The Effects and Determinants of Economic Diversity
and how it relates to Competition and Employment Protection Legislation
Christiaan Burggraaff
University of Groningen, MSc thesis
Abstract
The important role of Economic diversity in the economy is assessed for 271 European regions, looking at what causes certain regional economies to be more diverse than others, and what the effects of higher economic diversity are. As determinants of diversity, a higher Education level, Population level, and GDP per capita, fiercer Competition and more flexible Employment Protection Legislation are all concluded to have a positive effect on diversity. Islands are found to have a lower level of economic diversity.
Diversity, Competition and Employment Protection Legislation are argued to be interdependent and to have an effect on the same key economic variables. Their theoretic and empirical effects on Employment growth, Productivity growth and the Volatility of employment are analyzed and compared, which offers interesting insights into the interplay of these three variables and their spillover effects on the wider economy.
The dataset covers the period 1995-2008, and allows the study of economic diversity with a range of econometric methods on both the regional level (covering 41 sectors from agriculture to services) and for the manufacturing sector (20 sectors), which shows remarkable differences in spillover dynamics between these two levels of aggregation.
1. Introduction
Does it matter whether an economy is diverse or not diverse at all? Economic diversity is the diversity of economic activities, whether people (or firms) in an economy are pursuing activities different from each other or whether they are doing similar jobs and/or producing similar products or services. Some economists have argued that economic diversity is an important source of knowledge spillovers that lead to enhanced economic growth (e.g. Jacobs, 1969), but other economists have argued the exact opposite and stress the benefits of specialization (e.g. Romer, 1986). This study will assess the impact of economic diversity across 41 sectors on employment volatility, employment growth and productivity growth for 271 European regions for the years 1995 to 2008.
Next to the effects of economic diversity, we also will look at which factors cause some regions to be more diverse than other regions. Understanding why regions differ in terms of economic diversity and what the subsequent effects of economic diversity are is of great importance to both policymakers and (economic) scientists for a number of reasons.
1
become interested in economic diversity following the Great Depression in the 1930s (Dissart, 2003: 423; Rodgers, 1957) and is very relevant today as we are currently experiencing the Great Recession and unemployment rates are high in many countries. High unemployment can have a lasting impact on an economy and individuals when employees who are out of a job for a prolonged period lose skills and have trouble finding new jobs (Baldwin and Brown, 2004: 520). It would be a very powerful insight for policymakers if economic diversity can indeed help making an economy more resilient in the face of crises and shocks.
A second reason why economic diversity is important is that diversity may have several potential spillover effects and externalities on the overall economy. Economic diversity is one of the sources of spillovers identified by the new (endogenous) growth literature that are argued to stimulate innovation (Romer, 1986, 1990; Lucas, 1988; Krugman, 1991). Whereas traditional growth theory (Solow, 1957) sought to explain economic growth through the increase of factor inputs (such as labor and capital) and technological improvements, the new growth theory argues that amongst other factors the underlying pattern of economic activity may be an additional source of growth by stimulating innovation (Jacobs, 1969; Jaffe, 1986; Jaffe, Traijtenberg and Henderson, 1993). The related field of new economic geography attributes regional differences in growth to localized increasing returns arising from local spillovers such as the spillovers generated by economic diversity (Fujita, Krugman and Venables, 1999; Brackman, Garretsen and Van Marrewijk, 2001). There is however disagreement in the literature, whether the benefits of high diversity or of low diversity (specialization) are larger. A more precise understanding of which externalities are dominant in a specific situation can thus shed further light on the dynamics and causes of economic growth.
A third reason for why the study of economic diversity is relevant is that regardless of whether externalities of high or low diversity are dominant in the short-term, increasing economic diversity is crucial for long-term economic growth since productivity improvements and demand saturation will lead to a declining level of employment of current sectors over time and new sectors are thus required to create new jobs (Pasinetti, 1981, 1993).
Fourthly, economic diversity is also of interest as a phenomenon by itself. Why are not all organizations doing exactly the same thing? And if not, why do organizations not pursue more different things than they do at the moment? These questions are central to the field of organizational ecology (Hannan and Freeman, 1977) and are crucial for our understanding of the (economic) world we live in.
Finally, economic diversity could be an important indicator of the state of an economy. In the field of biology, biological diversity (the diversity of species) is a key indicator of the health of an ecosystem (Simpson, 2002: 4). Once our understanding of economic diversity has matured further, economic diversity could potentially be used in a similar way to assess the ‘health’ of regional economies.
While economic diversity is the key interest of the present study, two other concepts are investigated as well, which are Competition and Employment Protection Legislation (EPL).1 As will be more elaborately discussed in subsequent sections, competition and EPL are argued and found to have an impact on the same variables as economic diversity and it is therefore important to study these three variables simultaneously. Whereas economic diversity and competition have been studied together in many previous studies, the simultaneous study of economic diversity and EPL is a new step.
2
The determinants and effects of diversity are investigated for 271 European regions for the period 1995-2008 on both the regional economic level (measured over 41 sectors from agriculture to services, including the manufacturing sector) and zooming in on the regional manufacturing sector (20 sectors). The latter sector has been the focus of many previous studies of economic diversity. It is interesting to study the various (spillover) effects on these two different levels of aggregation and we find that the sign of certain effects can indeed differ between these two levels. We have employed several econometric models for the empirical analysis, such as a cross-section model, a first differences model and random and fixed effects panel models, which all capture different aspects of the dynamics of economic diversity. We now turn to the findings of the empirical analysis.
1.1 Findings of the study
This study thus looks at both the determinants and the effects of economic diversity. With regard to the determinants of diversity, we find significant results for a number of variables that can explain why some regions are more diverse than others. The variables Education, Population, and GDP per capita are all three found to have a positive effect on the level of diversity but have a negative effect on the growth of diversity, which indicates diminishing marginal returns of these variables on the level of economic diversity. Furthermore, whether a region is an island has a negative effect on the level of diversity, and whether a region is a major city agglomeration has a negative effect on the growth of diversity. There is only limited support for a positive effect of Population density on diversity.
