基于状态空间模型的中国上市公司信用风险研究

发布时间：2018-12-30 16:48

【摘要】：自2007年美国次贷危机以来，金融风暴全球蔓延，给全球的金融市场带来了恐慌和沉重的打击，一大批公司（尤其是银行和金融机构）由于资产质量急剧下降而遭到信用评级下调甚至破产，，再加上近期美国国债评级下调和欧洲主权债务危机蔓延，全球金融市场雪上加霜，许多发达国家经济萧条，失业率高居不下。这些事件表明，信用风险度量模型缺陷和信用衍生工具的快速发展以及信用风险监管的滞后成为这次金融风暴的主要原因之一。虽然中国受这次金融危机的影响较小，但目前市场上一些房地产公司资产负债率较高、地方政府融资平台现金流预期短缺、资金面缩紧的市场环境下公司以高成本从地下钱庄融资等等，这些迹象表明信用风险正在中国市场积聚。因此，有关信信用风险度量模型的改进和加强信用风险监管的措施已成为近期研究的焦点。信用风险度量问题一直是风险管理中的一个难题，信用风险模型从理论上可以分为两种，一种是简约化模型（Reduced-Form Model），一种是结构化模型（StructuralModel）。在实践应用中，由于结构化模型以公司的资本结构为基础，在评估违约风险时具有简约化模型无可比拟的数据获取优势，因而受到广泛推崇。针对以上问题，本文以结构化信用风险模型为基础，考虑资产价值在信息冲击下的跳跃行为和股票价格的交易噪音，构建状态空间模型分析上市公司的信用风险。运用沪深股市A股上市公司的数据进行实证分析，结果表明ST公司的资产价值在ST期间会存在显著的跳跃，并且相对于非ST公司，ST公司资产价值整体跳跃更明显。另外，所有上市公司股票价格都存在交易噪音，如果不考虑交易噪音的影响，会低估其信用风险。但也存在一些上市公司，即使被ST，其信用风险不大，相反，有些没有被ST的公司，其信用风险却比较大。这说明，在度量上市公司违约可能性大小时，不能仅仅依靠是否被ST作为评判的标准，还应该考虑资产价值跳跃和股价交易噪音的影响，这样才能更好地区分上市公司的真实风险。
[Abstract]:Since the subprime mortgage crisis in the United States in 2007, the financial turmoil has spread all over the world, causing panic and heavy blows to the global financial markets. A large number of companies, especially banks and financial institutions, have suffered credit ratings downgrades or even bankruptcies due to a sharp decline in asset quality, compounded by the recent downgrade of US Treasuries and the contagion of the European sovereign debt crisis. Many developed countries have depressed economies and high unemployment rates. These events indicate that the defects of credit risk measurement model, the rapid development of credit derivatives and the lag of credit risk supervision become one of the main reasons for the financial storm. Although China was less affected by the financial crisis, some real estate companies in the market now have relatively high asset-liability ratios, and local government financing platforms are expected to be short of cash flow. In a tight capital-market environment, companies are raising money from underground banks at high cost, and so on, suggesting that credit risk is building up in the Chinese market. Therefore, the improvement of credit risk measurement model and the measures to strengthen credit risk supervision have become the focus of recent research. Credit risk measurement is always a difficult problem in risk management. Credit risk models can be divided into two types theoretically, one is reduced model (Reduced-Form Model),) and the other is structured model (StructuralModel). In practical application, because the structured model is based on the capital structure of the company, it has the advantage of data acquisition which is incomparable to the simplified model when evaluating default risk, so it is widely respected. Based on the structured credit risk model, considering the jumping behavior of asset value under the impact of information and the transaction noise of stock price, this paper constructs a state-space model to analyze the credit risk of listed companies. By using the data of A-share listed companies in Shanghai and Shenzhen stock markets, the results show that there is a significant jump in the asset value of ST during the ST period, and compared with the non-ST companies, the overall jump in the asset value of ST companies is more obvious. In addition, trading noise exists in the stock prices of all listed companies, and credit risk is underestimated if the impact of trading noise is not taken into account. But there are also some listed companies, even if by ST, its credit risk is not big, on the contrary, some companies that have not been ST, its credit risk is relatively big. This shows that when measuring the probability of default of a listed company, it should not only depend on whether it is judged by ST, but also consider the impact of the jump of asset value and the noise of stock price trading. Only in this way can the real risks of listed companies be better distinguished.
【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F832.51;F224

【参考文献】