本文針對追蹤資料分量迴歸模型提出兩階段配適值法，探討央行獨立性和通貨膨脹之關係。本文有以下貢獻：首先，本文提出之模型說明央行獨立性的抗通膨效果在通貨膨脹較高時，效果較大，這可以解釋Franzese (1999) 的理論。其次，本文提出之方法可以解決用央行總裁更替率當央行獨立性變數的內生性問題。因此，本文能完整地分析央行獨立性和通膨之關係。 This paper investigates the empirical relationship between central bank independence (CBI) and inflation by proposing a two-stage fitted value approach to a quantile regression for panel data models. This approach has several advantages. First, Franzese (1999) proposes theoretically that the antiinflationary effect of CBI is heterogeneous and is stronger when inflation is higher. Our method,
which estimates the anti-inflationary effect of CBI for various rates of inflation, can expose the conditional heterogeneity of inflation. Second, a simple two-stage approach to the quantile regression for panel data models is proposed to solve the endogeneity problem by using the turnover rate of a central bank governor as a measure of CBI. Third, by exploiting an extensive panel data set, our
empirical findings show that the anti-inflationary effect of CBI is stronger in higher inflation episodes, and is weaker in lower inflation episodes. As we explore this method, the CBI--inflation relationship becomes more convincing.