Category: Gompertz–Makeham Model

  • Calculator methodology · Trust & method

    How Do Actuaries Validate a Gompertz-Makeham Fit Against Observed Population Death Rates?

    Actuaries validate a Gompertz–Makeham fit with three sequential checks: a chi-square goodness-of-fit test against observed death rates, an AIC comparison against rival models, and Cox–Snell residual plots.

    Get my survivorship-adjusted estimate See how the model is built
    Gompertz–Makeham method WHO & UN mortality data Standard actuarial tests
    p < 0.05
    Chi-square level flagging systematic misfit
    A/E = 1.00
    Actual-to-Expected ratio at a perfect fit
    ΔAIC > 2
    Gap that makes one model meaningfully better
    30–85
    Age range where the fit holds; degrades above 85
    On this page What validation means The chi-square test The A/E ratio AIC model comparison Cox–Snell residuals Where the model fails How this reaches your number FAQs

    What does ‘validating a mortality model’ actually mean in practice?

    Validation means comparing the model’s predicted death rate at each age against the observed rate from a reference life table, then applying formal statistical tests to decide whether the gap is random noise or systematic misfit.

    The quantity under test is the age-specific mortality rate, written q(x) — the chance of dying within a year at age x. The model produces a predicted q(x) at every age; the reference life table supplies the observed q(x). Validation puts the two side by side across all ages and asks one question: does the difference look like ordinary sampling scatter, or a consistent pattern? A consistent pattern means the model is wrong somewhere and gets rejected; random scatter means it is accepted. The reference tables here are country-sex period life tables — which mortality dataset the calculator validates against covers where they come from.

    The decision rule is directional: a systematic deviation between predicted and observed q(x) rejects the model, while random scatter around the observed rates accepts it. The benchmark is the observed q(x) column of WHO Global Health Observatory and UN World Population Prospects country-sex period life tables. WHO Global Health Observatory; UN World Population Prospects

    Every test below is a different lens on that one comparison. If you want the model those tests judge, see how the Gompertz-Makeham hazard function is constructed. The first and most common test measures the fit across all ages at once.

    How does the chi-square goodness-of-fit test detect a poor Gompertz-Makeham fit?

    The chi-square test compares observed deaths in each age band to the deaths the model predicts. A statistically significant result — p below 0.05 — signals that the model’s hazard curve deviates from the empirical life table by more than chance variation would explain.

    The test needs two counts per age band: observed deaths, O, and expected deaths, E, where E comes from the fitted model applied to that band’s population. It sums the squared gaps, scaled by the expected counts, into a single chi-square statistic. The larger the accumulated gap, the larger the statistic, and the smaller the p-value. Its degrees of freedom equal the number of age bands minus the number of estimated parameters — typically three for Gompertz–Makeham (the baseline level, the ageing slope, and the Makeham background constant). A significant result does not just say “misfit”; actuaries then inspect which bands contributed most to the statistic to see where the model is straining.

    The chi-square goodness-of-fit test rejects a mortality model when p < 0.05, using observed-versus-expected deaths per age band with degrees of freedom equal to the number of bands minus the three fitted Gompertz–Makeham parameters. Pearson chi-square goodness-of-fit; standard actuarial model-validation method

    The chi-square gives one verdict for the whole curve, but it cannot tell you the direction of a misfit. That is the A/E ratio’s job.

    What is the Actual-to-Expected (A/E) ratio and why do actuaries use it alongside chi-square?

    The A/E ratio divides observed deaths by model-predicted deaths in each age band. A ratio of 1.00 is a perfect fit; ratios consistently above or below 1.00 across adjacent bands reveal directional bias that a single chi-square p-value cannot locate.

    Where chi-square returns one number for the whole curve, the A/E ratio returns one number per age band, so it maps the misfit by age. A run of A/E values above 1.00 means the model is under-predicting deaths there; a run below 1.00 means it is over-predicting. That directional, age-located reading is exactly what the aggregate chi-square statistic hides. The Gompertz–Makeham model has a well-known A/E signature at the oldest ages: in high-longevity populations it tends to over-predict deaths above 85, pushing A/E below 1.00 there.

    A/E equals 1.00 at a perfect fit; a practical working rule treats A/E drifting outside a roughly 0.95–1.05 band across three or more consecutive age groups as a directional bias serious enough to re-estimate the parameters. Actual-to-Expected mortality experience analysis; standard actuarial practice

    Both chi-square and A/E answer “does this model fit?” Neither answers “is this the best model available?” — which is where AIC comes in.

    How does AIC comparison tell actuaries whether Gompertz-Makeham is the right model to use?

    The Akaike Information Criterion (AIC) rewards fit but penalises complexity. Actuaries fit Gompertz–Makeham alongside simpler models (plain Gompertz) and more complex ones (Kannisto, Beard); the model with the lowest AIC wins. A ΔAIC above 2 over the next-best model is treated as meaningful.

    AIC is defined as 2k − 2·ln(L), where k is the number of parameters and L is the maximised likelihood — how well the model reproduces the data. Adding parameters can only improve raw fit, so AIC docks 2 points per parameter to stop actuaries rewarding needless complexity. The model that balances fit against parsimony best has the lowest AIC. This is a different question from chi-square: chi-square asks whether a model fits adequately on its own; AIC asks which model fits best relative to its rivals. Adding the Makeham background term to a plain Gompertz model typically lowers AIC most in younger-adult populations, where age-independent background deaths are a meaningful share of the total.

