Previous studies have shown mixed results concerning the effects of participation in active labour market policy programmes (ALMPs) on the longer-term scars in the form of poor income development and low job stability following the end of an unemployment spell. Most previous studies, however, have been limited both in the time frame used and to particular programmes. We argue that human capital investments are long-term investments and should therefore also be investigated from a long-term perspective. ALMP training and ALMPs as subsidised employment also represent different types of human capital investments that may produce effects that are differently distributed over time. In order to handle these issues, this article uses a longitudinal register-based dataset in which all long-term (more than six months) unemployed Swedes in 1993, who had no labour market problems in 1992, were followed for ten years. We found positive effects of ALMP participation concerning both the probability of reaching pre-unemployment incomes and a reduction in the hazard of exiting the labour market, while the effect on the probability of having an unemployment-free year was mixed. The effects of the two forms of ALMPs were differently distributed over time, with ALMP employment having an immediate effect that decreased relatively quickly and ALMP training having a longer-term effect.