Using Machine Learning to Explain Violations of the 'Law of One Price'

Abstract

Substantial price variation for homogeneous goods in online markets is a well-known puzzle that has withstood attempts by empirical researchers to explain it. Economic theory suggests two possible sources of the dispersion: either market frictions are more important than previously thought, or there are subtle differences between product listings presented to e-commerce consumers that applied econometricians have failed to detect. We use a very detailed data set consisting of posted-price listings for new Kindle Fire tablets from eBay to determine if observable listing heterogeneity can explain the price dispersion of seemingly homogeneous products. By combining a richer set of variables than previous studies with more sophisticated machine learning techniques, we can explain 42% of the dispersion. We interpret this as a bound on the influence of market frictions on price dispersion. Variables describing the amount of information in the listing are good predictors of the price, but variables describing the style of a listing's text are good predictors as well. We identify readily interpretable groups of words that are also good predictors of price. We find a high degree of heterogeneity of the marginal effects of seller reputation and including an image in the listing, but the patterns of heterogeneity largely conform to economic intuition. A smaller, but non-trivial, latitude for market frictions remains, and we discuss their possible sources.