview of the major subjects of debate in the varieties of capitalism literature, illustrated and clarified by short rendering of the key insights provided by the most influential contributions in the literature. Followed by a short chronological overview of the development of the varieties of capitalism literature which shows the logic in the development of the literature and places the major approaches of the analysis in perspective.
The general objective of the VOC literature is the identification and analysis of the differences between the numerous characteristics of countries’ economic and institutional settings. Scholars continuously discover and add new subjects of comparison and attempt to describe and analyze distinctive forms of capitalism from a large number of different perspectives. The literature hardly agrees on a single trend in the changes of the scrutinized elements making the VOC literature a lively and diverse field of science. As Whitley (1999) remarks about future changes of forms of capitalism: ‘Convergence to a single most effective type of market economy is no more likely in the twenty-first century than it was in the highly internationalized economy of the late nineteenth century’.
1.2.1 Topics in the varieties of capitalism literature
The VOC literature can be roughly divided into groups of studies sharing a similar variable as subject to analysis. One of the fundamental debates in the VOC literature evolves around globalization. Do various forms of capitalism convergence towards a single form due to the pressure of global markets and increasing economic integration? (Maurice, Sellier and Silvestre 1986; Koechlin 1995; Hirst and Thompson 2005; Wade 1996; Kenworthy 1997) A parallel discussion evolves around ‘standard typologies of capitalism’ aiming at identifying one or multiple ‘ideal’ forms of capitalism, towards which eventually all economies should converge (Esping-Andersen 1990; Albert 1993; Soskice 1999; Amable 2003).
Contrary to the proponents of convergence, many contributors emphasize diverging trends and attempt to explain differences by pointing at the influence of diverse cultural and social contexts, the importance of the so-called ‘embeddedness of institutions’ (Whitley and Kristensen 1996; Hollingsworth and Boyer 1997; Held, McGraw, Goldblatt and Perraton 1999). Others point at the differences in the accumulation of historical events and choices, for example in the debate on the path dependence of the development of forms of capitalism (North 1990; Herrigel 1996). Many aspects of these cultural, societal and historical contexts are examined in detail by comparisons on regional, sectoral and industrial levels within and among countries.
Besides the literature on specific issues, many studies focus on certain global areas when comparing practices between economies. For example, variations in types of dominant firm, customer-supplier relations, employment practices, work systems, innovation systems and technological developments, coordinating institutions and the role and strength of the state have been compared between the regions and countries of the Western world (Whitley 1992b; Lane 1992; Nelson 1993; Hollingsworth, Schmitter and Streeck 1994; Kristensen 1997; Menz 2005).
countries, increasingly more attention is paid to new forms of capitalism in emerging countries in Asia and Eastern Europe.
1.2.2 Initial varieties of capitalism literature, state institutions
Below, a short chronological overview of the development of the VOC literature presents (examples of) the most influential contributions and concludes with the three most recent and influential approaches of analysis. The core of the varieties of capitalism literature consists of three subsequent perspectives on institutional variation that are related to the main problems economies encountered in three related periods after the Second World War. The first perspective developed in the post-war decades when economists were confronted with the challenge of how to modernize industries dominated by pre-war practices and how to achieve sustainable high rates of economic growth. Believed to be crucial was the set of actors that possesses the strategic capacity to devise and impress plans for industries and sectors. As a result the focus was on the institutional structures that give states leverages over the private sector, such as planning systems and instruments to inflict public influence over the flows of funds in the financial system (Hall and Soskice 2001).
The first influential scholar who elaborated on the importance of the coordinating role of the state was Shonfield (1965). The central claim of his work is that economic growth can be enhanced by technological innovations and the training of labor. The success of changes depends on the political ‘will and skill' of the state. Another often cited early contributor of this line of research is Chandler (1977) who was one of the first to do research on the coordinating institutions of industrial capitalism. The main rationale in his work is that the invisible hand of the market is replaced in many sectors by the visible hand of managerial coordination and control.
A series of studies investigated and compared the role and importance of institutional structures on economic functioning and performance through planning and coordination. Two examples are Estrin and Holmes (1983) and Cox (1986). Estrin and Holmes’ work examines the French national planning practices. They underline that planning has a basic role in all advanced industrial economies faced with uncertainties. However, planning should be limited to supporting the private market system which should remain the basic form of economic activity. Cox presents comparative analyses of the relationships between the state, financial and industrial development and investment decision-making in Japan, France, West Germany, Italy, the USA and the UK. The question is whether there exists a best model for these relationships. Cox’s conclusion reads that those countries that have worked out a national consensus over the need to underwrite the long-term financing of the industry have also been those with the most successful post-war record.
Related is the line of research that categorizes between ‘strong’ and ‘weak’ states to describe the structure and the power of the state. This is one of the early attempts to classify economies, which later led to the construction of ‘ideal’ typologies of capitalism. Johnson (1982) presents a history of Japanese industrial policy from 1925 to 1973 from an institutional perspective and describes the continuous increase of the power of the state. The study shows the increasing role of the industrial policy from inconsequential pre-war strategies to the successful central role in Japanese policy making in the 1950 and 1960. In this first perspective in the VOC literature coordinated economies as France and Japan were seen as models of economic success, while more liberal economies were seen as lagging behind.
