The Cyclicality of Entry and Exit: A General Equilibrium Analysis with Imperfect Information

The US establishment exit rate is acyclical. This poses a challenge to canonical models of industry dynamics-e.g., Hopenhayn (1992)-which imply a strongly counter- cyclical exit rate. To reconcile this gap between theory and data, imperfect information is introduced. Potential entrants have imperfect information about their productivity, leading to a signal extraction problem. When the volatility of idiosyncratic productivity dominates that of aggregate-as we observe in the micro data, potential entrants overestimate their productivity, and the value of entering, in booms. This amplified entry further increases factor prices and crowds out marginal incumbents, making the exit rate almost acyclical. The imperfect information mechanism proposed here also yields three testable implications: (i) entry is more cyclical in the industries where idiosyncratic components dominate; (ii) plant entry by new firms is more cyclical than that by existing firms; (iii) plants established by new firms during booms are more likely to exit rapidly. We show that all three predictions are consistent with the data.