China is ageing rapidly and is expected to become a super-aged society around 2035. As long-term care insurance expands nationwide, the coming decade will be critical for long-term care development and workforce planning. China is building a home- and community-based elderly care system with professional institutional support and medical-care integration, yet workforce shortages remain a constraint. Care needs and arrangements vary by disability level, and the two settings differ in recipient composition, service organization, and staffing mechanisms. Existing studies often treat care as homogeneous, obscuring the scale and structure of workforce demand.

This study addresses two questions: (1) What will be the scale of demand for China's long-term care workforce (LTC workforce) from 2020 to 2035? (2) How will this demand differ between institutional and home- and community-based elderly care settings? To address these questions, data are drawn from the China Longitudinal Aging Social Survey (CLASS), the Chinese Longitudinal Healthy Longevity Survey (CLHLS), the Seventh National Population Census, World Population Prospects 2024, and the China Civil Affairs Statistical Yearbook. Disability transition probabilities are estimated from 2014 and 2018 CLASS and CLHLS panel data and combined with census and population projection data to project disabled older adults by disability level. Two scenarios are constructed: a steady-state distribution scenario and a policy development scenario. Demand for the LTC workforce is estimated using staffing ratios for institutional elderly care and service time requirements for home- and community-based elderly care.

The results show that the number of disabled older adults will increase from 22.89 million in 2020 to 65.97 million in 2035, with older adults with mild disabilities remaining the largest group and those with severe disabilities experiencing the fastest growth. This expansion and changing composition increase LTC workforce demand, with distinct patterns across care settings. Under the medium specification, demand for the institutional LTC workforce reaches 0.68(steady-state distribution scenario)~1.53(policy development scenario)million in 2035; in the latter, care for older adults with severe disabilities accounts for more than 80% of institutional demand. Home- and community-based elderly care accounts for most LTC workforce demand, requiring 6.24(policy development scenario)~7.01(steady-state distribution scenario)million workers in 2035, with demand increasingly concentrated among older adults with mild and severe disabilities. Total demand for the LTC workforce reaches 7.69(steady-state distribution scenario)~7.78(policy development scenario)million in 2035. Although the policy development scenario raises institutional demand, the corresponding reduction in home- and community-based demand keeps the difference in total demand relatively small. Workforce supply growth lags markedly behind demand, with the supply-demand gap under the adjusted measure widening from 1.58(steady-state distribution scenario)~1.60(policy development scenario)million in 2020 to 3.12(steady-state distribution scenario)~3.16(policy development scenario)million in 2023.

These findings show that LTC workforce pressures reflect both expanding demand and structural differentiation across disability levels and care settings. Future workforce planning should therefore balance supply expansion with structural optimization: while broadening training and employment channels and improving occupational protection to enhance supply capacity, it should also allocate human resources rationally across disability levels and care settings, and improve service efficiency through digital and intelligent technologies. By distinguishing institutional elderly care from home- and community-based elderly care and their staffing mechanisms, this study extends workforce projection from aggregate estimation to structural analysis and provides targeted empirical support for long-term care system development and workforce planning in China.

Since 2026, China's long-term care insurance (LTCI) scheme has shifted from local pilots to nationwide implementation. Amid moderate population aging, nationwide LTCI is essential for equitable access to care for people with disabilities and the sustainable development of an aging society. The Urban and Rural Resident LTCI scheme covers non-employed residents and is financed by individual contributions and government subsidies. This study examines fund balances, regional disparities, and central-local fiscal responsibilities across 31 provincial-level regions in Chinese mainland during 2026-2036.

The data and methods are as follows. First, using China's 2020 Seventh National Population Census and statistical yearbooks, the cohort-component method is employed to project population by urban-rural residence, age, and sex, accounting for urbanization and interprovincial migration. Second, using the 2018 and 2020 waves of the China Health and Retirement Longitudinal Study (CHARLS), a Markov model is constructed based on six activities of daily living (ADLs). Transition probabilities are estimated among five states—independent, mildly disabled, moderately disabled, severely disabled, and deceased—to project disability prevalence among the older population. Third, based on China's existing LTCI pilot policies, annual benefits under the low, medium, and high scenarios are set at 10%, 20%, and 30% of each province's previous-year per capita disposable income, respectively. In accordance with Chinese policies, the 31 provinces are grouped into five tiers. The central government covers 80%, 60%, 50%, 30%, and 10% of government subsidies in these tiers, respectively, with local governments covering the remainder. Fourth, national and provincial policies guide enrollment and contribution rates. Fund revenue, expenditure, balances, and government subsidies are estimated to assess provincial financial sustainability and central-local fiscal responsibilities.

