Estonia lost its independence during the Second World War, experiencing both Russian (1940-41, 1944-91) and German (1941-44) occupation, which brought significant political violence, particularly in the 1940s and 1950s. This paper analyses the effect of these repressions on fertility, using two studies. First, we use data collected between 1994 and 2005 from surveys that included extended questions on the lives of respondents' parents, including their experiences of repression. Second, we use biographical data on Estonian men who volunteered in the Finnish army. Applying Poisson and logistic regression models, we demonstrate that severe political repression (death, long-term imprisonment, deportation, and/or exile) reduced the fertility of cohorts who were in their prime childbearing ages during the 1940s and 1950s. We discuss these results in the wider context of the absence of a post-war baby boom in Estonia in contrast to those in countries that were part of the 'First World' during the Cold War.
Spain's fertility is among the lowest in the world, due largely to an increasing number of childless women. Childlessness, whether a voluntary choice or not, often results from postponement of childbearing. This delay could be driven partly by the increased instability of partnerships and adverse labour-market conditions. Using data from Spain's Fertility Survey 2018, we use multichannel sequence analysis to identify the most common patterns of partnering and employment trajectories and then assess the likelihood of remaining childless at age 40 for each pattern. Results indicate that pathways deviating from the traditional marital pattern are highly associated with childlessness. When these trajectories are coupled with high attachment to the labour market, the likelihood of remaining childless increases further. These associations are moderated by education. Partnership status is a stronger determinant of childlessness than employment status, especially for highly educated women.
Research on kinship and survival is well established in modern societies, but less is known about historical populations. This study examines the presence of children, siblings, and grandchildren in later life and their association with survival in pre-industrial Finland. We use records of 6,900 Finns born in 1740-1859 to estimate discrete-time hazard models for mortality at ages 50-90. At age 50, most individuals had at least one living child or sibling, although only one-fifth had a grandchild. Merely having a child or sibling did not guarantee better survival. However, men with both sons and daughters or with at least one grandchild experienced lower odds of dying. Having siblings, particularly sisters, became beneficial for survival after age 70. These findings highlight the nuanced role of kin, suggesting that their historical association with longevity varied by kin type and sex and by the individual's age and sex.
Demographic studies of significant reproductive outcomes, such as infertility and unintended pregnancy, often use measures of perceived infertility, but there are varied understandings and conceptualizations of perceived infertility across studies. Some researchers appear to assume that perceived infertility is consistent with medically defined infertility and implicitly use it as a proxy for medically defined infertility, whereas other researchers do not appear to assume that perceived infertility will align with medically defined infertility, instead treating the two constructs as distinct concepts. Using the motivated reasoning perspective as a sensitizing framework, we contend that both understandings of perceived infertility may be valid if the meaning of the measure varies by age. Using the National Survey of Family Growth, we find support for this hypothesis, confirming that both uses of perceived infertility are valid under certain circumstances.
Full birth histories that enable analysis of mothers' ages at birth and birth intervals by parity are rare in historical data. Partial birth histories can be obtained from retrospective fertility surveys and census data that record the numbers of children ever born, children deceased, and ages of surviving co-resident children. In a 1988 article published in Population Studies, Luther and Cho proposed a method for reconstructing full birth histories by probabilistically imputing the ages of deceased and non-co-resident children, and in a 2020 article in Historical Methods, Hacker extended this method to historical census data from the United States. In this paper I further adapt the reconstruction method using a sample from the 1911 Census of England and Wales and census records linked from 1881 onwards; this reduces the number of missing observations in the partial birth histories. Birth histories based on linked data particularly improve estimates of duration-specific marital fertility for women who married and had children more than 15 years before the survey date.
In many countries, educational inequalities in mortality are estimated using unlinked data sources: death certificates and censuses. However, education misreporting in these sources can introduce biases of unknown direction and magnitude. This study comprehensively quantifies such biases by taking a scenario-based approach, where distortions in education reporting are introduced to evaluate their effect on education-specific mortality rates and educational inequalities in mortality. Our results show that education misreporting can greatly bias mortality estimates, with the direction and magnitude of the bias influenced by the composition of and inequality in the population. We identify two bias patterns in inequality measures, based on whether they account for the ordering of education groups. Additionally, we demonstrate how education overstatement on death certificates can lead to crossovers in education-specific mortality rates and reversals in inequality. This study highlights the need for improved data quality to estimate educational inequalities in mortality accurately.
This study examines whether social parenting-living with a child's biological parent but not being biologically related to the child-is linked to remaining childless among men and women in Norway. It offers a novel overview of the complexity of partnership histories that include social parenting and whether different types of such histories are associated with childlessness. Using Norwegian register data for the 1980 birth cohort, we reconstruct annual partnership histories from ages 25 to 41 and identify five trajectory clusters with sequence analysis. We then estimate linear probability models of being childless at age 42 and assess differences by sex and economic uncertainty, as measured by unstable employment and insecure income history. Individuals in social-parenting clusters experience a higher probability of childlessness, especially those with complex, less stable histories and later partnership entry. Men in social-parenting clusters show higher childlessness than comparable women, and economic insecurity further elevates childlessness, particularly among men.
