Project Topics
Get Complete Project Material File(s) Now! »
The identication strategy
From (1.6) we nd that the impact of EPLc varies according to the natural level of reallocation, i.e. the level in the absence of adjustment costs, of each sector, JF. The regulation is legally the same for all rms and sectors as it is established at national level, but its impact may be dierent because of technology characteristics or type of knowledge management required or dierent dynamics of aggregate shocks.
We use this dierence, to implement a dierence-in-dierence approach following the seminal paper by Rajan and Zingales (1998) and exploiting more the within-country dimension since not all industries are constrained in the same way by EPL as we have shown in the simple model. Bassanini et al. (2009) speak about EPL-binding industries to dene those industries that have a relatively high natural propensity to adjust their human resources through layos. Micco and Pages (2006) and Haltiwanger et al. (2014) comparing job turnover show that there is little cross-country 9This is valid insofar Cjc is not a completely dierent function of JFjc or EPLc, but there are no reasons to think so. variability in the ranking of industries with reference to their propensity to adjust on the external labor market, suggesting that some common features among industries hape this propensity. A dierence-in-dierence approach between countries and EPL-binding and EPL-non binding industries can help to minimize the possible endogeneity and omitted variable problems. A di-in-di approach allows controlling for all omitted policies that, either by aecting labor supply only or aecting homogenously labor demand in all industries, are unlikely to have a dierential eect on dierent industries. We estimate equation 1.6 by: WFjc = Xjc + Bj EPLc + Dj + Dc + jc (1.7).
where WFjc is a rate of worker reallocation (total worker reallocation, excess worker reallocation, etc. see next section) in industry j and country c, Bj is WF j in equation interest and gives the sign of the demand eect.
The identication assumption is twofold. On the one hand, the eect of EPL on worker reallocation is proportional to the natural propensity of each industry to adjust on the external labor market where their propensity is measured by estimated worker reallocation rate of each industry in the absence of EPL. On the other hand it is assumed that the impact of other policies and institutions is not proportional to theestimated worker reallocation rate, which is a priori plausible for most of the policies.
Including country- and sector-specic eects Dc and Dj , therefore, the eect of these other policies can be easily controlled for. In addition, for a number of policies, we control directly for their eect in a sensitivity analysis, by allowing their impact to interact with the natural propensity of rms to adjust on the external market.
Breaking down by workers’ groups
It is also quite unlikely that the direct partial-equilibrium impact of institutions on gross ows be the same for all demographic or skill group. In the case of the main gross ow measures, available data can be disaggregated further by gender, age classes and educational attainment for a number of countries, even if at the price of greater measurement error. The analysis of Table 1.6 can therefore be replicated by controlling more directly for these characteristics through a series of dummies and by checking cross-group dierences in the impact of those institutions that appear to be signicant in Table 1.6. Average estimated eects of EP in Table 1.14 remain consistent with those reported in Table 1.6. Reallocation patterns concerning highskilled workers appear to be somewhat less robustly aected by EP (Table 1.14).
This result might reect the fact that, in all countries, expanding industries tend to have a large, growing demand for skilled labour (see Bassanini and Marianna, 2009) and suggest that EP provisions have a smaller eect on these industries. Similarly, it appears that stringent regulations have a particularly depressing impact on gross worker ows involving youth and, to a more limited extent, women. These ndings appear consistent with the idea that EP for regular workers has a particularly negative eect on the job perspectives of outsiders, a group in which women and youth tend to be over-represented (see e.g. OECD, 2006).
Dening minimum wages
One of the key propositions of this second part of the thesis is that the concept of a minimum wage not only refers to statutory wage oors dened at the national level, but also extends to minimum wages that are dened at the sectoral or occupational level. It is unquestionably true that the national statutory minimum wage has received much more attention in the literature in disciplines such as labour economics or industrial relations, to such an extent that other types of wage oor are hardly ever analysed. Indeed, instead of examining the impact that the dierence between nationally and sectorally dened minimum wages might have on a range of labour market outcomes, the literature focuses almost entirely on data collected within the group of countries with national statutory minimum wages. This does not mean that no aspect of the process through which minima are determined has ever been scrutinised: Boeri (2012) for instance, has shown that national wage oors that are legislated unilaterally by the government are typically lower than those settled with closer involvement of the social partners.