The second perspective to VOC, which emerged in the 1970s, is based on the concept of neo-corporatism and is a reaction to the pre-eminent problem of high inflation which the developed economies had to face. Neo-corporatism is generally associated with the capacity of a state to negotiate durable bargains with employers and the trade union movements regarding wages, working conditions and social or economic policies. Literature in this perspective categorizes countries largely by reference to the organization of their trade union movement. Accordingly, a nation’s capacity for neo-corporatism was generally said to depend on the centralization or concentration of the trade union movement (Hall and Soskice 2001).
Olsen (1965) was the first contributor to the line of research associated with neo-corporatism. He specified that more encompassing unions are better able to internalize the economic effects of their wage settlements and ensure a more effective outcome. Research in the tradition of Olson argued that any institutionalized system of wage restraint is unstable as long as unions have the right to strike with the intention to improve the interests of their members. This presents the trade-off between the advantages of collective bargaining against the difficulties of its organization and the risk of free-riding practices.
Examples of studies in the Olsonian tradition are Berger (1981) and Goldthorpe (1985). Berger describes the neo-corporatist solutions and policies in the insecure pluralistic economic situation in the 1970. Since pluralism is regarded as an unstable and ungovernable system of representation, the trend to corporatist solutions becomes significant, bearing in mind that the timing and content of state intervention has influence on the role and organization of interests and, indirectly, the patterns of economic development. Goldthorpe (1985) examines what neo-corporatist arrangements made it possible for some Western European countries to succeed in defying neoclassical expectations. He finds that the institutional arrangements for managing economic life in countries have served to produce surprisingly good economic results, particularly the cooperative arrangements between business, labor and government in the management of collective bargaining.
The importance of the centralization or concentration of the trade union movement is investigated by Calmfors and Driffill (1988). It examines the importance of the structure of labor markets, especially the degree of centralization of wage setting, and finds that it is increasingly perceived as a major determinant of the economic performance of a country. The main finding is that either highly centralized systems with national bargaining or highly decentralized systems with wage setting at the level of individual firms seem to perform best. This suggests that the neo-corporatist approach is just one of the optimal strategies to work out optimal configurations and economic performance.
1.2.4 Business system and institutional system
1.2.4.1 Coordinated market economies versus liberal market economies approach
The first line of research evolves around the contrast between coordinated market economies (CME) and liberal market economies (LME) (Esping-Andersen 1990; Chandler 1990; Albert 1993) which culminated into the comparative framework of analysis of Soskice (Soskice 1997, 1999; Casper, Lehrer and Soskice 1999; Hall and Soskice 2001). The framework highlights how differences in institutional arrangements, such as welfare systems, employment laws and conventions, the organization of business associations, training systems, financial markets and legal systems, generate variations in innovation strategies, sectoral specialization and technological development, leading to the formation of distinctive forms of capitalism.
According to Soskice (1999) CME’s, such as Germany, Sweden and Switzerland, build institutions that facilitate the development of organizational competences in the coordination of knowledge and skills across organizational boundaries by encouraging long term cooperation and investments in firm and industry specific skills in order to develop continuous and incremental innovation strategies. LME’s, such as the UK and USA, promote short-term behavior of firms by facilitation of the rapid use of new knowledge and skills. Employees are highly mobile between firms and skills tend to be more generic across organizations. The institutions in LME’s support firms to develop radical innovations in newly emerging technologies in order to seize new opportunities and markets. Still, the variety of firm and innovation strategies within each type of market economy is often greater than the strict contrast between CME and LME would suggest, as underlined by empirical findings (Casper 2000; Hancke 2001). Therefore, several efforts have been made to further develop and differentiate this line of research by addition of more specified and detailed typologies (Schmidt 2002; Amable 2003).
1.2.4.2 The societal effects approach
The hiatus in the ‘CME – LME’ approach suggests the need for a more comprehensive framework for comparing varieties of capitalism that explains how distinctive systems of capitalism become established, reproduced and changed. The ‘societal effects’ approach (Maurice, Sorge and Warner 1980; Sorge 1991; Maurice and Sorge 2000) shows how differences in institutional arrangements has consequences for the ways in which firms organize and control work activities, establish networks and develop growth strategies. The major downside of this approach is the lack of a substantial amount of comparative studies that focus on how these differences in institutional arrangements led to the development of distinctive systems of economic organization.
In contrast, the ‘social systems of production’ approach does provide empirical evidence and focuses specifically on how different forms of market strategies and economic coordination have become established in particular institutional contexts, and examines variations in sectoral organization across countries (Campbell, Hollingsworth and Lindberg 1991; Hollingsworth, Schmitter and Streeck 1994). These studies distinguish a number of social systems of production, which are complexes of economic organization and societal institutions that vary in their economic outcomes (Hollingsworth and Boyer 1997).
associations (Campbell, Hollingsworth and Lindberg 1991). These studies further concentrate the scope of research through the focus on sectoral level analysis and explain how sectors develop and change, and how firms coordinate activities across sectors and alter sectoral boundaries.
