Fighting discrimination with reputation: The case of online platforms
创建于 更新于
摘要
Using data from a large French ridesharing platform, the paper shows that minority drivers face a revenue gap of 11.6% when they have no reviews, and that this gap declines as drivers accumulate reviews, nearly vanishing after around twenty reviews. The mechanism is twofold: passengers update beliefs about individual quality as reviews accumulate, and drivers actively invest in reputation by pricing strategically and exerting extra effort to overturn biased priors. A railway-strike natural experiment identifies a causal role for reviews in improving minority outcomes, with effects strongest for entrants. A structural model disentangles belief-based partiality from discrimination, quantifies welfare costs of incorrect priors, and evaluates counterfactuals on priors and rating-system informativeness. Sharper rating systems speed belief revision but do not fully erase the wedge; information alone helps platforms rely more on observed quality than group identity. [page::0]
