抛砖引玉一下
希望可以多交流
For the risk correlation matrix, CEIOPS denotes that:
Choosing the correlation coefficient, ρXY
s.t. Min |VaR(X +Y)^2- VaR (X)^2- VaR(Y)^2-ρXY ×VaR(X) ×VaR(Y)|
CEIOPS noticed that achieving this conceptual goal is likely to present a number of practical computational challenges, includes:
1. In most cases the VaR can’t be estimated directly.
2. The parameter, ρXY , is likely to depend on an insurer’s specific insurance risks and so vary across insurers.
3. When more than two risks are aggregated, the minimization of the target equation has to be extended to include all of the combinations of risks
so there may be some adjustments inside