基于中国1993年1月至2010年12月的月度数据,本文采用双变量GARCH模型测量了中国通货膨胀不确定性,并借助于传统的Granger因果检验以及基于马尔科夫区制转移向量自回归(MS—VAR)模型下Granger因果检验等方法对通货膨胀、通货膨胀不确定性与中国实际产出增长进行了实证研究。本文研究发现:中国通货膨胀和通货膨胀不确定性之间存在着Granger意义上的双向因果性,在作用机制上支持了Okun—Friedman理论假说以及Holland论断;传统的Granger因果检验表明中国通货膨胀、通货膨胀不确定性对中国实际产出增长之间存在Granger意义上的因果性,但是基于MS—VAR模型的Granger因果检验研究则表明上述因果性仅存在于特定的机制下。
This paper examines the relationships of monthly inflation, inflation uncertainty and China' s real output growth for the 1993 : 01 - 2010 : 12 period. Based on the bivariate GARCH model, we measure the China' s inflation uncertainty, then discuss different hypotheses on the relationships among inflation, inflation uncertaity and real output growth by using of standard linear and MS - VAR Granger - Causality tests. The results reveal that there are bidirectional Granger causalites beween inflation and inflation uncertainty in China, which supp. ort the Okun - Friedman hypotheses and Holland' s argument The linear Granger - Causality tests show that there are Granger - Causalities from inflation or inflation uncertaity to China' s real output growth, howerver the MS - VAR Granger - Causality tests show that the causalities exist in one regime.