Research chronicle

Many population projections used by international and government agencies assume that, in today's low-fertility societies, future fertility rates will stop declining and roughly stabilize. This paper assesses the evidence for an equilibrium point or rebound from low levels. We show that there is no robust evidence for a systematic fertility rebound, or for a stable equilibrium point at any fertility level conventionally assumed. Instead, where and when fertility has fallen low, it has tended to keep falling. This is true across a range of data sources, time periods, and fertility measures: United Nations data on period total fertility rates for all country-years from 1950 to 2023; Human Fertility Database data on completed fertility for cohorts born between 1935 and 1975 in 31 countries; and fertility rates in subnational geographies in Japan, India, and the United States. We reconcile these results with a prior literature that documented fertility reversals among some low-fertility countries in the early 2000s. Our findings provide an empirical basis for re-examining the equilibrium and rebound assumptions embedded in widely used population projections.

We show how fertility delays compound across generations to raise age at first grandmotherhood. Under the 1960 first-birth schedule median grandmotherhood came at 44. Under the 2024 schedule it is 57, cutting the healthy years with grandchildren from 28 to 15. Longer life does not make up the loss of years when the grandmother is healthy enough to supply money and care. Delay is most pronounced among the college-educated, with reproductive-control access accounting for over half the fall in grandmotherhood by age 55 among White college graduates. When a daughter times her first birth to maximize her own payoff, the loss her choice imposes on her mother is external to her calculation. We call this the Rotten Adult Kid Problem. Delay raises both the price of a grandmother's care and wealth per grandchild.

This paper sets up a model with endogenous fertility and urban migration. Children are costly in terms of both goods and space. Incomes depend on population density through an agglomeration effect. An increase in the strength of the agglomeration effect can set in motion a process whereby urban migration increases, and fertility rates fall in both urban and rural locations, similar to U.S. data. Urban fertility falls due to rising rents while rural fertility falls because incomes fall. The rural population vanishes in the long run but the urban location rebounds, ending up with the same population density that it started off with.

Child poverty under the Supplemental Poverty Measure fell 20 percentage points between 1967 and 2024, a decline conventionally credited to income growth and changing taxes and transfers. I show that 5.4 points—27% of the decline or 3.7 million fewer children in poverty in 2024—instead reflect a fertility dividend: smaller families spread finite resources across fewer children. Because equivalence scales and decompositions condition on family size, standard accounting records demographic change as income growth. I quantify the dividend with a rank-preserving transport of the family-size distribution, and corroborate the mechanism by showing that twin-induced parity transitions still raise child poverty 3–7 points.

In the late nineteenth century, Americans feared that crowded immigrant neighborhoods were spreading disease through industrial cities. We study this claim using newly constructed town-level panel data from Massachusetts, 1860 to 1915, a setting uniquely suited to this exercise because its early vital registration system allows consistent measurement of mortality before national registration became widespread. Using a shift-share instrumental variable strategy based on historical settlement patterns, we estimate the effect of local exposure to immigration on mortality. The relevant question is not whether immigrant families themselves experienced higher mortality, but whether immigration altered the local disease environment more broadly. We find that increases in the foreign-born share significantly raised infant and child mortality, with the infant effect falling disproportionately on the post-neonatal period. Mechanical changes in the composition of births by parent nativity account for only a portion of the estimated effect, implying that the results are not driven primarily by higher mortality among immigrant births. Instead, the evidence is consistent with increased exposure to immigrants, communicable disease, and spillovers to children of native-born parents. These findings highlight immigration as one factor shaping the mortality transition in crowded high-mortality environments.

We examine whether expanded legal access to the contraceptive pill for minors generated health benefits for subsequent generations. We use data from the restricted-use Natality files from 1970-2020, linking mothers to the level of access to the pill in their state and year of birth and the extent to which their mothers could access the pill as minors. Our results show that children whose mothers were born after expanded access to the pill as minors had better birth outcomes, with gains concentrated among Black families. Among affected births, legalization increased birth weight by 12-16 grams and reduced incidence of low birth weight by 2.7-3.4 percent. These magnitudes have economic significance, corresponding to approximately $113.8–$115.3 million in medical cost savings, measured in 2025 dollars, for the 2000 Black birth cohort alone. Our mechanism analysis points to a postponement of early childbearing by the generation directly affected by legal access to the pill as minors whose daughters subsequently attained higher levels of education, used more prenatal care, and had longer birth spacing. Taken together, our results suggest that reproductive health policy can generate multigenerational benefits and contribute to reductions in racial disparities in early-life health.

