1. Introduction
Unemployment has been a pressing problem in Europe for decades, and the high share of long-
term unemployment is a major concern. For example, according to OECD (2005), more than
40 percent of all European unemployment is classified as long-term with a duration of a year
and over. However, unemployment is far from uniform in Europe. There is indeed considerable
dispersion of both unemployment rates and the prevalence of long-term unemployment within
European countries. This variation across European countries, as well as the fact that the share
of long-term unemployment is much lower in the U.S. (12 percent) suggest that differences in
labor market institutions lie at the heart of the European unemployment problem (e.g. Siebert
1997).
Several empirical studies have emphasized that long-term shifts of OECD unemployment can
be best explained by changes in labor market institutions (rather than by macroeconomic shocks
or shifts of macroeconomic time-trends). See, in particular, Scarpettta (1996), Nickell (1997),
Elmeskov et al. (1998), Belot and van Ours (2001, 2004) and Nickell et al. (2005). The gen-
erosity and the duration of eligibility for unemployment benefits as well as labor taxes have
been identified as particularly strong predictors of unemployment rates in these studies. To
a somewhat lesser degree, employment protection indices and union densities were found to
play a role. Furthermore, the literature stresses interaction terms indicating that the effects of
a particular institutional feature may differ depending on the overall institutional framework.
A related empirical literature emphasizes the role of institutions in amplifying macroeconomic
shocks. Blanchard and Wolfers (2000), for example, find that shocks have larger effects on unem-
ployment when unemployment benefits are generous and granted over extended time and when
employment protection is strict. Yet, Nickell et al. (2005) demonstrate that shock-interaction
terms do not add much explanatory power when they augment the basic model in which unem-
ployment is caused by institutional factors.
A different view is taken by Ljungqvist and Sargent (1998) who argue that many of the labor
market institutions deemed to be the cause of high and persistent unemployment were already
in place when unemployment was still low in Europe. These authors use simulation techniques
to demonstrate the possibility that equilibrium unemployment rates can be similar in a generous
welfare state and in a laissez-faire economy in tranquil times with little aggregate fluctuation.
Yet, in turbulent times - engendering large skill losses of the unemployed - unemployment rates