【作者】 艾兴；

【导师】 陈守全；

【作者基本信息】 西南大学 ， 应用统计（专业学位）， 2022， 硕士

【摘要】 城市商业银行是我国经济体制从计划经济向社会主义市场经济转型时期的特殊产物。经过长时间的发展,已经是我们国家银行系统的重要组成部分。与大型国有银行和股份制商业银行相比,城市商业银行资本规模较小,发展时间短,收入来源相对更单一。当面临外部风险时,城市商业银行通常更容易陷入困境。在经济全球化的时代,随着我国金融体制的深化改革,城市商业银行面临的市场风险日渐复杂化。过去城市商业银行往往更加重视其所面临的信用风险,而对其所承受的市场风险缺乏定量化的分析。事实上,市场风险已经成为商业银行所面临的第二大风险。相比于国内城市商业银行对市场风险长期的不重视,在西方国家,商业银行在市场风险的计量与管控上明显走在前列。其中,VaR模型在西方国家是已经流行多年的经典的市场风险度量模型,在西方国家防范市场风险中取得有较好的效果。鉴于VaR模型在量化市场风险上的便捷与优良效果,本文将VaR模型运用到了我国城市商业银行的市场风险度量之中。本文介绍了VaR的概念和其计算方法。为了能便捷有效的计算VaR值,本文重点阐述了GARCH模型。同时,阐述了TGARCH模型和EGARCH模型。通过VaR模型结合GARCH类模型,实现市场风险的定量化计算。即是将历史数据所拟合求出的VaR值代表为市场风险的值。对于检测VaR值是否计算准确的问题,本文介绍了kupiec检验法。本文从汇率风险和股票价格风险这两大因素考量城市商业银行的市场风险。在汇率风险的实证分析中,以人民币兑美元汇率作为收益率序列。对数据进行了处理,和建模前的平稳性、自相关性等检验。对收益率序列进行了不同分布下的GARCH模型建模和TGARCH、EGARCH建模。发现尽管收益率序列有着“尖峰厚尾”、“杠杆效应”的特征,只有正态分布下的GARCH模型所计算出的VaR值是最优的。本文以VaR值来代表城市商业银行所面临的市场风险大小,并对城市商业银行未来一天的市场风险的大小进行了预测。在股票价格风险的实证分析中,以宁波银行在A股的日收盘价格为原始数据。同样对数据进行了处理和检验。进行了不同分布下的GARCH、TGARCH、EGARCH建模。发现收益率序列满足t分布下的GARCH类模型,由此计算出的VaR值,无法通过Kupiec检验,容易高估银行的市场风险。收益率序列满足正态分布下的的GARCH类模型拟合效果好,适合用于宁波银行的市场风险量化。无论是汇率风险还是股票价格风险的量化分析,GARCH模型的适用范围均要广于TGARCH和EGARCH模型。在实证分析中还发现,城市商业银行的股票价格风险远大于汇率风险。城市商业银行应该将市场风险的管理重心置于股票价格风险上。通过对城市商业银行的市场风险的定量化分析,希望能对我国城市商业银行的市场风险度量和管理有一些可供参考的建议。更多还原

【Abstract】 City commercial banks are a special product.It was born in the transition period of China’s economic system from planned economy to socialist market economy.After a long period of development,city commercial banks have become an important part of China’s banking industry.City commercial banks are different from large state-owned banks and joint-stock commercial banks.City commercial banks have small capital scale,short development time and single source of income.When facing external risks,city commercial banks are often more likely to fall into trouble.In the era of economic globalization,China’s financial system is deepening reform.City commercial banks’ market risks are becoming more and more complex.In the past,city commercial banks often paid more attention to credit risk,but lacked quantitative measurement of market risk.In fact,market risk has become the second largest risk faced by commercial banks.Domestic city commercial banks do not pay attention to market risk for a long time.In western countries,commercial banks are obviously in the forefront of market risk measurement and control.Among them,VAR model is a classic market risk measurement model that has been popular for many years in western countries.In western countries,it has achieved good results in preventing market risk.Considering the convenience and excellent effect of VAR model in quantifying market risk.This paper applies VAR model to the market risk measurement of city commercial banks in China.This paper introduces the concept of VaR and its calculation method.In order to calculate the VaR value conveniently and effectively,this paper focuses on GARCH model.At the same time,TGARCH model and EGARCH model are described.The quantitative calculation of market risk is realized through VAR model and GARCH model.The VaR value obtained by fitting the historical data represents the value of market risk.For the problem of whether the calculation of VaR value is accurate,this paper introduces kupiec test method.This paper considers city commercial banks’ market risk from two factors:exchange rate risk and stock price risk.In the empirical analysis of exchange rate risk,the exchange rate of RMB against US dollar is taken as the yield series.The data are processed and tested for stationarity and autocorrelation before modeling.GARCH model modeling,TGARCH and EGARCH modeling are carried out for the return series under different distributions.It is found that although the yield series has the characteristics of "peak and thick tail" and "leverage effect".However,only the VaR calculated by GARCH model under normal distribution is the best.This paper uses VaR value to represent the market risk faced by city commercial banks.This paper forecasts the market risk of city commercial banks in the next 24 hours.In the empirical analysis of stock price risk,the daily closing price of Bank of Ningbo in A shares is taken as the initial data.Similarly,the data are processed and tested.GARCH,TGARCH and EGARCH are modeled under different distributions are carried out.It is found that the VaR value calculated by class of GARCH model under t-distribution can not pass Kupiec test.It is easy to overestimate the market risk of banks.The GARCH model under the normal distribution of the yield series has a good fitting effect.It is suitable for market risk quantification of Bank of Ningbo.No matter the quantitative analysis of exchange rate risk or stock price risk,GARCH model is more applicable than TGARCH and EGARCH models.In the empirical analysis,it is also found that the stock price risk of city commercial banks is much greater than the exchange rate risk.City commercial banks should put the focus of market risk management on stock price risk.This paper makes a quantitative analysis of city commercial banks’ market risk.I hope to have some suggestions for reference for the market risk measurement and management of China’s city commercial banks.更多还原