Equity” William R. Kerrz

We examine how differences in labor regulations across European countries influence the development of private equity (PE) markets, comprised of venture capital and buy-out investors.

Recent theoretical models predict that countries with stricter labor policies will specialize in less innovative activities due to the higher worker turnover frequently associated with rapidly changing sectors (e.g., Saint-Paul 1997, 2002a; Samaniego 2006). We provide the first empirical evidence for this prediction at the industry level in the entrepreneurial finance literature. In the process, we also make a methodological contribution by demonstrating how jointly modeling the different policies for providing worker insurance delivers more consistent results than their individual relationships. Our techniques may find application in other settings, too.

We first observe that European countries empirically substitute between employment protection regulations (EPRs) and labor market expenditures (LMEs) in the provision of worker insurance. Figure 1 shows the cross-sectional relationship for 1998. The vertical axis documents the average LMEs per capita taken from the OECD Social Expenditures database. LMEs include both active and passive policies designed to facilitate job creation and transitions, with the majority of expenditures being unemployment insurance benefits. The horizontal axis provides an EPR index developed by the OECD. Higher EPR scores indicate more heavily regulated labor markets, factoring in a wide variety of legislation concerning the individual and collective dismissals of both temporary and regular workers.

This plot illustrates two important features. First, Anglo-Saxon countries provide lower worker insurance on both dimensions than Continental Europe. These differences in absolute levels of worker insurance provided by nations have been a frequent political-economy topic since at least de Tocqueville (e.g., Alexia et al. 2001, Kerr 2007). Second, the trend line, which is calculated only for Continental European nations, indicates that economies with higher LMEs per capita have weaker EPRs. These differences in the mechanisms used to provision worker insurance among Continental European nations have received less attention. Denmark provides the highest LMEs per capita but has the second-lowest employment protection in Continental Europe. This reflects the well-publicized Danish ‘flexicurity’ approach that emphasizes high job mobility facilitated by generous out-of-work benefits and active labor market programs to promote worker re-entry. Portugal, on the other hand, provides strong security to the employed but weaker benefits to the unemployed.

While employment protection and transition/re-entry assistance are perhaps substitutes for providing worker security, they have different implications for the costs firms face. Labor rigidities have a stronger impact on the adjustment margins of firms, especially those undertaking substantial restructurings. Even if general corporate or payroll taxation is higher to support LMEs, the direct incidence on the labor adjustments that firms wish to make is weaker in regimes favoring LMEs than in strict employment protection regimes. These taxes on labor adjustments are particularly pertinent for PE investments, which thrive in dynamic industries that require frequent labor adjustments. This PE focus on high-growth opportunities and rapid restructuring is necessary for achieving sufficient returns when portfolio companies offer the potential for exceptional investment returns but also carry a high risk of failure.

Combining these observations, nations emphasizing LMEs over employment protection should be more attractive for the development of PE financing, even after conditioning on the level of worker insurance provided. While labor market regulations do not specifically target the port-

folio companies of PE investors, these investors are seeking opportunities that are generally more sensitive to these taxes on labor adjustment. We investigate this hypothesis using PE surveys provided by the European Private Equity and Venture Capital Association and Venture Economics. Figures 2 and 3 show that policy choices are correlated with PE placement (trend lines are still for Continental Europe). European countries with stricter employment protection have lower PE investments per capita, while those favoring LMEs are more attractive to these financial forms.

While these correlations are suggestive, many other factors vary across countries besides labor market policies, and it is quite likely that omitted factors correlated with labor policies are important for PE formation. Labor market policies tend to evolve slowly in most countries, limiting the scope of panel estimation techniques at the country level for disentangling these effects. We thus test these predictions using a differences-in-differences approach similar to Rajan and Zingales (1998) that employs country-sector variation in PE market size over the 1990-2004 period. We specifically model whether countries that favor LMEs over employment protection for providing worker insurance develop relatively stronger PE markets in more volatile sectors. We calculate the volatility of sectors using US establishment-level data from the Census Bureau, which we take to be the unconstrained case.

Regression estimates find that the interaction of sector volatility and employment protection has a negative effect on PE formation, while the opposite is true for LMEs. While suggestive of labor regulations having an important bite, the coefficients are sometimes of borderline economic and statistical importance. As a methodological contribution, we show that the coefficients on the base policies are less informative than their joint effect. Studies that evaluate the impact of labor rigidities modeling one policy only will typically understate the impact of worker insurance policy choices for economic outcomes like PE investments.

This concept relates back to the policy decisions illustrated in Figure 1. The individual policies are simultaneously capturing both the level of labor market insurance provided and the mechanism used to provision the insurance. An empirical evaluation of an increase in employment protection will encompass both increases in insurance levels (e.g., Anglo-Saxon versus Continental Europe) and changes in policy mechanisms (e.g., Denmark versus Portugal). These two objects are distinct from a policy perspective, however, and it is important to distinguish their individual effects as much as possible. Indeed, the simple trend lines in Figures 1-3 can look quite different when the Anglo-Saxon economies are included in the calculation. Throughout this study, we assess the impact of adjusting worker insurance policies while keeping the overall level of insurance provided by a country constant. Substantial changes in insurance levels provided by countries are quite rare, but policy makers frequently contemplate moving towards or away from flexible labor markets with concomitant adjustments in other insurance programs (e.g., the recent interest in the Danish model).

We show two techniques to isolate the mechanism of worker insurance provision from the overall insurance level. One approach is particularly simple, just taking the linear difference of two policy coefficients after a multivariate regression. A second approach transforms the base policies into more intuitive indices. Both approaches find that policy mechanisms are robustly important for PE investment patterns, while the overall level of labor insurance provided is of much weaker importance. This is true on both the extensive margin (i.e., whether PE investments form at all in a country-sector) and the intensive margin (i.e., the volume of deals in the country-sector). The effects are particularly strong for US-sourced venture capital investments, and we show the sector-level patterns are generally robust to other policy characteristics and traits of countries.

Приложение