Liability-driven Investing in Defined Benefit Pension Plans under Loss-aversion using Genetic Algorithm

다국어 초록 (Multilingual Abstract)

This paper aims to suggest optimal dynamic investment strategies for DB pension plans when sponsors have differentiated attitudes regarding surplus and shortfall. Using Tversky and Kahneman’s prospect theory (1992), we set the sponsor’s objective function in terms of surplus and shortfall, and apply a differentiated sensitivity to surplus and shortfall. The main finding was that numerous optimal asset allocations exist depending on the sponsors’ attitude towards surplus and shortfall. Furthermore, it is shown that loss aversion to the shortfall does not completely dominate sponsors’ investment strategies. Thus, when determining the investment strategy in a DB plan, the DB sponsors’ attitudes towards both surplus and shortfall should be considered independently. Lastly, the optimal equity portion declines as the contribution rate increases.

목차 (Table of Contents)

• Abstract vi
• 1. Introduction 1
• 1.1 Introduction 1
• 1.2 Objectives of research 2
• 1.3 Outline of research 3
• 2. Literatures review 5
• 3. Assumptions and model construction 12
• 3.1 Assumptions 12
• 3.2 Sponsor’s loss aversion utility function 13
• 3.3 Stochastic liability model 15
• 3.4 Stochastic asset growth model 17
• 3.5 Optimization problem and solution method 18
• 4. Model parameters 23
• 4.1 Age-dependent loss aversion 23
• 4.2 Career salary profile 24
• 4.3 Growth rate of labor income 24
• 4.4 Interest rate 25
• 4.5 Immediate pension annuity factor 26
• 4.6 Asset parameters 27
• 4.7 Genetic algorithm parameters setting 29
• 5. Empirical application 30
• 5.1 Baseline case 30
• 5.2 Sensitive analysis 35
• 1) Age-dependent loss aversion ratio 35
• 2) Curvature parameter 38
• 3) Contribution rate 41
• 6. Conclusion 44
• References 46
• Appendix A: Minimizing shortfall risk 51
• Appendix B: Results for minimizing terminal utility 53
• Appendix C: Spread method 59
• Notation 61