具有杠杆效应的非线性SV模型及其应用.pdfVIP

　 18 1 系 统 管 理 学 报 Vol.18 No.1　 　2009 2 Journal of Systems Management Feb.2009　 　　:1005-2542(2009)01-0014-07 SV 1 2 孟利锋 , 　张世英 (1. , 300134;2. , 300072) 【】提出一类非线性SV 模型, 许多离散时间SV 模型都是它的 例。这类模型的优点在于用它可 以检验不同函数形式的随机波动, 该模型的检验仅基于一个单独参数 δ。在非线性SV 模型的基础上, 进 一 步把它扩展为具有杠杆效应的非线性SV 模型。使用沪、深股市的指数日收益数据进行了实证分析, 借助 BU S 软件, 利用 ibbs 取样的MCMC 方法对模型进行了贝叶斯参数估计, 证明了应拒绝对数正态SV 模 型, 而使用非线性SV 模型。 :非线性随机波动模型;杠杆效应;贝叶斯分析;MCMC 方法;对数正态随机波动 :F 062.9　　　:A Nonlinear Stochastic Volatility Model with Leverage Effect and its Application M EN G L i-f eng 1, 　Z H A N G S hi-y ing 2 (1.School of Business, Tianjin University of Commerce, Tianjin 300134, China; 2.School of Management, Tianjin University, Tianjin 300072, China) 【Abstract】A kind of nonlinear stochastic volatility (SV)modelwas proposed, many discrete time SV mod- els are its special cases.The advantage of this kind of model is the ease with which different specification on stochastic volatility can be tested.The specification test is based on a single parameter δ.On the basis of nonlinear SV model, the nonlinear SV model with leverage effect was expanded.Using BU S software, MCMC method with ibbs sampling is used to estimate the new SV model.We empirically tested log- arithm normal SV model against the new one using daily index return data in Shanghai and Shenzhen stock markets.The empirical test rejects logarithm normal SV model and favor the nonlinear SV model. Key words:nonlinear stochastic volatility model;leverage effect;Bayes analysis;MCMC method; logarithm normal stochastic volatility 2 　　(Stochastic Volatili