State legislatures enacted numerous public-health laws during the twentieth century, most of which likely had negligible effects on mortality. Using data on the universe of state-level public-health laws and regulations, I test whether these laws, taken together, account for a meaningful decline in mortality. To address endogeneity, I instrument for public-health legislation using the legislative calendar. Both OLS and IV estimates indicate that public-health laws reduced mortality, whereas other laws and repeals of public-health laws did not. Legislatures were most effective at reducing deaths from waterborne and drainage-related diseases, half as effective for airborne communicable diseases, and one-fifth as effective for noncommunicable diseases. I use text analysis to classify the laws into twenty topics and find that laws targeting particular causes of death produced larger mortality declines for those causes. The universe of public-health laws can account for approximately two-thirds of the mortality decline during the study period.

Highlights This is the first study of parenthood desires among LGB people across cohorts using a nationally representative sample. The prevalence of child-free desires among lesbian women has decreased across cohorts. Parenthood desires among gay and bisexual men have increased or remained stable across cohorts. Bisexual women show little cohort variation. Overall, desires for parenthood are higher when obstacles to accessing it are lower.

Highlights This study is the first in the San Diego region to evaluate population forecast error across multiple forecast vintages and increment years spanning more than five decades. Forecast accuracy declines with increasing forecast horizon, while forecast errors are also influenced by population growth rate, population size, and jump-off discrepancy. Findings confirm the generalizability of established relationships in the forecast evaluation literature while extending prior research through analysis of errors in population change rather than solely through population level. Simple naive extrapolation methods frequently matched or exceeded the accuracy of more complex forecasting approaches, demonstrating the importance of benchmarking forecast performance.

Abstract Fertility preferences offer insights into how social factors shape behavior. Despite their significance, it remains unclear whether preference measures capture distinct, firm constructs in respondents’ minds, or to what extent responses reflect survey design choices like question order. In two survey experiments, we tested whether the sequence of fertility preference measures affected responses. The first experiment manipulated the sequence of general ideal, personal ideal, wanted, and expected fertility among college students. Sequence changes caused no changes in means for these measures, except for expected, which declined when asked later. Personal ideal and wanted fertility were indistinguishable from each other and were fairly unresponsive to sequencing. The second experiment compared general and personal ideals among a sample representative of U.S. adults. Preferences were firm and distinct among reproductive‐age individuals. For older respondents, the measures were equivalent when asked first, but personal ideal declined when asked after general ideal. Across both samples, general ideal fertility remained similar and higher than other measures for reproductive‐age respondents. These findings suggest that most preferences are relatively unaffected by ordering. We also offer guidance for survey design; general ideals are not interchangeable with other preference measures, and fertility expectations may not be comparable across survey designs.

Highlights U.S. immigrants born in 1850–1890 had shorter lifespans than U.S.-born Whites. This mortality penalty was concentrated among specific origin groups. Migrants generally outlived origin populations and nonmigrant siblings. Migration–mortality patterns varied by origin, reflecting both origin and destination contexts.

We study the contribution of changing partnership patterns to recent fertility change. Using data for several European Union countries and the United States, we analyze fertility developments between 2005 and 2020. Applying a Kitagawa decomposition, we separate fertility changes into a composition component, i.e., shifts in partnership exposure, and a rate component, i.e.,changes in fertility within partnership statuses. We find that declines in marriage and cohabitation reduced total fertility rates in every country studied, including those where overall fertility increased. These results suggest that fertility trends cannot be understood solely as changes in childbearing behavior within unions, but also reflect a contraction in exposure to childbearing. Hence, the results highlight the importance of understanding union formation dynamics for explaining contemporary fertility change.

The fertility decline is everywhere in the world today and goes back decades for rich countries. Birthrates have been below replacement in the U.S. and Europe since the mid-1970s. Completed cohort fertility in the U.S. was lower for those born in 1955 than for 1975. The reasons for the initial declines involve greater female autonomy and a mismatch between the desires of men and women. Men generally benefit more from maintaining traditions; women often benefit more from eschewing them. When the probability is low that men will abandon traditions, some career women will not have children and others will delay.<br><br>Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at <a href="http://www.nber.org/papers/w35425" TARGET="_blank">www.nber.org</a>.<br>

The fertility decline is everywhere in the world today and goes back decades for rich countries.Birthrates have been below replacement in the U.S. and Europe since the mid-1970s.Completed cohort fertility in the U.S. was lower for those born in 1955 than for 1975.The reasons for the initial declines involve greater female autonomy and a mismatch between the desires of men and women.Men generally benefit more from maintaining traditions; women often benefit more from eschewing them.When the probability is low that men will abandon traditions, some career women will not have children and others will delay.

