The proverbial “law of one price” is virtually never empirically valid. Homogeneous goods are often sold at widely different prices by rival firms, even in environments that seem particularly conducive to economic competition. Following a seminal paper by Stigler (1961), several economists developed information-based models explaining this phenomenon (see, e.g., Salop and Stiglitz 1982; Burdett and Judd 1983; Stahl 1989). The principal success of this literature was to identify conditions under which price dispersion can arise as a stable equilibrium outcome. Generally speaking, price dispersion will arise when there is a positive (but uncertain) probability that a randomly chosen customer knows only one price. Thus, even in markets with symmetric firms selling homogeneous products, prices may differ in equilibrium if consumers must incur search costs to obtain price information.

This paper seeks to demonstrate the empirical importance of price dispersion that arises from imperfect information by examining the retail market for prescription drugs. Using data collected from individual pharmacies in upstate New York, I show that cash prices for equivalent prescriptions differ substantially across pharmacies within the same small town. On average, the highest posted price for a given prescription is over 50 percent above the lowest available price. The benefits of price shopping can be substantial even in absolute terms: the potential savings from finding the lowest-cost pharmacy (as measured by the price range across pharmacies) exceed $10 for over half of the prescriptions in the sample.

Differences in pharmacy service or location do not appear to explain fully the observed price variation. Pharmacies’ price rankings are inconsistent across drugs, and hedonic regressions using pharmacy service characteristics as explanatory variables are relatively unsuccessful in explaining price differences. I estimate that pharmacy effects account for at most one-third of the variation in drug prices about their means.

The central finding of this study is that observed price distributions are consistent with the predictions of models based on consumer search. The empirical approach hinges on the observation that incentives to price-shop are strongest for prescriptions that must be purchased frequently, such as medications used to treat chronic conditions. Consumers’ increased propensities to price-shop for frequently purchased prescriptions should lead to less absolute dispersion and lower markups for such prescriptions. This prediction is found to be true in the data: measures of both dispersion and markups are significantly lower for drugs that are purchased repeatedly. Price ranges for one-time prescriptions are estimated to be 34 percent larger than those for prescriptions that must be purchased monthly. Absolute markups (which are inferred from average wholesale price data) for one-time prescriptions are estimated to be 41 percent higher than those for prescriptions purchased monthly, other things being equal.

Two previous empirical studies have explicitly addressed the role of consumer search in explaining price dispersion. Dahlby and West (1986) use data on prices for auto insurance policies to test the predictions of a dispersion model due to Carlson and McAfee (1983) and ultimately conclude that the dispersion in premiums can be explained by costly consumer search. Their finding that premiums are least dispersed in driver classes in which search is most likely to occur is similar to the results discussed in the present study. However, they also find that premiums at a given firm are highly correlated across driver classes, in contrast to the finding here that prices at a given pharmacy are at most weakly correlated across drugs. In another study, Van Hoomissen (1988) examines price data from Israel during an inflationary period and attempts to distinguish between dispersion based on product differences and dispersion based on imperfect information. She finds that inflation (which is linked to reductions in the value of acquiring price information) is positively related to dispersion and interprets the finding as evidence that price dispersion arises from imperfect consumer information. Although this paper reaches a similar conclusion, it differs from Van Hoomissen’s study in that it exploits cross-sectional variation instead of time variation to identify the effects of search and information, and it directly addresses the role of product heterogeneity in generating dispersion. Moreover, price data in this paper come from stores competing within well-defined local markets, so the results can be more appropriately interpreted in the context of equilibrium price dispersion models.

Download
PDF Ebook Equilibrium Price Dispersion in Retail Markets for Prescription Drugs