We have hypothesized how the three variables of Diversity, Competition and EPL may relate to each other on the basis of the previous studies and new theoretical arguments (see section 2.6). High economic diversity, fierce competition and flexible EPL are argued to reinforce each other (see section 2.6), with on the one hand an extreme situation of high economic diversity, fierce competition, and flexible EPL and on the other hand, an opposite extreme situation with low economic diversity, weak competition and strict EPL. The first situation is argued to be characterized by radical innovations, strong employment growth and relatively weak productivity growth, whereas the second situation is argued to be characterized by incremental innovations, relatively weak employment growth, and strong productivity growth. These two situations correspond well to the concepts of an entrepreneurial and a routinised regime (Bassanini and Ernst, 2002). The results show a positive effect of fiercer Competition and more flexible EPL on Economic diversity, which supports the argument that these three institutions are mutually reinforcing. Altogether, we have identified a range of variables that influence diversity. This enhances our understanding of the factors driving economic diversity and its development path.
The results for the effects of Diversity, Competition and EPL on Employment and Productivity growth are however mixed and not fully in line with our hypotheses. We find that for the regional level, high Diversity, low Competition and strict EPL stimulate Employment growth and discourage Productivity growth. For the manufacturing sector, on the contrary, we find that low Diversity, high Competition and strict EPL stimulate Employment growth and discourage Productivity growth (though the effect of Diversity on the latter is inconclusive). It is interesting that the results thus differ between the regional level and the manufacturing sector, which is in line with the argument that different externalities dominate in different situations and that there is not one ‘true effect’ of diversity. This makes the study of economic diversity all the more interesting, but also makes it harder to generalize findings and to capture the effect of diversity with rules of thumb.
3
For Employment volatility, we did not find a significant effect of economic diversity, whereas this effect was hypothesized to be negative. This may be explained by the fact that both GDP per capita and Population have a significant positive effect on Diversity as well as a significant negative effect on Employment volatility, which may make it difficult to pick up the separate effect of Diversity on Employment volatility. The effect of stricter EPL on Employment volatility is negative for both the regional level and the manufacturing sector, which is in line with our expectation. The effect of fiercer Competition is also in line with our arguments and is found to increase the volatility of employment.
Finally, the study also makes a number of contributions to the measurement of economic diversity. This topic is discussed elaborately in Appendix B, which also includes an overview of a large number of different diversity indices used in economics and biology, as well as a proposed algorithm that can be applied for the selection of an appropriate diversity index for economic studies.
By studying economic diversity in a comprehensive way, looking at multiple determinants and effects of diversity, the relation of diversity with competition and EPL, for two different levels of aggregation (the regional economy and the manufacturing sector), and for a large and fine-grained dataset, this study brings our current understanding of economic diversity together and expands it in various directions, offering a good platform for further research into this fascinating topic.
1.2 Overview remainder thesis
In the remainder of this introductory section, a preview of the data is provided to give some indication of how the various regions differ from one another. Additionally, the conceptual model is presented which shows how the variables in the study relate to each other. This conceptual model forms the basis for the discussion of the literature in section 2, where the various determinants and effects of diversity are discussed (as well as the effects of competition and EPL).
The review of the literature, in turn, forms the basis for the hypotheses for the empirical part of the study. The methodology of the empirical study is discussed in section 3 as well as the selected data. The results of the empirical study are subsequently reported in section 4, after which section 5 draws conclusions, points to caveats of the study and offers recommendations for future studies.
There are also several appendices to this study: Appendix A provides an overview of the regions and the sectors included in the study; Appendix B discusses the measurement of economic diversity; Appendix C provides descriptive statistics for the variables used in the empirical study; and finally Appendix D presents detailed results which have been excluded from the main text for reasons of brevity.
1.3 Preview of the data
The current study is not the first study to investigate the causes of differences in employment and productivity growth between European regions. One particular study that has also looked into this topic is Cuadrado-Roura, Mancha-Navarro and Garrido-Yserte (2000), which examines data for 97 European regions for the period 1980-1993. Figure 1 is inspired by their study and plots all the regions of the current study and their employment and productivity growth rates. The symbol of each region corresponds to whether the region has above or below average economic diversity, competition and EPL. Figure 1 therefore gives a good initial overview of various characteristics of the regions included in the current study.
4
Figure 1. Employment and productivity growth 1998-2007 of European regions
At closer examination, the patterns of the various categories of regions (as determined by whether they have high or low diversity, competition or EPL) are interesting and provide first clues for our research. For example, the ‘black stars’ (low diversity; low competition; low EPL) are situated relatively more to the top-left and thus have relatively high productivity growth rates and low employment growth rates compared to the other regions. In the other corner, the ‘grey diamonds’ (low diversity; high competition; high EPL) are over-represented, which thus have above average employment growth and below average productivity growth.
Of course, we cannot draw any firm conclusions based on a simple cross plot and we will proceed to investigate the effects of diversity, competition and EPL in a more precise and sophisticated way to see whether we can explain some of the remarkable differences between regions. The next section introduces the conceptual model that will guide us in our next steps.
1.4 Conceptual model
The main variables that feature in the study as well as their relations are presented in the conceptual model in Figure 2. The three blocks in the middle of the figure represent the running thread of the study as we are mainly interested in the effect of the Determinants of Diversity on Diversity and in the effect of Diversity on the Effects of Diversity (relations A and B).
The potential determinants of diversity considered by the study are regional Population level; Population density; GDP per capita; Education level; and two geographical dummy variables, one indicating whether a region is an island, and one indicating whether a region is a major city.
-30% -20% -10% 0% 10% 20% 30% 40% 50% 60% -40% -30% -20% -10% 0% 10% 20% 30% 40% 50% Pro d u c ti vt y g ro w th 1 9 9 8 -2 0 0 7 Employment growth 1998-2007
High Diversity; high Competition; high EPL High Diversity; high Competition; low EPL High Diversity; low Competition; high EPL High Diversity; low Competition; low EPL Low Diversity; high Competition; high EPL Low Diversity; high Competition; low EPL Low Diversity; low Competition; high EPL Low Diversity; low Competition; low EPL
Vest (RO) Nord-Vest (RO) Sud-Vest Oltenia (RO) Bucuresti-Ilfov (RO) Zachodniopomorskie (PO) Latvia (LV) Flevoland (NL) Inner London (UK)
Murcia (ES) La Rioja (ES) Castilla la Mancha (ES) Reunion (FR) Stockholm (SE) Západné Slovensko (SK) Picardie (FR) Slovenia (SI)
Mellersta Norrland (SE)
5
The effect of diversity is tested for Employment volatility; Employment growth and Productivity growth. The selection of these variables is motivated in Section 3.
