Research Of Multi-Dimensional CvaR And Preliminary Discussion On Systemic Risk

This thesis mainly studies GARCH model in managing Value-at-Risk(VaR) in Chinese stock market. VaR (Value-at-Risk) is an important meas-ure of financial market risk, and VaR combined with statistical knowledgecan rapidly, accurately and comprehensively predict further quantitativemarket risk. Especially, VaR based on GARCH model has become thefocus of academic research in recent years. As to GARCH model, we usu-ally assume that residual term is a stochastic variable with mean is0andvariance is1. In this paper we just consider the situation that the residual isnormal distribution or t distribution.We not only care for the VaR or CVaR of an asset(asset portfolio),wealso be interested in their correlations of VaR or CVaR between nassets(asset portfolio). So this paper introduces the definition of ndimensional VaR and CVaR to describe their correlations. At last wechoose shangzheng (上证)index and shengzeng(深证) index to calculatetheir2dimensional VaR and CVaR.Finally, we all know that Risk consists two kind risks: systemic riskand un-systemic risk. Investor is also want to know that which part issystemic and un-systemic risk in CVaR. So the last part of the paper is toanalysis the systemic risk and un-systemic risk of CVaR. We define thesystemic risk of CVaR and prove definition reasonable.