Demography最新文献

We construct a unique sibling dataset by linking multiple comprehensive administrative data sources in Taiwan. Using data on one million siblings, we estimate the effect of parental divorce occurring at ages 13‒18 on children's university admission. Our approach leverages differences in admission outcomes between siblings who experienced parental divorce before the national college entrance test at age 18 and those who experienced it afterward. The mother fixed-effects estimates reveal a significantly negative impact of parental divorce on children's university admission. Adolescents who experienced parental divorce faced a 10.8% reduction in the likelihood of entering any university and a 15.9% reduction in the likelihood of being admitted to a first-tier university. Additional analyses show that younger adolescents are more vulnerable to the negative effects of parental divorce than their older counterparts. Furthermore, the study finds nonnegative effects of parental job loss on university admission, suggesting that the adverse impacts of parental divorce are unlikely to operate through income disadvantage.

This research note focuses on accurately documenting the trends in educational resemblance between husbands and wives in South Korea over six decades, from 1960 to 2020. Having undergone rapid social changes in recent history, including industrialization, economic development, and educational expansion, Korea offers a compelling context for studying long-term changes in educational assortative mating across different stages of development. Using 2% microsamples from 13 census datasets collected between 1960 and 2020, we construct marriage tables cross-classifying six educational levels of husbands and wives, both aged 25 to 45. Log-multiplicative layer effect models are applied to assess the husband‒wife association, controlling for changing marginal distributions of both spouses' educational levels. Our analysis of 843,527 married couples shows that the association between husbands' and wives' education increases to a peak around 1995, after which it continuously declines. The inverted U-shape trend remains robust whether analyzing current or first marriages of varying duration and across different types of log-linear models. We provide theoretical and empirical discussions of major macro-level trends, especially the timing and gendered patterns of educational expansion, in Korea to contextualize the observed patterns of educational assortative mating.

This study examines the relationship between community violence and the use and provision of contraception in Mexico, where family planning is a long-standing policy priority and the "war on drugs" has led to chronically high levels of violence. We adopt a two-step approach. First, we investigate the association between women's exposure to violence and first contraceptive use. Combining individual-level data (n  =  86,219) from two waves of the National Survey of Demographic Dynamics (ENADID) with information on monthly municipality-level homicides in event-history models, we analyze the timing and method of women's first contraceptive use and the source of first contraception. Second, leveraging rare data from Mexico's Ministry of Health in clinic fixed-effects models, we study the association between homicides and contraceptive provision from public clinics. Results show strong positive associations between community violence and both the transition to first contraceptive use and the contraceptive provision of reversible methods. These relationships are stronger in the long term; one more homicide per 10,000 population during the past five years is associated with triple the risk of initiating contraceptive use and two to three more reversible contraception users served in each public clinic per month. The findings suggest increasing contraceptive vigilance and fertility regulation preferences-but also healthcare system resilience-in times of insecurity.

Begum et al. (2018) examined gender bias in parental attitudes using an experimental approach in rural Bangladesh. Households were reported as randomly assigned to treatment conditions in a lab-in-the-field allocation task. We show that the group assignment was inherited from Islam (2019), a previous, nonrandomized experiment conducted in the same region. The lack of randomization contradicts the design descriptions provided by the authors in Begum et al. (2018) and elsewhere and raises concerns about the validity of comparisons across treatment groups. This also points to serious shortcomings in the reporting and transparency of the study design-issues that mirror those that led to the retraction of Islam (2019) from the European Economic Review.

The presence of exogamous consensual unions (marriages and cohabitations) is an important indicator of social closeness between groups. European research on intermarriage has focused on unions between minority groups and the native majority population, highlighting that religion, especially Islam, constitutes a bright boundary. In diversifying societies, however, opportunities for union formation between minority groups with distinct national origins are increasing. Yet, we know little about partnership patterns among these groups, particularly whether different Muslim groups coalesce through intergroup unions. Using full-population register data from the Netherlands (1999‒2023), we analyze union formation across any combination of the 21 largest national origin groups present on the partnership market. Our findings reveal that unions involving partners from Muslim groups originating from different countries remain strikingly rare overall compared with endogamy within national origin groups, indicating the persistence of national legacies rather than exclusively religious closure. Boundaries for union formation between Muslim groups do not weaken meaningfully across immigrant generations and time. Among non-Muslim groups, by contrast, unions are more commonly formed within broader panethnic "melting pots." Overall, these results underscore the need for a more nuanced understanding of religious boundaries and panethnicity in union formation in Europe.

