Increasing firms' motivation to train low-skilled youth: a factorial survey experiment

The German system of apprenticeship training is often claimed as an important driver for the good integration of youth in the labour market. Nonetheless, a steady share of 13% of school graduates who would like to start an apprenticeship are unable to find a training position. In this paper, a factorial survey experiment was used to analyze whether a financial or a non-financial bonus scheme could increase firms´ willingness to train low-skilled youth. In the experiment, firms were asked to rate the probability of providing a training position to hypothetical applicants that vary on cognitive and social skills and the possibility of receiving financial or non-financial support. Both applicants’ cognitive and non-cognitive skills are highly important for a firm’s decision to offer low-skilled applicants a training position. High non-cognitive skills can compensate for low cognitive skills. However, neither financial nor non-financial support can compensate for a low cognitive or non-cognitive skill level, as the effects of both support measures are rather weak.

There are two possible explanations for this finding. First, firms expect the overall training costs of weak applicants to be much higher than the average training costs of the apprentices they are used to recruit. Second, the core of German training firms invest in apprenticeship training to secure their need for skilled workers. The costs of training are therefore of less importance as in the long run the benefits of retaining the apprentices afterwards outweigh the initial training costs. For the low-skilled applicants – especially for those with low cognitive and/or non-cognitive skills – firms might expect that they will not finish the apprenticeship successfully or will not acquire the skills needed to be attractive for the firm as a skilled worker after graduation. Regarding the non-financial support, the same arguments hold. Additionally, the eligibility for a support scheme might be seen as an additional signal for the applicants’ low productivity.

Deadweight losses are a widely discussed issue when granting support, since it is always questionable whether the desired result would not have occurred even without the measure. This is also the case with firms’ decision to train low-skilled youth: the extent to which a subsidy is decisive for a company's decision to train a low-skilled young person seems debatable. By design, the scenario in our experiment prevents deadweight losses, as the application timing clearly implies that the young person would be recruited as an additional apprentice. Our analyses show that an active labour market policy introducing financial or non-financial support to recruit low-skilled youth does not affect the most firms’ decision to train an additional apprentice. The effects we found were small and remained insignificant in most cases. This suggests that the effectiveness of ALMPs meant to (re-)integrate low-skilled youngsters early in their working careers is low. This is particularly remarkable given the substantial volume of support and raises the question of whether the subsidy form is suitable for companies, or what other kind of support companies would require instead. Further research might therefore focus on identifying firms’ needs for support more closely to motivate them to train specific vulnerable groups.