While the process during which national minima are set seems therefore to lead to signicantly dierent outcomes in terms of the relative level of wage oors, most scholars have overlooked the much more fundamental issue of whether the minimum is dened for the entire labour force (in the case of a national statutory minimum wage with no exemptions) or only a part of it (for instance, by a sectoral minimum that binds only trade union members). This is particularly problematic because the impact of the dierence between national and sectoral minima on a range of labour market outcomes is potentially much more important, for instance when it comes to their respective inuence on the level of wage oors (an issue discussed by Grimshaw and Bosch, 2013) and the number of workers paid at or below it, but also the impact on more general issues, such as overall or inter-sectoral inequality. In this chapter, we show that the distinction concerning whether wage oors are set at the national or at the sectoral (occupational) level gives rise to a series of hypotheses regarding their impact on key labour market outcomes. The issues addressed by our hypotheses are admittedly not new: in fact, they correspond to the intuitive understanding of the wage setting process revealed by anecdotal evidence from conversations and published statements by trade unionists, employer representatives and policymakers in the context of the debate on a European minimum wage (see Introduction). For instance, many practitioners are convinced that trade unions are able to obtain higher minimum wages if they negotiate at the sectoral than at the national level. This intuition, however, may or may not be true and so far lacks any empirical proof that we are aware of, the sample used by Grimshaw and Bosch (2013) being too small to allow for econometrically sound conclusions. To be sure, the opposite relationship is also plausible: some trade unions may have lower bargaining power when they negotiate at the sectoral level, so that at least in certain parts of the economy the collectively negotiated wage oors might be lower compared to a situation in which the minimum is determined through a negotiation at the national level. By compiling a representative sample of sector- and national-level minima from dierent types of minimum wage systems, Chapter 3 of this thesis is the rst to be able to shed empirical light on these issues.4
The diversity of minimum wage systems in Europe
The literatures in industrial relations and labour economics provide rich traditions in the analysis of minimum wages. The former school tends to frame wage oors as institutions that can be analysed as the outcome of interactions between a set of actors (mainly the state, employer representatives and trade unions) and other institutions (wage setting processes, laws and so on). Parts of this literature also take into account the social norms and beliefs that the involved actors have about minimum wages and their impact on dierent labour market outcomes. By contrast, labour economics typically treats minimum wages as market imperfections that give rise to a deviation from outcomes that would come about in the absence of articial wage oors or ceilings. One of the basic predictions of these models is that a binding minimum wage fails to clear the labour market and thereby creates a situation in which more workers would be willing to work (and fewer employers willing to hire) compared to a situation without minimum wages (see our discussion on employment eects).
The approach in this second part of the thesis is closer to the tradition in industrial relations in that we distinguish between dierent institutional features of minimum wage systems and analyse their relationships with key labour market outcomes. While it is a priori also possible to formulate hypotheses on these relationships with the help of labour market models that treat minimum wages as market imperfections, we show that the observed outcomes can be accounted for by the institutional diversity between minimum wage systems.
The institutionalist literature on minimum wages so far has focused almost exclusively on systems in which wage oors are dened by national statutory minimum wage legislations. The focus of this literature has therefore been to account for different outcomes (such as the relative level of statutory minimum wages) by looking at dierences between statutory systems, either across countries or across time. Existing empirical evidence notably shows that in countries with statutory minimum wages a higher level of collective bargaining coverage is associated with relatively higher levels of minimum wages (Grimshaw and Bosch, 2013). This might be due to the fact that countries with higher collective bargaining coverage tend to have more egalitarian wage structures in which the median lies closer to the minimum wage, leading in turn to a higher Kaitz Index. Another explanation is that a higher level of collective bargaining coverage is associated with stronger trade-union inuence on the level of the statutory minimum wage (for example, in negotiations by tripartite commissions). On any account, the level of statutory minimum wages tends to be positively related to the degree to which trade unions are involved in the setting process at the national level (Boeri, 2012; Eyraud and Saget, 2005; Funk and Lesch, 2005).
Table of contents :
I Employment Protection Legislation
1 Dismissal Protection and Worker Flows in OECD Countries
1.1 Introduction
1.2 Theoretical and empirical framework
1.2.1 A simple theoretical framework
1.2.2 Adjustment costs
1.2.3 The identication strategy
1.3 The Data
1.4 Empirical Results
1.4.1 Cross-sectional results
1.4.2 Time-series results
1.4.3 Checking at individual level
1.5 Conclusions
II Minimum Wages in Europe
2 A Variety of Minimum Wages in Europe
2.1 Introduction
2.2 Dening minimum wages
2.3 The diversity of minimum wage systems in Europe
2.3.1 The sample of countries
2.3.2 Data sources
2.4 An overview of the EU variety
2.5 Comparison with other datasets
3 Institutional Diversity and the Sector-Level Minimum Wage Bite
3.1 Introduction
3.2 The minimum wage bite
3.2.1 The Kaitz index
3.2.2 The employment spike
3.2.3 The share of individuals below the minimum wage
3.3 Institutional diversity and the minimum wage bite
3.4 Empirical analysis
3.4.1 The minimum wage bite I: Kaitz index
3.4.2 The minimum wage bite II: employment spike
3.4.3 The minimum wage bite III: share of subminimum workers .
3.4.4 Robustness tests
3.5 Conclusions
4 Institutional Diversity and Earnings Inequalities
4.1 Introduction
4.2 Review of the literature
4.3 Analytical framework
4.4 Empirical analysis
4.4.1 Overall wage inequality
4.4.2 Inter-industry wage inequality
4.4.3 Share of workers paid below prevailing minima
4.4.4 Robustness tests
4.5 Conclusions
III Age, Gender and Education Diversity
5 The Heterogeneous Eects of Workforce Diversity in Belgium
5.1 Introduction
5.2 Review of the literature
5.2.1 Workforce diversity and rm productivity
5.2.2 Previous empirical studies
5.3 The theoretical framework and the estimation strategy
5.3.1 The theoretical framework
5.3.2 The measurement of diversity
5.3.3 The estimation method
5.4 Data and descriptive statistics
5.5 Empirical results
5.5.1 Benchmark specication
5.5.2 Does the technological/knowledge environment matter?
5.5.3 Does the diversity eect vary according to rm size?
5.5.4 Interdependencies between diversity dimensions
5.6 Discussion and conclusion
6 Workforce Diversity, Managers and Shareholders in France
6.1 Introduction
6.2 Empirical framework
6.3 Data and descriptive statistics
6.3.1 Managers’ and shareholders’ characteristics
6.4 Results
6.4.1 Baseline results
6.4.2 IV estimations
6.4.3 The role of managers
6.4.4 The proprietary structure of the rm
6.5 Discussion and conclusions
Conclusions
Bibliography