Unfortunately, the concern with industry-level coordination implies a disadvantage for analysis, because it tends to obscure more general features of economic organization institutionalized across sectors. The comparison of economies shows considerable national differences in markets, hierarchies, networks, states and institutions, and the distinctiveness of these general characteristics occurs to be substantial (Hollingsworth, Schmitter and Streeck 1994; Hollingsworth and Boyer 1997). This emphasizes the need of systematical identification of characteristics of economic coordination and control relationships across national boundaries. This would also further clarify the processes by which configurations of the state, financial systems and labor systems is established and changes, and how these systems structure the governance regimes and production systems.
1.2.4.3 Comparative business systems approach
The influential ‘comparative business systems’ approach attempts to develop a framework without these weaknesses that describes and explains differences in the nature of economic actors, forms of economic organization and changes in terms of variations in particular kinds of societal institutions (Whitley 1992a, 1992b, 1999a). The framework identifies the key differences in the organization of capitalist economies that constitutes distinctive business systems. These variations in forms of economic organization result from different institutional settings that encourage particular economic rationalities and discourage other ones (Whitley 2002).
The framework summarizes the ways in which business systems vary as coordination and control systems and identifies the key institutional features that interrelate with them. Business systems vary on two basic dimensions; the extent of ownership coordination of economic activities and alliance integration or non-ownership coordination. The key institutional arenas that govern the development of different business systems deal with; property rights and disputes about them, the organization and regulation of collective actors and competition and conflict between them, access to capital on particular terms, the development of skills and knowledge, the organization and control of skills in labor markets and authority and trust relationships. The business and institutional systems are organized in different ways in different societies, but develop interdependently to constitute distinctive institutional contexts at regional, national and international levels (Whitley 1999a).
A number of ideal types of combinations of institutional features can be identified that theoretically can be expected to generate distinctive kinds of business systems. Nevertheless most empirical social systems are comprised of conflicting features and contradictory tendencies (Whitley 2002).
1.3 Opening in the literature
and compares two extreme paths. If instead change is measured this opens up the possibility of comparing change among the sub spheres of the economy. For example, in line of the complementarity of institutions theory one could suggest that complementary paces of change would reinforce the strength of institutions or sub-spheres of the economy. The other way around, one could also suggest that the complementary forces of institutions to converge their change behavior.
In more concrete terms, what is missing is an empirical research investigating the complementarity, or match, of the change of the configuration of the financial system, a proxy for the institutional system, and the change of the economic concentration, a proxy for the business system and the effect of this match on economic performance.
1.4 Conclusion
Chapter 2
Comparative business systems model
Chapter 2 presents the selection and specifications of the theoretical basis for the quantitative analysis in chapter 3. Section 2.1 elaborates on the lapse in the literature by specifying the approach of this study to fill up this omission. Two examples are given that illustrate the importance of this approach. Section 2.2 first specifies the four requirements for the selection of the most optimal model out of the varieties of capitalism literature. Secondly, the most optimal basic model is selected. Section 2.3 presents the selected comparative business systems model. It further discusses the selection of what characteristics of the total model can best be used as a basis for the quantitative analysis. Finally, proxies are specified that best approximate values for those characteristics.
2.1 What makes it worthwhile to investigate the match of change?
The match of change concept is useful to indicate how well the business system and institutional system change adapted to each other. In general one could assume that a transition economy with a low match value is successful since this would imply an efficient and coordinated transition process, which would have a positive effect on the economic performance. As illustrations of this notion two examples of opposite transition scenarios are presented below which show how the methodology and terminology can be applied on transition economies.
The first example is the Russian attempt to shock-reform both the institutional and the business system in the early 1990s, in other words, to change both systems on a high pace. In order to change the business system quickly 14000 large and mid-sized state owned enterprises were privatized in a true mass privatization process, which turned 90% of the industrial workforce into the private sector (Puntillo, Ipsen and Dietrich 1996). Similarly, attempts were made to realize a large reform program entailing both macro economic and institutional changes, which failed to advance quickly and carried on slowly throughout the 1990s and into the 21ste century. Consequently the institutional system was very weak in the early 1990s which resulted in a strong recession between 1992 and 1997, and several forms of economic abuses, such as high levels of corruption and the emergence of so called oligarchs who managed to gain control over entire industries within years. In the last years, in an attempt to bring the economy back to a healthy balance and improve economic growth, the government increased the strength of the institutional system through continuous economic reforms and sometimes exceptional measures, such as, for example, the nationalization of some oil companies. The latter can thus be seen as an attempt to let the institutional system adapt more closely to the business system. (EIU Country Profile Russia 2005, Federal State Statistics Services Russia) The example that shows an almost opposite scenario is the continuous process of economic and political reform in China. From 1978 onwards the Chinese government directed the changes in both the institutional and business systems. Although three decades of economic growth showed strongly volatile rates of growth, on average the Chinese economy performed exceptionally well growing 9% annually. A closer look to the Chinese reform process shows that it entails an endless list of small steps that by now affected practically all economic sectors and led to some major advances towards the creation of a market economy. The style of the Chinese transition process can best be characterized by the Chinese phrase ‘crossing the river by feeling for the stones’. In terms of the methodology, the institutional changes are constantly and closely matched to the developments of the business system (EIU Country Profile China 2007).
