We explore the impact of randomly revealing or concealing researcher affiliations on the evaluation of submissions to an early career workshop in Economics. We find evidence that non-blind peer-reviewing leads to a premium for submissions coming from elite institutions, mainly driven by reviewers coming from similarly ranked universities. This points at a systemic affiliation bias or in-group favoritism among these establishments. The presence of affiliation bias might lead to self-reinforcing inequalities appearing very early in the career, reducing meritocracy, and penalizing researchers affiliated with less prestigious institutions, for whom network building, and consequently, conference attendance, is particularly important.