Fact-checked by Grok 2 weeks ago

Force of mortality

The force of mortality, denoted \mu_x, is a key concept in actuarial science and demography representing the instantaneous rate of mortality at age x, conditional on survival to that age.^[1] It is mathematically defined as \mu_x = \lim_{\Delta x \to 0^+} \frac{1}{\Delta x} \Pr(T \leq x + \Delta x \mid T > x), where T is the future lifetime random variable, providing a continuous analog to discrete mortality probabilities.^[2] This measure is fundamental to survival models, as the survival function from age x to x+t is given by S_x(t) = \exp\left(-\int_0^t \mu_{x+u} \, du\right), enabling precise calculations of life expectancies, annuity values, and insurance premiums.^[3] In practice, \mu_x is estimated from life tables or census data and often modeled using parametric forms like the Gompertz law, \mu_x = B c^x, which captures the exponential increase in mortality with age observed in human populations.^[4] For multiple lives, under independence, the joint force of mortality is the sum of individual forces, aiding in assessments of joint-life insurance and pension liabilities.^[5] Historically rooted in early life table constructions, such as those by Edmond Halley in the late 17th century, the force of mortality has evolved into a cornerstone of modern demographic forecasting and risk analysis, with applications extending to epidemiology and reliability engineering.^[6]

Definition and Interpretation

Mathematical Definition

The force of mortality, denoted as \mu_x(t), represents the hazard rate for an individual aged x after a duration t, formally defined as the limit

\mu_x(t) = \lim_{\Delta t \to 0} \frac{{}_{\Delta t}q_{x+t}}{\Delta t},

where {}_{\Delta t}q_{x+t} denotes the probability of death within the small interval \Delta t for a life aged x+t.^[7] Standard notation in actuarial contexts includes \mu(t) to indicate the force at attained age t for a general population starting from birth, while \mu_x(t) specifies the force at duration t for a life initially aged x; alternatively, \mu_{x+t} or \mu(x+t) may denote the force at attained age x+t.^[8] This formulation assumes a continuous-time model where the force of mortality is non-negative, \mu_x(t) \geq 0 for all t \geq 0, and integrable over [0, \infty) to ensure the corresponding survival probabilities are well-defined.^[7]

Intuitive Interpretation

The force of mortality represents the instantaneous rate at which mortality occurs at a given age, serving as the underlying "force" that propels individuals toward death over time. This concept captures the hazard or vulnerability to death at an exact moment, much like a force in physics denotes the rate of change in momentum, emphasizing the dynamic, point-specific risk rather than an aggregated measure. In demographic and actuarial contexts, it quantifies how quickly the population diminishes due to deaths at that precise instant, providing a continuous perspective on mortality dynamics.^[9] A useful analogy likens the force of mortality to the force of interest in finance, where both are continuous rates that compound over time to yield overall outcomes—here, survival probability instead of accumulated value. Just as a varying interest rate determines the growth of an investment through continuous compounding, the force of mortality integrates over an age interval to produce the probability of surviving that period, highlighting its role in modeling cumulative risk. This parallel underscores the force's utility in scenarios where mortality risk fluctuates smoothly with age, such as increasing vulnerability in later life stages.^[10] Unlike average mortality rates, which summarize deaths over fixed intervals like a year (e.g., the probability of death within that year), the force of mortality excels at depicting varying risks within those intervals, avoiding the smoothing effect that can mask age-specific peaks in hazard. For instance, while an annual rate might average a low early-year risk with a higher late-year risk, the force reveals the instantaneous escalation, enabling more precise predictions of survival trajectories in populations with heterogeneous mortality patterns. This distinction makes it particularly valuable for understanding why certain ages carry disproportionately higher risks compared to interval-wide averages.^[9]