    AIC = 2k − 2·ln(L), with k the parameter count and L the maximised likelihood; the conventional actuarial threshold is that a ΔAIC greater than 2 between two models justifies preferring the lower-AIC one. Akaike Information Criterion (Akaike, 1974); standard model-selection practice

    Chi-square, A/E, and AIC all work on grouped, population-level counts. To catch misfit hiding inside individual records — especially in the tails — actuaries turn to residuals.

    What do Cox-Snell residuals reveal that aggregate tests miss?

    Cox–Snell residuals transform individual survival times so that, if the model is correct, they follow a unit-exponential distribution. A quantile plot of these residuals exposes age-specific misfit — particularly at the tails — that stays invisible in chi-square or AIC summaries.

    The idea is a clever change of scale. Under a correctly specified model, running each person’s survival time through the model’s own cumulative hazard should produce residuals with a known, standard shape: a unit-exponential distribution. Plot those residuals against exponential quantiles and, if the model is right, they land on a 45-degree line. Systematic departures from that line reveal where the model misfits — and because this works at the individual-record level, it catches problems the grouped aggregate tests smooth over. The Gompertz–Makeham model’s residuals characteristically bend above the 45-degree line at the oldest ages, the visual signature of it underestimating survival at extreme old age.

    Under a correct model, Cox–Snell residuals follow a unit-exponential distribution and fall on a 45-degree line against exponential quantiles; a curve above that line at the oldest ages signals the Gompertz–Makeham model underestimating survival there. Cox & Snell, J. R. Statist. Soc. B, 1968

    That tail behaviour is not an accident of one dataset — it is a structural property of the model, with a specific age where it starts to break.

    At what ages does the Gompertz-Makeham model typically fail validation, and why?

    The model reliably fits ages 30–85. Above 85, its exponential hazard assumption over-predicts observed mortality — a known structural limitation called mortality deceleration — which causes systematic validation failures in the very oldest, supercentenarian data.

    Across the working adult range, roughly 30 to 85, the exponential Gompertz term tracks real death rates closely and the fit passes the tests above. Beyond about 85, observed death rates stop accelerating and begin to plateau, while the model’s exponential term keeps climbing — so the model predicts more deaths than actually occur. This gap is mortality deceleration, and it is the model’s most-documented weakness; it is why period tables are used as the validation benchmark rather than sparse extreme-age data. Actuaries handle it two ways: cap the hazard at a maximum yearly value, or switch above about age 90 to a logistic-type (Kannisto) model that lets the curve plateau.

    The Gompertz–Makeham model fits well across ages 30–85 but fails above 85 because observed mortality decelerates (the hazard plateaus) while the exponential term keeps rising — corrected in practice by capping the hazard or switching to a Kannisto logistic model above about age 90. Mortality-deceleration / late-life plateau; gamma-Gompertz and Kannisto extensions, demographic literature

    Knowing exactly where a model stops being trustworthy is itself a validation result — and it is what connects all this testing to the number the calculator actually shows you.

    How does this validation process connect to the numbers the calculator shows you?

    A parameter set only earns its place after passing these checks — a non-significant chi-square, an A/E near 1.00, and the lowest AIC among the models tested — against the reference life tables. When a fit fails those checks, the response is to re-estimate the parameters, not to ship the misfit.

    This is what “trust the number” means concretely: the estimate you see rests on a parameter set that reproduced real observed death rates closely enough to pass standard actuarial tests, within the age range where the model is known to hold. Outside that range — at the extreme oldest ages — the honest move is the correction described above, not a silent over-estimate. That is the whole reason a validation node exists: to show the checks behind the output rather than ask you to take the output on faith. To see how those validated parameters become a personal figure, read how validated parameters become your personalised estimate, and for the wider picture, how actuarial accuracy compares across calculator types. The practical takeaway is simple: treat a life-expectancy number as trustworthy only when its model names the tests it passed, the data it was checked against, and the ages where it stops working — then read your own result as a validated range, not a verdict. Get your survivorship-adjusted estimate.

    Frequently asked questions about validating a Gompertz-Makeham fit

    How do actuaries test if a mortality model fits real data?

    They compare the model’s predicted death rate at each age with the observed rate from a reference life table, then apply formal tests — chi-square goodness-of-fit, the Actual-to-Expected ratio, AIC comparison, and Cox–Snell residual plots — to judge whether the gaps are random noise or systematic misfit.

    What is a goodness-of-fit test in actuarial science?

    A goodness-of-fit test measures how closely a model’s predicted values match observed data. In mortality work, it compares predicted deaths to observed deaths across age bands and returns a statistic — such as chi-square — whose p-value indicates whether any mismatch is beyond chance.

    What is the chi-square test used for in life table validation?

    It tests whether a fitted model’s expected deaths match the observed deaths in a life table across age bands. A p-value below 0.05 signals systematic misfit; degrees of freedom equal the number of age bands minus the number of fitted parameters (three for Gompertz–Makeham).

    How is AIC used to compare mortality models?

    AIC = 2k − 2·ln(L) scores each model by fit while penalising extra parameters. Actuaries fit several models to the same data and prefer the one with the lowest AIC; a difference greater than 2 over the next-best model is treated as a meaningful improvement.

    What are Cox-Snell residuals in survival analysis?

    They are transformed survival times that follow a unit-exponential distribution if the model is correct. Plotted against exponential quantiles, they should fall on a 45-degree line; departures — especially at the tails — reveal where the model misfits at the individual-record level.

