Abstract: Using comprehensive administrative data on nearly 17 million U.S. business applications linked to subsequent outcomes, we examine how potential entrants’ expectations about entry as an employer business and first-year employment compare with realizations. On average, applicants overestimate employment, primarily because many anticipate entry that does not occur. Conditional on both expecting and realizing entry, however, applicants tend to underestimate employment. Expectations are informative but imperfect: higher expected employment is associated with a higher probability of entry, yet realized employment increases less than one-for-one with expected employment. Expectation errors are highly heterogeneous and systematically related to application characteristics and economic conditions. Moreover, they predict subsequent employment outcomes. A parsimonious model with heterogeneous priors, learning, and selection before entry can rationalize these patterns.
