Abstract
This study applied the Fourier nonlinear autoregressive distributed lag (FNARDL) model to investigate the bilateral J-curve effect between Korea and 18 trading partners. As a result, negative asymmetric effects of changes in the real exchange rate were detected in some cases in the short run. However, in other cases, these effects were positive or exhibited both positive and negative trends. This variation stemmed from the heterogeneity of import and export products and the diverse reactions of exporters to exchange rate fluctuations. By contrast, positive asymmetric effects were evident in most cases in the long run. Through the Fourier function, the FNARDL model better captured the J-curve effect compared to the standard NARDL model. Furthermore, by using bilateral trade data, this study mitigated the problem of aggregate misalignment, compared to using a country's gross trade data with the rest of the world. Therefore, the hidden effects of the real exchange rate were unveiled. This study's findings will enhance policymakers' awareness of formulating exchange rate management strategies.