摘要
本文主要目的系考量结构改变风险下,对台湾地区费雪效果之再检验。模型的使用选择ZivotandAd-rews(1992)对ADF单根检定之修正模型(简称Z&A修正模型)。该模型的优点,在于允许结构断裂点(break-point)由资料特性内生(endogenizing)决定,而非研究者主观判断。过去文献指出,检验费雪效果之结果,与选用之时间序列资料之定态与否有高度相关,因此本研究使用修正后之Z&A修正模型,更严密的检视资料分析过程之前置检验阶段,将有助于后续分析获得较精确之推论结果。Z&A修正模型发现名目利率、通货膨涨率之水平项即为定态序列,在修正冲击(shock)对序列的长期影响后,两总体变数在长期间是相当稳定的。而其进一步向量自我回归分析结果,显示台湾地区之费雪效果在此一期间并无法被支持。
The purpose of the study is to reexamine Fisher Effect in Taiwan. Considering the nonstationarity and structural break in time series data, we adopt Zivot & Andrews' s sequential ADF test, which is a modification model of traditional ADF unit-root test. The pros of Zivot & Andrews' s test are to determine the breakpoint via the endogenizing instead of subjectively decided by the researchers. The literature shows that the result of examining Fisher effect is highly correlated with the stationary of the time sequence observations. Thus, this research applies the Zivot & Andrews' s test to analyze the previous stage in the whole process to obtain more precise result. Zivot & Andrews' s test found that the level terms for both nominal interest and inflation rates are stationary. After modifying the long-term impact for the time sequence , both the nominal interest and inflation rates are stationary. The Vector AutoRegressive ( VAR) analysis shows that no Fisher Effect is supported in the period for Taiwan.
出处
《当代经济科学》
CSSCI
北大核心
2005年第4期1-10,共10页
Modern Economic Science
关键词
结构断裂
费雪效果
Z&A修正模型
名目利率
通货膨胀率
Structural-break
Fisher effect
Zivot & Andrews' s test
nominal interest
inflation rates