We examine whether or not unemployment early in an individual's career influences her later employment prospects. We use six years of the Labour Force Survey (LFS) to create pseudo-cohorts and exploit cross-cohort variation in unemployment at school-leaving age to identify this. We find heterogeneous responses: for the unskilled, there is evidence of an enduring adverse effect; for the more skilled, there is a small beneficial effect. (C) 2003 Elsevier Science B.V. All rights reserved.