Using Transaction Prices to Re-Examine Price Dispersion in Electronic Markets

Price dispersion is an important indicator of market efficiency. Internet-based electronic markets have the potential to reduce transaction and search costs, thereby creating more efficient, "frictionless" markets, as predicted by theories in information economics. However, earlier work has reported significant levels of price dispersion on the Internet, which is in contrast to theoretical predictions. A key feature of the existing stream of work has been its use of posted prices to estimate price dispersion. In theory, this can lead to an overestimation of price dispersion because a sale may not have occurred at the posted price. In this research, we use a unique data set of actual transaction prices collected from both the electronic and offline markets of buyers in a business-to-business market to evaluate the extent of price dispersion. We find that price dispersion in the electronic market is as low as 0.22%, which is substantially less than that reported in the existing literature. This near-zero price dispersion suggests that in some electronic markets the "law of one price" can prevail when we consider transaction prices, instead of posted prices. We further develop a theoretical framework that identifies several new drivers of price dispersion using transaction data. In particular, we focus on four product-level and market-level attributes-product cost, order cycle time, own price elasticity, and transaction quantity, and we estimate their impact on price dispersion. We also examine the electronic market's moderating role in the relationship between these drivers and price dispersion. Finally, we estimate the efficiency gains that accrue from transactions in the relatively friction-free market and find that the electronic market can enhance consumer surplus by as much as $97.92 million per year.