An investigation of the life history of three cohorts of exits based on longitudinal data covering the population of all manufacturing firms in a German federal state reveals three empirical regularities: (1) Although the probability of exit tends to decline with age ceteris paribus, only about a quarter of all exits is from the group of young firms aged five years or less. (2) Only a minor fraction of exits lived through a period of continous decline of employment for five years or more before closing down. (3) The "impact effect" of job loss due to exits can differ sharply from the "long run effect".