Study design. Retrospective, public-data observational study of billionaire deaths from 1 January 2015 through 31 December 2025. The unit of analysis is the individual person.

Data Collection and Verification

Case-finding. Candidates were identified iteratively from public billionaire directories (Forbes, Bloomberg, Hurun), yearly "billionaires in memoriam" compilations, and targeted media searches conducted in English and relevant local languages using standardized query templates (e.g., "<Name>" died <Year>, "<Name>" cause of death).
Death confirmation. Each death was classified as High confidence (supported by a primary statement or two independent reputable sources matching on identity and death year) or Medium confidence (one reputable source plus identity match on two or more independent attributes). Cases that could not reach at least Medium confidence were excluded.
Billionaire status. Inclusion required net worth ≥$1 billion at or near the time of death, verified by a dated estimate from Forbes, Bloomberg, or Hurun within 24 months before death. Wikipedia-derived net worth figures were not accepted for inclusion. Source hierarchy: Forbes > Bloomberg > Hurun, preferring the estimate closest to death within a tier. All wealth figures were additionally expressed in constant 2025 US dollars using the CPI-U annual average published by the Bureau of Labor Statistics.

Cause-of-Death Classification

Each case received a specific cause-of-death string (the most specific statement supported by sources) and a categorical classification into one of four buckets: aging-related disease, external cause, other medical (non-aging), or unknown/not disclosed.
Aging-related disease was operationalized following Le Couteur & Thillainadesan (2022) and Chang et al. (2019) as causes whose incidence rises strongly with age: cardiovascular/cerebrovascular disease, cancer, neurodegenerative disease, metabolic/renal failure, chronic respiratory disease, frailty/multi-organ failure, and terminal infections in the setting of advanced age or chronic illness. For decedents aged ≥70 whose obituaries used natural-cause phrasing ("passed away peacefully," "in their sleep," "natural causes") with no competing external-cause indication, the death was classified as aging-related with low specificity. "After a long illness" without a stated diagnosis was classified as aging-related for decedents aged ≥60. External causes (accident, homicide, suicide) overrode all other classifications.
Sensitivity analyses recomputed the aging-related share under stricter rules: (a) treating low-specificity natural-death phrasing as unknown, and (b) treating "long illness/complications" as unknown unless a diagnosis was stated.
Mortality Analysis: Person-Years and Standardized Mortality Ratios
Population data. Living billionaire counts were drawn from the Forbes Billionaires Evolution dataset (Kaggle, CC0 license) for 2015–2024 (10 annual snapshots; 97.4% birth-date coverage) and from the Forbes live website scraped in January 2026 as a proxy for 2025. Each unique individual appearing on a Forbes annual list contributed one person-year of exposure for that year.
Age banding. Age was derived from birth date (not the reported age field) and calculated at midyear. Standard 10-year bands were used (50–59, 60–69, 70–79, 80–89), with a combined 90+ bucket to improve statistical power.
Reference population. Hong Kong was selected as the benchmark because it has the world's highest life expectancy (~85.5 years). Sex-specific mortality rates (m_x per 1,000 person-years) were averaged over the 2015–2022 period from Hong Kong Census and Statistics Department life tables.
SMR calculation. For each age band:
SMR = Observed Deaths / Expected Deaths
where Expected Deaths = Person-Years × Hong Kong age-specific mortality rate. Confidence intervals were computed using the exact Poisson method; p-values from two-sided Poisson exact tests.
Results (males, N = 336 deaths; 19,527 person-years, ages 50+). Overall SMR = 0.71 (95% CI: 0.63–0.79), indicating 29% lower mortality relative to Hong Kong's general male population. The advantage was concentrated at ages 60–89 (SMR 0.48–0.70, p < 0.001 for each band) and vanished at 90+ (SMR 0.89, not significant). Female billionaires were excluded from SMR analysis due to small sample size (28 deaths).

Projected Life Expectancy

For living billionaires, we estimated projected life expectancy by applying the empirically derived, age-band-specific SMRs to a Gompertz mortality model fitted to Hong Kong male age-specific rates:
m_HK(x) = α · exp(β · x)
Parameters (α = 1.41 × 10⁻⁵; β = 0.0992 per year) were estimated by OLS on log-mortality at age-band midpoints. The billionaire hazard was computed as m_bill(x) = m_HK(x) × SMR(age band), and remaining life expectancy was obtained by numerical integration of the resulting survival function, truncated at age 110. These are period estimates assuming current mortality schedules persist; no projected medical advances are modeled.

Biases and Limitations

Several mechanisms likely inflate the observed billionaire mortality advantage:

• Survivor bias. Individuals must survive long enough to accumulate ≥$1 billion, excluding early deaths.
• Healthy-founder effect. Traits enabling wealth creation (discipline, education, stress tolerance) may independently promote longevity, conflating the wealth effect with trait selection.
• Late entry. Median age at first Forbes appearance is 57; 18% enter at age 70+. Wealth may not have influenced earlier health behaviors.
• Pre-death wealth loss. Individuals whose health deteriorates may lose billionaire status before death through asset sales, reduced business performance, or medical expenses, systematically excluding deaths preceded by health decline.
• Unstable threshold. Net worths fluctuating near $1 billion may drop below the threshold during illness, producing the same effect.

The SMR of 0.71 should therefore be interpreted as an upper bound on any causal effect of wealth on mortality.
Additional limitations include uneven public reporting across regions and languages, opacity of private wealth structures, imprecision in net worth estimates for private assets, and frequent absence of cause-of-death disclosure (30% of cases).

References

1. Le Couteur DG, Thillainadesan J. What Is an Aging-Related Disease? An Epidemiological Perspective. J Gerontol A Biol Sci Med Sci. 2022;77(11):2168–2174.
2. Chang AY et al. Measuring population ageing: an analysis of the Global Burden of Disease Study 2017. Lancet Public Health. 2019;4(3):e159–e167.
3. Gompertz B. On the nature of the function expressive of the law of human mortality. Phil Trans R Soc. 1825;115:513–583.
4. Chetty R et al. The Association Between Income and Life Expectancy in the United States, 2001–2014. JAMA. 2016;315(16):1750–1766.
5. Hong Kong Census and Statistics Department. Hong Kong Life Tables 2015–2022.