Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Do Interactions between Finance and Labor Market Institutions Affect Wage Distribution? Thibault Darcillon∗ October 16, 2013 Abstract This article analyzes the linkages between financial development, labor market institutions and wage inequality for 17 OECD countries over the 1989 to 2005 period. With the help of a fixed effect model with an interacted term, one crucial contribution of this article is to analyze the interacted impact of labor market institutions (i.e., workers’ bargaining power and employment protection legislation) on the one hand and financial development on the other hand on wage distribution. Our results indicate that changes in workers’ bargaining power and in employment protection affect wage distribution. Estimates of the marginal effects show that by increasing labor markers regulation (i.e., reinforcing workers’ bargaining power and increasing employment protection legislation) one also weakens the impact of financial development on the increase in wage inequality. Keywords: Financial Development, Labor market institutions, Political Economy, Wage inequality JEL Classification: F41 · I3 · P16 1 Introduction The increase in wage inequality in the last decades has been theoretically and empirical mainly explored. The role of foreign trade which is often assumed to increase the demand for highskilled workers or the premium paid to college graduates are the most popular explanations for the causes of increasing wage disparities. We focus in this article on the role of financial development and more specifically on the interactions between financial and labor markets to explain cross-national variations in wage disparities among OECD countries since the 1980s. Indeed, very recent papers (Jerzmanowski and Nabar, 2013; Philippon and Reshef, 2012) have shown that higher financial development contribute to the increase in the relative demand ∗ CES Centre d’Economie de la Sorbonne, Université Paris 1 Panthéon Sorbonne, Paris School of Economics, MSE, 106-112 Boulevard de l’Hôpital, 75647 PARIS Cedex 13. Email: [email protected]. This version is currently in revision in The Journal of Economic Inequality. Various versions of this paper have been presented at the 24th Annual Meeting of the ‘Society for the Advancement of Socio-Economics’ (SASE) in the Mini-Conference Socio Economic Conflict and the Dynamics of Institutional Change, in Cambridge (United States) on June 28-30, 2012; at the 20th International Conference of Europeanists of the ‘Council for European Studies’ (CES) in Amsterdam (Netherlands), on June 25-27, 2013; at the 25th Annual Conference of the ‘European Association of Labour Economists’ (EALE) in Turin (Italy) on September 19-21, 2013 and at the ‘Séminaire Interne des Institutions’ (SEI) on November 22, 2012 1 for skilled labor. In addition, we argue that the adoption of new corporate rules may have strong consequences on the wage distribution. Roe (2003) shows that countries with weak scores of Gini coefficient of national income inequality are more likely to have a high degree of ownership separation reflecting weakly developed stock markets. Beyond the direct impact of financial development on wage inequality, one major contribution of this article is to analyze the interactions between labor market institutions and the financial structures in a political-economy perspective. In other words, the purpose of this article is to show that changes in wage inequality in most OECD countries in the last decades can be interpreted as the result of the combined modifications in industrial relations and employment protection under the pressure of increasing financial markets. Our main argument is based on the stylized fact that some countries have experienced a larger increase in wage differentials, whereas all developed countries have been exposed to an important development of financial markets since the last decades. We provide some evidence of a reducing-effect of labor market institutions on wage dispersion when financial development is expanding. Our main hypothesis in this article is based on the concept of institutional complementarity: specific institutional forms are complementary if they jointly contribute to a higher economic performance (Amable, 2003). Several recent papers have shed some light on institutional complementarity between the financial systems and labor market institutions on specific economic performances. First, using data on output growth in 27 manufacturing industries in 19 OECD countries from 1970 to 1995, Ernst (2004) shows that concentration of ownership structures and employment protection act favorable on growth in bank-financed industries while ownership dispersion and labor market flexibility foster growth in equity-financed industries. More recently, Gatti, Rault and Vaubourg (2012) find that the interactions between labor and financial factors have a significant impact on unemployment in 18 OECD countries from 1980 to 2004: increasing stock market capitalization reduces unemployment with weak labor market institutions (union density and wage bargaining centralization) while enhancing intermediated credit increases unemployment with strongly regulated and coordinated labor markets. In the same vein, Wasmer and Weil (2004), in a general-equilibrium theoretical model, prove complementarity between increased competition in credit market and labor market flexibility in regard with unemployment. In other words, low banking concentration reduces unemployment when labor market institutions are weak. Our main results show that encompassing labor market institutions (i.e. workers with strong bargaining power and high employment protection legislation) are associated with a more compressed wage structure. In other words, our results indicate that strong encompassing labor market institutions contribute to the reduction in inequality in the era of financial development. The paper is organized as follows. Section 2 sets out our theoretical channels and the empirical related literature. In Section 3 we describe our dataset, the empirical model and the econometric results. Finally Section 4 provides some concluding remarks. 2 2 Theoretical channels and related literature 2.1 Financial development and wage inequality A growing literature has very recently investigated the linkages between financial development and wage inequality. The first explanation is that financial development contributes to the increase in the relative demand for skilled labor. In a recent presentation ,1 Daron Acemoglu argues that the deregulation in the financial sector since the 1980s in the United States has been responsible simultaneously for the increase in wage inequality at the top of the distribution and the financial crisis. Acemoglu shows that financial deregulation and the increase in the earnings inequality took place simultaneously. The literature has however shown mixed evidence on the relationship between financial development and wage inequality: a first series of papers (Jerzmanowski and Nabar, 2013; Philippon and Reshef, 2012; Larrain, 2013; Sjöberg, 2009) finds strong evidence that financial development has substantially increased the skill premium and thus the skilled/unskilled wage gap. By contrast, Beck, Levine and Levkov (2010) find that bank deregulation in the United States increased the relative wages of unskilled labor and is associated with a reduction in wage inequality at the bottom of the distribution. Using a two-sector model of endogenous growth with imperfect capital markets ,2 Jerzmanowski and Nabar (2013) demonstrate that financial development by improving the reallocation of workers by skill levels across firms (i.e. in the sense of an increase in the efficiency of the matching process for the allocation of skilled workers across sectors) is associated with an increase in the skill premium. First, the removal of financial constraints improve firm creations and increases the growth rate of the innovative economy productivity which increases the wages of skilled labor and then attracts skilled workers working in the manufacturing sector. Second, as the development of financial markets improves the matching process between financiers and skilled labor, some of skilled workers will move into the innovative economy. Thus, the number of skilled labor in the manufacturing economy decreases which reduces the relative wage of unskilled labor and then the skilled/unskilled wage differential. Empirical analysis on U.S. data at the aggregate level from 1977 to 2006 supports this hypothesis: overall, results indicate a positive effect of financial deregulation for skilled workers and a negative effect for unskilled workers. In this same line, Larrain (2013) finds strong evidence of a positive impact of capital account liberalization on wage inequality and more specifically strong support for the Ôcapitalskill complementarity hypothesis’ (Griliches, 1969) in 23 OECD countries from 1975-2005. According to this hypothesis, capital is more substitutable for unskilled labor and more complementary to skilled labor, financial development by facilitating the access to capital may be associated with an increase in the relative demand for skilled workers (Acemoglu, 2002). Results indicate that capital account liberalization increases wage inequality particularly in industries with high external financial dependence and strong capital-skill complementarity by 2.5% more than in the remaining industries. Alternatively, Beck, Levine and Levkov (2010) find that bank deregulation in the United 1 The presentation given by D. Acemoglu is available on http://economics.mit.edu/files/6348. The manufacturing sector produces final output and employs unskilled and skilled workers while the innovative sector produces a growing variety of intermediate and only employs skilled workers. 