Price gaps at the border: Evidence from multi-country household scanner data

Abstract

Studies employing micro price data to examine the extent of international goods market integration tend to find that borders imply arbitrage-impeding transaction costs, inducing market segmentation. Within monetary unions, these effects are found to be very minor though, at least when online prices are considered. However, analyzing household scanner price data from three euro area member states, Belgium, Germany and the Netherlands, we document that households face (and pay) significantly different prices for identical goods across these countries in the vast majority (around 75%) of cases considered. For regions within countries, as a counterfactual, however, no evidence for such effects exists. Employing cross-border shopping information, we are able to draw direct conclusions about the question of integration of markets for a substantial number of goods. In addition, we can also derive a measure of border costs for most of them. For goods, for which no cross-border shopping take place, we provide a lower bound for border costs in the majority of cases. In line with existing evidence, our findings suggest considerable heterogeneities in border costs. Median values generally range between 15% and 20% (of the price of a good) for exact border cost estimates and between 18% and 20% for the derived lower bounds. Considering the distribution of values obtained suggests that most values lie in a range of ±40%. Since our data set comprises purchase information from all major retailers present in a given market, we are also able to examine the role of retailer heterogeneity for cross-border price gap estimates. Differences in retailer composition turn out to be fairly small, however. Grouping goods by various characteristics reveals that goods purchased more often tend to exhibit somewhat bigger border estimates. Considering the price or the goods category, however, does not yield any conclusive insights.