The Cyclical Behavior of Labor Markets

University Of Chicago, Chicago IL

Investigators

Abstract

ABSTRACT PROPOSAL NO: 0351352 INSTITUTION: University of Chicago NSF PROGRAM: ECONOMICS PI: Shimer, Robert J TITLE: 0351352 This research seeks to improve our understanding of fundamental questions at the intersection of macro and labor economics---why aggregate shocks primarily affect employment rather than wages. Search models break that question into two: (i) why do so many workers lose their job during a recession rather than take a wage cut? (ii) why don 't firms create more new job openings during recessions in anticipation of unemployed workers be willing to work at a low wage? This research focuses on the latter question, which has received relatively little attention in recent years. This research first documents the low level of vacancy creation, and consequent high unemployment duration, during and immediately after a recession. Research confirms that high unemployment duration has become relatively more important than high unemployment incidence during recent recessions. A robust phenomenon that needs an explanation is that unemployment duration has increased and the vacancy-unemployment ratio has fallen in every recession since World War II. Current economic theory, such as labor market search and matching models, are incapable of explaining this phenomenon. This research will develop alternative theories that can explain why vacancy creation falls so much and wages fall so little during a recession. Preliminary research has uncovered four fruitful avenues for further work that will be pursued both theoretically and empirically in this research. First, if the matching function exhibits sufficient substitutability between unemployment and vacancies, strong procyclicality of vacancies may be an optimal response of the economy to fluctuations in labor productivity. Alternatively, if unemployment and vacancies are complements, fluctuations are suboptimal and potentially very costly at business cycle frequencies. Second, if recessions are periods with low average productivity but also a high degree of worker uncertainty about actual productivity at the firm level, workers may optimally refuse to accept wage cuts despite high levels of aggregate unemployment. Third, on-the-job search may amplify and propagate shocks by changing the composition of the searching population at business cycle frequencies. Finally, time variation in the matching of heterogeneous workers to heterogeneous jobs may significantly increase exogenous shocks or induce endogenous business cycle fluctuations. This research could yield new insights into matching models that are applicable in other areas as well as lead to new mathematical tools that will be useful in economics generally. The results of this research will also help in labor market policy evaluation.