名目利率恆定性再檢測：跨國的實證分析

本文採用 Tsong (2012) 所提出的共變量非對稱指數平滑轉移模型，以十七個APEC組織國家，探討名目利率是否為恆定。相較於以往傳統單根檢定方法在小樣本下會有檢定力不足的情形，本文採用的檢定量藉由經濟理論挑選相關共變量，以及模型考量數列在收斂過程中可能存在非對稱調整等方式，以提升檢定力。其中，根據經濟理論挑選的相關共變量為通貨膨脹率、實質利率、國內生產毛額平減指數、失業率，每一次的名目利率單根檢定只加入一個恆定的共變量，並在之後輔以t統計量來判定是否存在非對稱調整。實證結果顯示，當以實質利率取差分時，有十四個國家顯示名目利率為恆定，僅馬來西亞、巴布亞紐幾內亞、台灣等三國顯示名目利率存在單根。非對稱調整的檢定，除了巴布亞紐幾內亞、台灣因名目利率不為恆定無法做進一步檢定，其餘十五個國家名目利率皆存在非對稱調整。

關鍵字

單根檢定；共變量；非對稱ESTAR ；名目利率

並列摘要

This paper investigates the stationarity properties of nominal interest rates for 17 APEC countries by using the CF test advocated by Tsong (2012). The CF test exams the unit root null against the alternative of asymmetric STAR nonlinearity with correlated covariates for power boost. The selected covariates are based on economic theory, including – inflation rate, real interest rates, GDP deflator, and unemployment rate. When implementing CF test, we include a single covariate each time and reported the corresponding result. The empirical results reveal strong evidence that the nominal interest rates, are mean reverting, rejecting the unit root null for 14 out of the 17 countries when the covariate is the differenced nominal interest rate. Moreover, most of the nominal interest rates are tested to display asymmetries in the adjustment process towards their equilibrium values.