Many countries consider expanding vocational curricula in secondary education to reduce skill shortages and unemployment among non-college-bound youths. However, critics fear this could divert students from academic routes, whose focus on general, instead of occupation-specific, skills might be better suited for today's rapidly changing labour markets. The paper presented in this ZEW Research Seminar studies labour market returns to vocational education in England, where students choose between a vocational track, an academic track and quitting education at age 16. Identification is challenging, first, because self-selection is strong and, second, because students' next-best alternatives are unknown. The authors research design leverages multiple instrumental variables to estimate margin-specific treatment effects, i.e., causal returns to vocational education for students at the margin between vocational and academic education and, separately, for students at the margin between vocational and no post-16 education. Identification comes from plausibly exogenous variation in distance to the nearest vocational provider conditional on distance to the nearest academic provider (and vice-versa), while controlling for granular student-, school- and neighbourhood-level characteristics. The analysis is based on linked administrative education and earnings data that cover the universe of students in English public schools. The authors find that the vast majority of marginal vocational students are at the margin with academic education (instead of no further education). Hence, the first-order effect of expanding vocational-track access is diversion from the academic track. This diversion leads to substantial losses in earnings, which steadily increase through at least age 29, when they equal 9% for males and 7% for females. These effects are not driven by employment but due to wages (or working hours) and can partly be explained by worse educational attainment and progression. However, consistent with comparative advantage, the authors find that the earnings penalty is smaller for compliers with higher (unobserved) preferences for vocational education. Among males with the highest vocational preferences, the effect even turns positive. For the few compliers at the margin with no further education, the authors find tentative evidence of positive employment and earnings effects from vocational education but results are imprecise and insignificant. Altogether, their findings caution against an expansion of vocational upper secondary education in England in its current form. On the contrary, the current equilibrium seems to be tilted too far towards vocational enrolment.