Figure 2. Conceptual model
It would however not be prudent to only test for the relations between determinants and diversity and between diversity and its hypothesized effects, since the various blocks in the figure may be related in a more complex way. The Determinants of Diversity could have a direct effect on the Effects of Diversity (relation C), or Diversity may have an interaction or mediation effect rather than an independent effect (relations D and A & B). The current level of Diversity may also have an impact on the growth of Diversity (relation E). There could furthermore be feedback effects in the other direction (the dotted lines in the figure), but unfortunately the period for which data is available is too short to test for the direction of causality, which is why we will not investigate these relations.
Competition and Employment Protection Legislation (EPL) are incorporated in the study to test whether they have an effect on the Effects of Diversity (relation G), Diversity itself (relation F), or alternatively have an interaction (relation H) or mediation effect (relation F & B). Again, the data series is too short to test for the direction of causality; hence any significant results have to be interpreted with care and should perhaps be considered as evidence of correlation rather than of one-directional causation.
There are thus quite a large number of variables as well as relations between them, which means we will have to run a number of tests to examine all effects. This will provide us with a deeper understanding of the determinants and effects of diversity and of how diversity interacts with competition and with EPL. The next section will discuss what earlier studies have discovered regarding the various variables and relations in the conceptual framework of our study.
Determinants
of Diversity
Diversity
Effects of
Diversity
A E B D C Relations:A - The direct effect of the Determinants of diversity on Diversity B - The direct effect of Diversity on the Effects of diversity
C - The direct effect of the Determinants of diversity on the Effects of diversity
A&B - The mediation effect of Diversity on the relation between the Determinants and the Effects of diversity D - The indirect (interaction) effect of Diversity on the effect of the Determinants on the Effects of diversity E - The endogenous effect of Diversity on Diversity
F - The effects of Competition and Employment Protection Legislation (EPL) on Diversity
G - The effects of Competition and Employment Protection Legislation (EPL) on the Effects of diversity
F&B - The mediation effect of Diversity on the relation between Competition and EPL on the Effects of diversity H - The interaction effect of Competition and EPL on the effect of Diversity on the Effects of Diversity
Relations that cannot be investigated with the available data
Competition & Employment Protection Legislation (EPL)
G H6
2. Literature review
Along the lines of the conceptual framework presented in the previous section, this section will discuss the empirical findings and theoretical arguments regarding economic diversity (section 2.1), the determinants of economic diversity (section 2.2), the effects of economic diversity (section 2.3), competition (section 2.4) and Employment Protection Legislation (EPL) (section 2.5). Subsequently, the final section (2.6) will provide an overview of the various effects discussed in the previous five sections that will ease the comparison of these effects (e.g. the effects of diversity, competition and EPL on employment growth) and provide a basis for the empirical study in Section 3. We will not discuss the effects of the determinants of diversity variables on the effects of diversity variables for reasons of brevity, and since these effects have all been extensively covered in the economic literature.
2.1 Diversity
Diversity is a characteristic of a particular population. A population can be either diverse, or it can be homogeneous. In biology, diversity is determined by two factors: The number of different species in a population, and by how even the population is distributed over the various species. The more species in a population and the more even the distribution, the more diverse the population. The opposite of diversity is homogeneity, concentration or specialization. Within the context of the present research, the different species in the population are the different sectors in the economy (see Appendix A2 for an overview of the 41 sectors distinguished in this study). The distribution of the population is measured by the distribution of employment over the various sectors.
As discussed above, interest in economic diversity soared with the advent of the new (endogenous) growth literature (Romer, 1986, 1990; Lucas, 1988; Krugman, 1991), which argues knowledge spillovers are an important source of innovation. One of the sources of knowledge spillovers that were identified is economic diversity (Jacobs, 1969; Jaffe, 1986; Jaffe, Traijtenberg and Henderson, 1993). However, both economic diversity as well as concentration are considered to be potential sources of spillovers. With regard to this debate, Glaeser (2000: 92) notes that the roles of diversity and concentration have not yet been resolved, and that, since different time periods and samples give different results, there may not be a universal truth. Rather than a simple rule that diversity is always good or bad, it is more appropriate to focus on when diversity has a positive effect on a specific variable and a negative effect on other variables. Two reasons why different researchers have found different results is that the effect of diversity on a particular variable could be non-linear, or contingent on another variable. We will discuss these more complex dynamics where relevant and incorporate them in the empirical part of the study.
2.1.1 General externalities of diversity and competition
Before we go into the effect of diversity on specific variables, we will first discuss the generic types of spillovers of diversity which are usually discussed in the literature. These generic types of diversity differ with regard to whether diversity is argued to have a positive or a negative effect, and with regard to whether competition is argued to have a positive or a negative effect. Figure 3 provides an overview of these four generic types of spillovers. Following the seminal paper of Glaeser et al. (1992), three types of externalities are usually distinguished, which are MAR, Jacobs and Porter externalities. To complete the framework, it is proposed to add BEMA externalities as well.
7
Van Oort and Verburg, 2005). Competition is argued to discourage R&D investments, and to be hostile to growth. MAR externalities thus arise when both diversity and competition are low.
Figure 3. Spillover combinations of diversity and competition
Effect of diversity Positive Negative E ff e c t o f c o m p et it io n P o si ti ve Jacobs Porter N e g at iv e BEMA MAR
Knowledge spillovers within the same industry are also considered most important by Porter (1990). Hence, concentration is argued to facilitate growth. The difference with MAR externalities hinges on the role of competition, which is considered to foster growth (Porter, 1990), as it provides the incentives for continuous innovation. So-called Porter externalities thus arise when diversity is low, but competition is high.
Disagreeing with MAR and Porter externalities, Jacobs externalities regard inter-industry spillovers as the main source of new knowledge creation. Economic diversity is argued to facilitate the exchange of complementary knowledge from different industries, and to lead to new combinations of ideas, which results in innovation and growth (Greunz, 2003: 12). Like Porter (1990), competition is considered to be beneficial to growth. Jacobs externalities are thus the opposite of MAR externalities and arise when both diversity and competition are high.
8 2.1.2 Local nature of spillovers
As was implicit in the discussion of the general types of externalities of diversity above, and pointed out more directly in the discussion of new economic geography (Fujita, Krugman and Venables, 1999; Brackman, Garretsen and Van Marrewijk, 2001), the spillovers of economic diversity have a local character and their strength diminishes with distance. The knowledge spillovers are tacit and non-codified and flow through channels of interpersonal communication (Beaudry and Breschi, 2003: 326; Jaffe, 1986; Malizia and Ke, 1993: 222; Audretsch, 1998), and are therefore regional rather than (inter)national. The local character of these spillovers is also supported by empirical evidence (Jaffe, 1989; Jaffe, Traijtenberg and Henderson, 1993; Audretsch and Feldman, 1996).