Mathematical Relationships

Relation to Survival and Density Functions

In survival analysis, the force of mortality \mu_x(t), also known as the hazard rate at age x + t, is intrinsically linked to the survival function S_x(t), which represents the probability that an individual aged x survives an additional t years, or S_x(t) = P(T_x > t) where T_x is the future lifetime random variable. The relationship derives from the instantaneous nature of the force, yielding the survival function as S_x(t) = \exp\left(-\int_0^t \mu_x(u) \, du\right), assuming S_x(0) = 1. This exponential form arises because the probability of surviving a small interval du is approximately $1 - \mu_x(u) \, du, and integrating over [0, t] multiplies these infinitesimal probabilities.^[2]^[11] Equivalently, the force of mortality can be expressed as the negative logarithmic derivative of the survival function: \mu_x(t) = -\frac{d}{dt} \ln S_x(t). This characterization highlights how \mu_x(t) fully determines S_x(t) and vice versa in continuous-time models, providing a complete description of the lifetime distribution from the hazard perspective.^[12]^[2] The probability density function f_x(t) of the future lifetime T_x, which gives the probability density of death at exact time t after age x, is directly tied to both the force and survival functions via f_x(t) = \mu_x(t) S_x(t). This follows because the density is the product of the instantaneous hazard at t and the probability of having survived up to t, or equivalently, f_x(t) = -\frac{d}{dt} S_x(t). Substituting the survival expression yields f_x(t) = \mu_x(t) \exp\left(-\int_0^t \mu_x(u) \, du\right).^[2]^[11] A key intermediary is the cumulative hazard function H_x(t) = \int_0^t \mu_x(u) \, du, which accumulates the total risk exposure over the interval [0, t]. The survival function then simplifies to S_x(t) = \exp(-H_x(t)), underscoring the force of mortality's role as the derivative of the cumulative hazard, \mu_x(t) = \frac{d}{dt} H_x(t). This framework unifies the analysis of lifetime data in actuarial and demographic contexts.^[12]^[2]

Connection to Other Mortality Measures

The central death rate m_x, which represents the number of deaths between ages x and x+1 divided by the average population exposed to risk in that interval, serves as a discrete measure closely approximating the force of mortality. Precisely, m_x = \frac{\int_0^1 l_{x+t} \mu_{x+t} \, dt}{\int_0^1 l_{x+t} \, dt}, where l_{x+t} denotes the number of survivors at age x+t; this formula positions m_x as a weighted average of \mu_{x+t} over the year, with weights given by the survivors. A widely used approximation is m_x \approx \mu_{x+0.5}, valid under assumptions of near-constant or slowly varying force within the interval and introducing minimal error (typically less than 0.8% in parametric models like Gompertz).^[13]^[14] In life table construction, the one-year probability of death q_x connects directly to the force through the exact relation q_x = 1 - \exp\left( -\int_0^1 \mu_{x+t} \, dt \right), which follows from the survival probability {}_1p_x = \exp\left( -\int_0^1 \mu_{x+t} \, dt \right). This integral-based formula enables conversion from a specified continuous force to discrete probabilities, essential for building empirical life tables from parametric mortality laws. Furthermore, the number of survivors l_x in a life table relates to the force via l_x = l_0 \exp\left( -\int_0^x \mu_u \, du \right), where l_0 is the initial radix; inverting this yields \mu_x = -\frac{d}{dx} \ln l_x, facilitating derivation of the force from tabulated survivor data.^[13]^[15] The force of mortality differs from discrete measures like m_x and q_x by offering a continuous-time framework that inherently smooths irregularities in age-specific rates obtained from census or vital statistics data, where small event counts at certain ages can produce erratic patterns. By modeling \mu_x parametrically (e.g., via laws like Makeham or Gompertz), variability is reduced, enabling consistent interpolation between integer ages and more reliable projections beyond observed data.^[14]^[13]

Properties and Derivations

Fundamental Properties

The force of mortality, denoted \mu_x(t), is a non-negative function, satisfying \mu_x(t) \geq 0 for all t \geq 0, which ensures that it represents a valid instantaneous rate without implying negative probabilities of death.^[16] A key integrability property is that the integral \int_0^\infty \mu_x(t) \, dt = \infty, reflecting the certainty of eventual death over an infinite lifespan, while the integral over any finite interval remains finite.^[16] In human populations, the force of mortality typically exhibits a U-shaped pattern across the lifespan, with elevated levels in early infancy that decline to a minimum between ages 5 and 25, followed by a steady increase thereafter.^[17]^[18]

Derivation in Continuous Time

The force of mortality in continuous time arises from modeling the lifetime of an individual as a continuous random variable T, conditional on survival to age x. Let T_x denote the future lifetime random variable for a life aged x, with conditional survival function S_{T_x}(t) = P(T_x > t) and conditional probability density function f_{T_x}(t), both defined for t \geq 0. These functions describe the distribution of remaining lifetime given survival to age x.^[3] Under the continuous-time assumption, the force of mortality \mu_x(t) is derived as the instantaneous rate at which mortality occurs at time t after age x, given survival to that point. Specifically, it equals the ratio of the conditional density to the conditional survival function:

\mu_x(t) = \frac{f_{T_x}(t)}{S_{T_x}(t)}.

