Unemployment,institutions, and interdependencies: identifying successful reforms

Labor market reforms, which reduce institutional rigidities, are assumed to be a well-suited approach to lower unemployment. However, it is still not perfectly clear which reforms actually lead to a fall in unemployment. One crucial issue is that reforms do not work in isolation, but have labor market effects which depend on other institutional factors. Such institutional interactions have rarely been considered in empirical macroeconomic studies in a systematic way, mainly due to model uncertainty. As a solution to this problem, a Bayesian model averaging approach is adopted in this paper to identify robust and significant institutional interactions for unemployment. Using a panel data-set for 17 OECD countries from 1982 to 2005, five robust and significant interaction terms are identified, and country-specific reform effects for the institutional indicators are derived.