2.1.3 Different levels of sector aggregation
Similar to how diversity can operate at different levels of geographic aggregation (e.g. city, region, country) diversity also operates at different levels of sector aggregation within the economy. One classification of the various levels is as follows (Hoover, 1948; Isard, 1956; Forni and Paba, 2002: 152):
(1) Spillovers which are internal to the firm: Economies of scale and scope that arise from the size and diversification of an individual firm;
(2) Spillovers that operate within the same sector; (3) Spillovers that accrue to firms in related sectors;
(4) Spillovers which are available to all firms irrespective of their sector (i.e. unrelated sectors).
As is clear from this classification, there is some judgment involved in deciding whether firms are part of the same sector, part of related sectors, or part of unrelated sectors. When measured at a low level of aggregation three firms that are all part of the manufacturing sector may belong to very different sub-sectors (e.g. food processing, chemicals, and mechanics). Empirical studies distinguishing related and unrelated diversity have found results that differ between these two types of diversity (Frenken, Van Oort and Verburg, 2005; Forni and Paba, 2002; Bishop and Gripaios, 2007). Therefore, this distinction does have relevance, even though it will always be somewhat arbitrary what exactly is and what is not related. This complicates the comparison of the results of empirical studies, since it thus matters whether diversity was measured across related or unrelated sectors.
Another way of classifying sector aggregation is from the whole economy, to sector (e.g. manufacturing sector), to sub-sectors (e.g. the chemical manufacturing sector). At these different levels of aggregation, diversity may also operate in a different way. One could imagine that it is beneficial for a regional chemical manufacturing industry to be very homogeneous (promoting employment growth), but that the entire manufacturing sector as a whole would benefit from a diverse chemical industry.
9
aggregation (or a so-called cluster) as long as this specialization does not dominate the entire economy and is balanced.
Within the context of this study, we are interested in diversity and specialization as each other’s opposites on the same level of aggregation, and will therefore abstain from including the specialization in sub-sectors as well.
This discussion underlines that one has to be careful when comparing different empirical studies of diversity and specialization. Given that diversity can be measured at different levels of geographic and economic aggregation, it is no wonder different studies have reached different conclusions with regard to whether ‘the effect of diversity’ is positive or negative. Before we will look in more detail into what various studies have found, we will first discuss a number of possible determinants of diversity that have been proposed in the literature.
2.2 Determinants of Diversity
If economic diversity can have both positive and negative spillovers according to the four general types of externalities discussed in the previous section, economic diversity becomes an important factor to take into consideration when studying regional economic growth. Consequently, it becomes important to understand why regions differ in how diverse their economies are.
A general theory that explains diversity in an economy is the resource-partitioning theory within the field of organizational ecology (Carroll, 1985), which argues that the economy we observe is the product of the boundaries imposed, and the opportunities offered by the environment. The richer and the more diverse the environment – in terms of e.g. preferences, natural resources, cultures, technology – the more diverse the economy can be. The resource-partitioning theory predicts more competition will lead to more diversification into the niches of the resource space. This is also discussed below in section 2.4.5, which deals with the relation between competition and diversity.
The various variables which we will discuss next fit in the resource-partitioning framework since most make the environment richer, which relaxes some of the boundaries imposed on the economy and would be expected to stimulate economic diversity. Loosely based on the overview of various determinants of diversity by Bishop and Gripaios (2007), we will now discuss the following variables that are argued to have an influence on economic diversity: Population level; GDP per capita; population density; geography; knowledge-based activities; trade; and industrial structure.
2.2.1 Population level
In line with the resource-partitioning theory, a larger economy would be expected to be able to nurture a more diverse economy. This is also what Dissart (2003: 434) and Bishop and Gripaios (2007: 1740-41) conclude on the basis of their reviews of the literature. Early studies mainly reported empirical findings of this effect (e.g. Clemente and Sturgis, 1971; Crowley, 1973). Later, within the field of new economic geography, equilibrium models have been developed that explain cities as the outcome of forces of agglomeration and dispersion, with greater diversification in larger cities, as the product variety increases with the width of a market (population) and the depth of the market (the purchasing power per person) (Andel-Rahman and Fujita, 1993; Fujita and Krugman, 2000). Another theory arguing that larger cities will be more diverse is the central place theory (Christaller, 1966), which argues that cities that are higher in the urban hierarchy will provide a wider range of functions and therefore will be more diverse.
10 2.2.2 GDP per capita
Above, it was noted that, by widening the width of the market, population size has a positive effect on economic diversity, and that, in a similar way, increasing the purchasing power per person can deepen the depth of the market and thereby also enhance economic diversity. The Gross Domestic Product (GDP) per capita is therefore another determinant of economic diversity. This result appears to be less general then the effect of the population level. For example, Nissan and Caveny (1998: 168) find that economic diversity is not significantly greater for higher levels of development, or income. Furthermore, Imbs and Wacziarg (2003) propose a different relation between income and diversity, arguing that diversity first increases with economic development, but that countries tend to specialize again at a higher level of income.
2.2.3 Population density
Closely related to the population level is population density, which is argued to be a determinant of diversity as well. A high population density is thought to be able to support many niche functions that would not be economic in a sparsely populated region. Next to the population level of a region, the density of a region is thus a separate determinant of economic diversity. Frenken, Van Oort and Verburg (2005: 2, 16) note that there is wide agreement in the literature that diversity is positively related to the degree of urbanization, and that population density is a proxy for urbanization. Greunz (2003: 26) also reports a larger effect of diversity on innovation (measured by patent applications) for 153 European regions in ‘high density regions’ compared to the overall sample, which is an interaction effect between population density and diversity. Since the spillovers of economic diversity are localized, it makes sense that they are strongest when the population density is higher (and more people are closer to the source of the spillovers).
2.2.4 Geography
Regions compared in a given study do of course differ in terms of geography. Differences in geography may have an impact on economic diversity as they have an impact on the boundaries and opportunities offered by the environment (see the resource-partitioning theory above). While it is often difficult to introduce variables that capture geographic differences, two examples of studies that have investigated the influence of geographic variables are Nissan and Carter (1990: 206) who found that a variable indicating to which part of Africa a country belonged had an impact on the trend towards either diversification or specialization for 24 African nations between 1980 and 1990, and Bishop and Gripaios (2007: 1749), who found that the City of London was a significant outlier in their study compared to other regions (which was also concluded by Dewhurst and McCann, 2002).
