本文利用ARFIMA-APARCH模型對中國黑市匯率的非對稱性和杠杆效應進行了研究,發現黑市匯率取對數後形成的序列表現為偏度和峰度共存,如果在有偏學生分佈下對A-PARCH進行估計,參數檢驗的顯著性會大大增強,並證實了ARCH模型簇往往將對稱t分佈、GED誤差分佈與不對稱衝擊結合在一起來考察。
This paper estimates ARFIMA and APARCH models of the black month exchange rate in China against the USD for asymmetry and leverage effect, it is found that both leptokurtic and fat tail exist, if A-PARCH is estimated through a skew student's distribution based maximum likelihood estimation, the significance of the parameter test may greatly strengthen, and the characteristics of asymmetry and leverage effect of the mean process and volatility process of the black exchange rate of China are captured. Conclusions suggest that both stochastic volatility and a flexible series distribution have to be taken into account.