1、 摘要近几年的市场表现，充分证明互联网基金在中国的发展是非常成功的。中国的基金市场在1998年到2013年，积累了8500万累积户数（所有基金产品的开户数总和，同一客户有时会持有多只基金产品）。尤其是在2013年6月后，随阿里集团“余额宝”的推出，在接下来的4年多里，中国基金行业开户数有了飞快的增长。以“余额宝”为代表的互联网基金的开户数进入了爆发性增长期，至2017年底，累积户数已达6亿，仅余额宝的累积户数就高达3.69亿，几乎占到中国人口的四分之一。“余额宝”背后的天弘基金的服务质量和管理水平达到了中国公募基金公司的最高水平，在服务余额宝客户的同时也收获了可观的回报。随着互联网基金的蓬勃发
2、展，探究互联网基金的创新及风险特征有十分重要的理论与现实意义。研究大量的学术文献后，作者发现ARCH族模型主要用于研究债券、股票等传统金融产品的收益率特征，几乎没有学者将ARCH族模型套用到互联网理财产品的研究。在研究传统金融产品的收益波动规律时，GARCH模型的适用性较广。通过构建并拟合EGARCH, TARCH模型，研究传统金融产品的非对称性收益波动时的效果十分理想。另一方面，学术界在尝试风险测度与量化方面，发现VaR模型的结果更接近实际情况，且该方法在实际运用过程中灵活性较高。因此，本文运用ARMA和GARCH类模型来刻画传统的货币基金和互联网货币基金的风险即收益率的波动。并比较二者风险
3、的差别以及深入分析互联网货币基金不同模式下的风险表现。具体而言，本文先进行了基金样本的选取；然后对样本基金的序列特征进行分析，具体包括样本基金的基本统计分析、样本基金的平稳性分析、样本基金的相关性分析、样本基金的ARCH效应检验、样本基金的GARCH模型建立；最后计算理财产品的VaR值得出相应结论。本文实证得到以下结论最新的结论再斟酌一下：第一、传统的基金收益率较为温和，互联网货币基金的收益率变化相对激烈；第二、在对序列进行正态性检验时，无论是传统基金还是互联网货币基金，很明显地都属于偏态分布；第三、金融时序的“尖峰厚尾”特征在我们的序列中得到了充分的体现。从计算的VAR结果来看，传统货币基金
4、的风险值明显地高于互联网货币基金。但是互联网货币基金的风险值的变化范围要大于传统货币基金；第四、本文还对互联网货币基金三种模式的风险值的差异进行了探讨。在实证中确实看不出时序特征和风险值有什么显著差异。这可能在于互联网货币基金受到的大的宏观环境是相同的，并且基金的运作模式区别不大。大多数的货币基金理财产品都存在着严重的同质化现象。关键词：互联网货币基金；风险；VaR；GARCHAbstractThe market performance in recent years has fully proved that the development of Internet funds in Chin
5、a has been very successful. Chinas fund market accumulated 85 million cumulative accounts (the total number of accounts for all fund products from 1998 to 2013, and the same customer sometimes holds more than one fund product). However, in June 2013, with the launch of YuE Bao by Alibaba Group, the
6、number of Internet fund accounts opened into explosive growth for the next four years. By the end of 2017, the accumulated number of accounts reached 600 million. The balance of only YuE Baos household reached 369 million, accounting for almost a quarter of Chinas population. Although YuE Bao is an
7、internet fund, the service quality and management level of the celestial fund behind it have reached the highest level of Chinas raised funds and also rewarded considerable returns while serving the customers. With the rapid development of Internet funds, the innovation and risk characteristics of I
8、nternet funds has very important theoretical and practical significance.Combined with a large number of academic literature, it is found that the ARCH family model mainly studies the yield characteristics of traditional financial products such as bonds and stocks. Few scholars apply the ARCH family
9、model to the research of Internet wealth management products. In studying the volatility of returns of traditional financial products, GARCH model is more applicable. By constructing and fitting EGARCH and TARCH models, it is very effective to study the asymmetry of traditional financial products. O
10、n the other hand, in the field of risk measurement and quantification, the academic circles find that the result of VaR model is closer to the actual situation and the method is more flexible in practical application.Therefore, this article uses ARMA and GARCH models to characterize the risks of tra
11、ditional Monetary Fund and Internet Monetary Fund, and compare the differences between the two risks as well as in-depth analysis of the risk performance of the Internet Monetary Fund in different modes. Specifically, this article first selected the sample of the funds; Then we analyzed the sequence
12、 characteristics of the sample funds, including the basic statistical analysis of the sample funds, the stability of the sample funds, the analysis of the sample funds, the ARCH of the sample funds Effect test, and the sample fund GARCH model was established; Finally, the VAR of the financial produc
13、ts was calculated and the corresponding conclusions were drawn.This paper draws the following conclusions: First, the traditional fund returns are relatively modest, and the changes of internet monetary funds are relatively fierce. Second, when examining the normality of the sequence, it is obvious
14、that neither traditional funds nor Internet monetary funds, belongs to the skewed distribution; Third, the financial timing of the peak thick tail feature in our sequence has been fully reflected. From the calculated VAR results, the risk value of the traditional Monetary Fund is obviously higher th
15、an that of the Internet Monetary Fund. However, the range of the risk value of the Internet Monetary Fund is greater than that of the traditional Monetary Fund. Fourthly, this paper also discusses the differences between the three modes of Internet Monetary Fund. In the early stages of empirical evi
16、dence, we did not see any difference between timing features and risk values. This may be due to the fact that the vast macroeconomic environment that Internet monetary funds are subjected to is consistent with the mode of operation of the fund. Most of the financial products is with serious homogeneity.Keywords: Internet Monetary Fund; risk; VaR; GARCH目录摘要1Abstract2目录4第1章 绪论61.1研究背景61.2研究意义61.3研究内容71.4研究方法81.4.1文献研究法81.4.2模型实证分