Highlights This article presents causal effects of the Great Recession on U.S. fertility. Effects are identified using a new cohort discontinuity design. The design identifies effects for single-year cohorts of women, as well as for single-year cohorts of White and Black women. Estimated effects are negative for U.S. women born in the period 1964‒1992. Negative effects are largest for younger cohorts and for Black women.

BACKGROUNDEconomically insecure conditions may be driving young people's decisions to delay, forgo, or reduce childbearing.Yet establishing a causal link between economic resources and fertility is methodologically fraught.One approach has been to use survey experiments to manipulate hypothetical economic conditions and gauge respondents' fertility preferences.

BACKGROUNDThe US labor force and economic dependency ratios are projected to change significantly through 2060, driven by demographic shifts and persistent inequalities across race, ethnicity, and gender.The United States has lower participation rates than other industrialized countries and large participation gaps between population groups. METHODSWe use a dynamic microsimulation model, incorporating US population projections, to project labor force participation and economic dependency ratios.We assess multiple scenarios, simulating the effects of reducing inequalities in education and health and equalizing participation across demographic groups. RESULTSBaseline projections indicate a labor force increase of approximately 27 million people by 2060, driven primarily by population growth.Reducing health and educational disparities could increase the baseline labor force projection by about 10%, adding 2.6 million people.The projected effects of closing gender and racial/ethnic participation gaps are substantial: equalizing participation rates between demographic groups could add up to 14.3 million people to the labor force.

Abstract Demographers have struggled to make realistic population projections for some race groups. For example, the Census Bureau's 2023 national projection gives the unrealistically low estimate that the American Indian and Alaska Native (AIAN) population will be 8.7 million in 2050 (it was measured at 9.7 million in 2020). I argue that this disconnect occurs because the cohort component model ignores an important component of change: race response change (whether due to complex identities or administrative processes). This research note introduces a strategy for incorporating net race response change into cohort component model projections. I apply the strategy to the racially identified AIAN population in the United States from 2020 to 2050, concluding that it may grow from 9.7 million in 2020 to 19.8 million in 2050.

BACKGROUNDMortality is known to be higher in winter than in summer, with excess winter deaths ranging between 5% and 30% in Europe.A recent study conducted in the USA sought to quantify the excess winter mortality in terms of life expectancy reduction, by calculating the difference between summer and winter life expectancy. OBJECTIVESWe aimed to calculate Winter Life Expectancy Reduction (WLER) in Europe, illustrate the extent to which this indicator depends on definitions of summer and winter, introduce a novel indicator of WLER based on a statistical model accounting for country-specific seasonal cycles, and compare men and women in terms of WLER. METHODSWLER indicators were calculated from weekly mortality data in 24 European countries over the period 2000-2019. RESULTSOn average, WLER was a few months higher in Europe than in the USA, while depending heavily on country and sex, as well as the chosen indicator.Our model-based indicator measured the highest WLER values, ranging from 11 months (Finnish men) to 36 months (Portuguese women), with a European average of 18 months for men and 22 months for women.In most countries, WLER was higher for women than for men, regardless of the indicator used.

Abstract There are striking disparities in life expectancy across sociodemographic groups in the United States, shaped by structural forces such as racism, class inequality, and policy environments. To what extent do sociodemographic characteristics structure—or fail to structure—individual lifespans? Using U.S. Census data linked to administrative death records, we assess how well early‐adulthood social, economic, and demographic characteristics predict individual lifespan in a cohort of men born in 1910 and observed through their deaths between 1975 and 2005 ( N = 121,000). Despite large group‐level disparities, we find that sociodemographic characteristics measured in early adulthood explain less than two percent of the overall variation in individual lifespan. These findings reaffirm a central demographic regularity: variance in life expectancy between groups is small compared to variation in lifespan within groups. This highlights the fundamentally nondeterministic nature of how structural inequality shapes individual mortality.

