Worker separations in a nonstationary corporate environment

Assessing ten years of unique personnel data from a large firm, this paper investigates how determinants of worker turnover differ between periods of corporate expansion and decline. Performance and career progress considerably reduce the hazard of job separation during downsizing for blue-collar workers, but are less important for white-collar workers. Effects of firm-tenure, occupation, education, training, part-time status, and gender are largely in accordance with the implications from existing models. But no model generates all patterns observed in the data. Age effects are sensitive to retirement options and discontinuous, dropping in response to the anticipation of becoming eligible for early retirement during downsizing and jumping at the actual retirement age.