The Credibility Report
Edition 11: Foundation Models for Insurance, Neural Credit Risk, and Mortality Forecasting
Foundation models for insurance, neural credit risk, and mortality forecasting.
What’s in this edition
Primary-source market updates (no aggregator links) plus the latest actuarial-relevant arXiv papers (score ≥ 15, last 14 days).
📰 Headlines (primary sources)
S&P Global warns the Middle East war is intensifying cyber risk
S&P’s framing matters because it treats geopolitical escalation as a live cyber-insurance variable rather than just a background security story. For actuaries working near cyber pricing or accumulation, this is a reminder that event clustering can arrive through political channels just as easily as through technology alone.
Read source → • S&P GlobalBridging the short-term rental coverage gap
Short-term rental exposure keeps blurring the line between personal and commercial risk. The useful signal here is product design: when customer behavior outgrows traditional policy categories, pricing and underwriting often lag first and claims friction shows up later.
Read source → • Triple-IFarmowners’ insurance and the trends behind a challenging market
Farmowners is a useful example of a line where weather, inflation, property concentration, and rural exposure interact in messy ways. Even if you do not price farm business directly, it is a good case study in how niche lines can become stress points when multiple loss drivers move together.
Read source → • Triple-I🔬 Research Spotlight (arXiv)
General Bounds on Functionals of the Lifetime under Life Table Constraints
arXiv • Score: 44 • 2026-03-06
In life insurance, life tables are used to estimate the survival distribution of individuals from a given population. However, these tables only provide survival probabilities at integer ages but no information about the distribution of deaths between two consecutive integer values. Many actuarial quantities, such as variable annuities, are functionals of the lifetime and computing them requires full information about mortality rates. One frequent solution is to postulate fractional age assumptions or mortality rate models, but it turns out that the results of the computations strongly depend on these assumptions, which makes it difficult to generalize them. We hence derive upper and lower bounds of functionals of the lifetime with respect to mortality rates, which are compatible with the observed life table at integer ages. We derive two sets of results under distinct assumptions. In the first, we assume that each mortality trajectory is almost surely consistent with all the given one-year survival probabilities from the table. In the second, we consider a relaxed formulation that allows for deviations of the mortality rates while still being consistent in expectation with the given one-year reference survival probabilities. These distinct yet complementary approaches provide a new robust framework for managing mortality risk in life insurance. They characterize the worst- and best-case contract values over all mortality processes that remain compatible with the observed life-table information, thereby enabling insurers to quantify the impact on prices of deviations of the observed mortality rates from their mortality assumptions/models.
Open paper →Contributions of geolocated weather and building related data for insurance assessment of flood risks
arXiv • Score: 36 • 2026-03-02
Floods rank among the costliest natural hazards, causing over USD 100 billion in insured losses between 2013 and 2023. In France, persistent deficits in the natural catastrophe scheme highlight the need for accurate, building-scale flood risk assessment. Insurers typically rely on frequency-severity models supported by hazard maps and regional climate indicators. However, previous studies show that such large-scale variables explain only a limited share of the variability in individual flood losses. This study evaluates the marginal contribution of multiple georeferenced data layers to modeling flood claim occurrence and severity in a large French home insurance portfolio. Starting from a baseline model based on standard underwriting information, we sequentially introduce climate-expert variables, extreme rainfall indicators, and fine-scale geolocated building and environmental attributes. The analysis focuses on a practical setting in which insurers cannot deploy full hydrological or hydraulic catastrophe models because of budgetary, licensing, or operational constraints. Results show that rainfall-based indicators, particularly a newly constructed metric capturing intense local precipitation, substantially improve claim modeling performance. Building and environmental variables further enhance occurrence prediction. Overall, the findings demonstrate how high-resolution geolocated data improve exposure and vulnerability assessment, complement official flood maps, and provide insurers with an operational framework for refining flood risk evaluation and pricing.
Open paper →Asymptotics of Ruin Probabilities in a Subordinated Cramér-Lundberg Model
arXiv • Score: 26 • 2026-03-02
We study a dynamic model of a non-life insurance portfolio. The foundation of the model is a compound Poisson process that represents the claims side of the insurer. To introduce clusters of claims appearing, e.g. with catastrophic events, this process is time-changed by a Lévy subordinator. The subordinator is chosen so that it evolves, on average, at the same speed as calendar time, creating a trade-off between intensity and severity. We show that such a transformation always has a negative impact on the probability of ruin. Despite the expected total claim amount remaining invariant, it turns out that the probability of ruin as a function of the initial capital falls arbitrarily slowly depending on the choice of the subordinator.
