In this paper, we re-examine the impact of the exchange rate fluctuations on export and import prices in Japan, namely the exchange rate pass-through. Instead of using the overall average indices of export and import prices, we conduct vector autoregression analyses using goods-group level export and import prices. Considering the fact that major trade partners might differ by goods-groups, it might not be appropriate to use the standard effective exchange rate measure, which is a weighted average of currencies of trade partners based on their weights in total trade values, in such analyses. To resolve this potential problem, in this paper, we construct a new data set on the “goods-group level nominal effective exchange rate (trade value based)”. On the other hand, it is also possible that what is important is not with whom a country is trading but in which currency the prices are denominated in those transactions. To take account of this idea, we also construct a data set on the “goods-group level nominal effective exchange rate (contracting currency based)”, following the idea of Ito, Koibuchi, Sato, and Shimizu (2009). Using those two types of effective exchange rate indices, we find that the estimation results on pass-through are, sometimes, quite different from what we obtain when we use a more conventional measure of the effective exchange rate.
