Econometrics, model uncertainty and risk management in actuarial applications

点击翻译按钮获取中文摘要

Actuaries use models to quantify and manage risks in insurance and finance. Accurately modeling financial risks is challenging because financial processes exhibit complex dynamics which are difficult to reproduce with statistical models. When developing a model, actuaries face a tradeoff between its complexity and implementation. The first objective of my research program is to improve this tradeoff by investigating new model structures in finance and by developing original methods to implement them.****Models are often used to project financial and economic variables. For example, actuaries need to simulate returns on financial assets to illustrate, price and manage risks in the context of life insurance and pension plans. To generate realistic projections, the model must be calibrated to a data set. However, the choice of the model and data set can significantly influence the financial projection. To reduce the risk of using an inappropriate model, the output of several models can be compared. The second objective of my research program is to develop a simulation framework that is less dependent on the choice of a specific model or data set. This framework allows us to take model uncertainty into account when projecting financial scenarios.****The financial and economic world crisis of the late 2000s revealed that risk management practices must be improved. For example, insurance companies suffered large losses on their variable annuities business. Variable annuities are investment vehicles offered by insurance companies which combine a life insurance policy with financial guarantees. These guarantees expose the insurer to market risks which cannot be diversified, but only mitigated through a risk management strategy. This strategy is generally established based on a theoretical model, and is expected to reduce the risk of loss for the insurance company. Since the risk management strategy is based on model assumptions, its effectiveness in the real world will be determined by how well it can perform when these assumptions are violated. The third objective of my research program is to develop risk management strategies for variable annuities that can mitigate market risks effectively in the presence of model uncertainty.**