Distribution Costs and International Business Cycles

Backus, Kehoe, and Kydland (International Real Business Cycles, JPE, 100(4),1992) documented several discrepancies between the observed post-war business cycles of developed countries and the predictions of a two-country, complete-market model. The main discrepancy termed as the “quantity anomaly†that cross-country consumption correlations are higher than that of output in the model as opposed to data, has remained a central puzzle in international economics. In order to resolve this puzzle mainly two strategies: restrictions on asset trade, and introducing non-traded goods in the model, have been employed by researchers. While these extensions have been successful in closing the gap to some extent, the ordering of correlations has stayed unchanged: consumption correlations still exceed that of output. This paper attempts to resolve the quantity puzzle by introducing non-traded distribution costs in the retailing of traded goods. In a standard two-good model traded output and traded consumption, by definition, are identical goods. With distribution costs, traded output and consumption are two distinct entities as each unit of final traded consumption good incorporates a unit of traded good and a fixed amount of non-traded goods. Thus, effectively, the model with distribution costs can be viewed as a model without distribution costs but with a modified utility function that has a substantially stronger complementarity between traded and non-traded goods. In a simple two-good extension of the Backus, Kehoe, and Kydland model, it is shown that the cross-country consumption and output correlations are 0.55 and 0.30, respectively, whereas with distribution costs consumption correlation reduces to 0.09, output correlation to 0.23. Incorporating distribution costs, in addition, improves the model’s performance in matching the volatility of real exchange rates and the correlation of net exports with output. These improvements are achieved without sacrificing the model's performance in any other dimension.