From the Talmud and Markowitz to a holistic picture of risk

Most people are taught to break down a complex problem by immediately splitting it into smaller bite-size chunks, making it easier to analyze and then re-aggregate to understand the original complex problem. The actuarial control cycle is often performed in this manner.

Centuries of insurance business have allowed a wealth of data to be collected for various risks such as mortality, health, natural disaster, and motor accident, which allows statistical analysis and reaffirms that certain risk events become somewhat predictable when viewed in sufficient numbers. However, statistical models are arguably victims of their own success. In some cases there can be over-reliance on the model, which can lead to some significant disadvantages.

Systems thinking is a promising alternative to the reductionist approach. This article gives an example (asset allocation) of where statistical models often struggle, and then give an overview of the systems thinking approach.

-

This article was published by InsuranceERM (subscription required).