Recognizing that internal migration is an important source of error in population projections, we evaluate 21 forecasts from five classes of methods: (1) average of the most recent flows; (2) time-series econometric models with and without external factors; (3) machine-learning-based gradient boosting models; (4) multiplicative component and gravity-type models; and (5) average of all forecasts. We forecast bilateral interstate migration flows in Australia for the five years to mid-2016 and to mid-2023. No single model consistently outperforms the others in terms of bias, accuracy, and empirical coverage. Simple models perform as well or better than complex models. ARIMA models and their Bayesian equivalent perform well in both periods. Including control variables is no panacea, because their relationship with migration is flow and period specific. Similarly, globally trained machine learning models are not clear-cut alternatives, particularly when flow sizes vary widely. In contrast, the multiplicative component model suits contexts where the spatial structure of migration is stable, suggesting that internal migration forecasting requires context-specific approaches.
Life expectancy in the UK improved dramatically during the twentieth century. Whether more recent generations are also living longer in good health has serious implications for healthcare systems and the economy. Drawing on evidence from British birth cohort studies following people born between 1946 and 2000-02, we reviewed how physical and mental health have changed across post-war generations. We found little evidence for improving health across successive cohorts when compared at the same age. For several outcomes-particularly obesity, mental ill health, and diabetes-prevalence of poor health was higher in more recent generations, a pattern we term 'generational health drift'. There was little suggestion of improvements in health for cohorts born since 1946. More research is needed to understand the drivers of this trend, which has itself probably been shaped by changing exposure to preventable social and environmental risk factors experienced across the life course.
The intergenerational transmission of age at first birth is important for demographic dynamics and the reproduction of social inequality, yet it remains underexplored in China's rapidly transitioning society. Using data from the China Family Panel Studies, I examine cohort trends in intergenerational transmission of age at first birth for women born between 1941 and 1990. Results indicate that the intergenerational association has strengthened substantially over time, suggesting that family background is playing an increasingly important role in determining women's first-birth timing in the Post-Reform Era. Additionally, there is a non-linear, inverted-U-shaped pattern in the most recent cohorts (born during the 1970s and 1980s). Delays in first birth have been greatest for women whose parents had their first child at age 25-29, whereas those whose parents first gave birth after age 30 or before age 20 tend to enter parenthood earlier than this middle group.
This research examines the impact of population over-coverage on socio-demographic indicators for international migrants in Sweden. Over-coverage occurs when emigrants or deceased individuals remain in population registers, leading to overestimation of the resident population. This particularly affects migrants due to the increasing prevalence of return, onward, and circular migration. As register-based data collection becomes prevalent, addressing over-coverage bias is critical. Using multiple systems estimation models, we adjust for over-coverage in Swedish population registers and calculate accurate demographic and socio-economic rates. We use administrative data for 2003-16 to compute age-specific fertility rates, mortality rates, employment/unemployment rates, Total Fertility Rates, and life expectancy for migrants. Results demonstrate that over-coverage substantially affects demographic and economic indicators, particularly at typical migration ages, with mortality rates showing the largest bias at older ages. Findings highlight the importance of addressing over-coverage to improve the reliability of socio-demographic insights from register-based population systems.
Healthy working life expectancy estimates often exclude individuals working beyond retirement age, especially in countries with large informal economies. In South Korea, where old-age employment is highly prevalent, we need to identify who works past retirement, for how long, and in what state of health. Using the Korean Longitudinal Study of Aging (2006-22) and discrete-time multistate models, we estimate healthy and unhealthy working life expectancy at age 50 by gender, education, and urban/rural residence. On average, men work seven years past retirement age, and 28 per cent of their life expectancy at age 50 is spent working with at least one chronic health condition. These values increase to 10 years and 33 per cent for men in rural areas. Estimates for women are lower than those observed for men, and educational differences are minimal. Our findings highlight how gender and urban/rural residence may contribute to old-age health and employment by shaping opportunities and behaviours throughout the life course.
In the past, the choice of children's names was closely linked to life trajectories. This paper investigates the connection between children's names and child mortality during the demographic transition. Analysing data from the Veneto region (Northern Italy) and the Venetian parish of San Marco between 1816 and 1869, we find that traditional names are associated with higher neonatal mortality, particularly among girls. In contrast, children with unconventional names-those not named after grandparents or godparents or given the most frequent traditional names-face a lower risk of dying within the first month of life. Our findings underscore the importance of names in historical demography as predictors of demographic trajectories. Moreover, they suggest that naming practices may mirror varying levels of parental attention or a shift towards more modern concepts of childcare.
