Insurance Companies' Asset Liability Management Based On Multi-objective Programming

The asset liability management (ALM) is the core of insurance company management, including asset allocation, liability allocation and the match of asset and liability. Asset-liability management can deal with the relationship during investment strategy, product design and pricing and management, and fulfill with profit target and remain liquidity of asset under the condition of meeting the requirement of insurance regulatory. So, in the process of ALM, a lot of objectives have to be fulfilled and many of objectives are contradictory each other. It is a kind of multi-objective decision. And, the multi-objective programming is a branch of operation research and can fulfill multi-objective decision. So, this paper combines the ALM with multi-objective programming, so that the management level can be improved and the multi-objective decision can be fulfilled. Especially, if we always consider fulfilling multi-objective decision in the background of financial crisis, insurance companies'power of resisting risk can be improved.This paper is divided into seven chapters.The first chapter introduces background, meaning and reviewing literature. Because of the lack of asset liability, in the process of financial crisis insurance companies have large loss. Insurance companies should combine the ALM with multi-objective programming to improve decision level.The second chapter reviews theory and technique of ALM. And we pay more attention on model and recent advances of asset ability management. This paper divided ALM model into four:traditional ALM model, stochastic programming ALM model, stochastic control ALM models and multistage stochastic ALM programming with decision rules. And, we give the frame of ALM.The third chapter analyzes the multi-objective decision's characteristic of ALM, and introduces theory and resolving methods of multi-objective programming. ALM of insurance companies is a complicated problem and has to deal with many kind of objectives and coordinate benefit from different teams. So, ALM is multi-objective decision. The forth chapter build a multi-objective programming model of ALM. The multi-objective model is based on balance sheet of life insurance companies. The objectives include:minimum asset allocation risk, maximum asset allocation revenue, minimum liability allocation risk, maximum liability allocation revenue, maximum wealth of shareholders, maximum company value. And, we do the empirical by use of Chinese data. We take 18 funds from China Asset Management Co., Ltd. and 33 stocks from SSE 50 index as sample to research the effect of Multi-objective ALM model. We find that life insurance managers can do the right decision to fulfill with many kinds of contradictory or harmonious objectives. According to our sample, our model always chooses funds and stocks of low risk and high return and can structure efficient frontier of low risk and high return.The fifth chapter builds stochastic multi-objective programming. We consider the stochastic changing of asset, liability and equity in balance sheet. The simulation result shows that insurance companies should invest 20% ratio of asset into stock in order to fulfill maximum profit and policyholder benefit. And by the analysis of policy option influencing to asset and liability, the policy option will influence obviously asset, reserve and their changing.The sixth chapter analyzes how to allocate their asset in illiquid markets. Our analysis, based on genetic algorithm, reveals a clear diversification benefit in illiquid markets apart from the one introduced. And cash-flow matching is shown to be the optimal strategy for a large investor. In order to decrease fluency of stock price changing, insurance should sell cash to satisfy with claim demand, and when cash exhausts stock have to be sold. And, in order to decrease illiquid influence, insurance companies should diversify their asset portfolio and clear their one kind of assets.The seventh chapter is conclusion.