有鑒於以往探討拋補利率平價說時，未將交易成本納入考量，本篇將交易成本納入考慮，利用Tsay(1998)門檻模型，針對1992年至2005年底之新台幣與美元之拋補利率平價(Covered Interest Parity,簡稱CIP)進行分析，來重新驗證CIP是否成立。根據本實證研究，發現不論是單變量模型或是多變量模型來做分析，兩門檻模型可解釋拋補利率平價說。當偏離CIP的年百分比(z值)在所估計的門檻值內，CIP不成立；反之，在所估計的門檻值之外(不可套利區間外)，CIP成立。即一旦套利的利潤超過交易費用、違約風險、政治風險等所產生之交易成本，藉由市場套利力量的運作，讓遠匯貼水與兩國利差過大的差距逐漸縮小並朝向CIP所隱含的關係進行調整。因此在考慮交易成本之因素後，唯有遠匯貼水與兩國利差間的差異過大，兩者才會出現如CIP所指出的調整關係。反之，這兩者變化則無明顯的關係存在。 In this research deviations from covered interest parity are modeled nonlinearly. Unlike previous studies based on linear VCEM or VAR models, we employ a multivariate TAR model based on Tsay(1998)’s paper to investigate the dynamic behaviors of deviations form covered interest parity. The empirical results show a no-arbitrage band within which deviations are random, outside of which deviations revert to the edge of the band. Taking the analysis of double-threshold autoregressions into account, the theory of covered interest parity still stands.