    How do actuaries know when a Gompertz model fails?

    Failure shows up as a significant chi-square, an A/E ratio drifting away from 1.00 across consecutive age bands, a higher AIC than a rival model, or Cox–Snell residuals bending off the 45-degree line — most often at ages above 85, where mortality decelerates.

    What does A/E ratio mean in mortality experience studies?

    A/E is observed deaths divided by model-expected deaths in an age band. A value of 1.00 is a perfect match; consistently above 1.00 means the model under-predicts deaths, consistently below means it over-predicts — locating directional bias the aggregate chi-square cannot.

  • Calculator methodology · Trust & method

    Why Does the Makeham Term Matter More at Age 30 Than at Age 70 in a Life Expectancy Model?

    The Makeham term is the constant background-mortality risk in the model — accidents and infections that do not rise with age. At 30 it is a large share of your total risk; by 70 the age-driven term dwarfs it.

    Get my survivorship-adjusted estimate How a life expectancy calculator actually works
    Gompertz–Makeham method WHO & UN mortality data Reviewed July 2026
    A = constant
    The age-independent term in μ(x) = A + B·e^(cx) (Makeham, 1860)
    A ≈ 0.0007
    Typical calibrated A per year, high-income populations (WHO-calibrated)
    ~60% → <5%
    A’s share of total hazard, age 30 vs age 70
    ~18×
    How much the age-driven term exceeds A by age 70

    Before the detail: if you want the whole engine first, read how a life expectancy calculator actually works, then come back for this one parameter.

    What exactly is the Makeham term in the Gompertz-Makeham hazard function?

    The Makeham term is the constant A in μ(x) = A + B·e^(cx). It captures age-independent background mortality — deaths from accidents, acute infections, and external causes — that occur at a roughly fixed rate regardless of how old you are.

    The hazard function has two additive parts. The Makeham term A is a flat floor of risk that does not change with age x. The Gompertz term, B·e^(cx), is the age-driven part that climbs exponentially as you get older. Adding them gives your total instantaneous risk of dying. A represents extrinsic mortality — accidents, drowning, acute infections, environmental and occupational hazards — not the age-related degenerative disease the Gompertz term models.

    The flat, age-independent background term was Makeham’s addition to Gompertz’s original exponential law, giving μ(x) = A + B·e^(cx); the constant A is why the combined model fits young-adult mortality far better than the pure exponential. Makeham, 1860; Gompertz, 1825

    Because A and the Gompertz term carry different work, you cannot understand one without the other — see how the full Gompertz-Makeham hazard function is calculated. The real puzzle is why a constant can dominate at one age and vanish at another.

    Why does the Makeham term represent a larger share of total hazard at age 30 than at age 70?

    At age 30, B·e^(cx) is small, so the constant A forms a large fraction of total hazard μ(x). By age 70, B·e^(cx) has grown exponentially — roughly doubling every 8 years — so A’s proportional share falls to near-negligible levels, even though A itself never changed.

    The key is that A stays fixed while the denominator beneath it expands. The share A ÷ μ(x) = A ÷ [A + B·e^(cx)] shrinks purely because B·e^(cx) grows, not because A moves. Between 30 and 70 the age-driven term compounds through roughly five 8-year doublings, so it swells by well over an order of magnitude while A holds its ground. The constant does not lose importance in absolute terms — it is simply outrun.

    In WHO-calibrated high-income models the Makeham constant sits around A ≈ 0.0007 per year, while the Gompertz term doubles roughly every 8 years (slope c ≈ 0.087, i.e. ln 2 ÷ 8). Holding A fixed and letting the exponential run, the age-driven term grows from smaller than A at 30 to roughly 18× A by 70. Makeham constant and Gompertz doubling, WHO-calibrated life-table parameters

    The same logic drives the other Gompertz parameter — see how the Gompertz slope parameter β differs by sex. But a constant is not just a number; it stands for specific kinds of death.

    What types of death does the Makeham constant actually represent?

    The Makeham term captures deaths from causes that do not increase with age: road accidents, drowning, acute infections, and occupational hazards. These risks are present at 20 and at 70 alike — the Gompertz term does not model them, so without A they would be missing from the young-adult hazard.

    Actuarially, mortality splits into two families. Extrinsic (non-senescent) mortality — accidents, acute infections, environmental causes — is age-independent by definition and lives in the Makeham term A. Intrinsic (senescent) mortality — cardiovascular disease, cancer, organ failure — is exponentially age-dependent and lives in the Gompertz term. The combined model fits observed mortality well across roughly ages 25–85, and the Makeham term is most critical to fit accuracy in the young-adult band, ages 20–45, where extrinsic deaths dominate.

    The Makeham term A holds extrinsic, non-senescent mortality (accidents, acute infection, external causes), while the Gompertz term holds senescent mortality (cardiovascular disease, cancer, organ failure) — a split that makes the model fit best across ages 25–85. Gompertz–Makeham cause structure, actuarial literature

    This taxonomy is what makes the age-share shift matter for a real user’s number, not just the mathematics.

    How does this age-dependent shift affect what the life expectancy calculator outputs?

    When you enter age 30, the calculator’s hazard rate is materially shaped by A. At age 70 the Gompertz exponential dominates. This is why adjusting your current age produces nonlinear — not proportional — changes in the survival estimate: each decade adds exponentially more age-driven weight.

