Research On Catastrophe Bond Pricing Based On POT Model And Copula Model

In recent years,with the changes of the natural environment and the intensification of human social activities,various catastrophic events have occurred frequently,causing serious economic losses to all countries.Due to its vast territory,complex terrain and varied natural climates,China is one of the most affected countries with the largest number of natural disasters in the world.Once a catastrophe event occurs,the impact is often catastrophic.People may be displaced by catastrophe,and the government needs to bear the main responsibility of social assistance after the catastrophe,which leads to the national finances being tight and even overwhelmed.In the early days,countries mainly relied on insurance and reinsurance companies to share catastrophe losses.But with the sever losses increase,the insurance and reinsurance companies' underwriting capacity became increasingly inadequate,and even lead to bankruptcy for huge claim policies caused by catastrophe.In order to meet the growing demand for catastrophe insurance,people begin to look for new catastrophe risk management tools.Because capitals in financial markets have many advantages on quantity,as well as fluidity.The method to transfer risk from insurance markets to capital markets via catastrophe-linked bonds would become an important way to hedge catastrophe risk.Among them,catastrophe bonds are by far the most successful and the significant financial securities products.Therefore,pricing catastrophe bonds are not only of theoretical value,but of practical significance.It can provide technical support for the issuance of catastrophe bonds in China,thereby improving China's catastrophe risk transfer capacity and reducing government financial pressure.Considering the diversified needs of current capital market investors and different risk preferences,this paper constructs three different catastrophe bonds with high to low risk levels:single event triggered catastrophe bonds,double event triggered catastrophe bonds and multiple events triggered catastrophe bonds.They can effectively meet the market demand for catastrophe bonds of different risks,thereby increasing the market capacity of catastrophe bonds.The main work and conclusions are as follows:First,using the risk-neutral approach,the research constructs a single event triggered catastrophe bond pricing model by the POT model and the Vasicek interest rate model.The model makes up a shortcoming that the related research on the existing catastrophe bonds lacks the verification of the POT model with the catastrophe loss prediction effect.This paper conducts an empirical analysis of the VaR estimation and prediction effect of the death toll data caused by the global floods.The results of backtesting test show that the description of thick tail characteristics is the key to accurately estimate the quantile,and the POT model can provide more accurate catastrophe loss estimation.On this basis,the POT model is used to characterize the thick tail of the flood deaths,and the Monte Carlo simulation method is used to calculate the bond price without the analytical solution.Finally,the applicability and feasibility of the model are verified from the aspects of pricing examples and parameter sensitivity analysis.Second,using the POT model,Archimedean Copula model and CIR interest rate model,the pricing model of double event triggered catastrophe bond is constructed under the risk neutrality measure.Among them,the POT model is used to describe the thick tail characteristics of the catastrophe loss variable,and the Archimedean Copula model is used to describe the correlation between the catastrophe loss variables,then the CVaR estimation of the catastrophe loss is obtained.On this basis,the CVaR estimation is applied to the trigger value setting of the dual event triggered catastrophe bond,which makes up for the lack of correlation between the trigger values and the existing related research when setting the trigger values.Finally,an empirical analysis is carried out based on the economic losses and disaster area data caused by global floods.The pricing examples,parameter sensitivity analysis,and comparison with the price of dual event triggered catastrophe bonds without CVaR estimation are compared and verified.The model tests the necessity of triggering the correlation of values.Third,based on the vine Copula model and the Longstaff interest rate model,a multiple(three or more)event triggered catastrophe bond pricing model is constructed under the risk neutrality measure.Among them,the vine Copula model is mainly used to describe the correlation between multiple catastrophe loss variables,and then obtain more complicated catastrophe loss CVaR estimates.The empirical analysis of CVaR,CES estimates and predictions is performed combined with the disaster area,duration and economic loss data caused by global floods.The results of backtesting test show that the characterization of multivariate correlation is the key to accurately estimate the conditional quantile,and the vine Copula model can provide more accurate estimates of the catastrophe loss condition quantile.On this basis,the CVaR estimation based on the vine Copula model is applied to the trigger value setting of multiple event-triggered catastrophe bonds,which fills up the blank that there are no multiple event catastrophe bond pricing models in the existing catastrophe bond related research.Finally,the feasibility of the model is tested from the aspects of pricing example and parameter sensitivity analysis.And the dissertation carrys out comparative analysis between the single event,double event and multiple event triggered catastrophe bond pricing model in this paper.In summary,by considering the thick-tailed features and correlations of the catastrophe loss variables,the dissertation carrys out the pricing of catastrophe bonds according to the degree of risk from single event triggers,double event triggers,and multiple events through the POT model,the Archimedean Copula model,and the vine Copula model.The results is to enrich catastrophe bond products and meet the needs of investors with different risk preferences.The empirical results show that the catastrophe bond pricing model constructed in this paper has certain advantages over the existing pricing model in terms of the thick tail of the catastrophe loss variable,the correlation between the loss variables and the setting of the trigger value.The research results can provide new ideas for catastrophe-related management departments and increase the market capacity of catastrophe bonds.Therefore,the research in this paper has some application values and policy promotion prospects.