MIDAS-Expectile Regression And Risk Management

Potential financial risks may lead to the occurrence of financial crises,with catastrophic consequences.Throughout history,many large-scale financial crises have impacted the economy and society,and every corner of the society has been seriously affected.Driven by the globalization wave,countries are now more closely connected,and in particular,financial markets are becoming increasingly open and integrated.Although this change has promoted the efficiency of capital flows and resource allocation,it has also made the financial crisis spread faster among financial markets in various countries.The financial crisis in a single financial market may affect a wider range of financial markets.As a result,financial risk measurement and management is being valued by more and more regulators and financial practitioners.In the field of financial risk,how to identify and measure risk is the first problem to be solved.It is especially important to choose the appropriate risk measurement method to accurately measure risk.After years of development,some mature risk measurement methods have been widely used in academic research and the financial industry,such as VaR,volatility,etc.are often used to measure financial risk.However,these risk measurement methods still have limitations,so some new risk measurement methods are proposed,such as EVaR,ES,etc.As an alternative technology to the QVaR(based on quantile),the risk measure EVaR(based on the Expectile model)is simpler to calculate and can more accurately reflect the effects of extreme values.In order to make full use of the information contained in different frequency data,this paper introduces the MIDAS method into Expectile regression,and builds the MIDAS-Expectile regression model.In this paper,we estimate the parameters and conditional EVAR based on the nonlinear asymmetric least squares method,and give the asymptotic normality of the estimate and the coverage test of the conditional Expectile.In addition,this paper also gives the likelihood function and information criterion of the Expectile regression model from the perspective of maximum likelihood estimation,which can compare and test different models.This paper uses Monte Carlo simulation to test the validity of the estimation.The results show that the introduction of MIDAS method under a variety of model settings has a significant improvement over traditional Expectile regression.In addition,in the empirical research,this paper applies the proposed MIDAS-Expectile regression model to explore some empirical problems of financial risks.First,this article applies the MIDAS-Expectile regression model to the measurement of cryptocurrency returns risk,and at the same time explores the risk contagion of other traditional financial markets to cryptocurrency,an emerging financial asset.Then,based on the MIDAS-Expectile regression model,this paper studies the financial risks in the major stock markets,especially the risk contagion among markets.Empirical research shows that MIDAS-Expectile regression model can get accurate risk measurement,which is more suitable for empirical data than traditional equal-weight Expectile regression.