The findings are as follows. First, provincial financial sustainability under the medium-benefit scenario falls into five categories, whereas the high-benefit scenario increases the number of provinces experiencing deficits. Population aging, the size of the adult contributing population, enrollment rates, contribution rates, and benefit levels explain regional disparities. Second, central subsidies as shares of central general public budget revenue, expenditure, and outstanding central government debt peak in 2033 at 0.181%, 0.415%, and 0.046%, respectively. Tiered central subsidies alleviate local fiscal burdens but cannot eliminate structural pressures in regions experiencing advanced population aging and net population outflows.

Policy recommendations include shortening the contribution-rate transition period or temporarily increasing contribution rates in regions with persistent deficits. For provinces with limited fiscal capacity, annual benefit caps should initially be set at 30% of residents' previous-year per capita disposable income, while benefit levels should reflect care costs and actuarial balance. Fiscal burden-sharing should account for demographic structure and fiscal constraints while strengthening provincial pooling and interregional transfers.

This study makes three contributions. First, it extends analysis from local pilots to all provincial-level regions and identifies regional differentiation driven by demographics, financing capacity, institutional parameters, and fiscal capacity. Second, it addresses the understudied resident LTCI scheme and redefines its funding sources under current policies, aligning actuarial assumptions with institutional arrangements. Third, it develops an actuarial framework linking fund operations to fiscal responsibilities. These findings inform enrollment expansion, contribution rates, benefit levels, fiscal capacity assessment, and intergovernmental burden-sharing to promote equitable access to basic long-term care.

Against the backdrop of rapid population ageing and the rising demand for long-term care services, establishing a stable and sustainable long-term care system has become an important component of actively responding to population ageing. As an institutional arrangement designed to provide social risk-sharing protection for individuals with long-term care needs, long-term care insurance (LTCI) may not only alleviate households' financial constraints in obtaining care services but also influence the development of formal care supply. However, whether and how LTCI expansion generates supply-side responses in the elderly care service sector remains insufficiently understood. This study examines the impact of LTCI expansion on the supply of elderly care institution beds and explores the conditions under which such effects emerge.

Using an unbalanced city-level panel dataset covering 2012-2016 and 2021-2023, this study regards the nationwide expansion of LTCI pilot programs in 2020 as a quasi-natural experiment and employs a difference-in-differences model to identify the causal effect of LTCI expansion on elderly care institution bed supply. The main sample excludes cities that had implemented LTCI before 2020 and compares cities newly included in the 2020 expansion with cities that had not introduced LTCI by 2023.

The results show that LTCI expansion is associated with a statistically significant increase in the absolute number of elderly care institution beds. In the fully controlled specification, newly added pilot cities had, on average, about 3535 more beds during 2021-2023 relative to non-pilot cities. The results remain robust after the robustness tests including sample restrictions, inverse-probability weighting, placebo timing tests, and random reassignment of treatment status. Further analysis shows that the number of elderly care institutions increased, whereas average beds per institution did not change significantly. The expansion therefore mainly took the form of an increase in the number of institutions rather than the enlargement of existing institutions. Heterogeneity analyses show that the positive bed response is concentrated in cities with stronger pre-policy elderly care service foundations and greater fiscal capacity. Cities with higher levels of population ageing do not display a stronger response, indicating that demographic demand alone does not necessarily translate into additional institutional capacity.

This study contributes to the literature in several ways. First, it shifts the analysis of LTCI to city-level provider capacity, providing evidence on the supply-side consequences of a demand-side social insurance program. Second, by distinguishing between increases in the number of institutions and enlargement of existing institutions, the study identifies the main form through which elderly care bed capacity expands following policy implementation. Third, the study places demographic demand and local supply-side conditions within a unified analytical framework and shows that stronger ageing pressure does not automatically translate into a larger supply response. The effectiveness of LTCI in promoting elderly care supply depends on whether local governments and service systems possess sufficient institutional foundations and fiscal capacity to absorb the increased demand generated by insurance coverage. Finally, the findings provide empirical evidence for understanding the broader supply-side effects of LTCI and offer policy implications for improving long-term care systems in the context of population ageing.