Abstract Prevailing theories of family change and the relevance of marriage in the United States hinge on the steady rise in births to unmarried women that unfolded during the latter half of the twentieth century and into the 2000s. This increase was concentrated among individuals with lower education levels, raising concern about inequality in children's family circumstances. Despite theoretical expectations that this trend would continue, the proportion of births to unmarried women plateaued during the 2010s. By examining trends in union formation and childbearing patterns by union status using data from the 2006‒2023 American Community Survey, this study investigates the ongoing link between marriage and childbearing underlying this plateau. Birth and marriage rates fell throughout the 2010s. However, in a reversal, married women became increasingly likely to have children relative to their unmarried peers, particularly among those with a high school education or less. These findings challenge theories about the changing social meaning of marriage, suggesting that norms regarding marriage remain robust rather than becoming deinstitutionalized. Furthermore, this study highlights how the declining marriage rate has contributed to the ongoing decline in the birth rate in the United States, implying that barriers to marriage may also create barriers to childbearing.

Abstract Over the past 170 years, the United States has undergone demographic, structural, and cultural changes that are reflected in—and a reflection of—changes in living arrangements. In this research note, we link living arrangements and life expectancy to calculate expected life years spent across different living arrangements by sex for the U.S. population for the period 1850–2021. We decompose changes in this measure by age group and describe change across cohorts. We use harmonized data from the Integrated Public Use Microdata Samples, classifying living arrangements into alone, with primary kin only (partners, parents, and children), and in extended households; more detailed subcategories include, for example, single-parent households and extended families. Three historical ages of U.S. living arrangements emerge: a “large household” system (1850–1940) characterized by relative stability in the extended household, when primary kin arrangements incorporate the majority of the substantial gains in life expectancy; an era of “primary kin” dominance (1940–1980) when life years spent only with primary kin increase faster than life expectancy, while the prevalence of extended households declines; and a “diversified” phase (1980–onward) characterized by a decline in two-parent households in favor of greater diversity, including living alone and with extended family.

As Western populations face projected decline, understanding the demographics of high-fertility subpopulations becomes increasingly important. The Amish represent one rapidly growing North American subgroup, yet existing demographic studies are dated and narrowly focused. Here, we use a new population database of >50,000 households-the vast majority of Amish-to offer an up-to-date population-wide analysis that shows high fertility and low mortality and attrition. Specifically, women's median age at marriage is 20.9, and 87.1 per cent marry by age 50; premarital conceptions are low (4.30 per cent of first births); spacing between marriage and first birth is short (mean 17.2 months); the total fertility rate is 6.1; infant mortality is 5.9; life expectancy at birth is 81.16 years; attrition is low (84.46 per cent retention for those aged 40+); and in-conversion is very low (154 individuals across nearly a century). These definitive population-wide figures open the way for testing predictors of population change and charting how growing subpopulations are shaping regions.

Research indicates a significant slowdown in life expectancy growth in the United States (US) post 2010, marking a departure from the consistent progress in longevity throughout the twentieth century. We extend this understanding, tracing the deceleration of US life expectancy back to the 1950s, after which average decadal change dropped from 3.80 to 1.61 years. Surprisingly, these mid-twentieth-century shifts were consistent across race and sex in the US and also in other high-income countries. Using a simple approach of quantifying potential life expectancy gains by eliminating mortality at specific ages, we find that the potential gains in life expectancy from reducing midlife mortality have been larger in the US than in other countries since 1900. The findings suggest that US life expectancy is unlikely to progress at the high speed observed between 1900 and the 1950s, with future advancements hinging on the reduction of old-age mortality, particularly from cardiovascular diseases and mental and nervous system diseases.

Demographers use ratios, proportions, and rates-all calculated as counts in numerators divided by counts in denominators-as key research inputs. For age-specific death rates, the numerator is observed deaths and the denominator is person-years lived. Life expectancy summarizes those rates into one measure, and its changes convey messages of changing mortality across time. We examine the contributions from the two components of life expectancy change: the population growth rate (relative changes in person-years in the denominator) and growth rate of deaths (relative changes in number of deaths in the numerator). We name this the numerator-denominator decomposition method. Applying the method to high-longevity countries during 2009-19 shows increases in life expectancy driven by high population growth at older ages without comparable increases in deaths. The United States experienced little life expectancy increase, and subnational comparisons show stark differences between urban and rural areas.