2 3 States decreases wage inequality at the bottom of the distribution. Two different arguments are employed: first, bank deregulation increases the incentive for lower-income individuals to invest in human capital; second, bank deregulation reduces the cost of capital, which has two opposite effects: a substitution effect of capital for labor and an output effect. In the latter case, bank deregulation can increase the demand for labor only if the output effect dominates. However, to us, the substitution effect should prevail over the output effect, offsetting the positive effects of bank deregulation on the increase relative demand for unskilled labor. Very recently, some papers using micro data have paid particular attention to the impact of the growth of the financial industry on wage differentials. Philippon and Reshef (2012) argue that financial deregulation in the United States has caused an increase in skill intensity and in wages more particularly in the financial sector: this has resulted in excess wages in finance and then an increase in wage differentials between the workers working in the finance industry and those working in the rest of the economy. According to Phillipon and Reshef (2012), financial deregulation has strongly contributed to the increase in the demand for skill in the financial sector: they argue that deregulation in the financial sector has raised the productivity gap between the financial industry and the rest of the economy and has increasingly attracted high-skilled workers. Philippon and Reshef find that deregulation alone accounts for 23% of changes in wages in the financial sector. Whereas finance industry hires more proportionately high-skilled workers in the 1980s and 1990s when financial markets are strongly deregulated, wages in finance industry has grown faster than in the rest of the economy from 1995 to 2006. By this way, finance has contributed more proportionately to the increase in wage inequality at the top of the distribution: whereas finance contributes to 6.2% of the increase in the p90 /p10 ratio, this effect accounts for 15% of the increase in the p97 /p10 ratio. More generally, their analysis reveals that finance is responsible for 15% to 25% of the overall increase in wage inequality since 1980. These findings are consistent with the analysis by Moss (2010) who finds that the periods of financial liberalization are associated with an increase in the 10% top income share in the United States from 1917 to 2009. Top 10% income share substantially decreased after 1933 after the introduction of different regulatory measures to repress the financial system. Top 10% income share has then remained weak until the beginning of bank and finance deregulation in the early 1980s. More importantly, top 10% income share sharply increased from 1980 to nowadays. In the same vein, Godechot (2012) shows that half of the share of the top 0.1% in France is due to the increase in pay among top finance managers between 1996 and 2007. Another possible explanation is related to the nature of corporate governance with specific compensation arrangements. In this line, Sjöberg (2009) investigates the impact of more shareholder-oriented corporate governance on wage inequality. He argues that the development of more liquid financial markets has been translated into the adoption of new corporate rules. In many countries, new forms of incentive remuneration for executives (such as performance related salary or stock-options) were introduced in the 1980s and in the 1990s with the aim of aligning the minority shareholders’ interests with those of managers by proposing high rewards to top managers. This may have particular consequences at the top and at the intermediate levels of the wage distribution. Using a sample of 15 OECD countries from 1979 and 2000, Sjöberg (2009) finds that higher merge and acquisitions activities and a higher degree of minority shareholder protection are associated with increasing wage inequality. 4 2.2 Labor market institutions and wage inequality Labor market institutions are generally seen as powerful mechanisms of income insurance against shocks and in this sense may have strong impact on wages and employment. Empirical and theoretical work in labor economics has shown an ambiguous impact of unionization and centralized wage bargaining on wage dispersion. Unionization and centralized wage bargaining contribute to the reduction in wage differentials among union members (i.e. the inequality-reducing ‘between-sector’ effect) and among the individuals covered by union agreements. But, it may also participate to the increase in wage differentials between union members and nonmembers (i.e. the inequality-increasing ‘within-sector’ effect). Overall, Card, Lemieux and Riddell (2004) find a positive effect of unionization on the reduction in wage dispersion among men in the United States, the United Kingdom and in Canada from the early 1970s to 2001 because the former effect is smaller than the latter effect. As it has been frequently argued in the standard wage bargaining models, the effect of higher union bargaining power on wage dispersion can be ambiguous: the objective of the union is to maximize the expected wage of its members. Because unions tend to reduce wage competition, this reduces low-wage employment, which has two contradictory effects: it increases average wages for low-skilled, which tends to decrease wage dispersion, but this effect might be offset by a rise in unemployment rates caused by an increased average wage (Checchi and García-Peñalosa, 2010). But another explanation claims that unions may be willing to moderate their wage claims, and this more particularly when unions are more coordinated: union leaders take into account that higher wage may in-duce an increase in unemployment. Wage moderation has, in this case, a positive effect on aggregate employment (Checchi and Nunziata, 2011). Koeniger, Leonardi and Nunziata (2007) emphasize a strong complementarity between unions’ decisions and the degree of centralization/coordination of wage bargaining: a large agreement coverage increases the unions’ influence throughout the economy. In addition, the individualization of the wage bargaining reduces the unions’ capacity of maintaining a unilateral control of the interfirm and intersectoral wage differentials (Rueda and Pontusson, 2000). Empirical work in political economy has found strong evidence for a reducing-effect of unionization on wage dispersion: Pontusson, Rueda and Way (2002) find that unionization decreases wage inequality in the lower half of the distribution that tends to be more unionized than the upper half. Moreover, Western and Rosenfeld (2011) argue that unions have a positive effect on the reduction in overall wage dispersion because they “contribute to a moral economy that institutionalizes norms for fair pay, even for nonunion workers” (p.514). Finally, Blau and Kahn (1996) find that the decentralization of wage bargaining in the United States can explain the greater dispersion of wages for non-union workers than in other countries. The literature explaining the linkages between employment protection legislation (EPL) and wage inequality is quite sparse. Koeniger et al. (2007) argue that the strictness of employment protection legislation is more likely to protect the unskilled relative to the skilled workers and therefore to improve their bargaining position. It is implicitly supposed that firing taxes are more important for unskilled workers: dismissal costs create a hold-up problem in the sense that they reduce the producers’ outside options. Collective dismissal costs allow workers to collectively bid up their wage. For this reason, dismissal costs compress the wage 5 differential if they are relatively more important for unskilled workers. More generally, employment protection makes more difficult for a worker to be laid off even during economic downturns. According to the OECD, higher employment protection raises the costs of employment which moves downwards the labor demand curve particularly for lower-productivity workers. The effect on EPL on wage dispersion is a priori ambiguous because the impact of EPL on wages and employment is ambiguous. There is a general consensus on the fact that decreased unemployment is associated with a reduction in wage dispersion (see supra). However, there is a permanent debate among economists on the effect