基于GARCH的VaR模型在中国股票市场中的应用-数量经济学专业论文.docxVIP

华中科技大学硕士学位论文摘 华中科技大学硕士学位论文 摘 要 中国股票市场自1990年开市以来的十年间取得了超常的发展，各项指标趋于成 熟，但中国的股票市场毕竟比较年轻，加之体制的原因，市场风险也在不断加大。 特别是中国加入WTO之后，证券市场最终要向外资开放，而外资的流动性、投机性 很强，如不能有效控制势必对我国证券市场造成巨大的冲击。在这种情况下，找到 一个测量和控制风险的有效方法成为各方的当务之急。VaR由于统计意义清晰，使 用方便，受到了越来越多的关注。 VaR利用了统计意义上的自相似性，这不仅涉及到市场的条件问题，还涉及到 现在股价与历史股价相关时间长度以及样本选择等一系列问题。如不解决这些问题， 使用VaR就失去了理论基础。本文利用了分形理论与VaR模型的结合开创性的解决 了市场条件、资产持有期、样本选择以及相关度的问题，为VaR在各个领域内的使用 提供了理论基础。 在此基础上，本文提出了基于t—EGARCH的VaR模型并将其与基于t— GARCH、n·GARCH以及n--EGARCH的VaR模型在上升、下跌两个阶段，分别在检 验样本为500、250两种情况下做了比较。实证证明了通过分形理论确定的样本区间 的正确性，同时也证实了基于t--EGARCH的VaR模型在测量和控制风险方面最有 效。 关键词1 分形理论 R／S检验 VaR t-EGARCH kupiec检验 华中科技大学硕士学位论文Abstract 华中科技大学硕士学位论文 Abstract China’s stock market has achieved great SUCCESS during the last decade，with all indicatrix are getting improved．But for the market’s show history and the transforming regime，investors will face more risks．Especially after China’s entry imo WTO，the stock market will be open to foreign capital in the future．Foreign capital has strong liquidity and speculation．If it is not controlled properly,our stock market will suffer a great loss． So finding an emcient method to measure and control risk becomes emergent．Because of v￡武’s finite statistic meaning and convenience of use．people have paid more attention to V撩model． VaR bases on the autosimilarity in statistics．The method involves problems such as the market condition、the eorrelation period between stock price、sample selection etc．Without solving these problems，we can’t USC VaR properly．In this paper，fractal theory is combined with VaR model to solve these problems successfully．And this makes it possible to 1．kse VaR in other fields． Then fractal theory and VaR model are combincd to solve problems such as the market condition、the correlation period between stock鲥ce、sample selection etc．Then a VaR model based on t-EGARCH was proposed and compared with VaR models based on t-GARCH and n—GARCH．These models were tested in two specific stages，one was during which the stock price was in ascent tr