This paper contributes to the policy-relevant question whether self-employment is a way out of (long-term) unemployment. We estimate the relationship between the entry rate into self-employment and previous (long-term) unemployment on the basis of pseudo-panel data for Germany in the period 1996-2002. The estimation method accounts for cohort fixed effects and measurement errors induced by the pseudo panel structure. We find that previous (long-term) unemployment significantly increases entry rates into self-employment for both men and women. These effects are quantitatively important, both in absolute terms and compared to other potential determinants of self-employment transitions, such as age, the level of vocational qualification and certain household characteristics.