This paper uses population smallpox mortality rates in eighteenth-century Sweden and the death toll from the 1707-09 smallpox epidemic in Iceland to estimate plausible ranges for the case fatality rate (CFR) of smallpox (Variola major). We find that smallpox CFRs could be extremely high (43-55 per cent) when smallpox attacked a population where both children and adults were susceptible, as in Iceland. However, where smallpox was endemic and therefore a disease of childhood, as in Sweden, the estimated CFR is only 8-10 per cent: far lower than the consensus CFR of 20-30 per cent. We argue that social factors explain these differences. Where both adults and children were susceptible, smallpox epidemics fundamentally disrupted basic household tasks and nursing of the sick, dramatically increasing the CFR. Thus, when historians and epidemiologists give CFRs for smallpox, they should consider the population and context rather than relying on an implausible intrinsic CFR of 20-30 per cent.
In recent decades, a long-term internal migration decline has been observed in many countries, including Japan. To explore the dynamics of internal migration in Japan, we apply the age-period-cohort-interaction model to five-year inter-prefectural migration rate data from six nationally representative surveys conducted between 1991 and 2016. The results show regular age-specific migration patterns, fluctuating period effects, and inter- and intra-cohort differences. Age effects reflect general migration patterns and largely explain the decline, while period effects show no consistent trend. Relatively recent generations tend to exhibit statistically significant high or low migration rates, visible as inter-cohort differences. While some of these inter-cohort differences converge with age, cohorts that experienced socio-economic shocks in their early 20s display consistently low migration rates. This pattern indicates that such shocks impose long-term structural constraints to migration as a pathway to accessing improved employment, suggesting the need for targeted policy interventions to counter declining internal migration.
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.
In the absence of complete civil registration and vital statistics, Health and Demographic Surveillance Systems (HDSSs) are important sources of population-based data throughout sub-Saharan Africa. However, HDSS data on the vital status of newborns are often unreliable due to omission of those who were born and died between two rounds of data collection and are therefore never enumerated. This study investigates whether pregnancy registration improves estimation of under-five mortality (U5M) in three HDSSs in The Gambia, Kenya, and South Africa. We find that mortality is higher for children whose mother's pregnancy was observed than for children who were first registered after birth. Cox proportional hazards models with inverse probability weights further suggest that this difference is probably due to improved ascertainment of deaths in pregnancy cohorts and unlikely to be driven by a selection effect. These results highlight the importance of pregnancy registration in HDSSs for the estimation of U5M.
In historical populations, death rates for females often exceeded those of males in reproductive ages. However, childbirth-related deaths were not the only cause of excess mortality among adult women. This study expands on work by Schofield and colleagues by re-examining mortality after childbirth in the Cambridge Group Family Reconstitutions. Part 1 applies event history methods in a new way to focus on excess mortality following childbirth. Unlike previous methods, which assumed equal background mortality for wives and husbands, this method compares maternal and paternal mortality. The results show higher mortality for females than males even after removing deaths following childbirth. Part 2 explores the determinants of maternal deaths in the puerperal period. Deaths of new mothers rose when their husbands and children were more likely to die, but the risks of death for new mothers were much higher than for other married adults. These results highlight the extraordinary vulnerability of mothers in the weeks following childbirth.
This paper aims to assess the effect of the Great Recession on population-level immigrant mortality in Spain, highlighting the mediating effects of migration flows that shape the composition of immigrant populations. To investigate this, I use individual data from Spain's death and population registers for the period 2003-19. First, I find a significant mortality advantage at adult ages for non-Western immigrants compared with the native born. Second, I show that this mortality advantage for immigrants relative to the native born increased during the Great Recession, despite immigrants being affected more by the unemployment crisis. Unemployment-driven outmigration flows may have contributed to this change in relative mortality at ages 40-59, meaning that immigrants returning to their country were negatively selected. I argue that the crisis may have imposed a second selection on non-Western immigrants staying in Spain and that negative selection at departure-even if not due to ill health itself-affected mortality.
Despite its significance, men's fertility has been largely overlooked in demographic research. This study seeks to address this gap by conducting a systematic comparative analysis of men's and women's fertility using data from the Spanish ECEPOV-2021 survey, a large-scale data set (N = 424,493) from the Spanish national statistical office. Findings indicate that women generally exhibit slightly higher completed cohort fertility rates than men, with exceptions among remarried, college-educated, and immigrant men, who show higher fertility than their female counterparts. Childlessness emerges as a key factor underlying fertility differentials between the sexes, accounting for nearly half of the observed difference. After using matching techniques to control for compositional differences, the study concludes that adjusting for demographic and socio-economic factors significantly reduces, although does not entirely eliminate, the fertility differential. Residual differences may stem from measurement errors, selection biases, or unmeasured variables.
