我国股票、债券和期货市场波动溢出效应的实证分析——基于GARCH-BVAR模型
|
摘要
基于2002年12月-2013年2月股票成交额、期货成交额以及国债成交额等数据资料,运用GARCH(1,1)模型获取各市场的波动信息,通过贝叶斯VAR模型(BVAR)、脉冲响应函数(IRF)等方法模拟各市场波动的"传染"过程,研究发现:第一,股票市场与期货市场均对来自自身冲击影响的反应比较强烈,期货市场的冲击对股票市场的影响较小但是持续时间长,股票市场的冲击对期货市场的影响较弱;第二,股票和期货市场的冲击对国债市场影响在整个样本期间均不大。
Based on the data of stock turnover,futures turnover and bond turnover from December 2002 to February 2013 as well as the market volatility information obtained via GARCH( 1,1) Model,the infective process of the above-mentioned market volatility is simulated through Bayesian Vector Autoregressive Model( BVAR) and Impulse Response Function( IRF). The results show that both stock market and futures market respond strongly to their self-impact while the impact of futures market on bond market is weak but continues for a long time,with a weaker impact of bond market on futures market; and that the impact of bond market and futures market on the treasury market are quite weak in the whole sample period.
Based on the data of stock turnover,futures turnover and bond turnover from December 2002 to February 2013 as well as the market volatility information obtained via GARCH( 1,1) Model,the infective process of the above-mentioned market volatility is simulated through Bayesian Vector Autoregressive Model( BVAR) and Impulse Response Function( IRF). The results show that both stock market and futures market respond strongly to their self-impact while the impact of futures market on bond market is weak but continues for a long time,with a weaker impact of bond market on futures market; and that the impact of bond market and futures market on the treasury market are quite weak in the whole sample period.
引文
[1]王宝,肖庆宪.我国金融市场间风险传染特征的实证检验[J].统计与决策,2008(11):78-79.
[2]蒋序怀,吴富佳,金桩.当前资本市场的风险传导机制——基于传染效应的实证分析[J].财经科学,2006(2):16-24.
[3]陈燕武,黄静菲.利用非参分位数回归模型分析金融市场的风险传染[J].华侨大学学报(自然科学版),2013(2):215-219.
[4]陈庭强,何建敏,尹群耀,等.基于银行信贷行为的CRT市场信用风险传染评述[J].金融理论与实践,2011(11):95-99.
[5]兰鹏,李铭.波动溢出、风险传染与信息传递效应——基于亚洲股票市场的实证研究[J].哈尔滨商业大学学报(社会科学版),2012(2):16-24.
[6]华桂宏,周茂彬,成春林.银行体系、股票市场与固定资产投资——基于三元向量自回归模型的格兰杰因果关系检验[J].财贸研究,2008(1):90-94,146.
[7]张慧莲.证券投资基金和保险资金对股票市场影响的实证分析[J].金融理论与实践,2010(01):81-85.
[8]THEODOSSIOU P,KAHYA E,KOUTMOS G,et al.Volatility Reversion and Correlation Structure of Returns in Major International Stock Markets[J].Financial Review,1997,32(2):205-224.
[9]KING M A,WADHWANI S.Transmission of Volatility between Stock Markets[J].Review of Financial Studies,1990,3(1):5-33.
[10]HAHM S.Transmission of Stock Returns and Volatility:the Case of Korea[J].The Journal of Korean Economy,2003,5(1):17-45.
[11]PARK Y C,SONG C Y.Financial Contagion in the East Asian Crisis:With Special Reference to the Republic of Korea[R].Korea University Working Paper,2000.
[12]LESAGE J P.Applied Econometrics Using MATLAB[M].Department of Economics University of Toledo,1999.
[13]LITTERMAN R B.Forecasting with Bayesian Vector Autoregressions-Five Years of Experience[J].Journal of Business and Economic Statistics,1986,4(1):25-38.
