Shock Dependence of Exchange Rate Pass-through: A Comparative Analysis of BVARs and DSGEs

Abstract

In this paper, we make use of the results from Structural Bayesian VARs taken from several studies for the euro area, which apply the idea of a shock-dependent Exchange Rate Pass-Through, drawing a comparison across models and also with respect to available DSGEs. On impact, the results are similar across Structural Bayesian VARs. At longer horizons, the magnitude in DSGEs increases because of the endogenous response of monetary policy and other variables. In BVARs particularly, shocks contribute relatively little to observed changes in the exchange rate and in HICP. This points to a key role of systematic factors, which are not captured by the historical shock decomposition. However, in the APP announcement period, we do see demand and exogenous exchange rate shocks countribute significantly to variations in exchange rates. Nonetheless, it is difficult to find a robust characterization across models. Moreover, the modelling challenges increase when looking at individual countries, because exchange rate and monetary policy shocks (also taken relative to the US) are common to the whole euro area. Hence, we provide a local projection exercise with common euro area shocks, identified in euro area-specific Structural Bayesian VARs and in DSGE, extrapolated and used as regressors. For common exchange rate shocks, the impact on consumer prices is the largest in some new member states, but there are a wide range of estimates across models. For core consumer prices, the coefficients are smaller. Regarding common relative monetary policy shocks, the impact is larger than for exchange rate shocks in any case. Generally, euro area monetary policy plays a big role for consumer prices, and this is especially so for new member states and the euro area periphery.