Interethnic marriage is a relational phenomenon. It not only reflects individual marital choices but also forms a network of social ties among ethnic groups. Existing studies on interethnic marriage in China have mainly relied on intermarriage rates or regression models of marital choice. These approaches describe the level and determinants of interethnic marriage, but they do not fully reveal how the structure changes over time. An increase in the total volume of interethnic marriage may result either from more ethnic-group pairs becoming connected or from the accumulation of marriages within ties that have already existed. Distinguishing these two processes is essential for understanding the structural change of interethnic relations in China.

This study applies an intensity-connectivity framework to analyze changes in interethnic marriage. Drawing on Das Gupta's standardized decomposition method, it decomposes the total volume of intermarriage into two dimensions: intensity, defined as the average number of interethnic marriages per connected ethnic-group pair, and connectivity, defined as the proportion of connected ethnic-group pairs among all possible ethnic-group pairs. Based on interethnic marriage matrices for 56 ethnic groups from the 2000, 2010, and 2020 national population censuses, the study examines both aggregate network change and ethnic-group-level heterogeneity. Because the census data are stock rather than flow data, both measures describe accumulated relational structures rather than newly formed marriages.

The results show that China's interethnic marriage network underwent a structural change from breadth expansion to depth accumulation between 2000 and 2020. The total volume of interethnic marriage increased from 8.04 million in 2000 to 9.64 million in 2020, but the structural sources of this growth differed across periods. From 2000 to 2010, connectivity accounted for 93.8% of the increase, indicating that growth was driven primarily by the formation of marriage ties among previously unconnected ethnic groups. From 2010 to 2020, intensity contributed 120.9% of the increase, while the contribution of connectivity became negative. In this period, growth was driven by the accumulation of marriages within existing ties, whereas the decline in connectivity may reflect the disappearance of some weak or low-frequency ties. Accordingly, network connectivity followed an inverted U-shaped trajectory, rising from 2000 to 2010 and declining from 2010 to 2020.

At the ethnic-group level, changes in interethnic marriage were heterogeneous. The study identifies four types of structural change: depth accumulation, breadth expansion, simultaneous depth and breadth expansion, and marriage contraction. Most ethnic groups fall into the depth accumulation type, indicating that their growth in interethnic marriage mainly came from the deepening of existing ties. A small number of groups showed continued expansion in their relational scope, while others experienced both expansion and deepening. For a few groups with small populations and sparse ties, the loss of a few existing marriages may lead to contraction.

These findings show that changes in interethnic marriage cannot be adequately understood through intermarriage rates alone. The intensity-connectivity framework reveals the reorganization of the scope and accumulated depth of interethnic marriage ties, while the typology suggests that differing network positions should be fully considered when promoting interaction, communication, and integration among all ethnic groups in China.

The long-term projection of changes in the total fertility rate (TFR) is a crucial basis for assessing population development trends, as well as a persistent challenge in demographic research. This paper designs a Period-Cohort-Period (PCP) method for TFR projection: it utilizes period fertility data from successive censuses to construct cohort fertility distributions, and subsequently inverts these cohort distributions to project period fertility distributions, thereby deriving the TFR for future years. The PCP method requires the fertility pattern simulation function to have high precision and parameters that are explicit and flexibly controllable. This study discusses the relationship between the Gompertz fertility model and the standard Gumbel function, revealing shortages in the Gompertz model such as ambiguous parameter interpretations, parameter redundancy, and inflexibility in shape control. By addressing these limitations, we derive the Generalized Gumbel fertility model, which surpasses all previous models in goodness of fit.

Applying the Generalized Gumbel model, we fitted and analyzed data from successive population censuses since the 1980s. The results reveal that the peak fertility points across different periods exhibit a “C-shaped” trajectory, which can be broadly divided into three stages: the “early marriage and early childbearing” stage in the early 1980s driven by the post-peak effect, the rapid plunge stage of the fertility peak during 1986-2005, and the delayed childbearing stage with a broad peak since 2010. The period-to-cohort analysis reveals that the cohort total fertility rate (CTFR) experienced an initial increase followed by a decline, with the magnitude of change being much more moderate than that of the period TFR. Specifically, the CTFRs for 1965-1974 cohorts slightly increased from 1.93 to 2.17, while the CTFRs for 1974-1985 cohorts dropped from 2.17 to 1.87. In terms of the fertility distribution shape, this manifests as a vertical plunge in fertility peaks, with an ongoing trend of the right tail of the fertility curve lengthening and rising. For cohorts born since 1985, the rightward shift of the peak fertility age has accelerated, while the magnitude of the peak decline has narrowed.

