貨幣政策對所得不均之影響-空間分量迴歸之分析

造成所得不均持續擴大的原因其實有很多種，如科技進步、人力資本、全球化及勞動市場改變等原因，然而當景氣過熱或過冷時，政府常使用貨幣政策的變動來刺激景氣，學術界探討其影響時，卻得出正面及反面的結果，表示擴張型或緊縮型的貨幣政策對所得分配程度的影響非單一方向。故本文參考Davide Furceri et. al.,（2016）的文章，研究1998年至2012年全球56個國家其貨幣政策未預期的變動是否會造成所得不均的增加。本文使用吉尼係數的變化量作為被解釋變數，以未預期到的貨幣政策當作解釋變數，先後使用一般迴歸、分量迴歸、空間計量迴歸以及空間分量迴歸，討論貨幣政策的改變是否會影響所得分配程度、貨幣政策決策是否具有空間外溢效果以及會不會因為在不同的國家而有不同的影響程度。最後，本研究得到以下的結論。首先，未預期到貨幣政策緊縮時，短期會減少吉尼係數，中期而言會增加吉尼係數，與Davide Furceri et. al.,（2016）相比相差一期與相差二期的吉尼係數變化為負向效果，此結果與Draghi（2016）剛好相反；再者，使用區域型的資料時，需使用空間計量相關的迴歸，由一般迴歸與空間自相關迴歸的結果可以說明；最後，於空間分量的迴歸結果得知，低分量與中分量的國家其緊縮性的貨幣政策會減少吉尼係數，高分量則會增加吉尼係數，另外其影響程度在經濟發生重大不景氣時，影響特別顯著。

關鍵字

所得不均；貨幣政策；空間分量迴歸

並列摘要

There are many reasons can make income inequality expending, such as technological progress, human capital, globalization and changes in the labor market structure. However, when the economy depressed, the government often uses monetary policy to stimulate the economy, then many article show that the effect of monetary policy has positive and negative results. Therefore, this article follow Davide Furceri et. al., (2016), and studies whether the unexpected changes in the monetary policy of 56 countries in the world from 1998 to 2012 will increase the income inequality. In this paper, the variation of the Gini coefficient is used as the interpreted variable, and the change of the unanticipated monetary policy used as the explanatory variable. By using panel regression, quantile regression, spatial econometric regression and spatial quantile regression, we want to know three point. First, monetary policy change will effects income inequality; second, the monetary policy have the spatial effect; last, the effect of the monetary shock will have different effect in different country. Finally, this study leads to the following conclusions. First, when monetary policy tightening is not expected, the gini coefficient will be reduced in the short term, and increased in the medium term. Compared with Davide Furceri et. al., (2016), the gini coefficient is difference between the first and the second regression, the change is negative effect, and this result is just the opposite of Draghi (2016). Furthermore, when using regional data, spatial regression need to be used. The results of panel regression and spatial autocorrelation regression can be explained. Finally, The spatial quntile regression results show that in the low and middle quantile, the tight monetary policy will reduce the Gini coefficient, while the high quantile will increase the gini coefficient. The monetary shock will significantly effect gini coefficient with the economy in severe recession.