报告时间：2026年9月15日18:30-19:30

报 告人：关国卉，中国人民大学统计学院副教授

报告地点：#腾讯会议：362-683-9628

报告摘要：

This paper studies an insurer's reinsurance and investment problem when the risk aversion coefficient is random, and the insurer exhibits preference ambiguity toward this uncertainty. The insurer can purchase pre-loss reinsurance and allocate wealth between a risky asset and a risk-free asset. Preferences are modeled by the mean-variance criterion under ambiguity with random risk aversion, i.e., preference ambiguity, yielding a time-inconsistent control problem. Using an equilibrium approach, we establish a verification theorem and an extended HJB system that delivers closed-form equilibrium strategies. A key insight is that the preference ambiguity framework endogenously links reinsurance and investment decisions, even when insurance and financial risks are statistically independent. This breaks the separability implied by classical CARA, mean-variance, or robust preferences under model uncertainty. The framework thus rationalizes empirically observed ``risk-alignment effect" whereby insurers reduce investment risk following underwriting losses or heightened underwriting volatility. Ambiguity further induces a time-decaying distortion that vanishes at maturity, with all strategies converging to their ambiguity-neutral benchmarks. Comparative statics show that the strength of these effects is governed by the dispersion of the risk aversion distribution.

报告人简介：

关国卉，中国人民大学统计学院副教授，应用统计研究中心研究员，主要从事金融风险管理、保险精算与随机控制领域的教学科研工作，研究方向包括保险公司风险管理、再保险与投资策略、养老金资产配置以及金融市场中的随机控制与优化问题。在AAP、MOR、SIAM JFM、EJOR、JEDC、IME、NAAJ、SAJ、AOR、JOTA 等概率统计、运筹优化、金融数学与精算领域的重要国际期刊发表多篇论文，主持国家自然科学基金项目2项。