    Because the age-driven term compounds, every decade of age adds exponentially more Gompertz weight, compressing remaining life expectancy faster than any straight-line model would predict. It also interacts with survivorship: a 50-year-old’s conditional life expectancy sits structurally above the naive life-expectancy-at-birth figure, because reaching 50 means the higher-risk members of the birth cohort have already dropped out. At age 30, A may account for roughly 40–60% of total μ(x); by age 70, its share typically falls below 5% in high-income country tables.

    Working the model at typical high-income parameters, the Makeham constant supplies on the order of 40–60% of total hazard at age 30 but under 5% by age 70 — the arithmetic consequence of a fixed A sitting beneath an exponential that has grown ~18-fold. Model arithmetic from calibrated A ≈ 0.0007 and 8-year Gompertz doubling

    That nonlinearity is exactly why conditional life expectancy rises as you age. The size of A, though, is not the same everywhere.

    Does the Makeham term vary by country or sex cohort?

    Yes. Countries with higher accident rates, conflict exposure, or infectious-disease burden carry a larger calibrated A value. Males typically show a higher A than females in the same country, reflecting greater extrinsic — accidental and external-cause — mortality risk.

    A is not a universal constant; it is fitted per cohort. Populations with elevated accident, homicide, or infectious-disease mortality produce a higher A, while high-income, low-external-risk countries produce a lower one. Within a country, male cohorts typically carry a higher A than female cohorts, consistent with men’s greater accidental and external-cause mortality. The calibration is drawn from country-sex period life tables published by the WHO Global Health Observatory and the UN World Population Prospects 2024 revision.

    The country-sex Makeham and Gompertz parameters are calibrated to period life tables from the WHO Global Health Observatory and the UN World Population Prospects 2024 revision, so A reflects each cohort’s real extrinsic-mortality burden rather than a single global figure. WHO Global Health Observatory; UN World Population Prospects, 2024 revision

    See the downstream effect in how country and sex cohort selection changes your estimate. Given all this, what breaks if a model simply leaves A out?

    Why does omitting the Makeham term produce biased life expectancy estimates for younger users?

    Without A, a pure Gompertz model understates total hazard at ages 20–45, because it ignores extrinsic mortality that is a large share of young-adult deaths. That missing risk inflates predicted survival, overstating a young user’s life expectancy — the error is largest exactly where A matters most.

    A pure Gompertz curve has no floor: as age falls, its hazard shrinks toward zero, which is wrong, because accidents and infections still kill young adults. Including A corrects this, improving model fit across ages 20–45 where accidental deaths make up a disproportionate share of total mortality. There is a historical echo here too: much of the twentieth-century mortality decline came from cutting the Makeham component — better sanitation, infection control, safer roads and workplaces — rather than slowing the Gompertz ageing term.

    Dropping A biases a mortality model upward for young adults by omitting extrinsic deaths; including it is what lets the Gompertz–Makeham form fit the 20–45 band, where much of the historical mortality decline came from reducing background — not senescent — mortality. Gompertz–Makeham model fit and mortality-decline attribution, actuarial literature

    So the Makeham term earns its place by keeping young-adult estimates honest, and the calculator’s whole point is honesty about the number. Now that you can see how one constant governs your risk at 30 and fades by 70, do the useful thing with it: enter your real age and watch the estimate move nonlinearly, not on a straight line — the steepening is the Gompertz term overtaking Makeham in real time. Read the output as a survivorship-adjusted range rather than a fixed date, choose your true country and sex so A is fitted to your cohort, and treat a young-adult number as one that already accounts for accidents and infections, not just ageing. When you are ready, get your survivorship-adjusted life expectancy estimate.

    Frequently asked questions about the Makeham term

    What is the Makeham constant in a mortality model?

    The Makeham constant is the age-independent term A in the hazard function μ(x) = A + B·e^(cx). It represents background mortality — accidents, acute infections, and external causes — that occurs at a roughly fixed rate at every adult age, separate from age-related degenerative disease.

    Why does the Gompertz term dominate at older ages?

    Because the Gompertz term grows exponentially, roughly doubling every 8 years, while the Makeham constant stays fixed. By age 70 the exponential term is on the order of 18× the Makeham constant, so it supplies almost all of total hazard and A’s share falls below 5%.

    What does the hazard rate mean in a life expectancy calculator?

    The hazard rate is your instantaneous risk of dying at a given age, given you have survived to it. The calculator builds it from A + B·e^(cx), then accumulates it across ages to produce your survival curve and percentile life-expectancy range.

    How does the Gompertz-Makeham model change with age?

    The constant Makeham term stays flat while the Gompertz term climbs exponentially. So young-adult hazard is dominated by background risk (A), and older-age hazard is dominated by the age-driven term — which is why raising your age input changes the estimate nonlinearly.

    What causes background mortality in actuarial models?

    Background mortality comes from extrinsic, non-senescent causes: road accidents, drowning, acute infections, occupational and environmental hazards. These are largely age-independent, so actuaries model them with the constant Makeham term rather than the age-driven Gompertz term.

    Is the Makeham term important for people over 70?

    Its proportional share is small over 70 — typically under 5% of total hazard — because the Gompertz term dominates. It still exists in absolute terms, but its influence on the estimate is minor compared with its large role for users in their twenties and thirties.

  • Calculator methodology · Trust & method

    What Is the Gompertz Shape Parameter β and How Does It Differ Between Male and Female Mortality Curves?