2.2.1 Requirements model
To facilitate the selection of the appropriate model in the VOC literature, criteria are outlined below to serve as a guideline.
First, the model should outline the general characteristics of business and institutional systems. It must be possible to quantify the characteristics and data must be available for measurement. In order to facilitate the comparison and analysis of the performance of business and institutional systems in different countries the characteristics have to be of a general nature, because they have to apply to all countries and not only to specific groups of countries, such as high income countries or countries in specific parts in the world. The characteristics should be quantifiable, because of the quantitative nature of the framework of analysis that is constructed. Finally, data must be accessible and abundant to facilitate a valuable analysis of the characteristics.
Second, the model should integrate and link the business and institutional systems in general terms and on the level of the main characteristics. The model should structure and relate the characteristics in order to facilitate the construction of the framework of analysis which links the transition paces of the both systems.
Third, the model should be comparative by nature, because it has to serve as a basis for a comparative framework of analysis. It is likely that the VOC field offers suitable models, because comparative analysis is the major instrument and in the nature of the VOC tradition.
Finally, the model should be designed for comparative analysis at the state level, because the state level is the highest level at which relatively homogeneous business systems can be found, due to the common institutional, cultural and historical background. Rasanen and Whipp (1992) argue that a country may have multiple business systems if it exhibits institutional pluralism, different local social networks, cultures and political strategies towards the state, or if nation-wide institutions are locally implemented in different ways. Therefore Whitley (1992b) concludes that as long as the major institutions and the dominant culture retain their central position in a society considerable differences within and among business systems will not occur, which implies that the state is the optimal level of analysis.
2.2.2 Comparison models
The overview of the literature of the VOC field above presents three main approaches to the analysis of varieties of capitalism and the according models.
The first requirement demands that the model provides general characteristics of business and institutional systems that can be quantified and for which ample data is available. The ‘CME – LME’ approach makes use of several elements or characteristics to compare forms of capitalism as, for example, the nature of the financing and education systems, and the strength and role of coordinating institutions. Disadvantage of this approach is that not all characteristics are and can be applied to all cases, which implies that not all characteristics have a truly general nature and that comparisons are not made on the ‘characteristics-level’, but rather on the ‘typology-level’.
The ‘comparative business systems’ approach presents a framework that clearly lists the main general characteristics of business and institutional systems which are applicable to all comparisons of different forms of capitalism. This approach best suits the requirements for the basic model when the characteristics are considered.
The remaining problem for all three approaches is the availability and accessibility of data. Many of the characteristics are of a typical qualitative nature and the framework of analysis is applied to transition countries for which data are often non-existent or difficult to obtain.
The second requirement demands that the model integrates and links the institutional and business systems in general terms and on the level of the main characteristics. The most common way to integration the main characteristics is through the construction of typologies. In the ‘CME – LME’ approach the characteristics are related to each other by way of the CME and LME typologies. These two typologies are rather extreme and strict sets of characteristics. The question remains what has to be done with cases that do not fit one of these typologies, because, for example, characteristics of both typologies are mingled. As noticed above, this weakness is acknowledged in the VOC field and attempts have been made to refine the framework with a number of extra typologies (Schmidt 2002; Amable 2003). Still, the lack of a specific framework on the characteristics level makes specific analysis on both the business and institutional system and their interrelatedness difficult.
Again the same disadvantages apply to the ‘social systems of production’ approach, which provides six distinct types of governance regimes, but no typologies or frameworks that incorporate characteristics of both business and institutional systems, let alone on the characteristics level. The ‘comparative business systems’ approach has the major advantage over the other two approaches of having a single framework that relates the main characteristics of business and institutional system into one model. Furthermore, this approach presents linkages between the business and institutional systems by way of 6 ‘ideal’ types of business system correlated to their matching institutional settings. This model represents the best basis on which the framework of analysis can be constructed and the transition paces of the characteristics of both systems can be linked.
Regarding the third requirement all approaches apply since all originate from the VOC field and are set up to compare different forms of capitalism.
The fourth requirement is satisfied by the ‘CME – LME’ and the ‘comparative business systems’ approaches, but not by the ‘social systems of production’ approach since it focuses on sectoral analysis. This implies the disadvantage that a large amount of research is needed to compare on the country level properly, since multiple sectoral analyses must be executed to provide a country level overview.
The ‘comparative business systems’ model offers the best basis for the framework of analysis and further methodology. It offers the best general characteristics of both business and institutional systems, and it offers the best framework that links these characteristics together.