of stronger EPL on aggregate employment. Some papers have shown that this effect may also be conditional to the intensity in product market regulation. There are two competing conceptions in the literature. A dominant literature (Nicoletti and Scarpetta, 2002; Blanchard and Giavazzi, 2003) sees a positive effect of reducing labor market regulation on the reduction in unemployment, particularly in countries with strongly deregulated product markets. In other words, Blanchard and Giavazzi (2003) find that strongly deregulated product markets increase the incentive to reduce labor market regulation which will have a positive effect on aggregate employment. Thus, reducing employment protection in countries with deregulated product markets may decrease unemployment and then wage dispersion. A challenging view (Amable and Gatti, 2004; Amable, Demmou and Gatti, 2011) considers product market competition and labor market competition as substitutes. Using a dynamic efficiency model, Amable and Gatti (2004) demonstrate that higher product market competition by increasing job turnover and thus efficiency wage premium results in lower aggregate employment. Increasing employment protection has in this case a positive effect on aggregate employment by avoiding wage pressure. Amable, Demmou and Gatti (2011) find support for the substitution hypothesis between product market competition and labor market competition. Thus, increasing employment protection in countries with deregulated product markets may reduce unemployment and then reduce wage dispersion. Empirical work has found mixed evidence on the relationship between stronger EPL and wage dispersion. Checchi and García-Peñalosa (2008) find no robust relationship whereas Koeniger et al. (2007) find evidence on a sample of 11 OECD countries for the 1973-1998 period that the strictness of employment protection legislation is positively and significantly correlated with a more compressed wage structure. 2.3 Finance, labor market institutions and wage inequality It is often argued that the increasing financial development has caused a rise in wage inequality in the OECD countries during the three last decades (OECD, 2011). Some countries have experienced a larger increase in wage differentials than others whereas all countries were exposed to an important development of financial markets.3 We argue that some countries have weaker wage inequality because they are simultaneously stronger labor market institutions and more regulated financial markets. We suppose that the impact of finance on the wage distribution is conditioned by the existence of two central labor market institutions: collective bargaining institutions on the one hand and the strictness of 3 For instance, Figure 1 reveals that inequality is much weaker in France than in Germany while France has experienced higher financial development during the last decade: market capitalization ratio of listed companies increased from 27.5% in 1990 to 102% in 2007 in France and from 21% in 1990 to 57.1% in 2007 in Germany (Beck et al., 2010) 6 employment protection legislation on the other hand. We argue that weak labor market institutions contribute to the increase in wage inequality in countries where financial markets have been sharply liberalized. In that sense, we assume that strong labor market institutions have a reducing-effect on wage dispersion when the financial markets are expanding. First, we assume that an increase in financial development raises wage inequality, particularly when labor markets institutions are weak. Symmetrically, we argue that an increase in labor market regulation reduces wage inequality, especially when the influence of financial markets becomes larger. This section provides theoretical and empirical arguments on the combined effect of institutional arrangements on labor and financial markets on the wage distribution. The first channel describes the relationship between labor market institutions and the relative demand for skilled labor in a context of financial development; the second channel describes the relationship between market liquidity and the incentives for specific investments. The last channel looks at the impact of time horizon on corporate governance mechanisms and its effect on wage disparities. 2.3.1 Labor market institutions and the relative demand for skilled labor First, labor market regulation can mitigate the positive effect of financial development on the rise in the skill premium. Strong labor market institutions are powerful drivers for maintaining a control over intersectoral wage differentials (between skilled and unskilled labor), and will reduce the negative impact of financial development on the absolute level of the unskilled wage by offering higher wages (Koeniger et al., 2007) but also with collective agreements on a national minimum wage. Second, as argued by Rodrik (1997) and Sjöberg (2009), higher volatility on the labor market may be correlated with higher wage dispersion. Financial development at the international and domestic levels increases the volatility of hours worked and the volatility of wages, but that stronger labor market institution may also have an impact on mitigating labor-market volatility. Thesmar and Thoenig (2004) demonstrate that the rise in uncertainty caused by financial development is smaller in countries with strong labor market regulation. 2.3.2 Market liquidity and specific investments A second argument stresses the institutional complementarities between financial markets and labor markets institutions on the incentive to invest in specific assets: high liquidity on the financial and labor markets increases the outside option for workers and for financial investors which reduces their incentive to invest in specific resources (Hall and Soskice, 2001; Ernst, Amable and Palombarini, 2005). When financial and labor markets are both very liquid, firms are more likely to offer higher wages for trained and skilled workers (due to the wage efficiency premium) to prevent Ôskill poaching’. Competitive markets on labor markets combined with diffuse capital ownership may lead to higher wage differentials. According to Acemoglu and Pischke (1999), the countries with general skill training (such as the United States) in comparison with the countries with specific skill training (such as Germany) have weaker incentive to train unskilled workers in reaction to the introduction of new technologies. By contrast, in the economies with low liquidity of financial markets, stronger ownership concentration (i.e. insider monitoring) combined with strong union’s bargaining power improve 7 the evaluation of firm-specific assets and increases the incentive for workers to invest in specific assets: thus, the incentives for the development of internal labor markets are increased which induces a more compressed wage structure. In addition, employment protection appears as a powerful driver for specific skill investment (Estevez-Abe, Iversen and Soskice, 2001): in this case, workers provide higher effort (because they are more productive) and firms do not need to offer high efficiency premium to attract workers. 2.3.3 Corporate governance, time horizon and wage inequality It has been frequently argued that bank financing may reduce firms’ turnover and increase workers’ efforts. This also contributes to stabilizing the time horizon of management and to give greater incentive for workers to adopt a long-term strategy regarding wage bargaining (Ernst, Amable and Palombarini, 2005). By contrast, financial development may push managers to adopt short-term strategies and new corporate rules that include the use of individual bonuses for managerial employees and collective bonuses for all employees. In this case, the wage structure is more likely to become less homogenous because workers are paid their marginal product. As we have earlier argued, strong collective bargaining institutions can in this context avoid that wage dispersion becomes larger. 