2.2.5 Knowledge-based activities
Considering the fact that the spillovers of economic diversity are argued to have an influence on innovation, Bishop and Gripaios (2007: 1743) argue that knowledge-based activities could be an important determinant of economic diversity since these are an important input to the process of innovation and the generation of new opportunities (Baldwin et al., 2000). While there is substantial empirical evidence on the firm level (an overview is provided in Devine et al., 1979), Bishop and Gripaios (2007: 1753) only find evidence on the regional level for a significant relation between knowledge-based activities and (unrelated) diversity in one of their models.
11
education, research and development. These types of activities also expand the boundaries and increase the opportunities of the environment both in terms of the capabilities of customers and employees as well as the resource-space in terms of discovered and mastered technologies.
2.2.6 Trade
Another factor that is argued to have an effect on regional economic diversity is trade (Nissan and Carter, 2006: 186). Following the well-known theory of comparative advantage of Ricardo, trade will promote specialization of each region in the production of goods in which it has a comparative advantage. This may lead to a decrease in regional diversity. However, this argument only holds for the sectors that produce traded goods (e.g. manufacturing), and not for sectors which produce non-traded goods (many service sectors). More trade may lead to growth in service sectors that facilitate trade such as transport and communication. Furthermore, more trade also enlarges the market to which sectors in a given region can sell their goods, and this may relax the boundaries on the economy which may also lead to an increase in diversity. A priori, the net effect of trade on economic diversity is therefore ambiguous.
2.2.7 Industrial structure
Finally, the size of certain sectors may have a positive effect on economic diversity. Care should be taken when making these statements since the measurement of diversity (how employment is distributed over the various sectors) is often related to measuring the size of specific sectors (the relative employment of a certain sector). Bishop and Gripaios (2007: 1740) report a significant positive effect on economic diversity of the shares of the manufacturing sector, the private sector and the financial and business services sector, and a negative effect of a large public sector. Forni and Paba (2002: 170) report a special role for the metal products and machinery sector which was found to have many spillover links with other manufacturing sectors.
Thus, it is arguably the case that certain sectors have a larger spillover effect on the rest of the economy than other sectors. Well-developed financial services and an advanced machinery sector could indeed support the successful development of other sectors and thereby have a positive effect on economic diversity. This is another example of the discussion in section 2.1.3 of how relative specialization in a sector is compatible with higher economic diversity on a higher level of aggregation.
2.3 Effects of Diversity
Now we have presented the various determinants of diversity discussed in the literature, it is time to turn to the effects of diversity. We will discuss the effect of diversity on employment volatility, on employment growth, and on productivity growth. Subsequently we will discuss the impact of economic diversity on the type of innovation, as it features prominently in the spillover mechanisms advocated by the various externality theories. Finally, we will discuss to what extent diversity may be endogenous, i.e. whether the level of economic diversity has an effect on the growth rate of economic diversity.
2.3.1 Effect of Diversity on Employment volatility
As noted earlier, the Great Depression sparked interest in the study of economic diversity in order to identify which industrial composition experienced lower economic fluctuations. Similar to the investment portfolio theory, the general presumption was that a diversified economy would be more stable as different sectoral shocks would off-set each other (Rodgers, 1957).
12
who find that diverse metropolitan areas are more stable, and Nourse (1968), Richardson (1969), and Waikar et al. (1999) who find the same result for regions. The negative effect of economic diversity on employment volatility thus has broad theoretical and empirical support.
2.3.2 Effect of Diversity on Employment
Whereas the support for a negative effect of economic diversity on employment volatility is broad, the evidence for a positive effect of diversity on employment growth is mixed although predominately supportive. The latter is in line with Jacobs and BEMA externalities discussed earlier. In a review of the literature, Dissart (2003: 434) notes that more diversity is associated with lower growth of unemployment, but that the evidence is less conclusive than with regard to employment volatility. In another review of the literature, Bun and El Makhloufi (2007: 827) find six studies that report a significant positive effect of diversity on employment, and only one study that finds a mixed, or insignificant effect (Combes, 2000). Other studies which report a positive effect of diversity on employment are Glaeser et al. (1992), Simon and Nardinelli (1992: 384) and Simon (1988) for U.S. cities; Izraeli and Murphy (2003: 1) and Malizia and Ke (1993: 221) for U.S. states; Frenken, Van Oort and Verburg (2005: 21) for Dutch regions; and Peri (1992) and Forni and Paba (2002: 172) for Italian regions.
Furthermore, the compatibility of employment stability and employment growth is sometimes questioned. Policy makers might have to choose between employment growth and stability if stability is the result of economic diversity, as argued above, and if growth is derived from economic specialization based on comparative advantage (Wagner and Deller, 1998: 542). This conclusion may however not necessarily be correct. It was previously mentioned how diversity on the level of the whole economy does not rule out specialization in specific clusters. And also that diversity can be a source of (employment) growth by itself. Employment growth and stability are thus not necessarily incompatible. Additionally, as was noted above, Pasinetti (1981, 1983) argues that diversity is actually necessary for employment growth in the long run, as productivity improvement and demand saturation will lead to lower employment levels in the present sectors.
On the whole, the theoretical and empirical support thus points more towards a positive effect of economic diversity on employment growth.
2.3.3 Effect of Diversity on Productivity
Compared to the effect of economic diversity on employment growth, there is even less agreement in the literature with regard to the effect of economic diversity on productivity growth. On the one hand, Frenken, Van Oort and Verburg (2005: 7) argue that economic concentration will favor productivity most, as spillovers between firms in the same industry presumably lead to more incremental and process innovations, which would raise productivity. MAR and Porter externalities are thus argued to be dominant when it comes to the effect of economic diversity on productivity. On the other hand, Saviotti (1996) argues that economic diversity is perfectly complementary with productivity growth, similar to how according to Jacobs externalities the combination of different ideas leads to new innovations and productivity gains.
Unfortunately, empirical studies are not abundant. De Lucio, Herce and Goicolea (2002) for Spain, Cingano and Schivardi (2002) for Italy, and Almeida (2005: 5) for Portugal, do not find conclusive evidence for an effect of economic diversity on productivity, while Frenken, Van Oort and Verburg (2005: 20) find that economic diversity is negatively related to productivity growth for Dutch regions.
