巨災債券利差與巨災之關係

Abstract

Catastrophe bonds (CAT bonds) are one of the most important tools in transferring insurance risk to financial markets. This study uses the spread of CAT bond which is the difference between the yields of CAT bond coupon rate and LIBOR to represent the premium caused by the catastrophe. From previous researches, some researchers have found a connection between CAT bond spread and EL. Moreover, EL includes the information of frequency of catastrophe. For these reasons, in this study, we want to determine the representative factors affecting CAT bond spread and discuss whether there is a relationship between CAT bond spread and frequency of catastrophe from June 1997 to March 2013. By applying the most used OLS model, we try to figure out whether the main determinants of the spread at issuance are consistent with the previous researches. Furthermore, we use the Poisson regression and Negative binomial regression model to discuss if there is a strong connection between the frequency of catastrophe and the spread or the same determinants in OLS model. According to our empirical results, we find all specific factors of CAT bond are in line with the past papers except the perils, JP and SR. Besides, we actually can find there is a strong connection between the frequency of wind and spread, and the accuracy of model differs from the data we use in each model. This consequence shows the sponsors have the ability to predict the future disaster well.

Catastrophe bonds (CAT bonds) are one of the most important tools in transferring insurance risk to financial markets. This study uses the spread of CAT bond which is the difference between the yields of CAT bond coupon rate and LIBOR to represent the premium caused by the catastrophe. From previous researches, some researchers have found a connection between CAT bond spread and EL. Moreover, EL includes the information of frequency of catastrophe. For these reasons, in this study, we want to determine the representative factors affecting CAT bond spread and discuss whether there is a relationship between CAT bond spread and frequency of catastrophe from June 1997 to March 2013. By applying the most used OLS model, we try to figure out whether the main determinants of the spread at issuance are consistent with the previous researches. Furthermore, we use the Poisson regression and Negative binomial regression model to discuss if there is a strong connection between the frequency of catastrophe and the spread or the same determinants in OLS model. According to our empirical results, we find all specific factors of CAT bond are in line with the past papers except the perils, JP and SR. Besides, we actually can find there is a strong connection between the frequency of wind and spread, and the accuracy of model differs from the data we use in each model. This consequence shows the sponsors have the ability to predict the future disaster well.