Life expectancy is measured using a statistical tool called a life table, which tracks how many people in a population die at each age and converts that pattern into an average number of remaining years. The most commonly cited figure, life expectancy at birth, represents the average lifespan a newborn would have if today’s death rates at every age stayed constant for their entire life. In the United States, that number is currently 76.8 years for males and 81.7 years for females, based on the 2025 Social Security trustees report.
Period vs. Cohort Life Expectancy
There are two fundamentally different ways to measure life expectancy, and the distinction matters more than most people realize.
Period life expectancy takes the death rates from a single year (or a few years) and applies them to a hypothetical person passing through every age. If you see a headline saying “life expectancy in the U.S. is 77 years,” that’s almost always a period measure. It’s a snapshot of current mortality conditions, not a prediction of how long anyone alive today will actually live.
Cohort life expectancy follows everyone born in the same year and tracks their actual mortality over time. For people still alive, it fills in the future years with projected improvements in medicine, safety, and living conditions. Because it accounts for the reality that death rates generally fall over time, cohort life expectancy is almost always higher than the period number. The UK’s Office for National Statistics notes that period and cohort figures would only match if death rates never changed at all.
Nearly every national statistic you encounter uses the period method, because it doesn’t require guessing about the future. But that also means it systematically underestimates how long most people in developed countries will actually live, since medical advances tend to keep pushing mortality rates down.
How a Life Table Works
The core tool behind any life expectancy calculation is the life table. Building one requires two inputs: the number of people alive at each age and the number of deaths at each age. In practice, governments pull death counts from death certificates and population counts from census data or mid-year estimates. The CDC calculates age-specific mortality rates by dividing deaths in a given age group by the population of that age group at the midpoint of the year.
From there, the life table unfolds in a series of columns. The first tracks how many people out of a starting group (typically standardized to 1,000 or 100,000) survive to the beginning of each age interval. The next column records how many die during that interval. Dividing deaths by survivors gives the mortality rate for that age. A running total of all person-years lived from each age onward is then divided by the number of survivors at that age to produce the key output: expected remaining years of life.
This is why life expectancy “at age 65” can be surprisingly high even when life expectancy at birth seems modest. A 65-year-old has already survived the mortality risks of infancy, adolescence, and middle age. According to the 2025 Social Security projections, a 65-year-old American male can expect another 18.5 years on average, and a 65-year-old female another 21 years.
What Drives the Numbers Up or Down
Life expectancy is shaped by a web of socioeconomic, environmental, and health system factors. Research published in the Journal of Global Health identifies six major categories that consistently predict a country’s life expectancy: education levels (measured by average years of schooling), economic output (gross national income per capita), environmental infrastructure (like access to electricity), social stability (employment rates), urbanization, and public health capacity (vaccination coverage, HIV prevalence, and pandemic response).
These factors explain much of the gap between wealthy and poor nations. Global life expectancy rose from 66.8 years in 2000 to 73.1 years in 2019, according to the WHO, but that progress wasn’t evenly distributed. The COVID-19 pandemic then erased years of gains: global life expectancy fell to 72.5 in 2020 and further to 71.4 in 2021, rolling back to levels not seen since 2012.
The Genetics Question
How much of your lifespan is written in your DNA? Traditional twin studies placed genetic heritability of longevity at just 20 to 25%, and some large family-tree analyses suggested it could be as low as 6%. A more recent study published in Science, however, argues those numbers are misleadingly low because earlier generations died so often from external causes like infections, accidents, and poor sanitation that genetic differences were drowned out. Once those extrinsic deaths are statistically removed, the heritability of lifespan appears to be around 50%, similar to traits like height. That said, this figure reflects genetic influence under the specific historical conditions those cohorts lived in. The remaining variance comes from lifestyle, nutrition, medical care, and other environmental exposures.
Healthy Life Expectancy
Living longer doesn’t necessarily mean living well, which is why the WHO developed a companion metric called Healthy Life Expectancy, or HALE. It starts with the same life table framework but subtracts years spent in poor health, weighted by severity. A year lived with a mild limitation counts as nearly a full healthy year; a year with severe disability counts as much less.
The method, known as Sullivan’s method, divides each year of life in the table into a fraction of full health and a fraction of equivalent lost health, using country-level data on disease burden. In 2020, global HALE at birth was 62.8 years, roughly ten years less than overall life expectancy. That ten-year gap represents the average time people spend living with significant health problems. HALE is particularly useful for comparing countries that may have similar life expectancies but very different quality-of-life profiles in old age.
Estimating Individual Life Expectancy
Population averages tell you about groups, not about you. Researchers have developed individual prediction tools that use personal health data to generate more tailored estimates. One validated example, published in the Journal of General Internal Medicine, predicts remaining lifespan for older adults with diabetes using just 11 variables: age, sex, presence of heart failure, dementia, metastatic cancer, peripheral vascular disease, BMI, kidney function (serum creatinine), protein in the urine, home oxygen use, wheelchair use, and current smoking status. Each factor is assigned a point value, and the total score maps to a life expectancy estimate with reasonable accuracy (the model correctly ranks individuals about 79% of the time).
Online life expectancy calculators for the general population typically ask about similar categories: age, sex, smoking, BMI, exercise habits, chronic conditions, and family history. These tools are useful for spotting modifiable risks, but they carry real limitations. They’re built from population data, so they perform best for people who resemble the original study group and worst for outliers.
Why These Numbers Are Often Wrong
Period life expectancy, the figure you see most often, has a built-in flaw: it freezes today’s mortality rates in place and pretends they’ll never change. For most of modern history, that assumption has been wrong in the same direction, making the official number too pessimistic. A baby born today will benefit from decades of medical and public health improvements that aren’t captured in the period snapshot.
Cohort measures try to fix this by projecting future mortality improvements, but projections carry their own uncertainty. Research in the journal Demography found that calculating cohort life expectancy from period data introduces systematic errors, with estimates off by up to 4.8% in some cases. The authors warn that researchers can end up chasing trends that are actually artifacts of how the data was sliced rather than real changes in how long people live.
At the individual level, prediction tools face a different problem: they can’t account for rare events. A person with a perfect risk profile can still die young from an accident or an aggressive cancer, and someone with multiple risk factors can live to 95. These models describe probabilities across thousands of people, not certainties for any one person. Their real value is comparative: understanding which factors shorten or lengthen life, and by roughly how much, so you can focus on the ones within your control.