Evaluating Medical Underwriters in Life Settlements: Problem of Unreported Deaths

Abstract

The valuation of policies in the life settlement market hinges upon life expectancy assessments provided by medical underwriters. Because accurate assessments are essential to the market functioning well, determining so is of interest to various stakeholders, including regulatory bodies. As a third party to settlement transactions, these underwriters, and hence their auditors, have to rely on public sources in compiling mortality data on underwritten insureds. Such acquired data are known to contain a significant proportion of unreported deaths. Accurate accounting for their presence is crucial to a fair evaluation of an underwriter’s methodology. However, existing methods for doing so either rely solely on actuarial opinion or are based on assumptions untenable in the life settlement context. Toward developing a practical data-driven methodology for consistently estimating mortality in the presence of unreported deaths, we propose the IBNR (incurred but not reported) rate model. This adaptation of the cure rate model accommodates the heterogeneity in the rates of both mortality and unreported deaths seen in practice. Our methodology thus facilitates an accurate and objective evaluation of an underwriter.

ACKNOWLEDGMENTS

We thank the editor and two referees for their many insightful comments that led to a substantial improvement of the article.

Discussions on this article can be submitted until January 1, 2023. The authors reserve the right to reply to any discussion. Please see the Instructions for Authors found online at http://www.tandfonline.com/uaaj for submission instructions.

Notes

1 Interestingly, there is evidence that prices in the primary market went up during this period (Bauer et al. 2021).

2 Despite detailed and laborious underwriting, interestingly, Bauer, Russ, and Zhu (2020) found evidence of significant residual information asymmetry.

3 In practice the settlement occurs on average 3 months after the medical underwriter issues the LE assessment; an LE assessment older than 6 months would typically be annulled and replaced with a refreshed assessment (Xu 2020).

4 It is worth mentioning that because the underwriter’s assessments, through the transaction price, affects a policyholder’s decision on consummating the settlement, evaluating a medical underwriter based on only the subset of settled policies may deem it more aggressive. This is supported by Granieri (2016), which presented evidence to suggest the presence of a selection effect due to settlement. Furthermore, because the IBNR issue is moot in life settlement investment portfolios due to effective mortality tracking by the tracking agent, in this article we only focus on medical underwriter data.

5 Because life settlement mortality concerns settlement ages 60 and above, the mortality rates corresponding to T do not contain an accidental hump.

6 The bias goes from –0.36 to –0.34, and the root MSE decreases from 0.89 to 0.86.

Additional information

Funding

The first author acknowledges with deep gratitude the support from the Society of Actuaries' Hickman Scholars Program