This expression captures the hazard rate, representing the limit of the conditional probability of death in a small interval [t, t + h) divided by h, as h \to 0^+, conditional on survival to t. The derivation assumes the density and survival functions are differentiable, ensuring the force of mortality is well-defined and continuous except possibly at finitely many points.^[2]^[3] An alternative derivation links the force of mortality to the logarithmic derivative of the survival function. Integrating the force yields the cumulative hazard \Lambda_x(t) = \int_0^t \mu_x(u) \, du, and the survival function satisfies S_{T_x}(t) = \exp(-\Lambda_x(t)), so \mu_x(t) = -\frac{d}{dt} \ln S_{T_x}(t). This relationship highlights how the force encodes the entire survival distribution in continuous time.^[1] In stochastic modeling, the force of mortality can be interpreted as the intensity function of a non-homogeneous Poisson process governing death events. Here, the lifetime T_x is the waiting time until the first event in such a process with time-varying intensity \mu_x(t), where the probability of an event in [t, t + dt) is approximately \mu_x(t) \, dt given no prior event. This analogy underscores the deterministic nature of the intensity in standard continuous-time mortality models, without invoking stochastic extensions like Itô calculus.^[19]

Applications and Models

In Actuarial Science

In actuarial science, the force of mortality, denoted as \mu_x(t), serves as a core component in the valuation of life contingent products, particularly through its role in deriving the survival function S_x(t) = \exp\left(-\int_0^t \mu_{x+s} \, ds\right), which underpins premium calculations for annuities and insurances.^[10] The net single premium for a whole life annuity-due payable continuously to a life aged x, denoted \bar{a}_x, is computed as the expected present value of payments over the lifetime, given by

\bar{a}_x = \int_0^\infty e^{-\delta t} \, S_x(t) \, dt,

where \delta is the force of interest; here, \mu_x(t) influences the integrand via the survival probability S_x(t), enabling actuaries to price products that account for instantaneous mortality risk.^[20] This integral form allows for flexible mortality assumptions, such as constant force \mu yielding \bar{a}_x = 1/(\mu + \delta), which simplifies premium determination for term annuities under uniform demographic conditions.^[21] Reserve calculations for life insurance policies similarly integrate the force of mortality to ensure solvency and policyholder protection. Thiele's differential equation provides a retrospective or prospective framework for reserve evolution, expressed as

\frac{d}{dt} \, {}_tV = \delta \, {}_tV + \mu_{x+t} (b - {}_tV) - p_t,

where {}_tV is the prospective reserve at duration t, \delta is the force of interest, \mu_{x+t} is the force of mortality, b is the benefit payment upon death, and p_t is the premium income rate; this equation links \mu directly to reserve dynamics by balancing interest accrual, mortality costs, and premiums.^[22] Solving this first-order differential equation numerically or analytically yields reserve trajectories that reflect mortality-driven outflows, essential for statutory reporting and asset-liability management in insurance portfolios.^[10] Stochastic extensions of mortality modeling incorporate the force of mortality into projections for long-term liabilities, such as pension funding, where uncertainty in future \mu(t) affects valuation. The Lee-Carter model parameterizes the logarithm of the central death rate m_{x,t} (which approximates the force of mortality \mu_{x,t}) as \log(m_{x,t}) = a_x + b_x k_t + \epsilon_{x,t}, with k_t following a random walk to forecast declining mortality trends; this approach is applied in actuarial practice to estimate pension obligations by simulating survival probabilities and adjusting reserves for longevity risk.^[23] Such projections have been instrumental in quantifying the impact of mortality improvements on funding shortfalls, as seen in U.S. public pension analyses where Lee-Carter forecasts inform contribution rates and benefit adjustments.