13 2.3.4 Effect of Diversity on the Type of innovation
In the various theories regarding spillovers caused by economic diversity, innovation is the key mechanism, whether caused by low diversity such as according to MAR and Porter externalities or caused by high diversity such as according to Jacobs and BEMA externalities. Basically the argument is about whether firms will be more innovative when more firms are active in the same sector, or whether they will be more innovative when more firms in their region are active in different sectors. Some studies have therefore argued that one should focus on measuring innovation directly rather than measuring results of innovation (i.e. measuring employment and productivity growth) (Paci and Usai, 1999; Massard and Riou, 2002). Empirical results have been mixed with Paci and Usai (1999) for Italy, and Greunz (2003: 11) for European regions, reporting a positive effect of economic diversity on innovation, while Feldman and Audretsch (1999) for the U.S.A., and Massard and Riou (2002) for France, find no significant effect.
Other authors have argued however that it is not only the level of innovation which matters, but also the type of innovation (e.g. Frenken, Van Oort and Verburg, 2005: 7). They distinguish between radical and incremental innovations, where a radical innovation gives rise to a qualitatively different product or service with a completely different set of characteristics with respect to any previous technology. An incremental innovation would consist only of quantitative variations of existing characteristics (Nguyen et al., 2004: 2). A radical innovation is thus doing something completely new, while an incremental innovation is doing something old a little bit better. Frenken, Van Oort and Verburg (2005: 7) argue that MAR externalities (concentration of economic activity) will mainly lead to incremental innovations in existing sectors, which will increase productivity. Jacobs externalities (high diversity) are argued to promote radical innovations, which lead to new products and services and thereby to higher employment. This argument is entirely theoretical, and difficult to test with empirical data. The above results are in line with this argument though, since economic diversity in general has a positive effect on employment growth and a negative effect on productivity growth. Interestingly, we will see further on that EPL is also argued to have an effect on radical and incremental innovations.
2.3.5 Endogeneity of Diversity
Before turning to competition, we will briefly discuss whether economic diversity should also be expected to have an effect on itself, whether it is endogenous. If high economic diversity has a positive effect on radical innovation and the creation of new products (and sectors) as argued above, economic diversity can be argued to be self-reinforcing. In a very diverse economy there would be more radical innovations, leading to more new products and services, whereas in a low diversity economy there are fewer radical innovations and more incremental innovations. The latter improves the productivity of existing sectors, which may therefore become more dominant. Diversity growth would be expected to be higher in the first situation and thus a high diversity level would be argued to stimulate the growth of diversity.
14
externalities) and mechanisms that reward being different (competition effect, Jacobs and BEMA externalities).
A different development path of diversity is put forward by Boone, Wezel and Van Witteloostuijn (2007), who have formulated a diversity-dependence theory inspired by the density-dependence theory that argues that the benefits of increasing diversity are dominant when diversity is low, and that the benefits of decreasing diversity are stronger when diversity is high. Through the exit and entry of relatively (dis)similar firms, this leads to a cyclical pattern of population diversity.
The effect of diversity on diversity growth is thus not necessarily clear-cut and potentially non-linear or cyclical. In the empirical part of this study we will test what the effect of diversity on its own growth rate is for the economic diversity of European regions. In our discussion of the endogeneity of diversity, competition was already linked with economic diversity. We will come back to this point in the next section, in which we will discuss the effects of competition and its relation with diversity.
2.4 Competition
Theoretically, the effects of economic diversity and competition have often been linked together. In section 2.1.1, competition and economic diversity were both discussed as sources of spillovers of the four general types of externalities, and competition has been argued to have an important effect on the development of diversity in section 2.3.5. It is therefore important to incorporate competition in our study of economic diversity, as has been the practice of most studies of the topic, especially since the seminal work of Glaeser et al. (1992). We will discuss the effect of competition on the same variables as we discussed for diversity in the previous section, and will close with a discussion of the relation between competition and economic diversity, before turning to the effects of EPL.
2.4.1 Effect of Competition on Employment volatility
The effect of competition on employment volatility has been studied less extensively than the effects of economic diversity or EPL on employment volatility. To the knowledge of the author, there is not a study that has investigated this effect directly.
Intuitively, the effect of competition on employment volatility is expected be positive since in a competitive market firms will be inclined to hire the number of employees with which they can equate marginal costs to marginal revenues. In good times, firms will hire relatively more workers, and in bad times they will fire relatively more workers than comparable firms in less competitive markets. This increases the volatility of employment during the business cycle. The effect of competition on employment volatility is thus expected to be positive.
2.4.2 Effect of Competition on Employment
The effect of competition on employment growth has received more attention in the literature. Theoretically, the sign effect of competition on employment growth could be expected to be either positive or negative, depending on which one of the four general types of externalities dominates. Empirical studies report more evidence of a positive effect of fiercer competition on employment growth than of a negative effect though, which suggests that Jacobs and Porter externalities are more likely to dominate than MAR and BEMA externalities.
15
studies that report a positive effect of competition on employment growth are Peri (1992) and Forni and Paba (2002: 161) for the case of Italian regions. The effect of competition on employment growth could thus be both positive and negative, although a positive effect appears to be more common.
2.4.3 Effect of Competition on Productivity
The effect of competition on productivity has again received less attention in the literature, since most studies take employment growth or GDP growth as dependent variables. As with the effect of competition on employment, theoretically the effect could be both positive and negative, depending on which of the four general types of externalities dominates. One study that has examined the effect of competition on productivity is De Lucio, Herce and Goicolea (2002) for Spanish regions, which reports evidence of a negative effect of competition on productivity. The empirical support is thus meager compared to the studies which have investigated the effect on employment growth, but points towards a negative effect.
The relation between competition and innovation has received more attention in the literature. It is often argued that the effect of competition on innovation is non-linear (e.g. Aghion et al., 2002). When competition is very low, firms do not have enough incentives to innovate. Yet when innovation is very high, firms are not able to generate the funds required to invest in innovation. Considering the link between (radical) innovation and productivity gains, the relation between competition and productivity may very well also be non-linear.
2.4.4 Effect of Competition on the Type of innovation
There is thus a similarity between the empirical findings of the effects of competition and the effects of economic diversity, i.e. high economic diversity and high competition tend to stimulate employment growth whereas low economic diversity and low competition tend to stimulate productivity growth (although the empirical support for the effects on productivity is not extensive).
