Hi, I'm struggling to understand why we underestimate the variability when we look to combine our risks and there is correlation but are effected by similar events, legal, claims handling team and data issues versus the fact that we have diversification benefits when combining these segments of risk and so the variance of the sum of the risk is greater than the sum of the standalone risks. Could somebody explain how the two sections tie in together please? Much appreciated. Kind Regards, Joe
Hi Joe, I'm afraid that was rather a long sentence and it didn't quite make sense. I interpret your question to be: "Why do we need to allow for (positive)correlations as well diversification?" The overriding idea is that we should model all the interractions between risks, whether these increase the variability of the insurer's experience or decrease it. (This is discussed in Chapters 18 and 19.) Mind you, if we dig a little deeper we also need to consider the uses of the capital model. Most importantly, we would model the company's capital requirement for statutory reasons (allowing for both positive and negative correlations between risks), but then we'll also use the capital model for a variety of other purposes. For example, when we price a product we need to include a loading for the cost of capital. But, if the product brings a diversificaton benefit then this loading will be lower than if the product was written in isolation. Hence, the premium charged will be lower. The allowance that is made for diversfication will therefore depend on the purpose of the exercise, and this is discussed in Chapter 20.