Open paper →Association of Radiologic PPFE Change with Mortality in Lung Cancer Screening Cohorts
arXiv • Score: 18 • 2026-03-10
Background: Pleuroparenchymal fibroelastosis (PPFE) is an upper lobe predominant fibrotic lung abnormality associated with increased mortality in established interstitial lung disease. However, the clinical significance of radiologic PPFE progression in lung cancer screening populations remains unclear. We investigated whether longitudinal change in PPFE quantified on low dose CT independently associates with mortality and respiratory morbidity. Methods: We analysed longitudinal low-dose CT scans and clinical data from two lung cancer screening studies: the National Lung Screening Trial (NLST; n=7980) and the SUMMIT study (n=8561). An automated algorithm quantified PPFE volume on baseline and follow up scans. Annualised change in PPFE (dPPFE) was derived and dichotomised using a distribution based threshold to define progressive PPFE. Associations between dPPFE and mortality were evaluated using Cox proportional hazards models adjusted for demographic and clinical variables. In the SUMMIT cohort, dPPFE was also examined in relation to clinical outcomes. Findings: dPPFE independently associated with mortality in both cohorts (NLST: HR 1.25, 95% CI 1.01-1.56, p=0.042; SUMMIT: HR 3.14, 95% CI 1.66-5.97, p<0.001). Kaplan-Meier curves showed reduced survival among participants with progressive PPFE in both cohorts. In SUMMIT, dPPFE was associated with higher respiratory admissions (IRR 2.79, p<0.001), increased antibiotic and steroid use (IRR 1.55, p=0.010), and a trend towards higher mMRC scores (OR 1.40, p=0.055). Interpretation: Radiologic PPFE progression independently associates with mortality across two large lung cancer screening cohorts and with adverse clinical outcomes. Quantitative assessment of PPFE progression may provide a clinically relevant imaging biomarker for identifying individuals at increased respiratory risk within screening programmes.
Open paper →One-Shot Individual Claims Reserving
arXiv • Score: 18 • 2026-03-12
Individual claims reserving has not yet become established in actuarial practice. We attribute this to the absence of a satisfactory methodology: existing approaches tend to be either overly complex or insufficiently flexible and robust for practical use. Building on the classical chain-ladder (CL) method, we introduced a new perspective on individual claims reserving in Richman and Wüthrich [arXiv:2602.15385]. This manuscript has sparked considerable discussion within the actuarial community. The aim of the present paper is to continue and deepen that discussion, with the ultimate goal of advancing toward a new standard for micro-level reserving.
Open paper →Learning Optimal Individualized Decision Rules with Conditional Demographic Parity
arXiv • Score: 17 • 2026-03-05
Individualized decision rules (IDRs) have become increasingly prevalent in societal applications such as personalized marketing, healthcare, and public policy design. However, a critical ethical concern arises from the potential discriminatory effects of IDRs trained on biased data. These algorithms may disproportionately harm individuals from minority subgroups defined by sensitive attributes like gender, race, or language. To address this issue, we propose a novel framework that incorporates demographic parity (DP) and conditional demographic parity (CDP) constraints into the estimation of optimal IDRs. We show that the theoretically optimal IDRs under DP and CDP constraints can be obtained by applying perturbations to the unconstrained optimal IDRs, enabling a computationally efficient solution. Theoretically, we derive convergence rates for both policy value and the fairness constraint term. The effectiveness of our methods is illustrated through comprehensive simulation studies and an empirical application to the Oregon Health Insurance Experiment.
Open paper →The Theory behind UMAP?
arXiv • Score: 15 • 2026-03-02
In 2018, McInnes et al. introduced a dimensionality reduction algorithm called UMAP, which enjoys wide popularity among data scientists. Their work introduces a finite variant of a functor called the metric realization, based on an unpublished draft by Spivak. This draft contains many errors, most of which are reproduced by McInnes et al. and subsequent publications. This article aims to repair these errors and provide a self-contained document with the full derivation of Spivak's functors and McInnes et al.'s finite variant. We contribute an explicit description of the metric realization and related functors. At the end, we discuss the UMAP algorithm, as well as claims about properties of the algorithm and the correspondence of McInnes et al.'s finite variant to the UMAP algorithm.
Open paper →Until next time—stay credible.
— The Credibility Report
Edition 011 | Prepared March 15, 2026 (UTC)