This makes it tempting to extend the argument of Frenken, Van Oort and Verburg (2005: 7) to the role of competition. In their article, the authors mainly focus on the difference regarding the effect of economic diversity between Jacobs and MAR externalities. Since the effects of competition are also in line with their argument, the competition component of Jacobs and MAR externalities could very well play an important role in this mechanism as well. Higher competition would then stimulate radical innovation (leading to higher employment growth) and low competition would stimulate incremental innovation (leading to higher productivity growth). It is certainly intriguing that Jacobs and MAR externalities have such opposing effects on employment growth and productivity growth for both their economic diversity and competition dimensions.
The effects of diversity and competition on (type of) innovation are however difficult to measure directly. Therefore this remains outside of the scope of this study, in which we focus on the indirect effects on employment and productivity growth.
2.4.5 Relation between Competition and Diversity
16
Within the resource-partitioning theory discussed above, there is also a link between competition and diversity. When the number of generalists in a sector starts to increase and competition intensifies, new entrants will increasingly opt for niches at the boundaries of the sector (Carroll, 1985). Hence competition leads to an increase in diversity in future periods.
On the firm level, the empirical evidence of an effect of competition on diversity is mixed (Bishop and Gripaios, 2007: 1742), with both evidence of a positive effect of competition on diversification (Caves et al., 1980; Baldwin et al., 2000), and of a negative effect (Christensen and Montgomery, 1981). On a regional level, Bishop and Gripaios (2007: 1753) find evidence that suggests a positive link between the intensity of competition and the economic diversity of British regions. For this study, which also investigates dynamics at the regional level, we therefore expect a positive effect of competition on economic diversity. Next, we will turn to the discussion of the effects of Employment Protection Legislation.
2.5 Employment Protection Legislation
Next to competition, we will also include Employment Protection Legislation (EPL) in our study of economic diversity. As explained above, EPL refers to the entire set of regulations that place limits on the faculties of firms to hire and fire workers (Barone, 2001). These regulations may be grounded in laws, but can also originate from collective bargaining, or decisions of the judiciary, through evolving case law. Examples are unemployment insurance, a severance pay system and mandatory terms of notice. Broader definitions of EPL also encompass restrictions placed on utilizing labor, such as limitations on the use of fixed-term and temporary work agency contracts and the regulation of work hours, or even of labor standards, such as parental/maternity leave, health and safety regulations, mandatory sick pay, and minimum wages (Addison and Teixeira, 2003: 85).
The standard argument in favor of EPL is that these laws protect workers from unjust termination by employers, which imposes significant mobility costs on workers (Lazear, 1990: 699). EPL may, however, also have other effects that are of interest to workers and the economy. As we will see in the subsequent sections, EPL is argued and reported to have an effect on the same variables that we have discussed above in relation to economic diversity and competition. We will discuss these effects in the same order as above. This warrants the inclusion of EPL in our research, and raises interesting questions into how (the effects of) EPL, economic diversity and competition interact with each other.
2.5.1 Effect of EPL on Employment volatility
Since the aim of EPL is to protect workers from termination and limits the hiring and firing faculties of firms, one would expect that EPL has a negative effect on the volatility of employment. In periods of declining demand, the imposition of adjustment costs (because of EPL) will lead to fewer layoffs, as it creates a wedge between the firm’s marginal revenue product and the cost of changing employment. In times of rising demand, the firm will take the likelihood into account that severance pay has to be paid in the future, which will reduce hiring. Together, lower firing in bad times, and lower hiring in good times, result in less fluctuation of the employment level (Addison and Teixeira, 2003: 87).
17
F200) find a significant negative effect of the introduction of the wrongful-discharge protection exception on employment fluctuations for U.S. states for 1970-1999. This effect is more pronounced in the manufacturing sector, arguably because of its highly seasonal and highly cyclical nature (Autor, Kerr and Kugler, 2007: F202).
Based on the theoretical arguments and empirical findings, we thus expect a negative effect of EPL on employment volatility, which is more pronounced for the manufacturing sector than for the whole economy.
2.5.2 Effect of EPL on Employment
Whereas the effect of EPL on employment volatility is widely argued to be negative, the effect of EPL on the employment level is the subject of debate among economists, with opinions ranging from the absence of an effect, to both positive and negative effects.
Theoretically, the effect on firm employment is ambiguous, as has been argued by Bertola (1992), who argues that the sign of the effect hinges on the functional form of labor demand functions, the discount rate and labor turnover. Bentolila and Bertola (1990: 381) present a model in which, with realistic parameters, firing costs do not have large effects on hiring, nor do they reduce the average level of firm employment, which may even increase. Empirical support for this claim is provided by Anderson (2003: 1037), who studies the impact of the experience rating feature of the U.S. Unemployment Insurance, and finds a small increase of average firm employment over the business cycle.
The effect on firm employment is, however, not necessarily equal to the effect on the aggregate employment level, which is also affected by the entry and exit of firms. Haltiwanger, Scarpetta and Schweiger (2006) estimate that 30-40% of job flows in OECD countries are due to firm entry and exit. The effect of EPL on firm entry and exit can be ascertained along the lines of Dixit (1989), who argues that, since most investment decisions are made in an uncertain environment and are costly to reverse later, these decisions should be evaluated in a similar way as a call option (Dixit, 1989: 620). Subsequently, Dixit (1989: 630) shows that the larger the sunk costs associated with an investment decision, the lower both firm entry and exit will be, and also that a small increase in sunk costs can already have a relatively large effect. Since more stringent EPL is a sunk costs associated with the entry and exit of a firm, more stringent EPL should lead to lower entry and exit levels of firms. Koeniger and Prat (2007: F304) show that adjustments induced by firing costs at this margin are unambiguously detrimental to employment by reducing firm entry. For plausible parameters, this effect dominates the ambiguous effect at the hiring-and-firing margin discussed above and more stringent EPL is thus expected to reduce aggregate employment. Of course, an increase in self-employment would again counter this effect to some extent.
A negative effect of more stringent EPL on aggregate employment is also the conclusion which Addison and Teixeira (2003: 98) draw on the basis of a review of 15 studies, starting from the pioneer paper of Lazear (1990). Most studies report a negative effect on employment-population ratios, with only three studies finding no significant evidence, and not a single study reporting a positive coefficient estimate. A striking example of the negative effect of EPL on aggregate employment is the effect of the Americans with Disability Act, which protects the employment of disabled workers, but has, ironically, led to a sharp drop in the employment of disabled workers (Acemoglu and Angrist, 2001: 915). In one of the largest studies of the effect of EPL (covering 85 countries), Botero et al. (2004: 1339) find both a lower labor force participation rate and a higher unemployment rate as a result of stricter EPL.