3 3.1 Empirics Measuring wage inequality We use as the dependent variable the ratio of earnings at the 90th percentile to earnings at the 10th percentile (p9 /p1 ratio) that is a summary measure of the distribution of gross income from employment. Our measure ignores other sources of income, such as self-employment, income from capital, and government transfers. This measure of income inequality is limited to fulltime employees. More precisely, this indicator includes the annual base salary, overtime pay, some several periodic payments. However, some elements relative to executive compensation such as stock options can be excluded. Figure 1 provides a graphic summary of the wagedistributive outcomes and reveals important cross-national variations in levels and in trends: in the United States wage inequality is traditionally high and is constantly increasing whereas the Scandinavian countries (e.g., Sweden or Norway) have weak levels of inequality. Inequality has particularly risen in the United States, the United Kingdom but also in Austria or Germany where wage inequality was traditionally weak. This paper seeks to explain this cross-national diversity. [INSERT FIGURE 1] To look at the impact of financial development on the lower-tail inequality, we use the p5 /p1 earnings ratio. This ratio compares the median with the bottom of the wage distribution. We choose to use data on earnings ratio because they are available for a wide range of countries and across time. We consider that these data are consistent to conduct a cross-country comparison over time. Data on earnings are preferable to data on Gini coefficients because the variety of the methods of calculating these coefficients remains a major problem to conduct a crosssectional analysis over time. Figure 1 displays the wage distribution at the lower end: in most 8 OECD countries, low-tail inequality has remained stable (e.g. in Australia, Finland, Japan, Netherlands, Norway or Sweden) or has even slightly declined in France. As a mirror to our previous description of overall earnings inequality, we find that the United Kingdom and the United States are featured by very high inequality at the bottom of the distribution, and with an increasing trend for the United States. More interestingly, Germany experienced a deep decrease in low-tail inequality in the early of the 1990s. In the mid-1990s, inequality has, however, since then gradually increased. However, because data on earnings ratio do not include any forms of financial bonuses (such as stock options) for executives, we use the top 10% income share index proposed by Alvaredo, Atkinson, Piketty and Saez in their World Top Incomes Database to capture upper-tail inequality. This indicator presents some limitations: first, this measure provides no information about the evolution of inequality elsewhere in the distribution than at the top of the distribution. Then, the database is constructed using tax statistics: the series are concerned with gross income before tax and the definition of income can vary across countries. The picture depicted in Figure 2 is not very different in what is described in Figure 1: most countries, including the European countries, have experienced an in- crease in uppertail inequality. Top income share is very high in Canada, the United States and the United Kingdom and relatively low in Denmark, the Netherlands or in Sweden which is consistent with our description of overall inequality. [INSERT FIGURE 2] 3.2 Data on financial development and labor market institutions We use two measures of financial development: the stock market capitalization ratio to GDP and the stock market total value traded divided by GDP. The stock market capitalization ratio gives a measure of stock market activity, i.e. to what extent the stock market can efficiently allocate capital to investment projects. A more developed financial market is also assumed to increases the investors’ opportunities for risk diversification. Consequently, this indicator reflects the capacity of stock markets to provide external financing. In other words, the higher the ratio, the more likely firms in this economy would need of financing from external minority shareholders. The second indicator is the stock market total value traded divided by GDP. This indicator refers to the total value of shares traded during the period and is a common measure of liquidity of equity markets. Data on financial development is provided by Beck et al. (2010). Then, we are interesting in determining the impact of specific labor market institutions on wage distribution. First, we want to capture the impact of collective bargaining institutions on wage distribution. Collective bargaining institutions refer to the composition and the strength of trade unions and to the system of collective bargaining. To capture collective bargaining institutions it seems to us to be more appropriate to adopt a multidimensional definition because trade union density does not capture the strength of bargaining power but rather refers to the unions’ attractiveness to potential new members (Vernon, 2006). We have constructed a synthetic indicator Workers’ Bargaining Power by running a principal-component analysis (PCA) using three variables provided by Visser (2011): the degree of coordination of union bargaining, the mandatory extension of collective agreements by public law to non-organized firms, and union density (Gordon, 1994). Visser (2011) distinguishes five levels of coordination of wage 9 bargaining: economy-wide bargaining, mixed industry and economy-wide bargaining, industry bargaining with no or irregular pattern setting, mixed or alternating industry- and firm level bargaining, and fragmented bargaining. Extension of collective agreements to non-organized firms can be regularly applied and affecting a significant share of the workforce (>10%), can be not regularly or widely used (< 10%). In some countries, there is no legal provision for mandatory extension. High levels of wage bargaining centralization and extension of collective agreements are supposed to be correlated with a more homogenous wage distribution. Finally, we include in our composite index union density defined as the net union membership as a proportion wage and salary earners in employment. Looking at the impact of employment protection on wage distribution, we use the Employment Protection Legislation (EPL) indicator provided by the OECD available from 1985 to 2009. The OECD proposes a measure of employment protection strictness taking into account three different components: (i) protection of regular workers against individual dismissal, (ii) regulation of temporary forms of employment, and (iii) specific requirements for collective dismissals. 3.3 Econometric specification and methodology This study uses time-series cross-section data for 18 OECD countries (Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Spain, Sweden, United Kingdom and United States) over the 19892005 period. Table A1 in Appendix provides descriptive statistics for the variables used in the regressions. The aim of this article is to analyze the impact of financial development on wage distribution. More specifically, we want to test the argument that the impact of financial development on the wage structure is conditional to specific labor market institutions. We introduced an interactive term between financial development and labor market institutions. We will estimate the following relationship: log(Yit ) = αi + νt + β1 · F INit + β2 · LABit + β3 · F INit · LABit + βk · Σk Xk,it + it (1) where αi is the country i fixed effect, νt the time specific effect, Yit denotes the dependent variable capturing wage inequality, F INit denotes a set of two financial development variables (Stock Market Capitalization Ratio and Stock Market Total Value Traded ), LABit is a set of two variables accounting for labor market institutions P(BARGit for workers’ bargaining power and EP Lit for employment protection legislation), k Xk,it a set of control variables, and it an error term. We want estimate the joint effects of the degree of financial development and labor market institutions on wage distribution by using data on 18 developed countries over the 1989-2009 period. We control for different macroeconomic variables that may have consequences on the wage distribution: trade openness, unemployment rate and technological change. First, to estimate the impact of trade openness on the wage distribution, we compute the sum of the exports and imports as a percentage of current GDP (Trade Openness). It has been very frequently argued in the economic literature that economic openness contributes to the increase in wage differentials (p9 /p1 ratio) in the sense that globalization has brought higher rewards mainly for 10 high-skilled workers. The removal of barriers to trade and the workers’ shift from traditional low-productivity toward modern high-productivity activities has caused a rise in inequality in most industrial countries. As it is very frequently argued in the standard literature on inequality, increased trade integration would be associated with higher relative wages of skilled workers in richer countries. Symmetrically, low-skilled workers in those countries are directly competed with low-skilled workers from the emerging countries. Hence, trade integration is accordingly associated with a rise in wage inequality in developed countries. Then, we suppose that unemployment may also have a direct impact on the wage structure. An increase in unemployment rates may exert downward pressure on wages, particularly at the lower end of the wage distribution, because unemployment is more likely to affect low-paid and low-skilled workers and then to reduce the threshold earnings p1 . But it should be also noted that the effect of unemployment rate on the wage distribution is dependent on the proportion of unemployed, including in countries with generous unemployment subsidies (Checchi and García-Peñalosa, 2008). We compute the unemployment rate as a percentage of civilian labor force (Unemployment). In addition, we control for GDP growth. Theoretical literature has emphasized a mixed impact of GDP growth on earnings inequality. Aghion, Caroli and García-Peñalosa (1999) conclude that key factors generating