Is there a threat effect of mandatory activation programmes for the long-term unemployed?

Exploiting changes in Denmark’s unemployment insurance (UI) system—a ‘natural experiment’ that reduced the period during which unemployed individuals could receive UI benefits without being activated—Geerdsen (Econ J 116:738–750, 2006 ) finds that long-term unemployed individuals are more likely to find a job when facing the threat of having to participate in mandatory activation programmes. The threat effect estimated by Geerdsen is surprisingly large enough so to be comparable with the effect of benefits exhaustion found in studies of finite-duration UI systems. This article re-examines Geerdsen’s analysis. Using better data than were available for his study, we cannot confirm his findings. When we use the same sample period as Geerdsen, we find no significant evidence of a threat effect. Using more recent data, some of our estimates of the threat effect are significant, but these estimates are generally smaller than that found by Geerdsen. To explain the difference between our results and Geerdsen’s, we show that his estimate of the threat effect is seriously upward biased because of the shortcomings of his data and because of an important error in his code. Copyright Springer-Verlag 2013