Wednesday, November 12, 2014

Why Experts Buy Generic

You are buying an over-the-counter medication. You see several shelves of brand-name medications of the type you want, along with a generic store brand that is typically much cheaper. Do you buy the cheaper generic version? And perhaps more interesting, if you were standing beside an expert who really knew what was in all of the medications, would the expert buy the cheaper generic version?

Bart J. Bronnenberg, Jean-Pierre DubĂ©, Matthew Gentzkow, and Jesse M. Shapiro tackle this question in "Do Pharmacists Buy Bayer? Informed Shoppers and the Brand Premium." It's available as National Bureau of Economic Research Working Paper #20295, August 2014. (NBER working papers are not freely available on-line to everyone, but they are inexpensively available, and  many readers will have access through library subscriptions.) They start the paper by pointing out the example of aspirin as a generic equivalent (citations omitted).
A 100-tablet package of 325mg Bayer Aspirin costs $6.29 at cvs.com. A 100-tablet package of 325mg CVS store-brand aspirin costs $1.99. The two brands share the same dosage, directions, and active ingredient. Aspirin has been sold in the United States for more than 100 years, CVS explicitly directs consumers to compare Bayer to the CVS alternative, and CVS is one of the the largest pharmacy chains in the country, with presumably little incentive to sell a faulty product. Yet the prevailing prices are evidence that some consumers are willing to pay a three-fold premium to buy Bayer. 
A short readable overview of the paper is available from NBER here, and here are some notable findings from the overview.

In a detailed case study of headache remedy purchases, the researchers find that more-informed consumers are less likely to pay extra to buy national brands, with pharmacists choosing them over store brands only 9 percent of the time, compared with 26 percent of the time for the average consumer. Similarly, chefs devote 12 percentage points less of their purchases of kitchen staples to national brands than otherwise similar non-chefs. ...
Controlling for household income, other demographics, and the market, chain, and quarter in which the purchase is made, a household whose primary shopper correctly identifies all active ingredients in a national brand has an 85 percent chance of purchasing a store brand, 19 percentage points higher than a shopper who identifies none of the ingredients. ... When the primary shopper is either a pharmacist or a physician, the probability of purchasing the store brand is 91 percent, 15 percentage points higher than the probability of otherwise similar buyers who are not in these fields. Primary shoppers who were science majors in college buy more store brands than those with other college degrees. In a second case study of pantry staples such as salt, sugar, and baking powder, the researchers find that chefs devote nearly 80 percent of their purchases to store brands, compared with 60 percent for the average consumer.
The overall pattern is clear: those who are less knowledgeable are more likely to buy brand names, presumably because they feel that there is quality difference in doing so. Those who are more knowledgeable are more likely to go with generic equivalents, because they feel comfortable making their own judgements about quality--and then going with the lower price.

The data for the study comes from the Nielsen Homescan Panel. It includes information on  purchases made on more than 77 million shopping trips by 125,114 households between 2004 and 2011. People in the panel scan barcodes for all consumer packaged goods they buy. As a result, the data includes detailed information on the product and price, as well as when and where it was bought. The Nielson survey also has basic information about houeholds, like level of composition of household, education, income, race, age, homeownership, and area of residence. The researchers supplemented this data by doing their own survey to gather more information on the specific jobs held by the panelists, and whether panelists can name the specific ingredients in various products. The researchers can then use this mass of data to compare the buying patterns of different groups between branded and generic goods. Here's their bottom line:
We estimate that consumers spend $196 billion annually in consumer packaged goods categories in which a store-brand alternative to the national brand exists, and that they would spend approximately $44 billion less (at current prices) if they switched to the store brand whenever possible. If consumers are systematically misled by brand claims, this has clear implications for evaluating the welfare effects of the roughly $140 billion spent on advertising each year in the US, and for designing federal regulation to minimize the potential for harm ...