Nonlinear models for the sources of real effective exchange rate fluctuations: Evidence from the Republic of Korea

Abstract

The real effective exchange rate has exhibited sizable fluctuations from 1970Q1 to 2015Q1 in the Republic of Korea. This paper quantitatively investigates the sources of the fluctuations and the impacts of heterogeneous shocks on the conditional volatility of the real effective exchange rate through estimating a structural vector autoregressive (SVAR) model and an exponential generalized autoregressive conditional heteroskedasticity (EGARCH) model. The variance decomposition and historical decomposition consistently demonstrate that about 90% of the real effective exchange rate fluctuations is attributed to demand shocks, about 9.10% of the fluctuations is due to supply shocks, while only around 0.74% of the variations is because of nominal shocks. Moreover, the historical decomposition reveals that these contributions are time-varying. In addition, the estimated EGRACH model reveals that the conditional volatility of the Korean real effective exchange rate in response to negative shocks is much larger than that in response to positive shocks. Our analysis implies that temporary capital flow regulations and discretionary monetary policies should be combined to stabilize the real effective exchange rate fluctuations in the presence of negative demand shocks.