6 septembre 2017, 12h30–13h30
Salle MS 003
Reputation systems are a cornerstone of internet commerce: they discipline seller behavior in a setting of one-time, anonymous interactions. When moral hazard is present, reputation systems incentivize sellers to exert effort by creating a link between reviews by past clients and demand by new clients. The value of reputation is directly related to its ability to generate future demand, and when demand conditions fluctuate so does the value of reputation. This in turn affects the incentive to exert effort, and the effectiveness of the reputation mechanism may be compromised. We use data from an online marketplace for home services to document sellers' behavior when they expect a decrease in activity in the immediate future. Our strategy exploits the fact that some professionals are in highly seasonal lines of work while others can expect a more regular demand. We demonstrate that at the end of the high demand season, those professionals facing a drop in activity are less likely to receive a positive review.