2.3 Comparative business systems model
(1999) in which it is complemented with applications of the model on Asian and Eastern European cases. A general characterization and background of Whitley’s model is shortly outlined below. 2.3.1 Business system
Business systems can be differentiated according to differences in their economic organization. They vary in their degree and mode of coordination as well as in the configuration of economic activities. Central in the coordination and control systems are the relationships between five types of economic actors; the providers and users of capital, customers and suppliers, competitors, firms in different sectors, and employers and employees. Instead of assuming international similarities, the main characteristics outlined below serve to identify variations in the nature and behavior of aggregate firm behavior. Important differences are in the degree of organizational integration among economic actors and the way through which integration is achieved, such as ownership-based hierarchies, formal agreements, personal obligations and informal commitments. The general characterization of business systems suggests three characteristics and eight dimensions for comparing business systems which are outlined in figure 1.
Figure 1: characteristics of the business system Ownership coordination
1. Primary means of owner control (direct, alliance, market contracting) 2. Extent of ownership integration of production chains
3. Extent of ownership integration of sectors Non-ownership coordination
4. Extent of alliance coordination of production chains 5. Extent of collaboration between competitors 6. Extent of alliance coordination of sectors Employment relations and work management
7. Employer-employee interdependence
8. Delegation to, and trust of, employees (Taylorism, task performance discretion, task organization discretion)
Source: Whitley (1999)
The number of possible different types of business system, described in terms of the business system characteristics and dimensions outlined in figure 1 is restricted empirically by their interdependence with institutions. Distinctive forms of business system are shaped by their particular institutional context when institutional features are integrated into and mutually reinforcing with business systems. 2.3.2 Institutional system
suggests four characteristics and eleven dimensions for comparing institutional systems which are outlined in figure 2.
Figure 2: characteristics of the institutional system The state
9. Dominance of the state and its willingness to share risks with private owners 10. State antagonism to collective intermediaries
11. Extent of formal regulation of markets Financial system
12. Capital market or credit based Skill development and control system
13. Strength of public training system and of state-employer-union collaboration 14. Strength of independent trade unions
15. Strength of labor organizations based on certified expertise 16. Centralization of bargaining
Trust and authority relations
17. Reliability of formal institutions governing trust relations 18. Predominance of paternalist authority relations
19. Importance of communal norms governing authority relations Source: Whitley (1999)
2.3.3 Selection of measurable characteristics from the comparative business systems model
From the comparative business systems model presented above two of the characteristics, one representing the business system part of the model and one representing the institutional system part of the model, are selected and incorporated into the framework of analysis. The selection and incorporation of just two characteristics is due to constraints in time and resources. It is beyond the possibilities of this study to offer a sufficiently focused and detailed analysis of multiple characteristics when dealing with such broad and complex topics.
Four selection criteria are important to consider in the selection of the optimal characteristics. First, because the analysis is quantitative a methodology has to exist that points out how to measure the concepts in the analysis. Second, since the analysis is based on quantitative data and statistics, it is required that quantitative data and statistics exist that are sufficiently related to the core of the topic, available for all countries, available for sufficient years in order to conduct time series analysis, are of consistently good quality and from a reliable source. Third, the characteristic should have a clear direct correlation with economic performance so that changes in the investigated characteristics can be related to changes in economic development. Fourth, a related point is that the characteristic should have a relatively quick influence on economic performance for the same reason. The four criteria are taken through below for first the business system characteristics and next the institutional system characteristics.
Non-ownership coordination is harder to measure, since it cannot be described exactly in one dimension and is thus difficult to count. Hardly any data is available and certainly no comprehensive data for multiple countries or years, since it is not registered or measured easily. Its correlation to economic performance and speed of influence is likely to be equal to ownership coordination but harder to establish.
The employment relations and work management characteristic is difficult to measure directly due to the qualitative nature of the concepts. As a result data are not abundant and hardly comparable among studies. The direct influence of employment cultures and management trends on economic performance is difficult to establish and likely to be subject to substantial time lags.
The state characteristic focuses on the administration’s culture and attitude towards the business system, which are concepts with a qualitative nature and thus hard to measure. As a consequence, little data can be found, let alone of proper quality. Its influence on and correlation with the economic performance is likely to be quite indirect as well as slow and thus hard to establish or measure. The financial system characteristic is relatively easily measurable in a direct way, which is illustrated by the stream of literature and reports doing so. Data is therefore abundantly available and of proper quality. The influence of the financial system on economic development can be established directly, the availability of capital is crucial for any economic activity, and the influence is quick and relatively easy to relate.
The skill development and control system characteristic can be measured, for example by counting the amount of people involved in education and training programs or by calculating the relative strength of trade union in percentages of membership. Some data is available, but of inconstant quality, often incomparable and limitedly accessible. The relation to economic performance is merely indirect, although for example centralized bargaining can have large direct effect on economic performance. And the relation on economic performance is likely to have a time lag, since skill development institutions are not easily or quickly reorganized.