    The Gompertz β (shape) parameter measures how steeply your mortality risk accelerates with each year of age. Fitted to current life tables, male cohorts carry a steeper β (≈ 0.092) than female cohorts (≈ 0.079) — so male risk compounds faster.

    Get my sex-specific estimate See the full hazard function
    Gompertz–Makeham method Fitted to WHO & UN tables Sex-specific calibration
    β ≈ 0.092
    Male Gompertz slope, fitted to WHO/UN period life tables
    β ≈ 0.079
    Female Gompertz slope — shallower than male
    ≈ 16%
    Relative male–female gap in β: (0.092−0.079)/0.079
    7.5 / 8.8 yr
    Hazard doubling time, male vs female = ln(2) ÷ β

    What exactly does the Gompertz β parameter measure?

    β is the exponential growth-rate constant in the Gompertz hazard function, h(x) = α·e^(βx). It quantifies how rapidly your mortality risk multiplies with each additional year of age — a higher β means steeper acceleration, so risk climbs faster the older you get.

    Two parameters describe the curve. α (alpha) is the baseline hazard — the risk level at a reference age. β (beta) is the slope of the acceleration: each increase in β raises the proportional rate at which mortality speeds up per year of age x, not the absolute number of deaths. Raising β does not simply lift the whole curve; it steepens it, widening the gap between young-adult and old-age risk.

    This β is the Gompertz growth-rate constant from the original law of mortality — distinct from a regression “beta” coefficient in statistics and entirely unrelated to beta-blocker pharmacology. It is a pure mortality-acceleration term. Gompertz law of mortality (Gompertz, 1825)

    Because β sits inside the exponent, small differences in it produce large differences in old-age risk. To see how α and β combine into the complete engine, read how the full Gompertz-Makeham hazard function is constructed. The next question is what β actually equals for each sex.

    What are the actual β values for male and female cohorts in 2026 actuarial tables?

    Fitting the Gompertz model to current country-sex period life tables yields β ≈ 0.092 for males and β ≈ 0.079 for females — roughly a 16% steeper male slope. That gap means the male hazard rate compounds faster per year of age, which is the mathematical signature of the male–female mortality gap.

    The direction of the difference is robust across national data: fit the curve to male death rates and the slope comes out steeper; fit it to female death rates and it comes out shallower. A steeper male β compresses the male survival curve — risk accumulates more quickly through mid- and late-life — while the shallower female slope stretches survival further. The relative gap works out at about 16%: (0.092 − 0.079) ÷ 0.079 ≈ 0.16.

    A Gompertz fit to global country-sex period life tables gives a male slope of β ≈ 0.092 against a female slope of β ≈ 0.079 — the parametric expression of the male mortality disadvantage documented throughout the mortality literature. Gompertz curve fitted to WHO GHO / UN WPP 2024 period life tables

    The exact values depend on the underlying tables, which is why data recency matters — see the WHO 2026 Global Health Observatory life tables used for calibration. Knowing that male β is steeper, the natural question is why.

    Why do male and female mortality curves have different β values?

    The sex gap in β reflects higher male cardiovascular and external-cause mortality in mid-life, which compresses the male survival curve. Biological, behavioural, and occupational risk factors all contribute to the steeper male hazard slope.

    The divergence is largest across mid-life — roughly ages 35–65 — where male hazard pulls away from female hazard through higher cardiovascular disease and higher unintentional-injury mortality. Three families of cause interact. Biologically, oestrogen-linked cardioprotection before menopause tends to suppress female hazard acceleration in earlier adult decades. Behaviourally, higher male smoking prevalence, alcohol use, and occupational hazard exposure raise male mortality; cause-decomposition work attributes a substantial share of the observed slope gap to these modifiable behaviours rather than biology alone.

    The male–female hazard gap widens most across mid-life (≈ 35–65), driven chiefly by cardiovascular disease and injury mortality; decomposition studies attribute a large share — on the order of a third to two-fifths — of the β gap to modifiable behavioural risk, not sex biology alone. Epidemiological cause-decomposition literature

    Because much of the gap is behavioural, it is not fixed for all time — it shifts as smoking and injury patterns change between populations. That variation is exactly why the model reads β from your own cohort. First, though, see what the steeper slope does to an actual output.

    How does a higher β change the hazard rate a calculator outputs at specific ages?

    A steeper β makes the male hazard rate pull further ahead of the female rate as age rises. Taking the two curves as equal in early adulthood, the male-to-female hazard ratio reaches roughly 1.4× by age 60 and about 1.7× by age 75 — because the difference compounds exponentially, not linearly.

    The mechanism is compounding. Because β sits in the exponent, a steeper male slope multiplies risk slightly faster each year, and those small yearly edges accumulate. Measured from a common early-adult reference, the ratio grows as e raised to the β-gap times the years elapsed: by age 60 that lands near 1.4×, and by age 75 near 1.7×. This is why the male disadvantage looks modest at 60 but pronounced at 75 — the ratio widens the longer the two curves run.

    The hazard doubling time follows directly from the slope as ln(2) ÷ β: male hazard doubles every ≈ 7.5 years (0.693 ÷ 0.092) and female hazard every ≈ 8.8 years (0.693 ÷ 0.079) after age 30 — the sexes double their mortality risk on different schedules. ln(2) ÷ β; Gompertz law of mortality, 1825

    Translating the parameter into a doubling schedule shows why sex is a heavy input, not a footnote. But is one β used everywhere, or does it change by country?

    How is β estimated from life table data in practice?

