The thick market effect of local unemployment rate fluctuation

This paper explores the relationship between city size and the pattern of unemployment rate fluctuations. We present a model of the local labor market in which in a thicker market, when more workers are looking for jobs and more job openings are available, the matching probability between jobs and workers is higher. Workers incur search cost if they actively search for for jobs. A higher matching probability makes searching for jobs more desirable. Unemployed workers accumulate in a local market until the market reaches a critical size such that the expected wage is higher than the search cost. Since a given shock produces more unemployed workers in a larger city during a time period, it takes less time for a larger city to reach the critical size described above. As a consequence, the model predicts: (1)Unemployment rates are lower in larger cities. (2) The length of unemployment cycles decreases as city size increases. (3) The peak unemployment rate is negatively correlated with city size. Our empirical analysis utilizes data that covers 295 PMSAs in the U.S. over the years 1981--1997. After controlling for the effects of industry composition and risk diversification, we find that city size has a significantly negative effect on the mean unemployment rate. In particular, if city size increases by two standard deviations, the unemployment rate will be lowered by roughly a half percentage point. We also find that larger cities have shorter unemployment cycles. In particular, the unemployment cycle will be shortened by roughly one month if city size increases by two standard deviations. Finally, we find shallower recessions for larger cities. The peak unemployment rates are lower by .3 percentage points if the city size increases by two standard deviations. All these empirical results are consistent with the predictions of the model.