Finally, the trust and authority relations characteristic focuses on the nature of the relations of formal institutions with the business system. Because of the qualitative nature of relations it is hard to measure and data barely exists. Furthermore the influence on economic performance is indirect and slow, and thus hard to relate or compare to economic performance.
From the above it is clear that the ownership coordination characteristic for the business system and the financial system characteristic for the institutional system are the most suitable characteristics to incorporated into the framework of analysis.
2.3.3.1 Proxy for the business system characteristic ownership coordination, Hirschmann-Herfindahl Index
The analysis of business systems evolves around the control and coordination of economic activities. Fundamental in the analysis of different forms of business systems is who coordinates and controls property rights, economic resources and activities, and in what way their owners are organized. Whitley identifies three dimensions for this characteristic which are central in distinguishing different forms of ownership coordination: primary means of owner control; the extent of ownership integration of production chains; and the extent of ownership integration of sectors.
how the economic activity in the aggregate economy or in a specific market is divided among companies and thus in fact indicates how companies own and divide the economic activity.
The quantitative analysis evolves around one dimension, aggregate economic concentration, which combines both horizontal and vertical integration linkages. Aggregate concentration is expressed by the share of economic activity that the largest 100 companies hold in the aggregate economy. Ownership coordination on the aggregate level ranges from an economy completely owned and controlled by the largest 100 companies, to an economy in which the largest 100 companies do not have more coordination power than the other companies in the same economy.
The statistic used to calculate a value on the measurement scale is the Hirschman Herfindahl Index (HHI). The HHI is calculated by summing the squares of the shares of each individual firm in the aggregate economy or in a specific industry. The major benefit of the HHI in comparison with for example the normal concentration ratio is that it gives more weight to larger firms, because the market shares are squared prior to being summed, giving additional weight to firms with a larger share size, which better reflects the real competitive positions of companies in an economy or industry (Martin, 2002). Normally, the data of all companies in the aggregate economy would be required to calculate the HHI, but since those data are not available the statistic is calculated for the top 100 companies of the aggregate economy, which makes the statistic a ‘quasi’ HHI. The company group size is 100 because it simply is the largest commonly used amount for analysis of economic concentration (White, 2002) where company data of sufficient quality can be found for. Using a smaller amount of companies is also not desirable because it would reduce the prediction value of the statistic.
2.3.3.2 Proxy for the institutional system characteristic financial system, structure-size measure A critical part of the institutional context to the business system is the financial system that facilitates the exchange of financial assets and indirectly the reallocation of resources. According to Whitley (1999) the critical dimension by which financial systems are differentiated is the process by which capital is made available and priced. The two main forms by which capital is supplied is through capital markets and through the banking system. The former allocates capital by use of capital markets in which commodities are priced through a competitive market system (Zysman 1983). In credit-based financial systems this function is executed by a set of intermediaries, such as the large, universal banks in Germany, which deal directly with firms and other economic entities (Cox 1986).
The quantitative analysis evolves around one dimension, which focuses on the amount of financing through the capital market relative to the amount of financing through banks. This dimension ranges from a totally capital-market-based financial system, in which the entire economy is financed through capital markets, to a totally credit-based financial system in which all capital is made available to the economy by way of credit institutions such as banks.
The statistic used to calculate a value for this dimension is based on the work of Levine (2001). His ‘structure-size measure’ denotes the size of capital markets, an indication of the part of the economy financed by use of capital markets, relative to the size of banks, an indication of the part of the economy financed by use of banks.
The size of the banking sector is determined by use of the bank loan ratio, which equals the total amount of domestic bank loans divided by GDP. The domestic bank loans are the aggregate of all credit facilities offered by domestic banks to the private sector. The ‘structure-size measure’ is obtained by dividing the market capitalization ratio by the bank loan ratio. A large value indicates that the capital market is strong while a small value indicates that the bank system prevails (Demirguc-Kunt and Levine 2002).
2.4 Conclusion
Chapter 3 Methodology
After the discussion of the literature in the previous chapters, this third chapter outlines first the specific focus of this study and presents the research question. Subsequently the research question is translated into a main hypothesis and several specific sub hypotheses per country selection. Finally this chapter outlines the methodology according to which the hypotheses are tested in chapter four. The methodology specifies the estimation model and introduces the estimation variables.
3.1 Framework
3.1.1 Focus of the study and research question
The objective of this study is to investigate whether a relationship exists between the match of the changes of the configuration of the financial systems and economic concentration on the one hand and Gross Domestic Product per capita on the other. The configuration of the financial system is the proxy for the institutional system, the economic concentration is the proxy for the business system and GDP per capita is the proxy for economic performance. The institutional and business systems are the two main sub-spheres of the economy. The research question that needs to be answered is:
What effect does a match of the change of the configuration of the financial system and the change of economic concentration have on the economic growth of a transition economy?
