Foundations of Life Expectancy

Life Expectancy: Definitions and Measures What Is Life Expectancy? Life expectancy is a straightforward but often misunderstood concept. At its core, life expectancy is the statistical estimate of the average number of years a person is expected to live—either from birth or from any given age. The most commonly cited figure is life expectancy at birth, denoted $e0$. This represents the average lifespan of all people born in a particular year, calculated based on current mortality patterns. If a country's life expectancy at birth is 78 years, this means that on average, people born that year are expected to live 78 years—though of course, many will live shorter or longer lives. Period Versus Cohort Life Expectancy There's an important distinction between two types of life expectancy measurements: Period life expectancy is what we usually mean when we cite life expectancy statistics. It assumes a hypothetical group of people experiences the current age-specific mortality rates throughout their entire lives. This is a snapshot based on current conditions—not a prediction that everyone born today will actually experience these exact mortality rates (since mortality rates will likely change). Cohort life expectancy measures the actual mean lifespan of a real group of people born in the same year, calculated after all members of that group have died. This is the "true" average lifespan, but we can only calculate it long after the cohort is gone. The period measure is therefore a useful tool for comparing life expectancy across countries and over time, even though it's based on a hypothetical scenario. Why Infant Mortality Matters So Much Here's something crucial to understand: in populations with high infant mortality, life expectancy at birth becomes extremely sensitive to infant deaths. This creates a potential distortion. Consider two scenarios: A population where half of all infants die, and survivors live to age 60 A population where all infants survive but everyone dies at age 50 The first population has a lower life expectancy at birth (around 35-40 years), even though those who survive childhood actually have a longer life ahead of them. Because of this sensitivity, demographers often use life expectancy at age 5 ($e5$) when comparing mortality patterns across populations with very different infant mortality rates. By conditioning on survival to age 5, we get a clearer picture of mortality patterns after early childhood, free from the distorting effect of infant deaths. What Life Expectancy Is NOT Several important distinctions will help clarify this concept: Life expectancy is not a prediction for any individual. If life expectancy is 78 years, this doesn't mean you'll live to 78. Some people die at 40; others at 100. Life expectancy is an average, not a guarantee. Life expectancy differs from longevity. Longevity refers to the fact that some individual members of a population live relatively long lives. A society with high longevity has people who live very long lifespans. But this is different from having a high life expectancy—which measures the average. A population could have a few very long-lived individuals (high longevity) but a low average lifespan (low life expectancy) if many people die young. Life expectancy is not the same as maximum lifespan. Maximum lifespan is the age of the longest-lived person in a population. This is a hard upper bound that, some researchers argue, may be relatively fixed for humans. Meanwhile, period life expectancy reflects the average and can rise even if maximum lifespan doesn't change—by reducing deaths at each age and allowing more people to approach that maximum. The visualization above shows how global life expectancy has changed, illustrating that these improvements reflect broader population changes, not shifts in the maximum possible age humans can reach. How to Calculate Life Expectancy: Key Concepts Survival and Death Probabilities Life expectancy calculations rest on fundamental probability concepts. Demographers define: $px$ = the probability that a person aged $x$ survives to age $x+1$ $qx$ = the probability that a person aged $x$ dies during the year (before reaching $x+1$) These are complementary: $qx = 1 - px$. Together, they form the foundation of life table analysis. From Discrete Years to Expected Lifetime Life expectancy calculations typically work with discrete yearly intervals. The curtate future lifetime $Kx$ is a random variable representing the number of complete years a person aged $x$ is expected to live. The curtate expected lifetime $ex$ is simply the expected value of $Kx$—the average number of complete years remaining: $$ex = E[Kx]$$ However, in reality, people don't always die on birthdays. To convert this to a more realistic measure, demographers typically add 0.5 years (assuming people die on average halfway through a year). This gives us the complete life expectancy: $$ex^\circ = ex + 0.5$$ This 0.5-year adjustment is a convenient simplification that recognizes that deaths are distributed throughout the year, not concentrated on specific dates. Life Tables: The Foundation of Calculation