Abstract: | The intensity of competition among firms depends on commuting patterns, as has been noted, because commuters can reach any store located on their route to work without incurring any incremental travel costs. We incorporate this insight into our estimation of a retail gasoline price function for Lexington, Kentucky, by treating each commuter route as a separate market. Competition in these markets, however, displays an asymmetry because all the commuters travel to the Central Business District (CBD). To accommodate this asymmetry, we treat the market segments on each side of a firm as distinct submarkets and include independent variables (number of competitors and submarket length) from each submarket. Both sets of structural variables influence gasoline prices in the expected direction, but the variables representing the submarket near the CBD have significantly stronger effects. Refinements show that a firm's strongest competition comes from sellers of different brands and from its nearest neighbor. |