Goods market integration and border effects across European countries with a special focus on retail markets
Statistics and Econometrics
Final Report Abstract
The question to which extent markets are integrated between locations separated by national borders has long been a topic of innovative research in the field of International Economics. In this research project we have contributed to this literature along various dimensions: 1. We have introduced a new data set to the international economics literature. This data set, administered by commercial providers in the retail marketing industry, contains high quality information on shopping behavior of households from Belgium, Germany, the Netherlands and Poland. 2. A new approach to quantify the economic, rather than the statistical, importance of price differences with respect to market integration is developed. Using the concept of Compensating Variation from microeconomic theory, it is investigated if prices differences are associated with large welfare losses (or are washed out as soon as one aggregates to typical household shopping baskets). A major finding is that price differences are not just statistically important: German and Dutch household would need to be compensated with substantial amounts of money if they were forced to live with prices from Belgium. 3. A new method to specify spatial autoregressive models spanning observations from more than one country is proposed. The article proposes to adjust terms in the weighting matrix associated with cross-country location pairs by an adjustment factor c ∈ [0, 1] and then selecting the best adjustment factor using Bayesian Model Selection tools. For prices from Belgium, Germany and the Netherlands the finding is that the weights have to be adjusted downwards substantially - indicating a strong from of market disintegration at national borders. 4. We show that price dispersion for identical goods are sizable across most of our country pairs including the seemingly very highly integrated country pair Belgium and Netherlands. Border effects are largest for Belgium-Germany where prices discontinuously drop by around 20% when crossing the border from Belgium to Germany. For Belgium-Netherlands, this border effect amounts to around 13%. Interestingly, our results suggest only relatively moderate discontinuous price changes between the German-Dutch border. Overall, our estimates indicate larger border within-EMU border effects than those found by Cavallo et al. (2014) for online prices but somewhat smaller values than those documented found for U.S.-Canadian border. In conclusion, our results indicate mixed evidence on market integration. While large (and persistent) differences in prices exist for many of the goods contained in our panel data set, we find that the the u economic importance of these findings is less clear cut. We show on the one hand that the price differences between Belgium and Germany (and between Belgium and the Netherlands) have a sizable effect on the households in Belgium (i.e. Belgian households would benefit a lot from facing the prices of either Germany or the Netherlands). On the other hand the price differences between Germany and the Netherlands are of minor importance to the households in both Germany and the Netherlands.
Publications
- New Econometric Approaches to Measure Market Integration in European Retail Markets. PhD thesis, Goehte University Frankfurt (Germany), 2016
Christian Rühl
- Prices and consumer purchasing preferences at the border: evidence from a multi-country household scanner data set. Center for Financial Studies Working Paper Series, Goethe University Frankfurt, 2016
Günter W. Beck, Hans Helmut Kotz, and Natalia Zabelina
- Size and welfare costs of price differences across european countries. Center for Financial Studies Working Paper Series, Goethe University Frankfurt, 2016
Christian Rühl
