Abstract

Recent advances in time-series and cointegration analysis have allowed for the estimation of the nonlinear effects of currency depreciations on countries’ trade balances. This study examines Mexico’s bilateral trade with 13 trade partners, applying both the linear and the nonlinear versions of the Autoregressive Distributed Lag (ARDL) cointegration and error-correction methodologies, to test whether peso appreciations affect trade differently than do peso depreciations. The linear ARDL model generates significant long-run coefficients for six partners, implying that while peso depreciations improve Mexico’s trade balance with them, peso appreciation hurt it. However, the nonlinear ARDL model finds that peso appreciation hurts Mexico’s trade balance with three additional partners, and that peso depreciation improves it with one additional partner. Additional statistical tests reveal that indeed the effects of peso appreciation are significantly different than effects of peso depreciation, supporting the asymmetric effects of real exchange rate changes.