中国股市有多少风险源资料.pdf

中国股市有多少风险源？ ——基于时变风险溢价潜变量模型的研究 王美今1 李传乐2 内容摘要：时变风险溢价潜变量模型是一类重要的资产定价模型，具有很好的性质，它通过确定潜变量个 数来描述整个市场（Market-wide ）的风险源。本文对 Ferson，Foerster 和 Keim（1993）提出的时变风险 溢价潜变量模型的一般性检验方法作出严格证明，并以上证 50 为样本，用该方法对我国股市进行实证分析， GMM 估计的结果表明我国股市场不能拒绝 1 维潜变量模型；文中用 Block-Bootstrap 方法模拟检验统计量 的有限样本性质，并与其渐进分布特征进行比较，结果表明GMM 估计的 Block-Bootstrap 模拟具有相当的 稳健性。本文认为，这一研究结论揭示了我国股票市场风险－收益关系的实质，具有重要的政策涵义。 关键词：时变风险溢价 潜变量模型 Block-Bootstrap 方法 How Many Risk Sources in the Stock Market of China ？ —A Study Based on the Latent Variable Model with Time-Varying Risk Premiums Abstract ：The latent variable model with time-varying risk premium is an important model of the CAPM. With its excellent properties, the model can be used to describe the market-wide risk sources by identifying the number of the latent variables. In this paper the authors give a stringent proof of the general test of the latent variable model with time-varying risk premium raised by Ferson and Foerster. The shanghai 50 Index are selected as the samples and the general test is applied to empirical research of china’s stock market. The results of the GMM show that china’s stock market can not reject the 1 latent variable model. The Block-Bootstrap method is also adopted to study the finite sample properties of the general test and compare it with the result of asymptotic distribution. The results reveal that the Block-Bootstrap method of GMM is robust. The conclusion of this paper manifests the essence of the risk and return in china’s stock market and has great significance for the policy-making of the government. Keywords ：time-varying risk premium The latent variable model Block-Bootstrap method 一、文献综述 Hansen 和 Hodrick（1983），Gibbon 和 Ferson（1985）最先提出时变风险溢价的潜变 量模型，他们将资产的收益期望表达成和具体的经济状态相联系的条件期望，从而 beta 系 数、风险溢价的期望值都可以是时变的；此时决定期望收益的共同因子（时变的风险溢价）