基于波动择时绩效的已实现协方差预测模型比较
|
摘要
波动择时策略是一种根据资产波动以及相关性构建投资组合的方法,具有较为广泛的应用。鉴于此,提出以波动择时绩效的经济意义指标比较已实现协方差矩阵的预测模型。用高频数据构建股指期货、国债期货和黄金期货的已实现协方差矩阵,利用简单移动平均模型、指数加权移动平均模型和混合数据抽样回归模型对协方差矩阵进行一步向前滚动窗预测,然后在均值-方差框架下根据预测协方差构建动态投资组合,并通过经济效益指标对不同模型的预测进行比较评价。实证结果表明,在股市上升阶段,用简单长期移动平均模型预测已实现协方差矩阵时波动择时策略表现最好;在股市下跌阶段,用简单短期移动平均模型则更优;而用指数加权移动平均和混合数据抽样回归模型时波动择时策略表现则始终居中。
The widely adopted volatility timing strategies construct portfolios based on the volatilities and correlations of assets,and thus we propose to compare realized covariance forecasting models based on the volatility timing performance. High-frequency data are used to construct the realized covariance matrix of stock index futures,T-Bond futures and gold futures. Then the covariance matrix is forecasted using the simple moving average model,the exponential weighted moving average(EWMA)model and the mixed data sampling(MIDAS) regression model with one-step-ahead rolling windows. Portfolios are constructed using these models' forecasts in the mean-variance framework, and then the forecasting models are compared based on the economic values of the corresponding portfolios. Empirical results indicate that the volatility timing based on asimple long-term moving average model outperforms the others while the stock market is going up,the volatility timing based on a simple short-term moving average model outperforms the others while the market is going down,the volatility timing based on the EWMA model and the MIDAS model always has performance slightly inferior to that based on the best simple moving average model.
The widely adopted volatility timing strategies construct portfolios based on the volatilities and correlations of assets,and thus we propose to compare realized covariance forecasting models based on the volatility timing performance. High-frequency data are used to construct the realized covariance matrix of stock index futures,T-Bond futures and gold futures. Then the covariance matrix is forecasted using the simple moving average model,the exponential weighted moving average(EWMA)model and the mixed data sampling(MIDAS) regression model with one-step-ahead rolling windows. Portfolios are constructed using these models' forecasts in the mean-variance framework, and then the forecasting models are compared based on the economic values of the corresponding portfolios. Empirical results indicate that the volatility timing based on asimple long-term moving average model outperforms the others while the stock market is going up,the volatility timing based on a simple short-term moving average model outperforms the others while the market is going down,the volatility timing based on the EWMA model and the MIDAS model always has performance slightly inferior to that based on the best simple moving average model.
引文
[1]Andersen T G,Bollerslev T.Answering the skeptics:Yes,standard volatility models do provide accurate forecasts[J].International Economic Review,1998,39(4):885-905.
[2]Barndorff-Nielsen O E,Shephard N.Econometric analysis of realised covariation:Highfrequency based covariance,regression and correlation in financial economics[J].Econometrica,2004,72(3):885-925.
[3]Bauer G H,Vorkink K.Forecasting multivariate realized stock market volatility[J].Journal of Econometrics,2011,160(1):93-101.
[4]Chiriac R,Voev V.Modelling and forecasting multivariate realized volatility[J].Journal of Applied Econometrics,2011,26(6):922-947.
[5]Corsi F.A simple approximate long-memory model of realized volatility[J].Journal of Financial Econometrics,2009,7(2):174-196.
[6]Golosnoy V,Gribisch B,Liesenfeld R.The conditional autoregressive Wishart model for multivariate stock market volatility[J].Journal of Econometrics,2012,167(1):211-223.
[7]Clements A,Silvennoinen A.Volatility timing:How best to forecast portfolio exposures[J].Journal of Empirical Finance,2013,24(24):108-115.
[8]Fleming J,Kirby C,Ostdiek B.The economic value of volatility timing[J].Journal of Finance,2001,56(1):329-352.
[9]Fleming J,Kirby C,Ostdiek B.The economic value of volatility timing using"realized"volatility[J].Journal of Financial Economics,2003,67(3):473-509.
[10]Morgan J P.Riskmetrics technical document,4th edition[J],J.P.Morgan,New York,1996.
[11]Ghysels E,Santa-Clara P,Valkanov R.Predicting volatility:getting the most out of return data sampled at different frequencies[J].Journal of Econometrics,2006,131(1):59-95.
[2]Barndorff-Nielsen O E,Shephard N.Econometric analysis of realised covariation:Highfrequency based covariance,regression and correlation in financial economics[J].Econometrica,2004,72(3):885-925.
[3]Bauer G H,Vorkink K.Forecasting multivariate realized stock market volatility[J].Journal of Econometrics,2011,160(1):93-101.
[4]Chiriac R,Voev V.Modelling and forecasting multivariate realized volatility[J].Journal of Applied Econometrics,2011,26(6):922-947.
[5]Corsi F.A simple approximate long-memory model of realized volatility[J].Journal of Financial Econometrics,2009,7(2):174-196.
[6]Golosnoy V,Gribisch B,Liesenfeld R.The conditional autoregressive Wishart model for multivariate stock market volatility[J].Journal of Econometrics,2012,167(1):211-223.
[7]Clements A,Silvennoinen A.Volatility timing:How best to forecast portfolio exposures[J].Journal of Empirical Finance,2013,24(24):108-115.
[8]Fleming J,Kirby C,Ostdiek B.The economic value of volatility timing[J].Journal of Finance,2001,56(1):329-352.
[9]Fleming J,Kirby C,Ostdiek B.The economic value of volatility timing using"realized"volatility[J].Journal of Financial Economics,2003,67(3):473-509.
[10]Morgan J P.Riskmetrics technical document,4th edition[J],J.P.Morgan,New York,1996.
[11]Ghysels E,Santa-Clara P,Valkanov R.Predicting volatility:getting the most out of return data sampled at different frequencies[J].Journal of Econometrics,2006,131(1):59-95.