Asymmetric behavior of unemployment rates：Evidence from quantile regression analysis

[[abstract]]本文利用Koenker and Xiao (2004)所建構之分量迴歸單根檢定理論，探討20個主要已開發國家與部分發展中國家失業率之動態行為。面對外生因素影響時，在不同分量下，分量迴歸理論能夠藉由各分量所受衝擊的程度與方向展現出失業率潛在的反應情形。本文實證結果指出，整體而言，失業率不僅呈現出均數複歸性質，並存在非對稱的動態調整行為，亦即在景氣熱絡時的低分量下，負向外生因素傾向使得失業率產生均數複歸特性，而在景氣低迷時的高分量則具有磁滯性質。實證結果也將能夠對於失業率呈現出上升快速但下降遲緩的非對稱走勢提供一份有力的解釋。[[abstract]]This paper applies the regression quantile approach developed by Koenker and Xiao (2004) to investigate the dynamic behavior of unemployment rates in 20 developed countries and developing countries. This method allows us to show potential asymmetric responses of unemployment to exogenous shocks across various quantiles, depending on the sizes and sign of the shocks that hit the unemployment rate. Our results suggest that generally, the unemployment rates are not only mean-reverting but also exhibit asymmetries in their dynamic adjustments, in the sense that in the lower quantiles of economic booms, large negative shocks tend to induce strong mean reversion, while in the upper quantiles of economic recessions, large positive shocks do not, and hysteresis exists. These findings also can explain why unemployment rates display the dynamics of fast rises and slow falls.[[note]]碩