Marleen Marra - Website
Primary: Industrial Organisation, Structural Economics
Secondary: Applied Microeconomics
Abstract - Auction platforms are increasingly popular marketplaces that generate revenues from fees charged to users. As in other two-sided markets, due to strategic interactions between buyers and sellers both the profit maximizing fee structure and welfare impacts of these fees are ambiguous. I collect a data set of fine wine auctions and develop a structural model with endogenous bidder and seller entry, seller selection and costly listing inspection to examine user interactions in relation to fees. Using observed variation in reserve prices, transaction prices and the number of bidders, model primitives are nonparametrically identified for counterfactual fee structures that reduce expected seller surplus. A feasible estimation strategy combines methods from the auction and discrete choice literatures. I show that models which do not incorporate these features, commonly used in antitrust policy, may understate the welfare impacts of fee increases by over 50 percent and ignore that the economic incidence of either buyer or seller commission falls for 71-89 percent on sellers. I also show that an alternative fee structure that subsidizes bidders by discounting their transaction price, paid for by a higher seller commission, would improve platform profitability. While no auction platform has negative fees, this practice is adopted in other two-sided markets and is demonstrated to benefit both bidders and sellers on the platform.
- Andrew Chesher (UCL)
- Aureo de Paula (UCL)
- Adam Rosen (Duke University)
- Philip Haile (Yale University)