growth (essentially international, technological and organizational change) have an increasing-effect on inequality. In that sense, growth does not necessarily increase earnings inequality. We compute the growth of the real GDP, expressed in percentage changes from year (GDP Growth). Data on trade openness, unemployment and GDP growth are provided by Armingeon et al. (2012). Finally, we want to estimate the impact of technological change on the wage distribution. As noted by García-Peñalosa (2010), technological change is often seen as inherently ‘skill-biased’. Consequently, faster technology-driven growth may result in greater earnings inequality as technological change reduces the relative wage of unskilled or low- skilled workers. Following the OECD (2011), we take as a proxy of technological change the research and development (R&D) expenditures. The variable R&D Expenditures refers to business enterprise expenditures on R&D as a percentage of GDP and is provided by the OECD. In the first specification, we run OLS estimates with PCSE (Panel Corrected Standard Error) estimator in order to assess the combined impact on labor market institutions and financial liberalization on wage distribution. This estimator proposed by Beck and Katz (1995) is appropriate for TSCS (Time-Series Cross-Section) data in presence of heteroskedasticity and error dependence. In this first model, labor market institutions are considered as strictly exogenous. This first approach explores the correlations between financial development and wage inequality. Then, we address the issue of reverse causality by running IV-GMM estimations: labor market institutions and financial development can be endogenous to the level of wage inequality. Changes in the distribution of wage may affect the political economy in shaping labor market institutions and financing sector, especially by left-wing governments (Darcillon, 2013). Our financial variables, labor market variables and the interactive terms are thus all considered as endogenous. But finding good instruments is always a difficult task. On the basis on theoretical considerations and empirical analysis, we choose several instruments. Empirical and theoretical literature on the political economy of labor market institutions (Saint-Paul, 2000; Pontusson, Rueda and Way, 2002) has pointed out the role of replacement rate of unemployment, labor compensation and economic regulation. We include central determinants of financial development: two different financial liberalization indexes 11 and political-economy variables, including government ideological orientation, the nature of voting system and the political regime (Darcillon, 2013). Then, we submit those instruments to a series of tests to confirm its validity. Our principal argument is based on the idea that the effect of the financial liberalization on wage inequality is conditional on specific levels of workers’ bargaining power and employment protection legislation: ∂log(Yit ) ∂log(Yit ) = βb1 + βb3 · LABit and = βb2 + βb3 · F INit (2) ∂F INit ∂LABit The coefficient β1 can be interpreted as the effect of financial development on wage distribution but when LABit equals zero. Determining the significance of the effect of F INit on log(Yit ) conditional on LABit values, we compute the standard error of the sum (β1 +β3 ·LABit ) as follows: r se(β +β ·LAB ) = var βb1 + LAB 2 · var βb3 + 2LABit · cov βb1 βb3 (3) 1 3 it it Accordingly, the t values obtained from an interactive model indicate the effect of an independent variable on the dependent variable but depending on particular levels of another independent variable: hence, it is not surprising that insignificant variables can produce significant marginal effects (Friedrich, 1982). We report marginal effects of financial liberalization at different sample values of labor market institutions (minimum, mean minus one standard deviation, mean, mean plus one standard deviation, maximum) and vice versa. 3.4 Results Results are reported in Tables 1, 3 and 5 respectively for overall wage inequality, lowertail inequality and upper-tail inequality. PCSE/OLS and IV-GMM estimations give us very substantially similar results. ‘ First, the coefficients for control variables are not systematically significant. We find that, when they appear statistically significant, they have the expected sign: we find that trade openness and higher RD expenditures are all positively correlated with our three dependent variables. Our regressions give us mixed results on the impact of unemployment rate on wage inequality: PCSE/OLS estimations show a positive impact on wage inequality while IV-GMM estimations indicate a negative effect. We find no impact of GDP growth on our dependent variables. Then, focusing on the impact of financial variables, we find that higher stock market capitalization ratio and higher stock market total value traded have both a statistical significant and a positive impact on wage dispersion, more particularly on upper-tail inequality. Moreover, marginal effects given in Tables 2A indicate a decreasing relation between financial variables and the p9/p1 ratio when the level of labor market institutions is increasing: PCSE/OLS estimations indicate that financial development decreases wage inequality when labor market institutions are strong whereas IV-GMM estimations tend to show that financial development particularly increases inequality when labor market institutions are weak. Table 4A indicates very similar results on the impact of financial variables on lower-tail inequality: we find in PCSE/OLS estimations that financial development is associated with an increase in wage inequality when labor market institutions are weak and is correlated with a reduction when labor market institutions are strong. These results are supported in IV-GMM estimations that show 12 a negative relationship between financial development and wage inequality particularly when labor market institutions are strong. Finally, we find that at the lowest level of labor market institutions financial development is strongly associated with higher inequality, especially at the top of the distribution (Table 6A). However, we find a positive effect, albeit weaker, when the level of labor market institutions increases, except in model (2) where financial development is associated with a reduction in upper-tail inequality when EPL is higher. Hence, these results suggest that inequality generated by financial development is weaker in countries where labor market institutions are stronger. [INSERT TABLES 1 TO 6] Finally, if we look at the impact of labor market variables, our results show that labor market institutions have a strong negative impact on our three different dependent variables. Although the empirical literature on wage inequality has found ambiguous results, we find that strong labor market institutions clearly contribute to the reduction in wage inequality. Marginal effects in Tables 2B, 4B and 6B are significant and negative for all values of the financial variables. More importantly, we find a decreasing relationship between labor market variables and our three dependent variables when the degree of financial development is increasing. Nevertheless, we find in some IV-GMM specifications that labor market institutions are associated with an increase in lower-tail inequality when financial markets are weakly developed and with a reduction in inequality when liquidity on financial markets is higher (Table 4B). These results indicate that the effect of labor market variables on wage dispersion is very robust and would suggest that labor market institution have a stronger impact in the countries with strongly developed financial markets. Our results provide some support for our main hypothesis on the compensating effect of labor market institutions on wage dispersion: (i) first, we show that financial development increases wage inequality in countries where labor market institutions are weak and vice versa. Hence, the impact of financial development is conditional to the institutional arrangements on the labor markets. Henceforth, the development of financial markets has caused higher increase in wage inequality in countries with weak labor market institutions (e.g., the United States). (ii) Second, the impact of labor market institutions on the wage distribution is less driven by the degree of financial development. However, we find a strong impact on labor market institutions on our dependent variables: reducing labor market regulation would have strong consequences on wage disparities and this including in countries with traditionally weakly developed financial markets (such as in Germany). Hence, the rise in wage inequality experienced by the Anglo-Saxon countries since the three last decades can be seen as the result of the reduction in labor market regulation and the development of active financial markets. 