基于VaR和CVaR的我国开放式基金绩效评价

发布时间：2018-01-05 18:29

本文关键词：基于VaR和CVaR的我国开放式基金绩效评价　出处：《浙江工商大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着开放式基金的发展,开放式基金日趋成为主流的投资产品,加之投资开放式基金能够获得免税等有利条件,它逐渐获得大众的青睐。基金的产品种类繁多,我国目前已有的开放式基金更是数以百计。面对纷繁复杂的基金产品,如何做出合理的投资选择,基金的绩效评价就显得尤为的重要。现代的很多投资者都认为基金是一种不可靠的投资方式,这主要是投资者在对基金的认识上还有些偏差。他们认为基金投资基本上是无风险的投资,而事实上任何的理财产品或者投资的金融产品都是有风险的,只是分风险大小不同而已。目前我国现有的基金业绩评价指标过度侧重基金净值,把关注点都放在了基金的收益上,忽略了基金存在的风险。本论文引入了由传统Sharp比率演变而来的基金业绩评价RAROC模型,此模型综合考虑了基金的收益和风险,从而使基金的评价更为客观合理。同时,论文改进了RAROC模型中风险度量方法,从而使模型更加的精确。本论文首先回顾了国内外基金业绩评价的主要理论与研究成果,提出了以CVaR代替VaR作为RAROC模型的风险度量,并在此基础上进行基金业绩评价。然后论文对VaR和CVaR风险度量进行了理论介绍,在理论研究的基础上,论文选取了样本基金股票型基金、混合型基金和债券型基金各8只用于实证分析。这24只样本基金分别用残差假设为T分布和GED分布下的参数模型GARCH、EGARCH、 TARCH和基于峰度法的极值POT模型估计了VaR和CVaR值。实证结果表明,所选参数模型中GED分布下的GARCH模型拟合效果最好；同时,基于峰度法的极值POT模型也能够很好地拟合尾部收益分布。然后,通过Lopez(1998)提出的损失函数方法进行模型比较,发现GED分布下的GARCH模型在估计VaR时优于极值POT模型；而通过定义的ELC统计量的比较,发现极值POT模型在计算CVaR时优于GARCH-GED模型。最后,在模型比较之后,选出最优的风险度量结果计算RAROC值再进行基金的业绩评价。
[Abstract]:With the development of open-end fund, open-end fund becomes the mainstream investment product day by day. In addition, the investment open-end fund can obtain tax exemption and other favorable conditions, it gradually gets the favor of the public. There are many kinds of products of the fund. At present, there are hundreds of open-end funds in China. Facing the complicated fund products, how to make a reasonable investment choice. The evaluation of fund performance is particularly important. Many modern investors believe that the fund is an unreliable way of investment. This is mainly because investors still have some misunderstandings about the fund. They think that the fund investment is basically a risk-free investment, but in fact, any financial product or financial product is risky. At present, the existing fund performance evaluation indicators in China focus on the net value of the fund, focusing on the income of the fund. This paper introduces the RAROC model of fund performance evaluation, which is evolved from the traditional Sharp ratio, which considers the return and risk of the fund. Therefore, the evaluation of the fund is more objective and reasonable. At the same time, the paper improves the risk measurement method in the RAROC model, thus making the model more accurate. Firstly, this paper reviews the main theories and research results of fund performance evaluation at home and abroad, and puts forward the risk measurement of using CVaR instead of VaR as RAROC model. Then the paper introduces the risk measurement of VaR and CVaR theoretically, and on the basis of theoretical research, the paper selects the sample fund equity fund. Eight hybrid funds and eight bond funds are used for empirical analysis. The 24 sample funds use the parameter model GARCH-EGARCH under T distribution and GED distribution respectively. TARCH and the extreme value POT model based on kurtosis method estimate the VaR and CVaR values. The empirical results show that the GARCH model with GED distribution is the best one in the selected parameter model. At the same time, the extreme value POT model based on kurtosis method can fit the tail income distribution well. Then, the loss function method proposed by Lopeza 1998 is compared. It is found that the GARCH model under GED distribution is better than the extreme POT model in estimating VaR. By comparing the defined ELC statistics, it is found that the extreme value POT model is superior to the GARCH-GED model in calculating CVaR. Finally, after the model comparison. The optimal risk measurement results are selected to calculate the RAROC value and then evaluate the performance of the fund.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【参考文献】