Is There a Threshold in the Exchange Rate-Current Account Nexus? Evidence from Korea

This study tests whether Korea's aggregate current account responds to the exchange rate in a level-dependent way, unlike the directional asymmetry documented in bilateral trade balances. Combining Hansen's (2000) threshold regression with Hansen's (1996) residual bootstrap on quarterly data for 2000–2025, we find no threshold in the nominal KRW/USD rate. In the real effective exchange rate (REER), a single threshold near REER is marginally significant. Below that level, error correction is fast, and both appreciation and depreciation shocks move the current account; above it, the channel weakens sharply. The threshold is not sharply localized, and its significance depends on the trimming parameter, so we read the evidence as suggestive rather than confirmatory of regime dependence.本研究旨在检验韩国总体经常账户是否会以水平依赖的方式对汇率做出响应，这与双边贸易余额中已被文献证实的方向非对称性有所不同。我们结合汉森（Hansen）2000年提出的阈值回归模型与汉森1996年提出的残差自助法，对2000至2025年的季度数据开展实证分析。结果显示，韩元兑美元（KRW/USD）名义汇率不存在阈值效应；在实际有效汇率（real effective exchange rate, REER）层面，则存在一个接近该汇率水平的单一阈值，且其显著性处于边际水平。当实际有效汇率低于该阈值时，误差修正速度较快，升值与贬值冲击均会对经常账户产生影响；而当实际有效汇率高于该阈值时，这一汇率对经常账户的传导渠道会大幅减弱。该阈值并未呈现出清晰的局部性特征，且其显著性依赖于截尾参数，因此我们认为现有证据仅能提示存在状态依赖的可能性，而非证实该机制的存在。