4 Concluding remarks The aim of this article was to analyze the impact of the interactions between financial and labor markets on the wage distribution. We find that encompassing labor market institutions - strong workers’ bargaining power and the strictness of employment protection legislation contribute to the reduction in wage disparities in a strongly financially developed world. We find that, increasing the level of labor market institutions, one also weakens the negative impact of financial development on the increase in wage inequality accounting for the 13 interactions between these institutional arrangements and the degree of financial development. This analysis based on the interactions among different institutions confirms the existence of strong institutional complementarities between financial and labor markets. These complementarities can be used to explain cross-national variations in wage inequality: our results tend to demonstrate that financial development alone cannot be a source of the rise in wage inequality. We show that labor market institutions have been powerful drivers of changes in wage inequality in these last decades. However, we find contrasting evidence that financial development has exacerbated wage inequality through a significant impact on labor market institutions. Our results provide an explanation of the cross-national variations in wage inequality: increasing financial development implies larger inequality in countries with weak labor market regulation whereas countries with stronger encompassing labor market institutions have experienced a more modest increase in wage inequality. In our analysis, labor market institutions play a central role in explaining wage differentials: our results also indicate that labor market institutions decrease wage inequality more particularly when financial markets are strongly developed. There is, however, a wide variety of alternative causes of growing wage inequality: skill-biased technological changes, changes in social expenditure policies, changes in employment patterns, changes in family formation and household structures, changes in tax and benefit systems, more specifically changes in progressive taxation systems (OECD, 2011). References [1] Acemoglu, D.: Technical Change, inequality, and the labour market, Journal of Economic Literature, XL, 7-72 (2002) [2] Acemoglu, D., Pischke, J.-S.: Beyond Becker: Training in Imperfect Labour Markets, The Economic Journal, 109(453), 112-142 (1999) [3] Aghion, P., Caroli, E., García-Peñalosa, C.: Inequality and Economic Growth: The Perspective of the New Growth Theories, Journal of Economic Literature, 37(4), 1615-1660 (199) [4] Alvaredo, F., Atkinson, A.B., Piketty, T., Saez, E.: The World Top Incomes Database, http://g-mond.parisschoolofeconomics.eu/topincomes, Cited 1 Oct 2012 [5] Amable, B.: The Diversity of Modern Capitalism, Oxford: Oxford University Press (2003) [6] Amable, B., Demmou, L., Gatti, D.: The effect of employment protection and product market performance: substitution or complementarity?, Applied Economics, 43(4), 449464 (2011) [7] Amable, B., Gatti, D.: Product Market Competition, Job Security, and Aggregate Employment, Oxford Economic Papers, 56(4), 667-686 (2004) [8] Armingeon, K., Potolidis, P., Gerber, M., Leimgruber, P.: Comparative Political Data Set I 1960-2010, Institute of Political Science, University of Berne (2012) [9] Baum, C.F., Schaffer, M.E., Stillman, S.: Enhanced routines for instrumental variables/GMM estimation and testing, Stata Journal, 7(4), 465-506 (2007) 14 [10] Beck, T., Demirgüç-Kunt, A., Levine, R.: A new database on financial development and structure. World Bank (2010) [11] Beck, N., Katz, J.N.: What To Do (and Not To Do) with Times-Series-Cross-Section Data in Comparative Politics, American Political Science Review, 89(3), 634-647 (1995) [12] Beck, T., Levine, R., Levkov, A.: Big Bad Banks? The Winners and Losers from Bank Deregulation in the United States, The Journal of Finance, 65(5), 1637-1667 (2010) [13] Blanchard, O., Giavazzi, F.: Macroeconomic effects of regulation and deregulation in goods and labor markets, The Quarterly Journal of Economics, 118(3), 879-907 (2003) [14] Blau, F.D., Kahn, L.M.: International Differences in Male Wage Inequality: Institutions versus Market Forces, Journal of Political Economy, 104(4), 791-837 (1996) [15] Card, D., Lemieux, T., Riddell, W.C.: Unions and Wage Inequality, Journal of Labor Research, 25(4), 519-559 (2004) [16] Checchi, D., García-Peñalosa, C.: Labour market institutions and income inequality, Economic Policy, 23(56), 600-651 (2008) [17] Checchi, D., García-Peñalosa, C.: Labour Market Institutions and the Personal Distribution of Income in the OECD, Economica, 77(307), 413-450 (2010) [18] Checchi, D., Nunziata, L.: Models of unionism and unemployment, European Journal of Industrial Relations, 17(2), 141-152 (2011) [19] Darcillon, T. : Mesurer l’impact de l’effet partisan sur les réformes de corporate governance, Revue Economique, 64(4), 445-455 (2013) [20] Ernst, E.: Financial Systems, Industrial Relations and Industry Specialization: An Econometric Analysis of Institutional Complementarity, In: Oesterrichische Nationalbank, Workshops, Proceedings of the OeNB Workshops No.1, The Transformation of the European Financial System. Where Do We Go? Where Should We Go?, OeNB, Vienna, pp. 60-95 (2004) [21] Ernst, E., Amable, B., Palombarini, S.: Endogenous Shocks and Evolutionary Strategy: Application to a Three-Players Game, In: Andrzej S. Nowak and Krzysztof Szajowski (eds.) Advances in Dynamic Games. Annals of the International Society of Dynamic Games, vol. 7, pp. 369-390, Springer, Birkhäuser Boston (2005) [22] Estevez-Abe, M., Iversen, T. Soskice, D.: Social Protection and the Formation of Skills: A Reinterpretation of the Welfare State, In: Peter A. Hall and David Soskice (eds.) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, pp. 145-183 (2001) [23] Friedrich, R.: In Defense of Multiplicative Terms in Multiple Regression Equations, American Journal of Political Science, 26, 797-833 (1982) [24] García-Peñalosaa, C.: Income distribution, economic growth and European Integration, Journal of Economic Inequality, 8(3), 277-292 (2010) 15 [25] Gatti, D., Rault, C., Vaubourg, A.G.: Unemployment and Finance: How Do Financial and Labour Market Factors Interact?, Oxford Economic Papers, 64(3), 464-489 (2012) [26] Godechot, O.: Is finance responsible for the rise in wage inequality in France?, SocioEconomic Review, 10(3), 447-470 (2012) [27] Gordon, D.M.: Bosses of different stripes: A cross-national perspective on monitoring and supervision, American Economic Review, 64(2), 375-379 (1994) [28] Hall, P.A., Soskice, D.: An Introduction to Varieties of Capitalism, In: Hall, P.A., Soskice, D. (eds) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage, pp. 1-68. (2001) [29] Jerzmanowski, M., Nabar, M.: Financial development and wage inequality: theory and evidence, Economic Inquiry, 51(1), 211-234 (2013) [30] Koeniger W., Leonardi M., Nunziata, L.: Labor Market Institutions and Wage Differentials, Industrial & Labor Relations Review, 60(3), 340-356 (2007) [31] Larrain, M.: Does Financial Liberalization Contribute to Wage Inequality? The Role of Capital-skill Complementarity, Working Paper, June (2013) [32] Moss, D: Comments on Bank Failure/Regulation/Inequality Chart, Havard Business School, August (2010) [33] Nicoletti, G., Scarpetta, S.: Interactions between Product and Labour Market Regulations: Do They Affect Employment? Evidence of the OECD Countries, Unpublished paper, OECD Economics Department, Paris (2002) [34] OECD: Divided We Stand? Why Inequality Keeps Rising, OECD Publishing, Paris (2011) [35] Philippon, T., Reshef, A.: Wages and Human Capital in the U.S. Financial Industry: 1909-2006, The Quarterly Journal of Economics, 127(4), 1551-1609 (2012) [36] Pontusson, J., Rueda, D., Way, C.R.: Comparative Political Economy of Wage Distribution: The Role of Partisanship and Labour Market Institutions, British Journal of Political Science, 32(2), 281-308 (2002) [37] Rodrik, D.: Has Globalization Gone Too Far?, Washington, D.C.: Institute for International Economics (1997) [38] Roe, M.J.: Political Determinants of Corporate Governance. Political Context, Corporate Impact. Oxford University Press, Oxford (2003) [39] Rueda, D., Pontusson, J.: Wage Inequality and Varieties of Capitalism, World Politics, 52, 350-383 (2000) [40] Saint-Paul, G.: The Political Economy of Labour Market Institutions, Oxford: Oxford University Press (2000) [41] Sjöberg, O.: Corporate Governance and Earnings Inequality in the OECD Countries 1979-2000, European Sociological Review, 25(5), 519-533 (2009) 16 [42] Thesmar, D., Thoenig, M.: Financial Market Developments and The Rise in Firm Level Uncertainty, CEPR Discussion Paper, No. 4761 (2004) [43] Vernon, G.: Does Density Matter? The Significance of Comparative Historical Variation in Unionization, European Journal of Industrial Relations, 12(2), 189-209 (2006) [44] Visser, J.: Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention and Social Pacts, 1960-2010 (ICTWSS). Amsterdam Institute for Advanced Labour Studies AIAS, University of Amsterdam, Version 3.0 (2011) [45] Wasmer, E., Weil, P.: The macroconomics of labour and credit market imperfections, American Economic Review, 94(4), 944-963 (2004) [46] Western, B., Rosenfeld, J.: Unions, Norms, and