Population Characteristics and Price Dispersion in the Market for Prescription Drugs

Abstract

We examine the relationship between population characteristics and price dispersion for 75 prescription drugs in five markets. Based on models of price dispersion, we consider that search costs are likely lower for the elderly, who are repeat purchasers. Expected benefits from search are likely higher for low‐income households, who lack insurance. Our results are consistent with the hypothesis that for communities with a large percentage of elderly and poor population, search effort is greater for pharmaceutical drugs, causing lower price dispersion. By understanding the characteristics of who searches for low drug prices, we begin to identify the motives of consumers that might also lead to search for the lowest cost healthcare provider or lowest cost insurance. The results suggest that the Medicare legislation that attempts to close the pharmaceutical doughnut hole may reduce search by the elderly, increase price dispersion, and potentially increase the average price of prescription drugs. (JEL D12, D83, I1, I18)

Consumer Search, Market Characteristics, and Price Dispersion: New Evidence from the Retail Markets for Prescription Drugs

“…Following the literature, I develop three indirect measures of search intensity: purchase frequency, patient age, and drug type (life‐saving vs. life‐enhancing), and a direct measure, ‘seller density’ (number of pharmacies per zip code) . Consistent with previous studies (Sorensen, ; Ohler and Smith, ), the estimation results suggest a significant inverse relationship between search intensity and price dispersion for top‐selling prescription drugs. Taking into consideration both direct and indirect measures of search intensity, having an additional pharmacy within a given zip code decreases the range of per‐tablet prices by about $ 0.06; every additional refill required lowers the average range by $ 0.56; every additional year of mean age of the patients using a drug is associated with a $0.10 decrease in the range; and compared with life‐enhancing drugs, life‐saving ones lead to a lower dispersion by $ 0.82.…”

“…This data set is labeled ‘Sample 1’. Furthermore, because existing studies typically choose a handful of cities (e.g., Sorensen, , with two cities; Ohler and Smith, , with five cities), my main analysis focuses on the five largest cities in NH, namely, Manchester, Nashua, Concord, Dover, and Derry (the 2010 Census) . This final sample includes a total of 268 704 observations and is labeled ‘Sample 2’.…”

“…In the empirical literature, several indirect measures are used to gauge the extent to which search intensity affects price dispersion. Previous studies have employed purchase frequency to measure search intensity (Sorensen, ; Ohler and Smith, ). Intuitively, the benefit from comparison‐shopping rises with purchase frequency.…”

“…In relation, because all NH pharmacies are required by state law to report claims data annually, this study contributes to the literature by extending geographical area coverage and time horizon and is thus less prone to sample bias. Unlike previous studies (Sorensen, ; Ohler and Smith, ), my analysis uses multiple‐year data to control for any time‐invariant unobserved market characteristics that affect pharmacy pricing.…”

“… Ohler and Smith () include aggregated demographic information. However, such market‐level information fails to account for the differing price effects that the elderly population may have on drugs that they are likely to consume from those that they are unlikely to consume.…”