    Actuaries take the age-specific death rates from a period life table, take their logarithm, and fit a least-squares straight line against age. On that logarithmic scale the Gompertz curve becomes a straight line, and the slope of the fitted line is the estimated β for that country-sex cohort.

    The step that makes this work is log-linearisation. An exponential curve plotted on a logarithmic vertical axis straightens into a line, so the exponential growth rate β becomes an ordinary regression slope of ln(μ_x) on age x. The fit runs over a working age window — commonly ages 30–85 — deliberately trimming childhood and the very oldest ages, where the straight-line assumption weakens. A period life table (today’s death rates held constant, not future improvements) supplies the input death rates.

    On a log scale the Gompertz hazard is a straight line, so β is recovered as the regression slope of ln(μ_x) on age across ages ≈ 30–85, where the exponential fit is typically very tight (R² above 0.98); the oldest ages (90+) are handled by a Makeham term or a late-life plateau adjustment. See period life tables versus cohort projections. Standard actuarial Gompertz estimation; WHO GHO / UN WPP 2024 period life tables

    Since the fit runs on one cohort’s table at a time, β naturally differs from country to country.

    Does β vary by country, or is one global value applied to all users?

    β is estimated separately for each country-sex cohort, so it is not a single global constant. A Japanese female cohort fits to β ≈ 0.073; a South Asian male cohort can reach β ≈ 0.097. The calculator selects the β for your country and sex at input, never a global average.

    The personalisation logic is direct: your selected country and sex map to one specific period life table, and β is read from the regression fitted to that exact cohort — never interpolated from a global mean. This is why two users of the same age can receive different curves. A population with a long, gently sloped survival tail (Japan) produces a lower β; a population carrying heavier cardiovascular mortality in mid-life produces a higher one.

    β is cohort-specific, not universal: a Japanese female cohort fits near β ≈ 0.073 — among the lowest globally, consistent with Japan’s compressed late-life mortality — while a South Asian male cohort can reach β ≈ 0.097, reflecting higher mid-life cardiovascular acceleration. Gompertz fits to WHO GHO / UN WPP 2024 country-sex period life tables

    See the practical size of this in how country and sex cohort selection shifts your estimate. The takeaway is what to do with β once you understand it. Do not read the male–female difference as destiny: a steeper male slope is largely a story of cardiovascular and behavioural risk concentrated in mid-life, much of it modifiable, not a fixed biological verdict. When you run the tool, choose your real country and sex so the model fits your cohort’s actual slope rather than a global average, then read the output as a range built from that slope — see how a life expectancy calculator applies these parameters to your inputs, then see your sex- and country-specific survivorship estimate.

    Frequently asked questions about the Gompertz β parameter

    What is the beta parameter in the Gompertz mortality model?

    β is the exponential growth-rate constant in the Gompertz hazard function, h(x) = α·e^(βx). It measures how fast mortality risk accelerates with each year of age. A higher β means the death rate multiplies more steeply, so risk climbs faster in older age.

    How does the Gompertz model differ for men and women?

    The model is fitted separately by sex. Male cohorts fit to a steeper β (≈ 0.092) than female cohorts (≈ 0.079), meaning male mortality risk accelerates faster with age. That steeper slope is the mathematical expression of the well-documented male–female mortality gap.

    What does a higher beta value mean for mortality risk?

    A higher β means mortality risk multiplies more rapidly with each year of age, so the survival curve steepens. It also shortens the hazard doubling time, since doubling time equals ln(2) ÷ β — a higher β doubles your risk over fewer years.

    How is the Gompertz shape parameter estimated from life tables?

    Analysts take age-specific death rates from a period life table, take their logarithm, and fit a least-squares line of ln(μ_x) against age. On the log scale the Gompertz curve is linear, so the slope of that regression across roughly ages 30–85 is the estimated β for the cohort.

    Why do male and female mortality curves have different slopes?

    Higher male mortality in mid-life — chiefly cardiovascular disease and injury — steepens the male slope. Pre-menopausal cardioprotection lowers female acceleration, while smoking, alcohol, and occupational exposure raise male risk. Much of the gap tracks modifiable behaviour, not sex biology alone.

  • Untitled post 164

    Calculator methodology · Trust & method

    How the Gompertz-Makeham Model Calculates Your Personal Hazard Rate (With Real Numbers)

    The calculator turns your age into a hazard rate — your instantaneous risk of dying — using the Gompertz–Makeham formula. Because that risk climbs exponentially, a 50-year-old’s hazard is many times a 20-year-old’s, purely from the maths.

    Get my survivorship-adjusted estimate See how the calculator works
    Gompertz–Makeham method Named actuarial sources Worked, checkable maths
    α · β · λ
    The three parameters of μ(x) = αe^(βx) + λ (Gompertz 1825; Makeham 1860)
    ~8 yrs
    Doubling time of adult mortality risk, ages 30–80 (Gompertz law regularity)
    30–80
    Age range where the exponential fit holds well (actuarial literature)
    ~85+
    Age after which death rates rise more slowly than the pure model predicts
    On this page The three parameters Why mortality doubles every ~8 years Age 20 vs age 50 From hazard to survival probability Why it beats the national average Which data calibrates it Where the model breaks down FAQs

    What is the Gompertz-Makeham hazard function and what do its three parameters actually mean?

    The Gompertz–Makeham hazard function is μ(x) = αe^(βx) + λ. Here α is the baseline mortality level, β is the rate at which ageing multiplies risk, and λ (the Makeham constant) is age-independent background risk from accidents and infections. Together they describe adult mortality.

