互联网金融风险测度与数值分析——基于g-VaR模型

Risk evaluation and numerical analysis of Internet finance based on g-VaR model

与传统理财产品相比,以“余额宝”为代表的互联网金融产品面临的风险更具不确定性。为更好地测度风险,结合g-期望理论与在险价值(VaR),以互联网理财产品为例构建了可用于度量互联网金融风险的g-VaR模型。与传统风险度量模型相比,g-VaR模型在度量互联网金融的不确定性风险方面更具优势。利用求解倒向随机微分方程(BSDE)的Euler隐格式、Crank-Nicolson格式以及预估校正方法给出求解算法和数值分析。结果显示,Euler方法和Crank-Nicolson方法在数值求解BSDE的过程中运算时间长,而预估校正方法能够提升整体运算效率,且计算精度的稳定性较好。

Compared with traditional financial products,the risks of Internet financial products represented by“Yu′e Bao”are more uncertain.By combining g-expectation and value at risk(VaR)and taking Internet financial products as examples,this study established a g-VaR model,which could better measure the Internet finance risks and had more advantages in measuring the uncertainty risk of Internet financial products compared with the traditional risk measurement model.The solving algorithm and numerical analysis were given by using the Euler and Crank-Nicolson methods of backward stochastic differential equation(BSDE)and the prediction correction algorithm.The results show that Euler and Crank Nicolson methods take a long time in the numerical solution process of BSDE while the prediction correction algorithm can improve the overall operation efficiency relatively and has better stability of calculation accuracy.