Life expectancy is calculated using life tables, which organize mortality data in a systematic way. Here's the basic process: Calculate age-specific death rates from actual mortality data. For each age group, this is: (number of deaths) ÷ (person-years lived at risk during that age interval) Convert to probabilities. These rates become $qx$ values—the probability of death at each age. Build the life table columns. Starting with a hypothetical cohort (often 100,000), the life table tracks: How many people survive to each age How many die at each age interval Years lived at each age interval Sum remaining years. Life expectancy at any age is the total years expected to be lived by survivors, divided by the number surviving to that age. Life tables are powerful because they let us calculate life expectancy at any age, not just at birth. We can ask: "How many more years is a 50-year-old expected to live?" The answer comes from looking at the remaining years in the life table divided by the number of 50-year-olds. <extrainfo> Optional: Smoothing and Advanced Methods Raw mortality data is often noisy due to random variation. Demographers apply smoothing techniques—such as the Gompertz function (which models how mortality rises exponentially with age) or cubic splines (flexible curves fit to the data)—to create smoother mortality curves that better reveal underlying patterns. For forecasting future life expectancy, methods like the Lee-Carter model extrapolate trends in age-specific death rates, allowing governments and institutions to project future life expectancy under different scenarios. </extrainfo> Conditional Life Expectancy: A Person Who Has Survived to Age X An important practical concept: if you've already survived to a particular age, your remaining life expectancy may be quite different from life expectancy at birth. Consider this: life expectancy at birth might be 75 years. But a 60-year-old person has already lived through all the risks of infancy, childhood, and early adulthood. Their remaining life expectancy—calculated as $e{60}$—might be 20 years, meaning they're expected to live to age 80. This conditional life expectancy reflects the fact that having survived to age 60 means you've already made it through the years when mortality risk was highest. This is why demographers often specify both the measure and the age: "life expectancy at age 60" is a different (and more optimistic) number than "life expectancy at birth." The graph above shows how life expectancy has evolved in different African countries. Notice the dips in some countries around 2000—these reflect the impact of crises like the HIV/AIDS epidemic, which increased mortality at working ages and temporarily reduced life expectancy. Common Misconceptions "Life Expectancy Hasn't Changed" — A Confusion with Maximum Lifespan One persistent debate concerns whether life expectancy improvements represent fundamental biological changes or just better public health and living conditions. The evidence is clear: age-specific mortality rates have declined across the adult lifespan. This means fewer adults die at each successive age interval. For example, the percentage of people who die between ages 60 and 65 is lower today than it was in previous generations. What this does NOT mean is that humans can now live longer than ever before. Maximum human lifespan—the oldest anyone has ever lived—appears to have an upper bound that hasn't moved dramatically. The key distinction: Maximum lifespan (oldest anyone ever gets): relatively stable Life expectancy (average of all lifespans): increasing through reduced deaths at all ages Confusing these two leads to false conclusions. Someone might say "humans still can't live past 120, so life expectancy hasn't really improved"—but this misses the point. Life expectancy improves when millions of people avoid early and mid-life deaths, not when the oldest person gets older. The historical maps show dramatic improvements in life expectancy globally from 1800 to 2015, with the biggest gains in developed countries and increasing gains in developing regions more recently. Context: Life Expectancy in Broader Perspective Health Equity and Social Determinants Life expectancy is not evenly distributed within or across populations. Health equity—the study of fair distribution of health outcomes—examines how social and economic factors shape life expectancy differences. Factors like income inequality, access to healthcare, education, and environmental conditions create substantial gaps in life expectancy across communities. Understanding life expectancy therefore requires understanding not just the statistical measure, but the social forces that determine who lives longer and who doesn't. <extrainfo> Biodemography: Where Biology Meets Demography Biodemography is the field that studies how biological processes (aging, disease, reproduction) interact with demographic patterns (mortality, fertility, migration). It bridges biology and demography to understand longevity from both perspectives—the biological mechanisms of aging and the population-level patterns that emerge from individual mortality. </extrainfo>