Jens Robben, Torsten Kleinow
6 min
This paper by Jens Robben and Torsten Kleinow tackles a fundamental question in longevity research: does human lifespan have a fixed upper limit, and if so, does it vary across socio-demographic groups? Using unprecedented individual-level microdata from Belgium and the Netherlands (1995–2022) on everyone aged 90+, they apply extreme value theory (EVT) to the 'oldest-old' mortality patterns. Their findings confirm a finite lifespan ceiling—but one that's stratified by sex, origin, civil status, household type, and education. Men top out earlier than women; widowed people and those in institutional care have shorter maxima; non-Western Europeans lag behind natives; and higher education pushes the limit higher. This challenges the idea of a universal lifespan cap and highlights persistent inequalities even at extreme ages.
EVT is the statistical powerhouse here, designed for modeling rare extremes like supercentenarian deaths. The authors fit a generalized Pareto distribution to the upper tail of age-at-death data, estimating a covariate-dependent scale parameter that reveals group-specific upper endpoints. Imagine lifespan distributions as curves approaching a wall: EVT measures how close different groups get to their wall. For Belgians and Dutch alike, the data robustly support finite limits (e.g., confidence intervals around 113–124 years from prior studies), but the walls shift based on demographics. This isn't aggregate stats—it's complete population records, handling left truncation (entering at 90) and right censoring (still alive) via likelihood methods.
The results paint a clear picture of inequality at the longevity frontier:
These hold across countries and time, surviving robustness checks. Even among the oldest-old, where mortality improvements slow dramatically past 100, social determinants don't fade—they stratify the ceiling.
The dataset is a game-changer: full microdata on 90+ year-olds, tracking covariates over time. No more lumping ages into crude bins or relying on scarce centenarian records. This granularity exposes variation overlooked in prior debates (e.g., Dong et al. vs. Lenart & Vaupel on limits' existence). Implications ripple outward: with centenarians projected to hit 25 million by 2100, understanding these 'social determinants of extreme longevity' informs policy on health equity, pension risks, and inequality. It also reframes the lifespan limit debate—from monolithic constant to heterogeneous landscape—urging nuanced models of human mortality tails.
The existence of an upper limit to the human lifespan has been widely debated, with studies offering both supporting and opposing evidence. Using unique individual-level death and population records for individuals aged 90 and older in Belgium and the Netherlands between 1995 and 2022, we provide statistical evidence supporting the existence of an upper limit. A related yet unexplored question is whether this life span limit differs across socio-demographic groups. Our microdata include information on the sex, origin, civil status, type of household, and education level of each individual. Using tools from extreme value theory, we quantify and compare the upper tail of human lifespan distributions across these socio-demographic characteristics. We find that men have a statistically lower maximum lifespan than women and that individuals who are widowed or live in institutional households have a clearly lower maximum lifespan. Finally, individuals of non-Western European origin and those with higher educational attainment exhibit longer maximum lifespans.
Alex: Okay, so everyone's tail has the same ending type, but social factors stretch or shrink how far it reaches before stopping. Like rivers with the same water physics but different channels from the landscape?
Sam: Precisely—shared physics in the shape parameter, but terrain like sex or living setup changes the flow rate via the scale. They link traits to the log of that scale with a simple equation: baseline plus adjustments for each factor.
Alex: And does this fit the actual ages well?
Sam: Yes, checks against the data show close matches for extras up to 20-plus years over 100 in both countries. The negative shape confirms light tails with upper bounds. Covariate tweaks to scale then shift group maxima meaningfully—for instance, favoring longer reaches for certain profiles.
Alex: Huh. So inequalities don't just slow the average—they redraw the finish line itself for the very oldest.
Alex: So if those social traits tweak the scale like that, which ones push the cap highest—and which drag it down?
Sam: In both countries, women reach further than men. Living alone at age 100 gives the biggest boost—far higher than in group homes like nursing facilities, which serve as the baseline. Unmarried or married folks also edge out widowed ones.
Alex: Household setup stands out then. And education or background?
Sam: Education effects are smaller overall; in Belgium, higher levels like university tie to longer reaches, but not so clearly in the Netherlands. For origins, non-Western backgrounds sometimes show modest lifts. These shifts add up across traits.
Alex: Right, so no two people at 100 have the exact same projected end, based on their setup. How wide does that get across common combinations?
Sam: They map out hundreds of trait mixes and compute caps for those seen often enough. In Belgium, frequent ones range from about 116 years for a widowed woman in a group home, up to 122 for another widowed woman living solo with higher education—a gap of roughly six years. The Netherlands shows similar spreads of five to seven years among top profiles, with medians near 117 in both places.
Alex: That's a clear difference even for the elite survivors. But they caution those highs are from common profiles?
Sam: Yes—the model fits the seen data solidly, but extreme combos beyond common ones are extrapolations. Still, among everyday profiles at 100-plus, the five-to-seven-year gaps highlight how factors like living alone or being female extend the upper edge, without changing the overall bounded shape.
Alex: So inequalities shape distinct paths to the end, even past 100. Makes the "one limit for all" idea feel too simple now.
Alex: Those gaps add up to real differences—like switching from a group home to living alone could stretch the limit by around four years in Belgium. What does the paper make of all this in the end?
Sam: The study confirms a finite upper limit to lifespan across both countries. But the key advance is showing how social factors create meaningful spreads in those limits—five to seven years among common profiles at age 100—without needing separate shapes for each group. It points to chances for targeted steps—like easing household isolation or supporting marital ties—to close some of those gaps. The evidence suggests inequalities stick right to the edge, so addressing them early could shift tails for the longest livers. Overall, this moves us past a single-number view of human limits toward one shaped by everyday life.
Alex: That's a grounded takeaway—lifespan's end isn't fixed for all, but molded by social setup even past 100. Thanks, Sam, for breaking it down so clearly. Thanks for listening to ResearchPod.