摘 要

我国证券市场经历了二十六年的发展，现代的股票投资理论也在我国日益普及并得到改善和补充，为证券市场的健康发展解决了一些问题。 但对广大投资者来说依然面临一些问题：怎样才能有效地降低股票投资带来的风险；股票投资风险究竟能够降低到何种程度；如何根据不同的风险偏好选择股票投资额度等。因此，本文根据中国股票市场的发展规律，运用金融财务知识解释其发展趋势，并对证券市场中存在的问题提出针对性的建议，这对中国股票市场的稳定发展以及广大投资者的实际利益来说具有现实及深远的意义。

首先，本文从沪深股市整体的角度研究收益率与风险的关系，以2015年7月1日至2016年3月31日沪深300指数的超额日收益率序列作为研究对象，采用基于Gibbs抽样的MCMC方法对其建立SV-MN以及杠杆SV模型，并通过WinBUGs软件对模型中的参数进行估计，结果发现风险溢价参数的估计值接近于0，经分析可知：波动性对超额收益率影响较小，即波动性并不能很好地诠释股市风险。

其次，为了研究单支股票的收益率与风险之间的关系，本文从沪深市场银行、房地产、计算机和新能源汽车四个板块引入100支样本股，收集这些股票的超额日收益率数据，并运用Matlab软件分别计算他们与沪深300指数超额日收益率序列相关系数的平方，结果发现每个板块中，系统性风险占比超过50%的股票数都不过半。于是考虑加入非系统性风险影响因素：从金融风险、经营风险以及流动性风险等方面搜集100支样本股在研究时段内的净利润率、净资产收益率、市值、流通股本、账面市值比、资产负债率、换手率的季度数据，结合沪深300指数的季度超额收益率，以股票季度超额收益率为因变量，其余为自变量，运用Stata软件分不同板块建立面板数据的多元线性回归模型，最后发现不同板块的回归结果差异较大，这说明投资者对不同板块进行投资时需要关注不同板块的非系统性风险指标。

本文研究结果表明：对整体股市而言，波动性并不能完全代表风险，只用资产收益率的波动来表示风险是不全面的；对个股而言，系统性风险在总风险中占比较小，且对于不同板块系统性风险影响力不同，投资者可通过板块挑选来降低系统性风险；在面板数据的研究中，股票投资收益率与风险呈现出正相关的关系,不同板块的回归结果相差较大，其中银行板块拟合效果最好，而新能源汽车板块的非系统性风险因素均未通过显著性检验，但每一个板块的系统性风险因素都很显著。

关键词：SV模型；相关系数平方；面板数据；多元线性回归

Abstract

Chinese stock market has developed for about 26 years. Modern stock investment theory has been popularized, improved and replenished, which contributes to the healthy development of our stock market. However, investors still face a series of questions: How to reduce the risk of stock investment efficiently; What extent the risk of stock investment can be reduced to; How to choose the limit of stock investment according to personal risk preference. Therefore, according to the law of Chinese stock market development, this paper utilizes financial and economic knowledge to explain its developing trend and proposes targeted suggestions to the problems existing in the stock market, which have realistic and long-term meanings for the stable development of Chinese stock market and the practical interests of investors.

Firstly, from the perspective of the relationship between investment return and risk in Shanghai and Shenzhen stock market, this paper takes the sequence of daily excess return rate of CSI 300 Index from1^st July 2015 to 31^st March 2016 as the research object. It uses MCMC method based on Gibbs sampling method to establish SV-MN model and leverage SV model and estimates the parameters with WinBUGs software. The result shows that the estimated parameter of risk premium is close to 0, which means fluctuation ratio has a relatively small effect to the excess return rate, that is, fluctuation ratio cannot perfectly explain the risk in stock market.

Secondly, in order to research the relationship between investment return and risk of single stock, this paper introduces 100 sample stocks from four sections in Shanghai and Shenzhen stock market, namely Bank, Real Estate, Computer and Green Car. It collects their daily excess return rate and uses Matlab software to calculate the square of the correlation coefficient with the daily excess return sequence of CSI 300 Index, respectively. The result shows that the number of stocks with the proportion of systematic risk over 50% is less than the half. And then, this paper adds some non-systematic factors. From the perspective of financial risks, business risks and liquidity risks, it collects seasonal data of the 100 sample stocks about Net Profit Ratio, ROE, Market Value, Equity of Tradable Shares, Market-to-Book Ratio, Asset-Liability Ratio and Turnover Rate. Combining with the seasonal excess return sequence of CSI 300 Index, this paper uses Stata software to build multiple linear regression models of panel data on different sections with seasonal excess return as the dependent variable and others as independent variables. Finally, the results of regression on four sections are significantly different, which means investors should pay attention to different non-systematic risk factors when they invest stocks in different sections.

Results in this paper reflect that fluctuation ratio cannot perfectly explain risks for the whole stock market, so it is incomplete to represent risk with only volatility of asset returns; For the

certain stocks, the systematical risk takes a relatively low proportion in general risks and has varied effects on different sections, so investors can reduce the systematical risk by selecting different sections; According to the study of panel data, a positive relationship exists between stock investment return and risk. In addition, the regression results on different sections are different: the fitting effect of Bank Section appears to be the best and non-systematical risk factors on Green Car Section cannot pass the significance test, while the systematical risk factors on every section appear to be significant.

Key Words: SV Model; Square of Correlation Coefficient; Panel Data; Multiple Linear

Regression

目 录

摘 要 I

Abstract II

第1章 绪论 1

1.1 引言 1

1.2 本文的研究背景与意义 1

1.3 国内外研究现状 2

1.4 本文的研究思路与内容 3

1.4.1 本文的研究思路 3

1.4.2 本文的研究内容 3

1.5 本文的创新点 4

第2章 沪深股市的风险与收益研究 5

2.1 沪深300超额日收益率序列描述性统计 5

2.1.1 样本数据的选取及处理 5

2.1.2 数据分析 5

2.2 沪深市场收益率与波动性实证分析 6

2.2.1 SV族模型建立 6

2.2.2 实证分析结果 7

2.3 本章小结 8

第3章 各股的风险与收益研究 9

3.1 股票投资风险构成分析 9

3.1.1 样本股票的选择及时限的确定 9

3.1.2 单支股票中系统性风险的研究 9

3.2 基于面板数据的股票投资收益及风险的回归分析 11

3.2.1 指标选取 11