2.5.3 Effect of EPL on Productivity
18
sanctioning shirking workers) and a positive effect (by increasing the loyalty and dedication of the employee, by encouraging the firm to invest in the training of employees, and by increasing the willingness of employees to accept technological change and internal job mobility). In addition, stricter EPL is arguably causes firms to substitute capital for labor, which raises the capital per labor ratio (Autor, Kerr and Kugler, 2007: F189). Furthermore, stricter EPL may favor human capital accumulation through longer job tenure and through a positive effect on the returns to education (Saint-Paul, 2002: 387). These additional effects are all beneficial to labor productivity.
Next to the effect on labor productivity, EPL is also argued to have an effect on total factor productivity. Assuming employees value EPL less than EPL costs employers, stricter EPL inhibits efficient job separations and relocations. This distorts production choices and will reduce allocative efficiency, which has a negative influence on total factor productivity (Autor, Kerr and Kugler, 2007: F195; Hopenhayn and Rogerson, 1993; Bertola, 1994). But, this negative effect could be off-set by a positive effect of EPL on labor productivity. Hence, as the theoretical net effect on labor productivity is ambiguous, the theoretical net effect on total factor productivity is ambiguous as well.
In an attempt to measure the impact of EPL on both types of productivity, Autor, Kerr and Kugler (2007) have studied the impact of the adoption of the wrongful-discharge protection by U.S. courts on U.S. firms for 1970-1999. They find evidence that in states which adopted the protection, firms engaged in capital deepening, which on the one hand led to a concurrent rise in labor productivity. Total factor productivity, on the other hand, declined, which indicates that the rise in labor productivity was not sufficient to off-set the decrease in allocative efficiency (Autor, Kerr and Kugler, 2007: F190, F212).
Several authors have argued in favor of different effects of EPL on labor productivity though. For example, Micco and Pagés (2002: 23) report a decline in average labor productivity as a result of stricter EPL for a sample of both developed and developing countries. Furthermore, Sargent (1978: 1012) argues, supported by empirical evidence based on U.S. data, that productivity evolves independently from employment and that EPL has no effect on productivity at all.
2.5.4 Effect of EPL on the Type of innovation
Similar to economic diversity (Frenken, Van Oort and Verburg, 2005) and competition (see the argument put forward in section 2.4.4), EPL is linked to the type of innovation in the literature. Since EPL hinders labor adjustment, EPL is argued to hinder innovation by increasing the costs in case the innovation is either a failure (when jobs are lost and severance pay has to be paid), or, in case the innovation is a success, in lower post-innovation profits, because stricter EPL gives employees more bargaining power (Bassanini and Ernst, 2002: 11). EPL may thus reduce the appetite for innovation altogether, and especially the appetite for risky innovations.
Bassanini and Ernst (2002: 15) proceed by making a distinction between two industry-specific innovation regimes: An entrepreneurial regime and a routinised regime (based on Audretsch, 1995). In an entrepreneurial regime, firms often undertake sequences of short-lived projects, based on general knowledge, and relying on a one-shot match of human and physical capital. In a routinised regime firms undertake incremental innovations along an existing technological trajectory. For this regime, internal labor markets are important, in which both workers and firms invest in firm-specific competencies. Bassanini and Ernst (2002: 15) argue, based on empirical support, that flexible EPL favors entrepreneurial regimes, and thus risky innovations in new goods and technologies, while strict EPL favors routinised regimes, and thus innovations in relatively secure goods and technologies.
Saint-19
Paul argues that countries with a rigid labor market will specialize in the latter type of innovation, and, hence, in the production of low-risk, mature goods. This is roughly consistent with innovation patterns between the U.S. on the one hand, and France and Germany on the other in 1993, with the U.S. having more patents in new industries such as biotechnology and computers, whilst Germany and France have more patents in traditional industries such as transportation and construction (Saint-Paul, 2002: 376).
The distinction between primary and secondary innovation by Saint-Paul, and the distinction between an entrepreneurial regime and a routinised regime by Bassanini and Ernst are both very similar to the distinction between radical and incremental innovation made in section 2.3.4. Strict EPL can thus be argued to tend to stimulate incremental innovation (i.e. secondary innovation and a routinised regime), and flexible EPL to tend to stimulate radical innovation (i.e. primary innovation and an entrepreneurial regime). Like economic diversity and competition, EPL is thus argued to have an effect on the innovation process and the type of innovations that are favored.
2.5.5 Relation between EPL and Diversity
Although economic diversity and EPL are argued to have an impact on the same variables (employment stability, employment growth and productivity growth) and are both argued to have an impact on the innovation process and whether radical or incremental innovations are stimulated, there has not been any research into how economic diversity and EPL are related to each other. There is also not much overlap between the determinants of both variables that are discussed in the literature. For economic diversity, on the one hand, we have discussed in section 2.2 the following determinants: Population level; GDP per capita; population density; geography; knowledge-based activities; trade; and industrial structure. EPL on the other hand, is argued to be determined by cultural variables, political variables (e.g. power of the left) and legal variables (e.g. legal origin, general preference for regulation) (Botero et al., 2004).
But perhaps culture could also be a determinant of economic diversity, especially considering a cultural continuum of opportunity-security. Security-oriented cultures might prefer strict EPL, and also be more risk-averse, which could lead to lower entrepreneurship and relatively less attempts to try new things, which will ultimately result in lower economic diversity. Opportunity-oriented cultures, on the contrary, might prefer flexible EPL and might be less risk-averse, which would result in higher economic diversity in the long-run. This suggests the pairs ‘low economic diversity, strict EPL’ and ‘high economic diversity, flexible EPL’ as two different institutional settings of an economy that would each be compatible with either a security-oriented culture or an opportunity-oriented culture. In total, one can make four extreme combinations of economic diversity and EPL, as depicted in Table 1.
Based on the various effects of economic diversity and EPL discussed in sections 2.3 and 2.5, one can make an argument what the dynamics will be in each of the four quadrants of Table 1. Starting in the top-left corner (1), employment would be expected to be very volatile because of both low economic diversity and flexible EPL. This is likely to create political support for stricter EPL, which would move the economy to situation (2).
Table 1. Four (extreme) combinations of economic diversity (ED) & EPL
Flexible EPL Strict EPL
Low economic diversity
(1) Employment is volatile because of low ED and flexible EPL, which will generate support for stricter EPL
(2) Strict EPL and low ED stimulate incremental innovation, which keeps ED low
High economic diversity
(3) High ED ensures employment stability; both high ED and flexible EPL stimulate radical innovation
