Active labour market policies and job tenure

In this paper, we study the effect of subsidised on-the-job training, training for the unemployed and pure wage subsidies on the probability of leaving an employer. We base the analysis on a sample of unemployed workers who have been hired during the 1991-92 period. Since individuals benefiting from the policy were over represented in the sample, we face an endogenous sampling problem apart from the well known selectivity problem in evaluation analysis. The analysis deals with these two issues simultaneously. We find that each of the labour market policies increases the length of job tenure. Yet, in line with the literature (Lalonde 1986, Fraker and Maynard 1987), the magnitude and the significance of this effect depend crucially on the parametric assumptions in the model. Nevertheless, one robust conclusion is that subsidised on-the-job training schemes significantly decrease the incidence of job termination. We claim that this result supports Stevens' (1994, 1996) hypothesis of transferable training and consequently of the underprovision of training by the market. Training programs for the unemployed and pure wage subsidies only have an important positive but non significant effect on job tenure. We argue that the stated effect of training programs provides weak support for human capital theories as opposed to matching theories in the explanation of job turnover.