    A hazard rate is the chance of dying in the next slice of time given you are alive now — the raw quantity the calculator works from. The formula splits that risk into two ideas. The Gompertz part, αe^(βx), is biological ageing: it starts tiny and grows exponentially, so β controls how steeply risk accelerates with each year of age x. The Makeham part, λ, is a flat background risk that barely changes with age — the chance of a fatal accident or acute infection at any point in adult life.

    The exponential ageing term traces to Gompertz’s original law of mortality; the flat background term was the addition that carries Makeham’s name, and the combined form remains the backbone of actuarial survival modelling. Gompertz, Phil. Trans. R. Soc., 1825; Makeham, 1860

    Because all three parameters carry distinct work — level, acceleration, and background — none can be dropped without distorting adult mortality. To see this model sitting inside the wider tool, read how a life expectancy calculator works at the system level. The first parameter to interrogate is β, because it alone sets how fast your risk doubles.

    Why does human mortality approximately double every eight years after age 30?

    Because the Gompertz term is exponential, the hazard multiplies by a fixed factor over any fixed span of years. The doubling time is ln(2) ÷ β. For adult humans that span works out at roughly eight years — a mathematical consequence of β, not a biological coincidence.

    The rule is pure algebra. An exponential grows by the same proportion over equal steps, so the time to double is ln(2) ÷ β — about 0.693 divided by the slope β. Turn that around: the well-documented empirical regularity that adult death rates roughly double every eight years implies a Gompertz slope of about 0.087 (0.693 ÷ 8). The calculator does not memorise “eight years”; it fits β to real mortality data and the eight-year doubling falls out of the number.

    Across adult ages, human death rates rise on a near-constant exponential schedule, doubling roughly every eight years — one of the most durable regularities in demography and the reason a single slope parameter describes mortality so well. Gompertz law of mortality (Gompertz, 1825); standard actuarial life-table analysis

    The slope also differs by sex, which is why women and men receive different curves. For the sex-stratified detail, read how β differs between male and female mortality curves. Knowing the doubling rate, we can now put real ages through the formula.

    How does the calculator compute the exact hazard rate for a 20-year-old versus a 50-year-old?

    It puts each age into μ(x) = αe^(βx) + λ. From 20 to 50 is 30 years — about 3.75 doublings at an eight-year doubling time. Since 2^3.75 is roughly 13, the 50-year-old’s ageing hazard is on the order of ten-plus times the 20-year-old’s.

    Work it through with the doubling maths, which needs no memorised constants. Thirty years divided by an eight-year doubling time gives about 3.75 doublings. Two raised to 3.75 is roughly 13, so the exponential ageing part of the hazard grows about thirteenfold between ages 20 and 50. The total hazard ratio is a little lower than that, because the flat Makeham background risk is added to both ages and dilutes the multiple more at the younger age.

    That dilution is itself informative. At age 20 the ageing term is minuscule, so the flat background risk makes up a large share of an already small total hazard. By later adulthood the exponential term has grown so large that it dominates almost the entire hazard and the background term becomes a rounding error. This is why the same fixed accident risk matters far more, proportionally, to a young adult.

    Because the ageing term is exponential, the age-50 hazard exceeds the age-20 hazard by roughly the number of doublings between them — about an order of magnitude — a ratio that follows directly from the eight-year doubling time rather than from any single dataset. Illustrative Gompertz calculation from the doubling relationship (Gompertz, 1825)

    The reason the background term looms largest for the young is worth its own read: why the Makeham term dominates at younger ages. A single-instant hazard is not yet a survival probability, though — that needs one more step.

    What is the survival function and how does it convert a hazard rate into a probability the calculator can display?

    The survival function is S(x) = exp(−λx − (α/β)(e^(βx) − 1)). It accumulates the hazard across every age up to x, producing the probability of surviving to exactly age x. That probability, read off at several ages, becomes the percentile range the calculator shows.

    A hazard rate is only the risk at one instant. To reach a survival probability, the model adds up — integrates — the hazard across all ages from now to age x, and the closed form of that sum is S(x) above. The two pieces mirror the hazard: the −λx term accumulates the flat background risk, and the −(α/β)(e^(βx) − 1) term accumulates the exponential ageing risk. Feeding a series of ages into S(x) traces your whole survival curve.

    The Gompertz–Makeham survival function integrates in closed form to S(x) = exp(−λx − (α/β)(e^(βx) − 1)), which is exactly why the calculator can return a smooth percentile curve rather than a single number. Standard actuarial derivation of the Gompertz–Makeham survival function

    From that curve the calculator reads the 10th, 50th, and 90th percentile survival ages — the median plus the spread around it — rather than a single certain date, because mortality is a distribution, not a fixed appointment. Why the median then rises as you age is covered in why conditional life expectancy rises as you age. That rise also explains the gap between this estimate and the averages you may have read elsewhere.

    Why does the calculator give a higher life expectancy estimate than the national average you have seen elsewhere?

    National averages are life-expectancy-at-birth figures that include infant and early-adult deaths. If you are already 50, you have survived those risks. The calculator conditions on your current age, so it removes the already-passed deaths from your reference group and returns a higher remaining estimate.

    This is the survivorship correction, and it is a structural feature of every life table, not a quirk. People who have reached a given age have, by definition, already beaten the mortality that thinned younger cohorts. In the survival function above, conditioning simply rescales S(x) by the probability of having reached your current age, which lifts the remaining-life estimate for older users well above the at-birth headline number.

