Financial Risk Management Based On Quantile Regression

In recent years, as countries to speed up the pace of economic globalization, the relevance of financial markets are constantly reinforced. The fusion of financial markets accelerated capital flow and enhanced the efficiency of resource allocation. It also brought the unprecedented risk to foreign financial markets at the same time. Under the circumstances, the financial crisis in some countries will rapidly infect other countries, and trigger the global financial crisis. As the financial market plays a growing role in the economy, the financial risk management has become the focus of financial institutions and academic circles. How to control risks effectively has become the core content of financial research and practice.In the field of financial risk management, VaR is the universal indicator. This paper puts forward the quantile regression EGARCH model (Q-EGARCH model),which does not only have the advantage that is not affected by the yield distribution hypothesis, but also can reflect the effects of asset returns volatility to VaR directly. The Q-EGARCH model is able to accurately calculate the VaR under different levels of quantile, and makes the calculation of VaR value is more close to the real market situation.The conventional parameter estimation methods are useless in nonlinear quantile regression parameter estimation. Therefore, this paper introduces the application of MM algorithm in nonlinear quantile regression parameter estimation detailedly in the third chapter. And demonstrates the robustness and superiority of MM algorithm in nonlinear quantile regression parameter estimation by the monte carlo simulation. The simulation test results show that MM algorithm is more superior than the conventional interior-point method of nonlinear quantile regression under the low quantile level.In view of the advantage of MM algorithm in nonlinear quantile regression parameter estimation. This paper discusses the application of the MM algorithm Q-EGARCH model. And the MM algorithm is used to estimate the Q-EGARCH model thus the MM algorithm is used to estimate the parameters of Q-EGARCH model, and then calculate the corresponding value at risk in the fourth chapter. In this paper there is an empirical study based on the csi300 index futures. The result of study shows that the Q-EGARCH model is more accurate than other VaR calculated methods on the VaR calculation under the confidence level of 95% and 99%. Especially, the Q-EGARCH model is more obviously advantaged under the 99% confidence level.Empirical analysis result shows that the new structure of Q-EGARCH model is more accurate to calculate VaR. And the financial institutions can be more effective in risk management by using the Q-EGARCH model, especially for the aversive risk of financial institution, this article has a very valuable reference value.