Training, occupations, and the specificity of human capital

Eckardt, Dita (2020) Training, occupations, and the specificity of human capital. PhD thesis, London School of Economics and Political Science.

Text - Submitted Version Download (950kB)

Abstract

Individuals are often trained in a specific filed but work in another. This thesis analyses the reruns to different training-occupation combinations. To this end, I use an administrative employment panel which contains the apprenticeship training for a large sample of workers in Germany. In this context, 70% of individuals with at least upper-secondary education hold apprenticeships, and 40% of these work in occupations they were not trained for. I combine the administrative data with historical data on occupation-specific vacancies to control for selection into trainings and occupations, and causally identify the returns. To implement the identification strategy, I set up an augmented Roy model and extend existing control function approaches to deal with selection in a two-stage, high-dimensional setting. The results suggest that workers trained in their current occupation earn 10-12% more than workers trained outside their occupation, and that not controlling for selection leads to substantial negative bias. Intuitively, individuals who choose to work outside their training are positively selected since their occupation-specific ability needs to compensate for their lack of training. I find considerable heterogeneity in the estimated returns and use task content data to provide a microfoundation for the results in this thesis. My analysis shows that returns across training-occupation combinations are decreasing in the task distance between the training and the occupation, suggesting that workers are trained to carry out a specific mix of tasks and receive larger wage penalties the less applicable the acquired skills are in their current occupation. Finally, argue that ex-ante imperfect information may lead to training choices that are suboptimal ex-post and find that, as a result, 4-6% of wages are foregone for the average worker. Back-of-the-envelope calculations suggest that retraining programmes could be very effective in addressing this friction.

Item Type:	Thesis (PhD)
Additional Information:	© 2020 Dita Eckardt
Library of Congress subject classification:	H Social Sciences > HC Economic History and Conditions H Social Sciences > HD Industries. Land use. Labor
Sets:	Departments > Economics Collections > Economists Online
Supervisor:	Spinnewijn, Johannes and Manning, Alan
URI:	http://etheses.lse.ac.uk/id/eprint/4055

Actions (login required)

Record administration - authorised staff only