Price Discovery in Online Markets: Convergence, Asymmetries and Information

Abstract

Prices tend to converge rapidly to competitive prices in traditional markets for non-durable goods. An open question has been whether and why the same would apply in online markets in light of the increase in size, anonymity and information decentrality that characterizes the online setting. To address these questions we built an online trading platform to conduct controlled experiments. In terms of prices, we find that convergence does occur, but not necessarily fast. Moreover, aggregate equilibration dynamics consistently favor buyers over sellers. Regarding subjects' individual updating behavior, we identify a simple baseline rule, whereby agents with successful bids/offers become more greedy, unsuccessful ones less. As long as the available information allows subjects to improve their `guess' of at which price trade will occur convergence to equilibrium prices is fast. We link our empirical findings with theoretical conditions under which fast convergence is proven to occur. In addition, we provide a behavioral explanation for why price convergence typically occurs via rising prices, thus favoring in buyer-optimal prices, a phenomenon that has been observed but not explained.