by Gita Gopinath, Pierre-Olivier Gourinchas, Chang-Tai Hsieh, Nicholas Li
To what extent do national borders and national currencies impose costs that segment markets across countries? To answer this question the authors use a dataset with product-level retail prices and wholesale costs for a large grocery chain with stores in the United States and Canada. They develop a model of pricing by location and employ a regression discontinuity approach to estimate and interpret the border effect. They report three main facts: One, the median absolute retail price and wholesale cost discontinuities between adjacent stores on either side of the U.S.-Canadian border are as high as 21 percent. In contrast, within-country border discontinuity is close to 0 percent. Two, the variation in the retail price gap at the border is almost entirely driven by variation in wholesale costs, not by variation in markups. Three, the border gaps in prices and costs co-move almost one-to-one with changes in the U.S.-Canadian nominal exchange rate. They show these facts suggest that the price gaps they estimate provide only a lower bound on border costs.
The authors build a two-country model of horizontal differentiation based on Salop’s (1979) circular city model in which degree of segmentation is endogenous and depends on the sizes of countries, distance from the border and substitutability across locations. The model itself provides numerous novel predictions and can account for a number of properties of prices across borders. The authors then discontinuity regression approach on a scanner price dataset of a large US & Canadian retailer to document how the model can match several interesting properties of prices across borders. A very interesting contribution.