Is devaluation expansionary or contractionary: Evidence based on vector autoregression with sign restrictions

The purpose of the paper is to examine the impact of real exchange rate changes – real devaluation or real depreciation – on outputs in 16 countries that fall within one of the three groups: Latin American countries, Asian countries, and non-G3 developed countries. For the first time in the contractionary devaluation literature, the analysis is based on a Vector Autoregression (VAR) model with sign restrictions method by Uhlig (2005) and Fry and Pagan (2011). The exchange rate shock is identified by imposing restrictions on the signs of impulse responses for a small subset of variables. The findings are as follows: (1) whether output increases after a real devaluation or not has little to do with whether the current account improves or not; (2) Latin American countries are quite homogenous in that the current account generally improves while output decreases after real devaluation; and (3) contractionary devaluation could happen in developed countries as well as in developing countries.