In recent years we have witnessed a growing literature on the impact of labour market institutions on labour market performance. This has resulted in a number of multi country models aiming at estimating an empirical relationship between institutions, differently measured by a number of aggregate variables and indicators produced by the OECD and other researchers, and a labour market performance variable that has generally taken the form of the standardized unemployment rate. The labour market institutions most widely studied in this setting are unemployment benefit provision, employment protection regulations, the tax wedge and trade union bargaining power. Recently Bassanini and Duval (2005) have reviewed this empirical evidence, stressing that unemployment benefits and tax wedges are the most significant institutions affecting the unemployment rate, even when considering subgroup of the population. They cast some doubts on the robustness of previous results in the literature, which attributes an unemployment enhancing role to trade unions.
All multi country models estimating unemployment as a function of labour market institutions we are aware of, implicitly assume labour market institutions as super exogenous. This is to say that institutions do not depend on the state of the labour market and estimated parameters are independent of changes in policy. However this simplifying assumption is contradicted by our knowledge of how institutions are determined. If institutions are coordinating devices that are introduced and modified as optimal answers to market failures, they cannot be taken as fully exogenous to market outcomes. In general, changes in institutions respond to changes in political equilibria, to macroeconomic shocks, or a combination of the two. Botero et al. (2004) suggest that the institutional framework of a country (including labour regulations such as employment protection, collective bargaining and social security) depends on the stage of development, as well as on the legal tradition of the country, while the political orientation of the government seems to play no role. Other authors have focused on the potential endogeneity of single institutions. Saint Paul (1996) explores the role of the median voter in determining the degree of employment protection while Di Tella and MacCulloch (2002) analyze the determinants of unemployment benefits. Checchi and Lucifora (2002) consider union density as endogenous and analyse how other institutions play a role of union complement or union substitute in European countries. Bertola and Koeninger (2004) find that a more reduced dispersion and volatility of labor income (through employment protection, unemployment benefits and wage compression) is more prevalent in countries where ineu cient legal systems restrict borrowing opportunities. In each of these frameworks, unemployment (or more precisely unemployment risk) is considered as one of the determinants of institutions.