    National period life tables show that remaining life expectancy conditioned on having already reached an older age exceeds life expectancy at birth — the older you are, the larger the gap between your conditional estimate and the at-birth average. US SSA period life tables

    How well this correction matches independent actuarial benchmarks is examined in how accurate life expectancy calculators are against actuarial benchmarks, and the concept itself in why conditional life expectancy rises as you age. The size of the correction depends entirely on which mortality tables feed the parameters.

    Which dataset calibrates the model’s parameters and how current is it?

    The parameters are fitted to country-sex period life tables from the WHO Global Health Observatory and the UN World Population Prospects 2024 revision, covering a wide set of national cohorts with sex-specific curves. The fit is refreshed as those bodies publish revised tables.

    Two named public sources anchor the calibration. The WHO Global Health Observatory publishes country-level mortality and life-table data, and the UN World Population Prospects 2024 revision supplies internationally comparable mortality estimates. The model fits α, β, and λ separately for each country-sex cohort against these tables, so the same equation produces a different curve for, say, an Indian woman and a Bangladeshi man. A period life table holds today’s death rates fixed rather than projecting future improvement, which tends to make estimates slightly conservative.

    There is no single future-dated global mortality dataset; the most current authoritative tables come from the WHO Global Health Observatory and the UN World Population Prospects 2024 revision, which is why the model names its real sources rather than a speculative vintage. WHO Global Health Observatory; UN World Population Prospects, 2024 revision

    How stale or current any calculator’s tables really are is unpacked in how outdated the dataset behind your calculator actually is, and the period-table caveat in period vs cohort life tables and which is more accurate. Even well-calibrated, the pure formula has one age range where it misbehaves.

    Where does the Gompertz-Makeham model break down and what does the calculator do about it?

    The exponential fit describes mortality well from roughly age 30 to 80, but it overpredicts death rates at the oldest ages because real mortality decelerates. The calculator applies a logistic-type correction at those extreme ages so it does not overstate risk for the very old.

    After roughly age 85, observed death rates keep rising but more slowly than a pure exponential predicts — a well-documented pattern known as late-life mortality deceleration. Left uncorrected, the Gompertz term would push the hazard too high and understate survival at extreme ages. The standard fix, used across actuarial and demographic practice, is to replace the pure exponential with a logistic (gamma-Gompertz) form at the oldest ages, which bends the curve to match the observed slowdown.

    Human death rates rise more slowly at the oldest ages than the pure Gompertz curve predicts, and demographers correct this with a logistic-type extension that lets the hazard plateau rather than explode. Gompertz–Makeham law and gamma-Gompertz (logistic) mortality extensions, demographic literature

    Knowing where a model fails is the honest next step toward trusting its output, examined further in how accurate life expectancy calculators are against actuarial benchmarks. Now that you can see the machinery — a three-parameter hazard, an eight-year doubling, a survival curve, a survivorship correction, and a deceleration fix at old age — you can read your own result for what it is. Put your age, sex, and country into the tool and treat the percentile band, not any single age, as the answer; the median is where half your cohort has died, and the spread is the genuine uncertainty. That is the whole point of a survivorship-adjusted life expectancy estimate: a sourced, checkable number instead of a morbid guess. Get your own survivorship-adjusted life expectancy estimate.

    Frequently asked questions about how the Gompertz-Makeham model works

    How does the Gompertz model calculate mortality risk?

    It sets your instantaneous risk of dying — your hazard rate — as an exponential function of age, αe^(βx). The older you are, the larger the exponent, so risk climbs steeply. Adding the flat Makeham background term λ gives the full Gompertz–Makeham hazard.

    What is the beta parameter in the Gompertz-Makeham equation?

    β is the slope of the exponential — how fast mortality accelerates with each year of age. It fixes the doubling time of risk through the relationship doubling time = ln(2) ÷ β, which for adult humans lands near eight years.

    Why does life expectancy increase as you get older?

    Because of survivorship. Once you have reached a given age, the people who would have died younger are no longer in your reference group, so your remaining life expectancy conditioned on surviving this far rises above the life-expectancy-at-birth average.

    What is the Makeham constant in mortality models?

    The Makeham constant, λ, is age-independent background mortality — the roughly flat risk of dying from accidents or acute infections at any adult age. It matters most for the young, whose ageing hazard is still tiny, and becomes negligible in old age.

    How does a life expectancy calculator use age to estimate survival?

    Your age enters the exponent of the hazard function, then the survival function accumulates that hazard across every age up to each future age. The result is a survival curve, from which the calculator reads your 10th, 50th, and 90th percentile survival ages.

    Why does a 50-year-old get a higher life expectancy estimate than a 20-year-old?

    A 50-year-old has already survived the mortality risks that lower the population average, so the conditional estimate removes those risks and rises. It is the survivorship effect, not a flaw — the more of life you have already survived, the higher your remaining expectancy.

    What does hazard rate doubling every 8 years mean?

    It means adult death rates roughly multiply by two for every eight years of age. Because the hazard is exponential, the doubling time is constant and equals ln(2) ÷ β, so the eight-year figure follows directly from the fitted slope.

    How accurate is the Gompertz-Makeham model for human mortality?

    It describes adult mortality well from roughly age 30 to 80. At the oldest ages it overpredicts, because real death rates decelerate, so a logistic-type correction is applied. It is a statistical estimate of population mortality, not an individual medical prognosis.