In Demographic Analysis

In demographic analysis, the force of mortality, denoted as μ(x, t), serves as a key tool for understanding population dynamics by distinguishing between period and cohort perspectives. Period analysis employs a cross-sectional approach, estimating μ(x, t) at a fixed time t across different ages x to capture the current mortality regime experienced by the population at that moment, which reflects contemporaneous health and environmental conditions.^[24] In contrast, cohort analysis tracks μ(x, t) for individuals born in the same year (fixed cohort), incorporating both past and projected future rates as they age, thereby providing insights into the lifetime mortality experience of a specific generation.^[25] This distinction is crucial because period measures often underestimate longevity improvements compared to cohort measures, as the latter account for ongoing declines in mortality over time.^[26] Decomposition of the force of mortality into cause-specific components, where the total μ(x, t) equals the sum of cause-specific forces μ_i(x, t) under competing risks assumptions, enables demographers to quantify the contributions of individual diseases or factors to overall mortality.^[25] This approach is particularly valuable in health policy, as it allows assessment of the impact of interventions, such as reductions in cardiovascular disease-related μ_i(x, t), on population-level survival and life expectancy trends.^[27] For instance, decomposing changes in μ(x, t) has revealed how declines in infectious disease causes have driven broader mortality improvements in low-mortality populations.^[28] For forecasting population dynamics, demographers apply smoothing techniques to estimate μ(x, t) from observed age-specific death rates, ensuring realistic projections free of erratic fluctuations. Kernel estimation, a non-parametric method, is commonly used to smooth these rates into a continuous force of mortality curve, facilitating accurate extrapolation in models like those employed by the United Nations for global population projections.^[29] Such techniques help predict future cohort trajectories by integrating historical patterns with assumed improvements, informing policy on aging populations and resource allocation.^[30] The force of mortality also underpins the construction of life tables, which summarize these demographic patterns.^[31]

Historical Development and Examples

Historical Context

The concept of the force of mortality, denoting the instantaneous rate at which individuals succumb to death at a given age, traces its early roots to the mid-18th century in the context of probability and annuity calculations. Leonhard Euler's 1760 paper, "Recherches générales sur la mortalité et la multiplication du genre humain," implicitly incorporated hazard-like rates by modeling survival probabilities over continuous time to value life annuities, laying groundwork for later formalizations in demography.^[32] Similarly, Pierre-Simon Laplace's 1778 memoir on probabilities extended these ideas by applying analytical methods to survival contingencies, influencing the mathematical treatment of mortality risks in probabilistic frameworks. In the 19th century, the force of mortality received its first explicit mathematical formulation through Benjamin Gompertz's seminal 1825 work, where he proposed an exponential model for the increasing intensity of mortality with age, derived from empirical life tables to enhance actuarial computations for life contingencies.^[33] This Gompertz law marked a pivotal advancement, shifting mortality analysis from discrete tables to continuous functions that captured the accelerating risk of death. William Matthew Makeham extended this in 1860 by introducing a constant term to the model, accounting for age-independent mortality causes like accidents, thereby improving fits to observed data across diverse populations.^[34] The 20th century saw the force of mortality integrated into standardized actuarial practices, with the Institute of Actuaries adopting it in the construction of English Life Table No. 8 during the 1920s, where it was used to graduate mortality data for assured lives and derive precise decrement measures. Post-World War II, stochastic demography further formalized its role, incorporating probabilistic variations in mortality processes, as exemplified in Nathan Keyfitz's contributions to mathematical demography that modeled random fluctuations in population dynamics.

Classical Examples

The Gompertz model describes the force of mortality as \mu_x(t) = B c^t, where B > 0 is a scale parameter representing the initial mortality level and c > 1 governs the rate of exponential increase with age t. This formulation captures the accelerating risk of death observed in adult human populations, primarily driven by biological aging processes after infancy. The model's simplicity and empirical robustness have made it a cornerstone in actuarial and demographic analyses, with the exponential term ensuring the force remains non-negative for all ages.^[35] The Makeham model builds upon the Gompertz form by incorporating an age-independent component: \mu_x(t) = A + B c^t, where A > 0 represents a constant hazard from extrinsic factors such as accidents or environmental risks that do not vary with age. This addition improves fit for populations where non-age-related deaths contribute significantly to overall mortality, particularly in younger adults. Like the Gompertz model, it maintains non-negativity and focuses on post-infancy patterns, with parameters estimated via maximum likelihood from life table data.^[35] Empirically, these models align well with human mortality data, illustrating the force's rise from approximately 0.001 at age 20 (reflecting low baseline risk in young adulthood) to around 0.1 at age 90 (indicating sharply elevated senescence-related hazards). For instance, U.S. actuarial tables show death probabilities q_x near these values, approximating the force of mortality under small-interval assumptions. However, both models exhibit limitations in childhood, where mortality rates peak in infancy due to congenital defects and infections before declining, deviating from the assumed exponential trajectory and requiring separate parametric adjustments for early life stages.^[36]^[37]

