基于分位数回归的存货质押融资市场风险度量研究

发布时间：2018-04-16 07:31

本文选题：存货质押融资 + 市场风险　；参考：《安徽财经大学》2012年硕士论文

【摘要】：近些年来,物流与供应链金融创新活动在我国的物流业界和金融界都得到了一定的发展,它是一种关于银行、物流企业和企业之间的业务整合的创新模式,它已经是生产流通中的一个潜力非常大的新的经济增长点。而存货质押融资业务是这种创新模式的最核心的形式,所以本文研究的就是存货质押融资模式的市场风险。本文先阐述了存货质押融资的概念、运作模式和面临的风险,并且对质押物样本的对数收益率进行了统计描述和特征研究,结果发现数据有波动成群的特征,而且不服从正态分布和具有平稳性的特征,这些为后面的风险度量在度量方法的选择上提供了重要依据。在研究了存货质押融资的市场风险之后,进一步对市场风险进行了定量研究。VaR方法是目前金融市场风险的测量中运用最普遍的方法,VaR是指在一定置信水平下,在正常市场情况下,所持有资产或者组合在未来特定持有期内的最大损失值。这种方法简单且仅使用一个特定的数字来表示未来的风险值,因此便于管理当局分析和研究,但由于实际金融市场的复杂性,以及近几年来频发的国际性金融危机,单纯的VaR方法已经不能充分并准确地度量市场风险了因此在研究的方法上,本文采用分位数回归来计算VaR值。由前面提到的VaR的定义,可以将VaR看成是金融收益分布的某一分位数。分位数回归在进行回归分析的时候不需要对序列的分布做出某种特定的假设；分位数回归可以实现对于不同的分位点分别进行回归分析,为研究提供一个更为全面的数据分析,特别是对于数据中出现异常点的拟合方面,其作用更为重要。本文在详细讨论了分位数回归的模型,并将其运用到VaR值的计算当中。在不同的置信水平和不同的分位点上,基于分位数回归的VaR模型对市场风险进行了度量,并得到分位数回归的VaR模拟序列,从回归的VaR模型可以明显得出：市场的系统风险和非系统风险在分位数回归的VaR模型中都得到了体现。
[Abstract]:In recent years, the innovation of logistics and supply chain finance has been developed in the logistics and financial circles of our country. It is an innovative mode of business integration among banks, logistics enterprises and enterprises.It has been a new economic growth point with great potential in production and circulation.Inventory pledge financing is the core of this innovative mode, so this paper studies the market risk of inventory pledge financing mode.In this paper, the concept, operation mode and risk of inventory pledge financing are introduced, and the logarithmic return rate of pledge sample is statistically described and studied. The results show that the data have the characteristic of fluctuating in groups.Moreover, we do not agree with normal distribution and stationary characteristics, which provide an important basis for the choice of measurement methods.After studying the market risk of inventory pledge financing, this paper makes further quantitative research on the market risk. VaR method is the most common method used in the measurement of financial market risk at present, which means that VaR is at a certain confidence level.Under normal market conditions, the maximum loss value of an asset or portfolio held in a specific future holding period.This method is simple and uses only one specific number to represent future risk values, thus facilitating management analysis and research, but due to the complexity of actual financial markets and the frequency of international financial crises in recent years,The simple VaR method is no longer sufficient and accurate to measure market riskTherefore, in this paper, the quantile regression is used to calculate the VaR value.By the definition of VaR mentioned above, VaR can be regarded as a sub-digit of the distribution of financial returns.The quantile regression does not need to make a certain hypothesis about the distribution of the sequence, and the quantile regression can carry out the regression analysis for different loci, which provides a more comprehensive data analysis for the study.Especially for the abnormal points in the data fitting, its role is more important.In this paper, the quantile regression model is discussed in detail and applied to the calculation of VaR value.At different confidence levels and different loci, the market risk was measured by VaR model based on quantile regression, and the VaR simulation sequence of quantile regression was obtained.From the regression VaR model, it is obvious that the market system risk and non-system risk are reflected in the quantile regression VaR model.
【学位授予单位】：安徽财经大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.91;F224

【参考文献】