By referring to the changing trends of indicators such as lifetime childlessness rates and TFRs in other countries, this study analyzed the parameter variation trends of China's future fertility patterns. We projected the fertility distributions for 1991-2030 cohorts by parity under high, medium, and low scenarios, thereby extrapolating the potential evolutionary trajectories of the period TFR from 2025 to 2050. The findings indicate that China’s TFR may exhibit a slow recovery trend during the latter half of the 15th Five-Year Plan period. On the one hand, this is a result of the diminishing tempo effect of fertility; on the other hand, it reflects the self-correction of diverse groups within a massive population. To escape the “low fertility trap” and return to a normal fertility pattern by the mid-21st century, the state must not only actively implement fertility support and pro-natalist policies but also trust and protect the rich diversity of fertility values, maintaining confidence in the future and avoiding extreme policies driven by excessive anxiety.

In settings where marriage and childbearing are tightly linked, marriage is a prerequisite for having children. Women take on more home care duties after marriage, and their labor market outcomes often begin to shift even before a first birth. If changes in women's work are put down to childbearing alone—overlooking the gendered redivision of labor that marriage itself brings—the total family formation penalty will be seriously underestimated. Using CFPS data, this paper adopts a joint event study design built around first marriage and first childbirth. It estimates recent marriage, childbearing, and family penalties on women's work for those first married between 2010 and 2022, uses Oaxaca decomposition methods to gauge how much these penalties add to the gender gap in work, and asks whether they arise from a gendered split of home labor rooted in social norms.

Key findings are as follows. First, Chinese women face a 17.4% marriage penalty and a 16.2% child penalty, yielding a total family penalty of 33.6%; both contribute almost equally to women's work losses, yet the child penalty is larger in the short run. Second, standard single-event studies typically leave out the marriage penalty and thus underrate the harm family formation inflicts on women's work. Third, the family penalty makes up 74.1% of the 2020 gender gap in paid work, and the child penalty's share of this gap has risen over time. Fourth, these penalties persist even when women have no less schooling than men. Long-held gender norms keep driving how home roles are split, which is very likely the main reason women end up with a larger family penalty.

The paper adds to the field as follows. First, by turning seven waves of CFPS cross-sections into panel data via personal IDs and a joint event study design, it pins down both the marriage and child penalty for Chinese women. Second, for cases with no pre-marriage data, it uses pseudo-event study matching on sex, age, year, schooling, and urban/rural status (or province) to enlarge the sample and run robustness checks. Third, mechanism analysis further shows the old gendered split of home labor is the main path through which both penalties operate. No matter how spouses' pre-marriage edges stack up, women's housework time rises sharply after marriage versus men's, strongly hinting that old social norms may lie at the root of gender gaps in China today.

These empirical results point to several clear and actionable policy steps. First, formally recognizing domestic services and offering subsidies for their use can transform gendered unpaid care into paid services that families can readily purchase. Second, introducing a dad-only, non-transferable paid parental leave would send a strong and clear signal that care is a joint responsibility, thereby helping to break the deeply entrenched image of women as sole carers. Third, establishing an inclusive child care system based on full-day care with extended-hour add-ons can ease family care burdens and enable both parents to participate more fully in paid employment.

Family support for older parents is often treated as an aggregate transfer from children to parents, although financial support and housework assistance are distributed among siblings within multi-child families. Using the 2014, 2016, 2018, 2020, and 2023 waves of the China Longitudinal Aging Social Survey (CLASS), this study matches respondents aged 60 or older with their two to five living adult children, yielding 39850 family-year observations and 114023 parent-child-year records. It distinguishes participation breadth, within-family share evenness, primary-provider separation, and share differences between financial support and housework assistance. A support-sharing index combines the proportion of participating adult children with a number-of-children-adjusted measure of within-family share evenness. The index describes observed behavioral distributions rather than normative obligations, negotiation, cooperation, or perceived fairness.

