基于MCMC算法的人民币汇率市场的分析——双门限非对称GARCH模型的应用

本文旨在考察,汇改后美元/人民币汇率前期收益的影响下,人民币汇率市场上非美元/人民币汇率收益均值和波动不对称的程度。为了捕捉非美元汇率收益的均值和波动不对称的特点,我们设定双门限非线性的GARCH模型,结合GJR效应(即加入非美元收益利空或利好消息的影响),利用基于MCMC算法的贝叶斯推断来完成。应用中我们选取了美元(欧元、日元、港元)/人民币日汇率数据进行分析,发现了门限非线性的结果,表明在美元和非美元汇率本身双重变化的影响下,非美元汇率收益的均值和波动同时表现出非对称的特点。并且在美元收益利好消息的影响下,美元汇率对非美元汇率的溢出效应明显增强,非美元表现出低均值回归的特点。

Analysis of RMB Exchange Rate Market Based on MCMC Algorithm——Application of Double Threshold Non-linear GARCH Model

This paper examines the extent to returns on RMB exchange rate exhibit mean and volatility asymmetries,as a response to past return of dollar/RMB exchange rate.In particular,we wish to assess the asymmetric effect,we use the Bayesian Inference and propose a double threshold non-linear GARCH model,combined with a GJR effect(i.e.,influenced by past return of non-dollar),to capture both mean and volatility asymmetric behaviours.In an application to four returns of exchange rate,clear evidence of threshold non-linearity is discovered,which shows the asymmetric behaviours under the influence of two factors.This stronger spillover effect from the U.S.and lower mean persistence are following a good news of US dollar.