Periodic evaluation of defined-contribution pension fund: A dynamic risk measure approach

2025-08-07

AI Summary

This paper addresses dynamic tail-risk management for defined-contribution pension funds during both the accumulation and decumulation phases. Methodologically, it proposes a periodic evaluation–based risk control framework that integrates model-free reinforcement learning—designed to handle market stochasticity—with an enhanced Lee-Carter mortality model for calibrated and forward-looking life expectancy forecasting, jointly optimizing investment and annuitization decisions. Its key contribution lies in departing from the conventional single-point (retirement-time) assessment paradigm by introducing a lifecycle-wide dynamic risk measure, enabling real-time tail-risk identification and adaptive response. Empirical results demonstrate that periodic evaluation significantly enhances strategy conservatism, while projected increases in life expectancy induce a moderate upward adjustment in risk tolerance; together, these effects improve long-term welfare outcomes and intertemporal sustainability of retirement income.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Reasoning under Uncertainty: Stochastic Optimization
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements

Abstract

This paper introduces an innovative framework for the periodic evaluation of defined-contribution pension funds. The performance of the pension fund is evaluated not only at retirement, but also within the interim periods. In contrast to the traditional literature, we set the dynamic risk measure as the criterion and manage the tail risk of the pension fund dynamically. To effectively interact with the stochastic environment, a model-free reinforcement learning algorithm is proposed to search for optimal investment and insurance strategies. Using U.S. data, we calibrate pension members' mortality rates and enhance mortality projections through a Lee-Carter model. Our numerical results indicate that periodic evaluations lead to more risk-averse strategies, while mortality improvements encourage more risk-seeking behaviors.

Problem

• Dynamic risk measure for pension fund evaluation
• Optimal investment and insurance strategy search
• Mortality rate calibration and projection enhancement

Innovation

• Dynamic risk measure for pension evaluation
• Model-free reinforcement learning algorithm
• Lee-Carter model for mortality projections