the Rise in U.S. Wage Inequality, American Sociological Review, 76(4), 513-537 (2011) 17 Appendix. Summary statistics Variable Mean Std. Dev. Min. Max. N n p9 /p1 ratio p5 /p1 ratio Top 10% income share Stock market capitalization ratio (stmktcap) Total value of shares traded on the stock market/GDP (stvaltraded) Workers’ bargaining power Employment Protection Legislation (EPL) Unemployment rate Openness Government ideological position GDP growth R&D expenditures/GDP 1.063 0.486 31.251 66.492 57.905 0.00 1.954 5.721 67.611 1.38 2.346 1.198 0.216 0.126 5.359 39.25 61.401 1.297 0.972 3.633 34.434 15.973 0.266 0.607 0.662 0.25 21 5.556 1.679 -2.465 0.21 0.00 16.106 -49.839 0.304 0.182 1.606 0.847 45.96 246.196 401.233 2.54 3.82 16.817 184.308 45.6 2.988 3.199 389 361 377 368 368 538 428 518 540 540 539 463 17 17 17 17 17 17 17 17 17 17 17 17 18 Table 1. The impact of financial development on overall inequality Table 1. The impact of financial development on overall inequality Dependent variable : p9/p1 ratio Stock Market Capitalization Ratio (1) (2) (3) (4) -0.0004 (0.0003) 0.0007* (0.0004) 0.0004*** (0.0001) 0.0007* (0.0004) Stock Market Total Value Traded Workers' Bargaining Power Employment Protection Legislation (EPL) Workers' Barg. Power x Stock Market Cap. -0.0807*** (0.0146) -0.0980*** (0.0109) -0.0004*** (0.0001) Empl. Protection Leg. x Stock Market Cap. -0.1038*** (0.0090) -0.0637*** (0.0142) -0.0132 (0.0150) -0.1075*** (0.0242) -0.0001 (0.0001) -0.0006*** (0.0002) -0.0205* (0.0111) -0.0833*** (0.0267) (5) (6) (7) (8) -0.0002 (0.0002) -0.0864*** (0.0091) -0.0927*** (0.0093) 0.0008*** (0.0002) -0.1020*** (0.0081) -0.0719*** (0.0091) 0.0002** (0.0001) -0.0362** (0.0141) -0.0789*** (0.0254) 0.0004* (0.0002) -0.0359*** (0.0122) -0.0787*** (0.0255) -0.0002 (0.0002) Workers' Barg. Power x Stock Market Value -0.0004*** (0.0001) Empl. Protection Leg. x Stock Market Value Trade Openness Unemployment GDP Growth (log) R&D Expenditures Constant 0.0013*** (0.0003) 0.0090*** (0.0021) 0.0216 (0.0396) 0.0846*** (0.0119) 0.9603*** (0.1135) 0.0013*** (0.0003) 0.0081*** (0.0021) 0.0035 (0.0360) 0.0786*** (0.0123) 0.9486*** (0.1180) -0.0025*** (0.0009) -0.0038* (0.0020) 0.0047 (0.0183) 0.0042 (0.0279) -0.0016* (0.0009) -0.0045** (0.0021) -0.0050 (0.0183) -0.0052 (0.0171) 0.0013*** (0.0003) 0.0096*** (0.0024) 0.0177 (0.0374) 0.0787*** (0.0092) 0.9439*** (0.1061) -0.0000 (0.0001) -0.0005*** (0.0002) 0.0013*** (0.0003) 0.0089*** (0.0022) -0.0015 (0.0328) 0.0693*** (0.0099) 0.9619*** (0.1015) -0.0009 (0.0007) -0.0050* (0.0026) 0.0045 (0.0141) -0.0071 (0.0237) -0.0001 (0.0002) -0.0007 (0.0009) -0.0057* (0.0031) 0.0052 (0.0146) -0.0065 (0.0196) Observations 250 250 183 193 250 250 237 237 R-squared 0.7010 0.7004 0.7148 0.7113 Country and year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Estimator PCSE/OLS PCSE/OLS IV-GMM IV-GMM PCSE/OLS PCSE/OLS IV-GMM IV-GMM Hansen J-statistic (p-value) 0.1194 0.1006 0.3028 0.3103 Kleibergen-Paap LM χ²-statistic (p-value) 0.0024 0.0011 0.0000 0.0002 Anderson-Rubin χ² test (p-value) 0.0000 0.0000 0.0000 0.0000 Anderson-Rubin F-test (p-value) 0.0000 0.0000 0.0000 0.0000 Kleibergen-Paap LM F-statistic 5.127 5.215 9.922 4.714 Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 Endogenous variables: Stock Market Capitalization Ratio, Stock Market Traded Value, Workers’ Bargaining Power, Employment Protection Legislation, Workers' Barg. Power x Stock Market Cap., Empl. Protection Leg. x Stock Market Cap., Workers' Barg. Power x Stock Market Value and Empl. Protection Leg. x Stock Market Value Instruments: EPL (lag), Workers’ Bargaining Power (lag), labor compensation, replacement rate of unemployment benefit, economic regulation, capital account openness, financial liberalization index, national minimum wage, presidential political system, voting system STATA compute four different validity tests: (1) overidentifying restrictions test (Sargan-Hansen-J statistic): the null hypothesis of this first test is that the instruments are valid instruments, i.e., uncorrelated with the error term, and that the excluded instruments are correctly excluded from the estimated equation. If the null is not rejected, the overidentification restrictions are valid; (2) Underidentification test (Kleibergen-Paap LM χ²-statistic): the underidentification test is relevant to verify whether the equation is quite identified: in this case, instruments are relevant in the sense that they are correlated with assumed endogenous regressions. The null hypothesis cannot be rejected; (3) Endogenous regressors tests (Anderson-Rubin χ² test and Anderson-Rubin F-test): the Anderson-Rubin tests are tests of the significance of endogenous regressors in the structural estimated equation: the null hypothesis tested is that the coefficients of the endogenous regressors in the structural equation are jointly equal to zero; (4) Weakidentification test (Kleibergen-Paap F-statistic or Cragg-Donald): the Kleibergen-Paap (or Cragg-Donald) test of the weak-instruments problem that arises when the correlations between the endogenous regressors and the excluded instruments are nonzero but small. The null hypothesis is that the estimator is weakly identified in the sense that it is subject to bias that the investigator finds unacceptably large. See Baum et al. (2007) for more details on testing the relevance and validity of instruments. 19 Table 2. Marginal effects (overall wage inequality) A. Marginal effects of financial development conditional to labor market institutions Employ. Protection Leg. Workers’ Barg. Power Stock Market Capitalization (1) (3) Min Mean_less_1sd Mean Mean_plus_1sd Max Min Mean_less_1sd Mean Mean_plus_1sd Max 0.0005 (0.0004) 0.0001 (0.0003) -0.0004 (0.0003) -0.0009*** (0.0003) -0.0013*** (0.0004) 0.0006** (0.0003) 0.0005*** (0.0002) 0.0004*** (0.0001) 0.0003 (0.0002) 0.0002 (0.0003) Stock Market Total Value Traded (5) (7) 0.0007*** (0.0002) 0.0003 (0.0002) -0.0002 (0.0002) -0.0006** (0.0003) -0.0011*** (0.0003) 0.0003*** (0.0001) 0.0003*** (0.0001) 0.0002 (0.0002) 0.0002 (0.0003) 0.0002 (0.0003) (2) (4) (6) (8) 0.0005 (0.0003) 0.0000 (0.0003) -0.0005* (0.0003) -0.0011** (0.0005) -0.0017*** (0.0006) 0.0007** (0.0003) 0.0005*** (0.0002) 0.0003* (0.0002) 0.0001 (0.0003) -0.0001 (0.0005) 0.0007*** (0.0002) 0.0003 (0.0002) -0.0002 (0.0002) -0.0008** (0.0004) -0.0012** (0.0005) 0.0003** (0.0001) 0.0003*** (0.0001) 0.0002 (0.0002) 0.0001 (0.0004) -0.0000 (0.0006) B. Marginal effects of labor market institutions conditional to the degree of financial development Workers’ Bargaining power (1) (3) Stock Market Cap. Min Mean_less_1sd Mean Mean_plus_1sd Max -0.0827*** (0.0141) -0.0908*** (0.0123) -0.1054*** (0.0097) -0.1199*** (0.0090) -0.1721*** -0.0136 (0.0145) -0.01500 (0.0127) -0.0176* (0.0101) -0.0218** (0.0093) -0.0294 (5) (7) Stock Market Value Min -0.0846*** -0.0362*** (0.0089) (0.0140) Mean_less_1sd -0.0828*** -0.0361** (0.0091) (0.0143) Mean -0.1049*** -0.0373*** (0.0067) (0.0117) Mean_plus_1sd -0.1270*** -0.0384*** (0.0068) (0.0103) Max -0.2286*** -0.0436** (0.0239) (0.0202) Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 20 Employment Protection Legislation (2) (4) -0.0682*** (0.0122) -0.0806*** (0.0101) -0.1032*** (0.1000) -0.1257*** (0.0144) -0.2064*** -0.0845*** (0.02635) -0.0893*** (0.0255) -0.0980*** (0.0263) -0.1067*** (0.0298) -0.1376*** (6) (8) -0.0728*** (0.0086) -0.0701*** (0.0088) -0.1022*** (0.0100) -0.1342*** (0.0171) -0.2813*** (0.0580) -0.0788*** (0.0254) -0.0783*** (0.0226) -0.0843*** (0.0267) -0.0903*** (0.0332) -0.1178 (0.0849) Table 3. The impact of financial development on lower-tail inequality Dependent variable : p5/p1 ratio Stock Market Capitalization Ratio (1) (2) (3) (4) 0.0000 (0.0001) 0.0015*** (0.0003) -0.0004 (0.0003) 0.0008** (0.0004) Stock Market Total Value Traded Workers' Bargaining Power Employment Protection Legislation (EPL) Workers' Barg. Power x Stock Market Cap. -0.0125 (0.0090) -0.0480*** (0.0039) -0.0004*** (0.0001) Empl. Protection Leg. X Stock Market Cap. -0.0394*** (0.0028) -0.0002 (0.0116) 0.0345** (0.0145) -0.0067 (0.0190) -0.0003** (0.0002) -0.0008*** (0.0001) 0.0167** (0.0071) 0.0053 (0.0114) (5) (6) (7) (8) 0.0001 (0.0001) -0.0369*** (0.0040) -0.0511*** (0.0039) 0.0006*** (0.0001) -0.0456*** (0.0023) -0.0407*** (0.0052) -0.0006* (0.0004) 0.0235** (0.0111) -0.0133 (0.0188) 0.0003 (0.0002) 0.0250*** (0.0067) 0.0028 (0.0183) -0.0007*** (0.0002) Workers' Barg. Power x Stock Market Value -0.0002*** (0.0000) Empl. Protection Leg. X Stock Market Value Trade Openness Unemployment GDP Growth (log) 0.0005*** (0.0001) 0.0017** (0.0008) 0.0563* (0.0295) 0.0005*** (0.0001) 0.0013* (0.0008) 0.0379 (0.0294) R&D Expenditures Constant 0.0006 (0.0014) -0.0048*** (0.0012) 0.0118 (0.0081) 0.0696** (0.0306) 0.0006 (0.0005) -0.0061*** (0.0014) 0.0167 (0.0109) 0.0422*** (0.0111) 0.0006*** (0.0001) 0.0016* (0.0010) 0.0450 (0.0298) -0.0004** (0.0002) -0.0003*** (0.0001) 0.0006*** (0.0001) 0.0016* (0.0009) 0.0368 (0.0300) 0.0016 (0.0013) -0.0058*** (0.0012) 0.0051 (0.0088) 0.0511** (0.0246) -0.0003*** (0.0001) 0.0019* (0.0011) -0.0055*** (0.0009) -0.0011 (0.0085) 0.0328* (0.0178) 0.3921*** 0.3416*** 0.4215*** 