The shift of immigration enforcement to local communities has amplified anti-immigrant sentiment and enabled the unprecedented expansion of mass deportation systems. This study examines one such effort-the 287(g) program, which empowers local law enforcement to enforce U.S. immigration laws-and its effects on the residential segregation of Latinos. Using a difference-in-differences design that compares counties that implemented 287(g) with applicant counties that did not implement, we find that these policies significantly slowed declines in Latino‒White segregation, particularly in early-adopting Southern counties, where enforcement was most aggressive. We detect no effects on Black‒White or Latino‒Black segregation. Our analysis also suggests that these effects are driven by shifts in residential mobility and migration behaviors among Latino households, as enforcement amplified economic vulnerability, fear, and spatial isolation. These results indicate that interior immigration enforcement functions as a structural barrier to Latino integration, with downstream consequences for racial stratification and neighborhood inequality.

Estimates of future migration patterns are of broad interest in demography. Forced migration, including refugee and asylum seekers, plays an important role in overall migration patterns but is notoriously difficult to forecast. Focusing on refugees and asylum seekers, we propose a modeling pipeline based on Bayesian hierarchical time-series modeling for projecting refugee population official statistics by country of origin using data from the United Nations High Commissioner for Refugees. Our approach is based on a conceptual model of refugee and asylum seeker populations following growth and decline phases, separated by a peak. The growth and decline phases are modeled by logistic growth and decline through an interrupted logistic process model. We evaluate our method through a set of validation exercises that show it has good performance for forecasts at 1-, 5-, and 10-year horizons, and we present projections for 35 countries of origin of large refugee and asylum seeker populations.

A sizable number of children in low- and middle-income countries experience separation from their parents due to labor migration. While scholarship has examined parental migration effects during school years and adolescence, few studies have examined the effect of migration on developmental outcomes in early childhood, a critical period characterized by a high degree of developmental plasticity and sensitivity. In addition, almost all previous work has relied on cross-sectional data, treating migration-induced family living arrangements as a static, time-invariant measure. However, parental migration is often temporary and circular, exposing children to multiple transitions that vary in timing and duration. To address these two limitations in the literature, the present study follows a panel of more than 4,000 young children in the China Family Panel Survey from 2010 to 2018 and examines early childhood outcomes using a counterfactual causal-inference framework with time-varying exposure. My results show that the effect of parental migration on children's outcomes depends on whether one versus both parents migrated and at what child age the parental migration happened. Specifically, one-parent out-migration is less detrimental to child development than both-parent out-migration. Parental migration impacts specific developmental domains depending on children's sensitive periods of development, with early absence being most consequential for physical growth and later migration being most consequential for noncognitive skills. The study offers a causal framework to better understand migration as a time-varying process, contributes to family demography by theorizing parental migration as a unique source of family complexity, and deepens our understanding of the toll that parental absence takes on human capital.

China has the largest number of patients with dementia in the world, and the rate of growth is expected to escalate further as the population ages. The majority of dementia patients rely on their families for care and assistance. Using demographic models of kinship, we provide quantitative estimates of the burden of dementia, from 1990 up to 2050, by illustrating the number of kin accessible to dementia patients, the dementia prevalence among kinship networks, and the dependency ratio of kin with dementia to working-age kin without dementia. We then compare the estimates of dementia burden across 194 countries and territories, accounting for historical trends in, and future projections of, mortality, fertility, and dementia prevalence. Our findings suggest that, unlike in other aging societies, China's aging crisis is exacerbated by the fact that, in addition to the alarming rise in the number of older adults in need of care, the number of potential family caregivers is also dropping at an unprecedented pace. The increase in dementia dependency ratio is expected to exceed the increases in most other countries across East Asia, Western Europe, and the United States. These findings have important implications for understanding the evolution of care networks for older adults in China over time and from a cross-country comparative perspective.

In this research note we address measurement challenges in the study of residential segregation to advance our understanding of the segregation of Latino and Asian ethnic subgroups. Our primary purpose is to address some of the key methodological barriers to advancement in this area, and we also answer this question: Do we miss important patterns of segregation when we study Latino and Asian panethnic groups versus ethnic subgroups? Research has been hindered by problems with segregation index bias that are exacerbated when studying the segregation of smaller groups, as well as by a limited understanding of different patterns of uneven distribution that certain segregation indices, such as the dissimilarity index, cannot capture. Using a carefully chosen segregation index corrected for index bias, we find some variation in Latino and Asian ethnic subgroup segregation that warrants disaggregating panethnic groups, but more importantly we find that segregation of these groups is much lower than previously understood. This latter finding is because index bias and the choice of segregation index can have major impacts on our understanding of these patterns, with the separation index emerging as a superior method of measurement. These findings support the study of ethnic subgroup residential segregation, so long as researchers make careful decisions about segregation measurement.