Why do firms train apprentices? The net cost puzzle reconsidered

This paper analyses the impact of increasing the share of apprentices at the cost of the share of unskilled or semi-skilled employees on establishment performance. We use representative matched employer-employee panel data and correct for estimation biases. We show that an increase of the apprentice share in trade, commercial, craft or construction occupations has a positive impact on establishment performance. Establishments that increase the apprentice share in manufacturing occupations face a negative impact on performance, however. These results shed a new light on the stylised fact that apprenticeship training always leads to net costs during the apprenticeship period in Germany: we argue that establishments only hire apprentices at a cost if their skills are relatively specific, their retention rate is high and skilled employees are hard to hire.