This paper estimates an oligopoly model of consumer choice and supermarket profitability using a survey of consumer choices and a dataset of store characteristics. The survey data includes expenditure as well as choice of supermarket. Unobserved characteristics associated with the operating firm are incorporated through firm dummies in the utility function. Predicted Nash price–cost markups are close to observed markups. We examine the social e¢ciency of store numbers and store characteristics by making minor changes to the characteristics and solving for new profits and consumer surplus. We find no evidence of excess entry for large stores. There is some excess entry of smaller stores depending on location and operating firm. Increases in store size usually increase welfare. Ownership conversions can increase welfare signi…cantly. Location inefficiencies, in contrast, are small.