References

[1]
[PDF] Stat 344 Life Contingencies I Chapter 2: Survival models
Force of Mortality. We will define the force of mortality as. µx = lim dx→0+. Pr [T0 ≤ x + dx|T0 > x] dx. Then for small values of dx, we can use the ...
[2]
[PDF] Chapter 2 - Survival Models
Therefore the force of mortality function characterizes the survival function. Note also that. Sx (t) = S0(x + t). S0(x). =.
[3]
[PDF] Force of Mortality - Manual for SOA Exam MLC.
The force of mortality is also called the force of failure. An alternative notation for the force of mortality is µx . Force of mortality and force of failure ...
[4]
[PDF] A Note On Force Of Mortality - SOA
In actuarial literature there exists many assumptions regarding the distribution of deaths in unit intervals. The trio, namely, uniform distribution of death, ...
[5]
[PDF] Multiple Life Models
The force of mortality of the joint life status is the sum of the individuals' force of mortality, when lives are independent. Lecture: Weeks 9-10 (STT 456).<|control11|><|separator|>
[6]
https://www.degruyterbrill.com/document/doi/10.1515/9780773562301-028/pdf
[7]
[PDF] LIFE ASSURANCE MATHEMATICS W.F.Scott - Sushil Parajuli
... (q-type) rate of mortality at age x. (The m-type, or central, rate of ... This may be defined as the “instantaneous rate of mortality”, i.e.. µx = lim h ...
[8]
[PDF] Notation and Terminology Used on Exam LTAM - SOA
Aug 1, 2021 · The force of mortality may be represented by μx or μ(x) or μx+t or μ(x + t) where x and x+t are attained ages. The symbol μ[x]+tindicates ...
[9]
[PDF] Actuarial Mathematics and Life-Table Statistics
Finally, using equation (6.21) and the definition of q once more, we calculate. A1 x:ne äx:ne. = v³P t−1 k=0 vk kpx qx+k + P n−t−1 j=t vj jpx qx+j. Pt−1 k=0 vk.
[10]
[PDF] SUPPLEMENTARY NOTES FOR ACTUARIAL MATHEMATICS FOR ...
This note is provided as an accompaniment to 'Actuarial Mathematics for Life Contingent Risks' ... where the force of mortality in future years follows a ...
[11]
[PDF] Chapter 8 - Survival Distributions - UCSD Math
The next example shows that if a death may happen from several independent causes, then the force of mortality is the sum of the hazard rates corresponding to ...
[12]
Survival Analysis Part I: Basic concepts and first analyses - PMC - NIH
Survival data are generally described and modelled in terms of two related functions, the survivor function and the hazard function. The survivor function ...
[13]
Definition of , the central death rate
The force of mortality, being a continuous function, gives the instantaneous death rate at each separate age within this interval. More specifically, gives the ...
[14]
[PDF] A Practitioner's Guide to Statistical Mortality Graduation - SOA
4 The central death rate 𝑚𝑥 is closely related to the force of mortality 𝜇𝑥, and thus models for 𝑚𝑥 can be considered to be ones in continuous time. The ...
[15]
https://edfurman.info.yorku.ca/files/2020/09/MATH3280-2020-lec2.pdf
[16]
[PDF] Survival Models
force of mortality. 2.3 Force of mortality of Tx. In deriving the force of mortality, we can use the basic definition: µx(t) = fx(t). Sx(t). = f0(x + t). S0(x).
[17]
[PDF] How long do we live? - Knowledge Commons
... U-shaped with a minimum between ages 5 and 25. In contemporary populations with high levels of life expectancy, however, the force of mortality has a more ...
[18]
Explaining the Optimality of U-Shaped Age-Specific Mortality - PMC
Mortality is U-shaped with age for many species, declining from birth to sexual maturity, then rising in adulthood, sometimes with postreproductive survival.
[19]
[PDF] MLC 28-08 Study Note: Poisson Processes - Purdue Math
Aug 29, 2007 · ⇒ Poisson process”, while “non-homogeneous Poisson process” has been used to indicate a rate function that is not a constant.
[20]
[PDF] Educación para todos - actuarialestarea
Actuarial mathematics / by Newton L. Bowers . . . [et al.]. p. cm. Includes bibliographical references and index. ISBN 0-938959-46-8 (alk. paper). 1 ...
