Forecasting Industry-Level CPI and PPI Inflation: Does Exchange Rate Pass-Through Matter?

In this paper, we examine whether industry-level forecasts of CPI and PPI inflation can be improved using the ``exchange rate pass-through" effect, that is, when one accounts for the variability of the exchange rate and import prices. An exchange rate depreciation leading to a higher level of pass-through to import prices implies greater expenditure switching, which should be manifested, possibly with a lag, in both producer and consumer prices. We build a forecasting model based on a two or three equation system involving CPI and PPI inflation where the effects of the exchange rate and import prices are taken into account. This setup also incorporates their dynamics, lagged correlations and appropriate restrictions suggested by the theory. We compare the performance of this model with a variety of unrestricted univariate and multivariate time series models, as well as with a model that, in addition, includes standard control variables for inflation, like interest rates and unemployment. Our results indicate that improvements on the forecast accuracy can be effected when one takes into account the possible pass-through effects of exchange rates and import prices on CPI and PPI inflation.