Longevity Risk Management

My objective is to apply new state-of-the-art Bayesian econometric tools to the area of longevity risk management. In this regard, my focus is on three main areas, (1) Mortality modelling, (2) Risk management, and (3) Risk transfer. The three areas chosen covers a wide range of application to be used in the management of longevity risk. Furthermore, developments in these areas can be extended to a broader class of risk management.

To tackle the first issue, I investigate how modern Monte-Carlo Markov-Chain (MCMC) algorithms can be applied efficiently in mortality modelling. In particular, we develop an efficient MCMC algorithm for the linear and nonlinear dynamics of two renowned mortality models, namely the Lee-Carter and Cairns-Blake-Dowd model by casting them in a Bayesian state-space framework.

Secondly, under Article 101 of the Solvency II framework, an insurance company is required to hold a Solvency Capital Requirement that ensures all quantifiable risks that the insurance company is undertaking, are taken into account and it should correspond to a 12-month 99.5% Value-At-Risk. While many researchers have tackled the issue of uncovering the ``correct” amount of capital required, there is currently a lack of development in the area of backtesting the required solvency capital. Thus, we combine the area of Bayesian Decision Theory and the backtesting literature of Value-At-Risk to Solvency Capital Requirements. We develop a new method for Bayesian Value-At-Risk backtesting that can simultaneously act as a tool for model choice.

Recently, the high upfront cost of longevity risk transfers have plagued the insurance companies due to the mortality improvements experienced. Previously, risk transfers were dominated by Buy-in and Buy-out plans. However, due to the mortality improvements in the recent decade, it has pushed insurance companies and pension plans to focus on cheaper alternatives. This gave rise to a longevity derivative market, where risk taking behaviours from the financial sector tries to profit off longevity risk. While these products are heavily sought after, the lack of available data on the market is a major problem when pricing. Without available market data, it then begs the question: How do we determine a correct risk premium? To answer this question we apply Bayesian methods to obtain a risk premium density to reflect the uncertainty within an incomplete market.