Who makes, who breaks: Which scientists stay in academe?

•We examine the survival of young scientists in the academic labor market.•The propensity to leave academe follows an inverse u-shape.•Publishing one additional paper increases the chance of survival by 6%.•Patenting scientists are less likely to stay in academe.•Preferences for doing business increase the propensity of leaving academe.

We examine the survival of young scientists in academe. The propensity to leave follows an inverse u-shape. Publishing increases and patenting decreases the chance of survival. Scientists with strong preferences for business are less likely to stay in academe.