0.4182*** (0.0753) (0.0793) (0.0773) (0.0795) Observations 260 260 241 241 260 260 241 241 R-squared 0.6225 0.6305 0.6159 0.6136 Country and year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Estimator PCSE/OLS PCSE/OLS IV-GMM IV-GMM PCSE/OLS PCSE/OLS IV-GMM IV-GMM Hansen J-statistic (p-value) 0.1255 0.1639 0.1632 0.2232 Kleibergen-Paap LM ²-statistic (p-value) 0.0008 0.0016 0.0011 0.0046 Anderson-Rubin c² test (p-value) 0.0001 0.0000 0.0013 0.0000 Anderson-Rubin F-test (p-value) 0.0037 0.0001 0.0000 0.0000 Kleibergen-Paap LM F-statistic 2.503 2.911 1.899 5.041 Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 Endogenous variables: Stock Market Capitalization Ratio, Stock Market Traded Value, Workers’ Bargaining Power, Employment Protection Legislation, Workers' Barg. Power x Stock Market Cap., Empl. Protection Leg. x Stock Market Cap., Workers' Barg. Power x Stock Market Value and Empl. Protection Leg. x Stock Market Value Instruments: EPL (lag), Workers’ Bargaining Power (lag), labor compensation, replacement rate of unemployment benefit, economic regulation, capital account openness, financial liberalization index, national minimum wage, presidential political system, voting system 21 Table 4. Marginal effects (lower-tail wage inequality) A. Marginal effects of financial development conditional to labor market institutions Employ. Protection Leg. Workers’ Barg. Power Stock Market Capitalization (1) (3) Min Mean_less_1sd Mean Mean_plus_1sd Max Min Mean_less_1sd Mean Mean_plus_1sd Max Stock Market Total Value Traded (5) (7) 0.0011*** (0.0002) 0.0006*** (0.0001) 0.0000 (0.0001) -0.0005*** (0.0001) -0.0011*** (0.0002) (2) 0.0004* (0.0002) 0.0000 (0.0002) -0.0003 (0.0003) -0.0007 (0.0004) -0.0011* (0.0006) (4) 0.0005*** (0.0001) 0.0003*** (0.0001) 0.0001 (0.0001) -0.0001 (0.0001) -0.0003*** (0.0001) (6) 0.0003 (0.0002) -0.0001 (0.0002) -0.0006* (0.0003) -0.0011* (0.0006) -0.0015** (0.0008) (8) 0.0013*** (0.0002) 0.0007*** (0.0001) -0.0001* (0.0001) -0.0010*** (0.0002) -0.0017*** (0.0003) 0.0007* (0.0004) 0.0001 (0.0002) -0.0006*** (0.0002) -0.0013*** (0.0004) -0.0019*** (0.0006) 0.0005*** (0.0001) 0.0003*** (0.0001) 0.0001 (0.0001) -0.0002* (0.0001) -0.0004*** (0.0001) 0.0024 (0.0002) -0.0000 (0.0002) -0.0003 (0.0002) -0.0006** (0.0003) -0.0009** (0.0004) B. Marginal effects of labor market institutions conditional to the degree of financial development Workers’ Bargaining power (1) (3) Stock Market Cap. Min Mean_less_1sd Mean Mean_plus_1sd Max -0.0102 (0.0077) -0.0201*** (0.0059) -0.0381*** (0.0029) -0.0561*** (0.0025) -0.1204*** (5) 0.0327** (0.0139) 0.0261** (0.0103) 0.0140 (0.0103) 0.0020 (0.0119) -0.0412 (7) Stock Market Value Min -0.0354*** 0.0229** (0.0035) (0.0109) Mean_less_1sd -0.0345*** 0.0248** (0.0036) (0.0114) Mean -0.0454*** 0.0019 (0.0020) (0.0112) Mean_plus_1sd -0.0563*** -0.0209 (0.0022) (0.0191) Max -0.1064*** -0.1260* (0.0116) (0.0677) Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 22 Employment Protection Legislation (2) (4) -0.0049 (0.0108) -0.0224*** (0.0081) -0.0544*** (0.0041) -0.0863*** (0.0050) -0.2006*** (6) 0.0013 (0.0108) -0.0144 (0.0101) -0.0429*** (0.0149) -0.0713*** (0.0230) -0.1732*** (8) -0.0404*** (0.0051) -0.0391*** (0.0053) -0.0549*** (0.0036) -0.0706*** (0.0045) -0.1431*** (0.0187) 0.0022 (0.0182) 0.0039 (0.0183) -0.0157 (0.0190) -0.0354 (0.0223) -0.1255** (0.0506) Table 5. The impact of financial development on upper-tail inequality Dependent variable : Top 10% Income Share Stock Market Capitalization Ratio (1) (2) (3) (4) 0.0158*** (0.0050) 0.0650*** (0.0068) 0.0582*** (0.0127) 0.1060*** (0.0207) Stock Market Total Value Traded Workers' Bargaining Power Employment Protection Legislation (EPL) Workers' Barg. Power x Stock Market Cap. -2.1798*** (0.3398) -0.8554*** (0.1455) -0.0039 (0.0029) Empl. Protection Leg. X Stock Market Cap. -2.2347*** (0.1881) 0.8067** (0.3690) 0.5696 (0.5058) -0.9487*** (0.3583) -0.0133** (0.0052) -0.0290*** (0.0041) -0.2937 (0.3978) -0.7046 (0.4384) (5) (6) (7) (8) 0.0223*** (0.0043) -2.4185*** (0.2523) -0.8973*** (0.1102) 0.0299*** (0.0048) -2.4596*** (0.1930) -0.6619*** (0.0774) 0.0304*** (0.0056) -0.6993** (0.3345) -0.7647** (0.3491) 0.0424*** (0.0095) -0.7427** (0.3267) -0.7781** (0.3257) -0.0187* (0.0097) Workers' Barg. Power x Stock Market Value -0.0018 (0.0017) Empl. Protection Leg. X Stock Market Value Trade Openness Unemployment GDP Growth (log) 0.0035 (0.0059) -0.0092 (0.0551) 1.4809 (1.5733) 0.0108 (0.0066) 0.0564 (0.0554) 1.2061 (1.2622) 26.5754*** (3.9734) 23.0667*** (3.5618) R&D Expenditures Constant -0.0778*** (0.0232) 0.0544 (0.0590) 0.1168 (0.6862) 0.5503 (1.0173) -0.0794*** (0.0265) 0.1171* (0.0657) -0.6957 (0.7572) -1.4790* (0.7555) -0.0035 (0.0025) 0.0134* (0.0069) 0.0364 (0.0548) 1.1889 (1.6769) -0.0055*** (0.0015) 0.0137** (0.0068) 0.0334 (0.0513) 1.2811 (1.5665) 26.8915*** (4.0507) 26.1690*** (3.9382) -0.0807*** (0.0240) -0.0988 (0.0738) 2.3110*** (0.5991) -0.0982 (0.8150) -0.0079 (0.0073) -0.0636** (0.0295) -0.0910 (0.0680) 2.2516*** (0.5988) -0.6104 (0.6380) Observations 247 247 200 200 247 247 200 200 R-squared 0.5979 0.6276 0.6224 0.6253 Country and year fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Estimator PCSE/OLS PCSE/OLS IV-GMM IV-GMM PCSE/OLS PCSE/OLS IV-GMM IV-GMM Hansen J-statistic (p-value) 0.5343 0.5282 0.1375 0.1047 Kleibergen-Paap LM ²-statistic (p-value) 0.0027 0.0002 0.0000 0.0004 Anderson-Rubin ² test (p-value) 0.0000 0.0000 0.0000 0.0000 Anderson-Rubin F-test (p-value) 0.0000 0.0000 0.0000 0.0000 Kleibergen-Paap LM F-statistic 2.140 1.205 9.490 3.299 Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 Endogenous variables: Stock Market Capitalization Ratio, Stock Market Traded Value, Workers’ Bargaining Power, Employment Protection Legislation, Workers' Barg. Power x Stock Market Cap., Empl. Protection Leg. x Stock Market Cap., Workers' Barg. Power x Stock Market Value and Empl. Protection Leg. x Stock Market Value Instruments: EPL (lag), Workers’ Bargaining Power (lag), labor compensation, replacement rate of unemployment benefit, economic regulation, capital account openness, financial liberalization index, national minimum wage, presidential political system, voting system 23 Table 6. Marginal effects (upper-tail wage inequality) A. Marginal effects of financial development conditional to labor market institutions Employ. Protection Leg. Workers’ Barg. Power Stock Market Capitalization (1) (3) Min Mean_less_1sd Mean Mean_plus_1sd Max Min Mean_less_1sd Mean Mean_plus_1sd Max 0.0253*** (0.0082) 0.0208*** (0.0058) 0.0157*** (0.0049) 0.0107 (0.0066) 0.0059 (0.0094) 0.0909*** (0.0141) 0.0754*** (0.0119) 0.0581*** (0.0127) 0.0410** (0.0165) 0.0245 (0.0215) Stock Market Total Value Traded (5) (7) 0.0266*** (0.0047) 0.0246*** (0.0040) 0.0223*** (0.0043) 0.0200*** (0.0056) 0.0178** (0.0072) 0.0390*** (0.0065) 0.0349*** (0.0052) 0.0304*** (0.0056) 0.0259*** (0.0075) 0.0216** (0.0100) (2) (4) (6) (8) 0.0589*** (0.0061) 0.0365*** (0.0042) 0.0084* (0.0049) -0.0197*** (0.0073) -0.0456*** (0.0106) 0.1021*** (0.0191) 0.0877*** (0.0143) 0.0695*** (0.0128) 0.0514*** (0.0173) 0.0347 (0.0241) 0.0277*** (0.0043) 0.027*** (0.0037) 0.0186*** (0.0035) 0.0135*** (0.0040) 0.0087* (0.0049) 0.0408*** (0.0082) 0.0347*** (0.0053) 0.0270*** (0.0082) 0.0194 (0.0144) 0.0123 (0.0206) B. Marginal effects of labor market institutions conditional to the degree of financial development Workers’ Bargaining power (1) (3) Stock Market Cap. Min Mean_less_1sd Mean Mean_plus_1sd Max -2.1148*** (0.3378) -2.2027*** (0.2903) -2.3619*** (0.2274) -2.5211*** (0.2164) -3.0907*** 0.4959 (0.04840) 0.2082 (0.4071) -0.3124 (0.3248) -0.8330** (0.3567) -2.6961*** (5) (7) Stock Market Value Min -2.3765*** -0.7051** (0.2616) (0.3328) Mean_less_1sd -2.3678*** -0.6872** (0.2687) (0.3382) Mean -2.4705*** -0.8997*** (0.1958) (0.3048) Mean_plus_1sd -2.5731*** -1.112*** (0.1642) (0.3439) Max -3.044*** -2.0876** (0.5134) (0.9146) Note: (Panel Corrected) Standard errors in parentheses; *** p < 0.01, ** p < 0.05, * p < 0.1 24 Employment Protection Legislation (2) (4) 0.6123** (0.3097) 0.0090 (0.2347) -1.0919*** (0.1384) -2.1928*** (0.1794) -6.1323*** -0.8083* (0.4220) -1.2131*** (0.4228) -1.9456*** (0.6372) -2.6782*** (0.9613) -5.2996** (6) (8) -0.6627*** (0.0798) -0.6355*** (0.0803) -0.9587*** (0.1214) -1.2818*** (0.2080) -2.7656*** (0.6586) -0.7913** (0.3261) -0.7505** (0.3263) -1.2344** (0.5368) -1.7118* (0.9329) -3.9399 (2.9473) Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands Norway New Zealand Sweden United Kingdom United States 2 5 4 p9/p1 ratio 3 2 90 00 10 80 90 00 10 80 90 00 10 2 3 4 5 80 80 90 00 10 80 90 00 10 year Graphs by country Australia Austria Belgium Canada Denmark Finland France Germany Ireland Italy Japan Netherlands New Zealand Norway Spain Sweden United Kingdom United States 0 100200300400 0 100200300400 0 100200300400 Figure 1: p5 /p1 ratio in 17 OECD countries 100200300400 1980 1990 2000 2010 1980 0 d9/d1 ratio 3 4 5 2 3 4 5 Australia 1980 1990 2000 2010 1980 1990 2000 2010 1980 1990 2000 2010 year stmktcap stvaltraded Data source: 'Financial Structures dataset' from Beck et al., 2012 Figure 2: p5 /p1 ratio in 17 OECD countries 25 1990 2000 2010