It is theorized that a close match of change leads to better aggregate economic performances in the form of higher GDP per capita growth. In the section below the main and sub hypotheses are stated. The exact determination of the country selections mentioned is presented in chapter below ‘sampling’. 3.1.2 Hypotheses
Having outlined the relevant literature and the focus of this study the next step is to state hypotheses which specify the relationship between the match of change of the business and institutional system on the one hand and economic performance on the other. In general it can be expected that a closer match has a positive effect on economic growth, as is explained in section 1.3 and illustrated in section 2.1. The main hypothesis specifies the relationship between the match of change and economic performance:
(1) H: A match of the change of the configuration of the financial system and the change of economic concentration has a positive effect on real GDP per capita growth.
This hypothesis can be further specified in several sections of the dataset. The total set is the total EU, the EU27, for which the main hypothesis applies:
(2) HEU27: A match of the change of the configuration of the financial system and the change of
transition countries. Considered the fact that changes in the transition economies are expected to be more pronounced and resulting from more fundamental changes, it can be expected that the match effect is also more prominent in the transition economies. In the more advanced economies growth rates are less high so the urgency of a very close match is less apparent. It could very well be that the systems match in the long term but deviate more in relative terms in the short term in the advanced economies than is the case for the transition economies.
(3) HEU15: A match of the change of the configuration of the financial system and the change of
economic concentration has a negative effect on real GDP per capita growth in the EU15.
(4) HEU12: A match of the change of the configuration of the financial system and the change of
economic concentration has a positive effect on real GDP per capita growth in the EU12.
Another specific group within the EU are the 13 countries that participate in the Monetary Union and have the Euro as their common currency. The EU13 form a sort of advanced elite group within the EU. It could be expected that the EU13 selection shows a similar effect as the EU15 selection, but than even more pronounced. Because their long term development is relatively very stable there is no need for any short term coordination of the systems.
(5) HEU13: A match of the change of the configuration of the financial system and the change of
economic concentration has a negative effect on real GDP per capita growth in the EU13.
Finally, another division can be made between the so called CMEs (Coordinated Market Economies), the LMEs (Liberal Market Economies) and a mixed form which is sometimes called Mediterranean Economies (Akkermans, Castaldi and Los, 2005). The coordinated economies are most likely to have strong matched business and institutional systems and are thus expected to have a similar relation to economic performance as the EU12 countries have.
(6) HCME:: A match of the change of the configuration of the financial system and the change of
economic concentration has a positive effect on real GDP per capita growth in the EUCME.
3.2 Methodology 3.2.1 Estimation model
A second similar assumption goes for the economic concentration as a determinant of the business system and the financial system as a determinant for the institutional system. Since those characteristics are just parts of the total systems the assumption has to be made that they cover a large part of the variation in the total systems.
A third important assumption states that no country or time specific effects are considered within a selection of countries analyzed. The coefficients of both the intercepts and slopes of the panel data are considered constant. This yields the main advantage that all data can be pooled and an ordinary least squares regression model can be run. Disadvantage is that the model is less specifically adapted to the data which might influence the outcomes. Sill, the main objective of this study is to detect the main effect and country or time specific outcomes are not likely in the first place due to lack of sufficiently detail and quality.
For the quantitative analysis of the relationship between the proxies and economic performance is the following linear equation formulated:
YGDPct = C + ctmZctm+ ctn Xctn + (3.1)
YGDPct is the value of GDP per capita growth of country selection ‘c’ in time ‘t’.
C is the constant.
ctmZctm is the coefficient ‘ ’ for country selection ‘c’ in time ‘t’ for the explanatory variable ‘m’ times the
value ‘Z’ for country selection ‘c’ in time ‘t’ for the explanatory variable ‘m’.
ctnXctn is the coefficient ‘ ’ for country selection ‘c’ in time ‘t’ for the control variable ‘n’ times the value
‘X’ for country selection ‘c’ in time ‘t’ for the control variable ‘n’. And is the error term.
The ctm coefficients are expected to be either positive or negative depending on the relative trend of
the business versus the institutional system. The ctn coefficient can also be expected to be either
positive or negative depending on the respective control variable. 3.2.2 Dependent variable
To test the hypotheses real GDP per capita growth is used as the dependent variable in the estimation model as a proxy for economic performance.
Instead of real GDP per capita growth it could be suggested that the natural logarithm of GDP per capita is used as a dependent variable. Unfortunately this implies that the difference statistics are in fact based on nominal instead of real GDP data, which alters the outcome of the analysis severely and is therefore not applied.
3.2.3 Explanatory variable
The explanatory variable is the match indicator which is the absolute difference between the two absolute change values in percentages of the structure-size measure and the HHI100:
ABS(((ABS( S(%)))-(ABS( H(%))))tc (3.2)
In equation 3.2 the absolute difference is calculated over the absolute values between the two sub variables. The reasons for taking the absolute values is that this approach does attempt to measure change and does not attempt to measure a specific course of action, which would be the case if the sign was taken into measurement and discussion too. A judgment of the rightness of the sign of a certain change is outside the scope of this study. The sole focus is to notice whether a change occurs in system a when system b changes and the other way around. Perhaps the first time a negative change as a reaction to a positive change is the right course of action while the next time a positive change might be the right reaction to a similar positive change of the other system.