[14]SIMS C A,ZHA T.Bayesian Methods for Dynamic Multivariate Models[R].Federal Reserve Bank of Atlanta Working Paper,1996:96-13.
[15]ENGLE R F.Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation[J].Econometrica,1982,50(4):987-1008.
1用eviews软件,运用X12方法去除季节性。
2在进行GARCH(1,1)分析之前,我们用三个变量lngp、lnqh、lnzj分别对常数进行回归,然后对残差进行ARCH效应检验,lngp与lnqh回归后的残差均在滞后一期resid^2(-1)的1%水平显著;而lnzj回归后的残差在滞后1-4期均在5%水平下显著。可见,都存在arch效应,可以进行后续的GARCH建模分析。
1按照很多文献的做法,进行VAR建模时一般要求所有变量平稳。此处我们发现gp_ch和qh_ch均是平稳的,只有zj_ch不平稳。原变量是表示波动情况,差分一次后经济意义解释可作相近理解,不影响分析。
[2]蒋序怀,吴富佳,金桩.当前资本市场的风险传导机制——基于传染效应的实证分析[J].财经科学,2006(2):16-24.
[3]陈燕武,黄静菲.利用非参分位数回归模型分析金融市场的风险传染[J].华侨大学学报(自然科学版),2013(2):215-219.
[4]陈庭强,何建敏,尹群耀,等.基于银行信贷行为的CRT市场信用风险传染评述[J].金融理论与实践,2011(11):95-99.
[5]兰鹏,李铭.波动溢出、风险传染与信息传递效应——基于亚洲股票市场的实证研究[J].哈尔滨商业大学学报(社会科学版),2012(2):16-24.
[6]华桂宏,周茂彬,成春林.银行体系、股票市场与固定资产投资——基于三元向量自回归模型的格兰杰因果关系检验[J].财贸研究,2008(1):90-94,146.
[7]张慧莲.证券投资基金和保险资金对股票市场影响的实证分析[J].金融理论与实践,2010(01):81-85.
[8]THEODOSSIOU P,KAHYA E,KOUTMOS G,et al.Volatility Reversion and Correlation Structure of Returns in Major International Stock Markets[J].Financial Review,1997,32(2):205-224.
[9]KING M A,WADHWANI S.Transmission of Volatility between Stock Markets[J].Review of Financial Studies,1990,3(1):5-33.
[10]HAHM S.Transmission of Stock Returns and Volatility:the Case of Korea[J].The Journal of Korean Economy,2003,5(1):17-45.
[11]PARK Y C,SONG C Y.Financial Contagion in the East Asian Crisis:With Special Reference to the Republic of Korea[R].Korea University Working Paper,2000.
[12]LESAGE J P.Applied Econometrics Using MATLAB[M].Department of Economics University of Toledo,1999.
[13]LITTERMAN R B.Forecasting with Bayesian Vector Autoregressions-Five Years of Experience[J].Journal of Business and Economic Statistics,1986,4(1):25-38.
[14]SIMS C A,ZHA T.Bayesian Methods for Dynamic Multivariate Models[R].Federal Reserve Bank of Atlanta Working Paper,1996:96-13.
[15]ENGLE R F.Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation[J].Econometrica,1982,50(4):987-1008.
1用eviews软件,运用X12方法去除季节性。
2在进行GARCH(1,1)分析之前,我们用三个变量lngp、lnqh、lnzj分别对常数进行回归,然后对残差进行ARCH效应检验,lngp与lnqh回归后的残差均在滞后一期resid^2(-1)的1%水平显著;而lnzj回归后的残差在滞后1-4期均在5%水平下显著。可见,都存在arch效应,可以进行后续的GARCH建模分析。
1按照很多文献的做法,进行VAR建模时一般要求所有变量平稳。此处我们发现gp_ch和qh_ch均是平稳的,只有zj_ch不平稳。原变量是表示波动情况,差分一次后经济意义解释可作相近理解,不影响分析。