Skewed bivariate models and nonparametric estimation for the CTE risk measure (Q2518541): Difference between revisions

The paper illustrates a numerical method to estimate the Conditional Tail Expectation (CTE) for the sum of two risks. The authors first fit several parametric joint distributions to a bivariate data set consisting of motor insurance claims. Then, they estimate the CTE for the sum of the components. A different nonparametric estimation of the joint distribution function is obtained using a symmetric kernel density estimation method. Numerical issues are also discussed.