The causal effects of internships on academic outcomes: A shred of evidence from the UK

Many student internships are short term and do not delay graduation. Some internships however are lengthier and delay graduation by a whole year. These internships, in addition to facilitating the transition from school to work, are expected to have educational benefits. This paper estimates the causal effects of year-long internships (known as work placements in the UK) on post-internship academic performance and degree classification. These effects are identified using instrumental variable methods that rely on detailed institution-level information. We find that the average internship effect on the post-internship grades is in effect zero, but the effect on the probability of a degree in one of the top two classes is in the order of 10 percent. Using (unconditional) quantile treatment effects we find two features that reconcile these estimates. First, internship effects decrease as one moves up the grade distribution. Second, internship effects are positive just below the threshold of a 'good' degree. These findings do not support skill-based arguments in support of subsidizing internships. Instead they suggest that students use work experience to update their beliefs on the market value of educational attainment and adjust their efforts accordingly.