The absolute percentage change value of the structure-size measure for the present period is calculated as follows:
ABS( S(%))ct = ABS((((structure-size measure)ct-1) / (structure-size measure)ct)-1)*100) (3.3)
where ‘t’ is time in years and ‘c’ is country selection.
The structure-size measure itself for the present period is calculated as follows:
structure-size measurect = market capitalization ratioct / bank loan ratioct (3.4)
where ‘t’ is time in years and ‘c’ is country selection.
The market capitalization ratio for the present period is calculated as follows:
market capitalization ratioct = total stock market capitalisationct / GDPct (3.5)
where ‘t’ is time in years and ‘c’ is country selection.
The bank loan ratio for the present period is calculated as follows:
bank loan ratioct = total bank loansct / GDPct (3.6)
where ‘t’ is time in years and ‘c’ is country selection.
The absolute percentage change value of the Hischmann-Herfindahl Index of the top 100 companies (HHI100) for the present period is calculated as follows:
ABS( H(%))ct = ABS(((HHI100ct-1 / HHI1000ct)-1)*100) (3.7)
where ‘t’ is time in years and ‘c’ is country selection. The HHI100 for the present period is calculated as follows:
HHI100ct = ( 100 (turnover companyn)2) / aggregate turnoverct (3.8)
where ‘t’ is time in years and ‘c’ is country selection.
A closer match of the business system and the institutional system is expected to have a positive effect on economic performance. The relative behavior of the proxy for the business system, HHI100 and the proxy for the institutional system, structure-size measure determine the value of the match variable. The HHI100 is calculated from the turnover of the 100 largest companies, individually squared, then summed and finally divided by aggregate turnover. When the turnover of the top 100 companies goes up or the aggregate turnover goes down, the HHI100 goes up. Whether this creates an increases or a decrease in the value of the match variable is unsure, since that depends on the position of the HHI100 relative to the structure-size measure. A similar reasoning applies to the influence of the structure-size variable in the match variable, since its influence also depends on the value of the HHI100.
To ensure that the variance in YGDPct is the result of Zctm it is necessary to exclude other factors that
might have an impact. Therefore control variables Xctn are included in the regression analysis. These
control variables are expected to contribute to the variance in the value of the dependent variables. Barth, Li, McCarthy, Phumiwasana and Yago (2006) explain: “There is no consensus regarding an appropriate theoretical framework to guide cross-country empirical studies on economic growth”. However, although empirical work explaining variation in these types of variables is quite diverse in terms of the particular specification used, they do use a similar set of control variables. The set of control variables for real GDP per capita growth are presented below.
Population growth
If the population grows it is to be expected that the labor force grows as well which increases the productivity power of the population and therefore the total product they produce. On the other hand the total population can also grow without adding extra labor force, babies for example do not work in Europe. In the latter situation the GDP per capita growth might be influenced negatively. If the former situation applies the GDP per capita growth could be positive. Population growth is added as an control variable to prevent the increase or decrease in the GDP per capita growth is the effect of a change in the average population.
Inflation rate
Inflation reduces the relative value of money and the products expressed in terms of money. If inflation is higher than the percent increase of the GDP, the GDP per capita decreases, assuming other variables ceteris paribus. If on the other hand the percent increase in the GDP is larger than inflation the GDP per capita increases. To make sure changes in the GDP per capita are not the result of inflation, the inflation rate is added as a control variable.
Gross capital formation as a percentage of GDP
Gross capital formation as a percentage of GDP is the amount of investment by the private sector relative to the GDP. In general in can be expected that a higher amount of capital formation yields more products afterwards and thus increases GDP per capita growth afterwards. Gross capital formation is thus added as an control variable to make sure a change in the GDP per capita is not the result of an increase of decrease of the amount of investments made.
Trade from the previous period as a percentage of GDP
Three other control variables that are regularly included to control for real GDP growth are real GDP per capita from the previous period initial income and secondary education. Real GDP per capita from the previous period is excluded from the regression because it combines the variable real GDP per capita and inflation which are both already included in the analysis. Initial income and secondary education are dropped from the analysis because they are time invariant.
3.2.5 Interaction variables
To investigate whether the explained variance in YGDPct by the independent variables can be reinforced
it is tested whether interaction effects exist among the match variable and the control variable by way of interaction variables. A significant interaction variable would show that the match variable is dependent on the level of one of the control variables. This would reinforce the explanatory power of the overall estimation regression. The interaction variables can be computed as follows:
Interaction variable = control variable * explanatory variable (3.9)
3.3 Conclusion
This chapter presents the framework and methodology that can answer the research question in the subsequent chapters. Seven hypotheses are set up which each express a relationship of the match variable for a specific selection of EU countries with economic performance. The estimation model that can measure these relationships is introduced together with the dependent, explanatory and control variables.
