Research

Extensive research documents ethnicity- and gender-based disparities in judicial decision-making, a puzzling finding given legal professionals' rigorous training in evidence-based reasoning. Whether these disparities originate in professional practice or already exist during legal training remains unknown. We address this via a survey experiment with about 3,000 law students across 19 German universities. We randomly assign students to treatment or control versions of five criminal-law cases in which only the main subject's name is varied to signal ethnicity or gender. We find that subjects with ethnic-minority names receive significantly harsher, faster, and more confident evaluations than those with German-sounding names, while female subjects are evaluated more leniently than men; both patterns carry through to sentencing recommendations. Effect sizes vary by offense type, and students exhibit in-group biases. Critically, bias magnitude does not decline with training, indicating current legal education does not correct these disparities and curriculum reform may be needed.

This paper studies how access to the parental workplace affects career trajectories of children. The setting is the German vocational training system, where many young workers enter the labor market through firm-based apprenticeships. Using matched employer-employee administrative data and a novel household-linking algorithm, we identify approximately 700,000 children and their parents and examine whether starting an apprenticeship at the primary earner's employer improves subsequent labor market outcomes. To address endogenous sorting into connected firms, we implement an instrumental variables strategy exploiting transitory, firm-specific variation in competition among coworkers' children for a limited number of apprenticeship slots. Conceptually, we compare children whose parents work at the same firm but who enter the labor market in different years. We find that starting at the parent's employer increases earnings substantially at labor market entry. One year after apprenticeship start, connected entrants earn 15 percent more than comparable non-connected entrants; ten years after entry, the earnings advantage remains 11.5 percent. Dynamic estimates show that earnings gains are largest shortly after apprenticeship completion and are partly driven by access to higher-paying firms, higher retention at the training firm, and faster within-firm career progression. Over time, non-connected workers catch up through more frequent firm switching, and earnings differences disappear after 15 years.