The effect of social capital is often overrated because contacts and centrality can be a consequence of success rather than its cause. Randomized or natural experiments are an excellent way to assess the real causal effect of social capital, but these are rare. This paper relies on data from one such experiment: recruitment at the EHESS, a leading social science institution in France, between 1960 and 2005. The EHESS recruitment process uses an electoral commission to produce a first-stage ranking of applicants, which is then provided to the faculty assembly for final voting. The commission is partly composed of faculty members drawn at random, a feature that this article exploits in order to compare the chances for success of applicants whose contacts have been drawn to sit on the commission (treated) versus those whose contacts have not been drawn (control). It shows that a contact such as a PhD advisor has a causal impact, especially for assistant professor hiring exams: it doubles the chance of being ranked and increases the share of votes by 10 percent. This phenomenon may explain part of the classic "academic inbreeding" issue.