新興國家短期名目利率之均數復歸

This paper investigates the stochastic properties of short-term nominal interest rates for several emerging countries, and the OECD is a benchmark for comparisons. For that purpose, we employ univariate unit root tests as well as panel unit root tests. Empirical results indicate that univariate unit root tests fail to reject the unit root null at conventional levels; however, the panel test proposed by lm et al. (2003) rejects the unit root null. Because unaccounted cross-sectional dependency may cause a serious over-size problem, we first carry out a formal test for the presence of cross-dependency and, to account for cross-sectional dependency we employ the two panel unit root tests which explicitly allow for cross-sectional dependency: the instrument generating function method developed by Chang (2002) and common factor method proposed by Pesaran (2007). For emerging countries, we find overwhelming evidence that rejects the unit root null, favoring the notion that some countries are mean reverting. Evidence of mean reversion indicates that, instead of VECM-based models, VAR in levels is more appropriate for interest rates modeling (Wu and Chen, 2001). In terms of policy implications, the stationarity of interest rates is consistent with uncovered interest parity (UIP), which implies that a central bank's ability to determine national interest rates will be restricted by the international flows of capital.