Journal of Risk and Insurance最新文献

Many Americans remain financially underprepared for retirement. While automatic enrollment in employer-sponsored retirement plans has helped, target-date funds (TDFs) used as default investments have limitations. We propose target-date registered index-linked annuities (TD-RILAs) as a transparent, cost-effective alternative providing decreasing equity exposure over time. A theoretical analysis explores TD-RILAs' optimal structure and compares them to TDFs. A lab experiment examines investors' preferences, product transparency, default choices, and the role of information. We find that TD-RILAs are a suitable addition to retirement plans, potentially rivaling TDFs.

Limited research has been conducted on the optimal public–private risk-sharing for catastrophe risks. This paper develops a theoretical framework to study the risk-sharing decisions and interactions of three types of catastrophe-market participants: a large number of individuals, a large number of private insurers in a competitive market, and a government that can choose between alternatives of re/insurance or ex post relief. Our analysis shows that the optimal government intervention varies depending on the correlation levels among individual losses. For moderately positive levels of loss correlation, it is optimal for the government to offer an ex post relief program to supplement private insurance. However, for higher levels of loss correlation, government reinsurance becomes optimal, although not to the extent of replacing private insurance if the government is less efficient than private firms. In sum, as catastrophe-loss correlations increase, that is, as the risk becomes more catastrophic, more risk-sharing tools and funding are needed to maximize social welfare.

In many jurisdictions, insurance companies are prohibited from discriminating based on certain policyholder characteristics. Exclusion of prohibited variables from models prevents direct discrimination, but fails to address proxy discrimination, a phenomenon especially prevalent when powerful predictive algorithms are fed with an abundance of acceptable covariates. The lack of formal definition for key fairness concepts, in particular indirect discrimination, hinders effective fairness assessment. We review causal inference notions and introduce a causal graph tailored for fairness in insurance. Exploiting these, we discuss potential sources of bias, formally define direct and indirect discrimination, and study the theoretical properties of fairness methodologies. A novel categorization of fair methodologies into five families (best-estimate, unaware, aware, hyperaware, and corrective) is constructed based on their expected fairness properties. A comprehensive pedagogical example illustrates the implications of our findings: the interplay between our fair score families, group fairness criteria, and discrimination.

We investigate whether the price paid for insurance explains dishonesty in reporting an insurance claim. In our laboratory experiment, participants earn money in a real-effort task but risk losing some of this income through one of four randomly assigned, privately observed loss amounts. Before observing their loss, participants indicate their reservation price for insurance that pays an indemnity equal to their stated loss. Participants are insured if their randomly assigned premium is less than their stated reservation price. This mechanism provides data on each participant's consumer surplus from insurance. After receiving their cash earnings minus their assigned loss in private, participants report their loss. We find that the insured report modestly but statistically insignificant larger losses than the uninsured. Among the insured, we find no clear evidence that their reporting of excess losses increases in the randomly assigned price of insurance or decreases in the consumer surplus from insurance.

Accurately estimating loss reserves is critical for the financial health of insurance companies and informs numerous operational decisions. We propose a novel neural network architecture that enhances the prediction of incurred loss amounts for reported but not settled claims. Moreover, differing from other studies, we test our model on proprietary datasets from a large industrial insurer. In addition, we use bootstrapping to evaluate the stability and reliability of the predictions and Shapley additive explanation values to provide transparency and explainability by quantifying the contribution of each feature to the predictions. Our model shows superiority in estimating reserves more accurately than benchmark models, like the chain ladder approach. Particularly, our model exhibits nuanced performance at the branch level, reflecting its capacity to effectively integrate individual claim characteristics. Our findings emphasize the potential of using machine learning in enhancing actuarial forecasting and suggest a shift towards more granular data applications.

Many regulated health insurance markets with community-rated premiums rely on risk adjustment (RA) to mitigate insurer-incentives to risk select. However, insurers remain typically undercompensated for chronically ill enrollees. We use historical data on health spending and risk adjuster information to identify individuals undercompensated by the Dutch RA model of 2021 and find a selective group (1% of the population) with an average annual undercompensation of €6,050. We supplement the RA model with a risk sharing modality called high-risk pooling (HRP) to organize residual-based compensations towards insurers for the identified group to reduce the mean undercompensation to zero. The effects are evaluated on subgroups defined by chronic disease, finding a 42% reduction of their average undercompensation. Therefore, through compensating 1% of the population, the insurer-incentives to select against chronically ill individuals substantially diminish. These results are compared to outlier-risk sharing (reinsurance), proving HRP to be more effective at reducing selection incentives.