The results show that participation broadens and support-score shares become more dispersed for both forms of support. From 2014 to 2023, the support-sharing index for financial support rises from 0.757 to 0.884, while the corresponding index for housework assistance increases from 0.520 to 0.703. After adjustment for family characteristics, the differences between 2023 and 2014 in both component indices and the overall index are positive in families with two, three, four, and five adult children, although they are smaller in five-child families. Family-level total scores for both forms of support show no sustained decline. The findings remain robust to a 500-yuan participation threshold for financial support, alternative score assignments, individual fixed-effects models based on five waves linked through verified respondent identities, and fractional logit models. Estimates within each fixed child-number group provide the most direct evidence against a purely mechanical denominator effect. Hence, declining numbers of adult children alone therefore do not account for the observed aggregate trend.

Financial support consistently shows a higher level of sharing than housework assistance, consistent with the greater divisibility, additivity, and transferability of money and goods across distance. Unadjusted primary-provider separation fluctuates across survey years. The within-family distributions of support shares across siblings become more similar, whereas the unadjusted proportion of families with no overlap between primary providers of the two support types increases by about 1.4 percentage points between 2014 and 2023. After adjustment for family characteristics, however, primary-provider separation increases significantly by 7.2 percentage points between 2014 and 2023.

Within families containing at least one son and one daughter, son-daughter differences vary significantly across survey years in both financial support and housework assistance. In all eligible families, birth-order differences also vary significantly across survey years in both support types. These patterns are nonmonotonic: neither gender differences nor birth-order differences consistently widen or narrow over time. The findings describe observed behavioral distributions. The survey cannot determine whether broader participation and more dispersed shares arise from negotiated coordination, parental mobilization, low-intensity participation, or changes in individual contributions. Policy interventions should identify actual primary support providers and address time-intensive tasks rather than assign responsibility on the basis of gender or birth order. Future research should combine sibling-level longitudinal data with measures of negotiation, fairness, personal care, support quality, and caregiver burden.

China's total fertility rate has declined to ultra-low level. Against this backdrop, this study investigates the determinants of subsequent childbearing intentions for a second or third child among Chinese individuals of reproductive age.

The existing literature typically assumes, either explicitly or implicitly, that individuals of reproductive age have a son preference—specifically, a desire to have at least one son—and accordingly treats the number of existing children and whether they have a son or not as separate core explanatory variables in estimating the likelihood of subsequent childbearing. However,currently more than half of the reproductive-age population has shifted their ideal family structure toward having one son and one daughter. To address this gap, this paper contributes to the literature by conceptualizing “fertility intention” as a construct that encompasses both subjective goal orientation and realized outcomes. Specifically, we investigate how fertility support and fertility desire interact to shape subsequent childbearing intentions, and examine which dimension of unfulfilled preference—quantity or sex—exerts a stronger influence on the decision to have another child. Particular attention is paid to complex scenarios in which the quantity preference is satisfied but the sex preference is not, or vice versa.

Drawing on data from the 2021 Chinese General Social Survey (CGSS) and employing an ordered logit model, this empirical analysis reveals that household annual income exerts an inverted U-shaped effect on higher-order fertility intentions, peaking at approximately RMB 400,000. Both grandparental care and childcare services facilitate subsequent childbearing intentions. Reproductive-age individuals exhibit no significant preference regarding the public or private nature of childcare institutions, yet remain sensitive to the pricing of high-quality services. Under specific economic constraints, individuals are generally unwilling to exceed their intended family size solely to satisfy a sex preference. In contrast, when the actual number of children falls short of the desired level, having “more girls than desired” promotes the intention to have a third child. Conversely, having “more boys than desired” suppresses the intention for a third child, regardless of whether the existing number of children is below, equal to, or above the desired level. Furthermore, among those desiring at least one son and one daughter, the gender of the firstborn does not significantly affect second-child fertility intention.

Based on these findings, the following policy implications are proposed. First, expanding the coverage of childcare subsidies and educational burden-reduction initiatives is necessary to alleviate the economic pressures of childrearing for ordinary families. Second, reforms targeting marriage customs are required to promote a more equitable sharing of marital responsibilities between the families of the bride and groom, thereby alleviating the long-term financial burden on families with sons. Finally, public health and social interventions should aim to cultivate a positive and supportive marriage and childbearing environment, progressively building a favorable landscape where people of childbearing age are willing, confident, and eager to expand their family.