[21]
[PDF] Chapter 5 - Annuities
Example 5-5: Important Setting Assume a constant force of mortality. µ∗ x and a constant force of interest δ. Find ax and ax:n|. 5-23. Page 24. Example 5-6 ...<|control11|><|separator|>
[22]
[PDF] life contingency models ii - vladislav kargin - math.binghamton.edu
Under constant force of mortality tVx = 0. Proof. Under constant force ... This is called Thiele's differential equation. It has a clear intuitive mean ...
[23]
[PDF] Testing Deterministic versus Stochastic Trends in the Lee-Carter ...
Mortality assumptions are crucial to many areas of actuarial practice such as life insurance financial reporting and maintenance of private and public pension ...
[24]
Period and cohort-specific trends in life expectancy at different ages
Let μ(x,t) be the force of mortality at age x and time t for some population and let HA(x,t) and SA(x,t) be the cumulative hazard and survival functions, ...Article · Results · Percentiles Of The...
[25]
[PDF] Decomposing change in life expectancy - MPIDR
Decomposing Life Expectancy by Cause of Death. Let ¹i(a; t) be the force of mortality from cause of death i at age a and time t. The chance of surviving ...
[26]
Period and cohort life expectancy explained
Jan 19, 2023 · Period life expectancies tend to be lower than cohort life expectancies because they do not include any assumptions about future improvements in ...
[27]
A Cause-of-Death Decomposition of Young Adult Excess Mortality
Jun 4, 2018 · The method that we propose takes into account the contribution to the deviation of the force of mortality, rather than its absolute level.
[28]
Decomposing changes in life expectancy: Compression versus ...
Sep 1, 2015 · Methods: We introduce a decomposition method using newly developed parametric expressions of the force of mortality that include the modal age ...
[29]
Longevity studies based on kernel hazard estimation - ScienceDirect
We propose a non-parametric smoothing method to visualize the evolution of mortality rates. This is used to compare the mortality experiences of Denmark and ...
[30]
[PDF] World Population Prospects 2024: Methodology of the United ...
Jul 10, 2024 · For estimates, complete annual series of fertility, mortality and net international migration for calendar years 1950 through 2023 were ...Missing: kernel | Show results with:kernel
[31]
Mortality as a Function of Survival | Demography
Feb 1, 2023 · Function x ( s ) indicates the age at which survivorship is s, and μ ( s ) is the force of mortality at that specific survival level. We start ...Introduction · Survivorship Ages · Results · Discussion<|control11|><|separator|>
[32]
"Recherches générales sur la mortalité et la multiplication du genre ...
Sep 25, 2018 · Euler deals with questions of mortality and life expectancy. In particular, he addresses six questions: (1) A certain number of men, of whom all are the same ...Missing: 1760 paper<|separator|>
[33]
https://doi.org/10.1098/rstl.1825.0026
[34]
On the Gompertz-Makeham law: A useful mortality model to deal ...
As it is well-known, Gompertz's ideas can be properly expressed in terms of what we now call instantaneous intensity (or “force”) of mortality and constitute ...Missing: original | Show results with:original<|separator|>
[35]
Deciphering death: a commentary on Gompertz (1825) 'On the ... - NIH
In essence, the equation represents a force of mortality that increases progressively with age in such a manner that log μ(x) grows linearly. The term β might ...Missing: definition | Show results with:definition
[36]
Actuarial Life Table - Social Security
For this table, the period life expectancy at a given age is the average remaining number of years expected prior to death for a person at that exact age ...
[37]
How do we age? A decomposition of Gompertz law - ScienceDirect
Gompertz's law is a well-established regularity of human life, describing an almost perfect exponential increase in mortality with age.