Regressions of price differences between locations in different countries without controlling for the local market structure and the location of origin of the product will lead to a biased estimate of the impact of national boundaries. We demonstrate that non-classical measurement error in distance and unaccounted mark-up differences across countries are responsible for these biases. In a quantitative exercise based on our previous work (Cosar et al., 2014), we show that the estimated border effect with price difference regressions overstates the true border effect by a factor of two or more.