Essays on the macroeconomics of labor market institutions

Abstract

This thesis contributes to furthering the understanding of the macroeconomic impact of two types of labor market institutions: temporary help service agencies and temporary contracts. In the rst chapter, I depart from the observation that employment in the temporary help service industry in the United States has seen a secular rise in recent decades. The chapter provides a theory of the temporary help service industry within the steady state version of a random search model of the labor market with endogenous job destruction and a second sector in which employment relationships are intermediated. In this framework temporary jobs are endogenously of short duration and recruitment is fast. Conditions are provided under which intermediated employment relationships exist in equilibrium. The implications of the model for two possible explanations of the secular rise of employment in the temporary help service industry, technological progress and a rise in rm-level uncertainty, are such that technological progress as an explanation is favored. In the second chapter, I investigate the impact of uncertainty shocks on a dual labor market using the Spanish economy as a case study. In an empirical analysis, I nd that, given my identi cation strategy, uctuations in uncertainty cause a signi cant drop in temporary employment, a non-signi cant reaction in permanent employment and a signi cant decline in GDP. Since in the data the responses to a second-moment shock are similar to the responses to a rst-moment shock, a quantitative labor demand model of the Spanish labor market is built and calibrated. I use this model to generate simulated response functions to a (pure) second-moment, a (pure) rst-moment and a combined rstand second moment shock. I nd that the empirical impulse responses can only partially be rationalized by the model when considering a (pure) second-moment shock. A (pure) rst moment shock in the model generates impulse response functions similar to the empirical ones. A combined rstand second moment shock can not improve on the rst-moment shock in replicating the data.