The impact of intergenerational support on multidimensional poverty in old age: empirical analysis based on 2018 CLHLS data

China’s rapidly aging population and the transition to a post-poverty-alleviation era have shifted policy focus from absolute to relative and multidimensional poverty. Older adults face high poverty risks not only in income but also in health, psychological well-being, and rights. Traditional Confucian filial piety historically positioned children as primary supporters of older parents, but demographic change, family miniaturization, and high mobility have altered intergenerational dynamics. The study asks whether children’s intergenerational support affects multidimensional poverty among older adults, how different support types (financial, caregiving, emotional) influence different poverty dimensions (economic, health, spiritual, rights), and how social security moderates these relationships. The purpose is to inform multidimensional poverty governance for the elderly in China.

Prior research evolved from single-dimensional (income/consumption) to multidimensional poverty perspectives, influenced by Sen’s capability approach, incorporating health, education, housing, and subjective well-being. Measurement frameworks range from static indices (e.g., Alkire-Foster) to dynamic measures distinguishing temporary versus chronic poverty. Determinants of elderly poverty span individual (gender, age, marital status, education, health), family (income, number of children), and social factors (rural/urban, regional disparities, social insurance participation). Micro-level studies on intergenerational support’s effects are fewer: theoretical work suggests children’s support can reduce elderly poverty but may be declining with fewer children; unmet support can worsen mental health. Empirical studies show mixed effects of financial, instrumental, and emotional support on specific dimensions such as health and mental well-being. Gaps identified include limited attention to filial-level intergenerational support effects on multidimensional poverty, lack of detailed mapping of support types to poverty dimensions, methodological limitations (few large national samples and single-method designs), and limited consideration of social security’s moderating role.

Data: 2018 Chinese Longitudinal Healthy Longevity Survey (CLHLS), covering respondents aged 65+ across 22 provinces. After excluding missing/inconsistent responses, N=8,061. Measures: - Outcome: Multidimensional poverty defined as having poverty in at least two of four dimensions (binary=1): • Economic poverty: “Are all your sources of livelihood sufficient?” (No=1) • Health poverty: self-rated health Poor/Very poor=1 • Spiritual poverty: loneliness Always/Often/Sometimes=1 • Rights poverty: decision-making on own affairs Seldom/Never=1 - Key explanatory variables (intergenerational support from children): • Economic support: total cash/in-kind from children in past year (log transformed) • Emotional support: main talk partner includes children (1/0) • Care support: caregiver when ill includes children (1/0) - Moderator: Social security participation (any program=1; none=0). - Controls: gender, age, age squared, marital status, household registration (urban/rural), total number of children, number of boys, proportion of boys, total household income (log), number of co-residents, region (East vs. non-East). Models: - Baseline: Binary logit regressions for (a) multidimensional poverty and (b) each dimension separately. - Moderation: Split-sample regressions by social security participation (yes vs. no) to assess moderating effects. - Robustness: Propensity score matching (nearest neighbor) for emotional support effect; sub-sample regression (East region); model replacement using probit. Model fit assessed with likelihood ratio and Hosmer–Lemeshow tests.

Descriptives: - Overall incidence of multidimensional poverty: 15.7% among elderly; spiritual poverty highest at 26.1%; economic, health, and rights poverty each ~14%. Baseline regressions (logit): - Emotional support: significantly reduces multidimensional poverty and three dimensions—economic (coef ≈ -0.251, p<0.01), health (≈ -0.239, p<0.01), and spiritual poverty (≈ -0.245, p<0.01); significant for overall multidimensional poverty (≈ -0.283, p<0.01); no significant effect on rights poverty. - Economic support: no significant effect on multidimensional, economic, health, or spiritual poverty; significantly reduces rights poverty (coef ≈ -0.033, p<0.01). - Care support: no significant effect on multidimensional, economic, health, or spiritual poverty; significantly increases rights poverty (coef ≈ 0.236, p<0.05). Moderation by social security (split samples): - With social security (N=7,510): Emotional support significantly reduces multidimensional poverty (coef ≈ -0.240, p<0.01); economic support not significant. - Without social security (N=551): Economic support significantly reduces multidimensional poverty (p<0.05); emotional support not significant; care support not significant in either group. Robustness: - PSM: Emotional support effect on multidimensional poverty remains significant; ATT before matching -0.027 (p<0.05), after matching -0.034 (p<0.01). - Subsample (East region): Emotional support remains significant and negative (p<0.05); directions and significance largely consistent with full sample. - Probit models: Results consistent; emotional support remains significantly negative (p<0.01).

Findings confirm that as China’s elderly population grows and family structures change, multidimensional poverty increasingly reflects non-material deprivation, particularly spiritual loneliness. Emotional support from children emerges as the most potent protective factor against multidimensional poverty, improving subjective economic adequacy, health status, and reducing loneliness, consistent with social support theory’s main and buffering effects. Financial transfers and caregiving from children do not broadly alleviate multidimensional poverty; rather, financial support principally enhances autonomy and decision-making (lower rights poverty), and caregiving may inadvertently reduce perceived autonomy (higher rights poverty), possibly due to health declines prompting care and role constraints in co-residence. Social security moderates these relationships: when social protection is present, emotional needs dominate and emotional support has greater salience; absent social security, financial transfers become more consequential for preventing multidimensional poverty. Overall, results partially support H1 and H2 (effects differ by dimension), strongly support H3 (emotional support alleviates multiple dimensions), and support H4a and H4c (moderation for financial and emotional support) but not H4b (no moderation for care support).

This study uses national CLHLS 2018 data to map how distinct intergenerational supports from children relate to different dimensions of elderly multidimensional poverty and how social security moderates these effects. Contributions include: (1) documenting high and shifting multidimensional poverty with spiritual poverty most prevalent; (2) showing diminished breadth of financial and caregiving support effects, limited to rights poverty with opposite directions; (3) identifying emotional support as pivotal in reducing multidimensional, economic, health, and spiritual poverty; (4) establishing moderation by social security—emotional support more salient with social protection, financial support more salient without it. Policy recommendations: promote and incentivize children’s emotional engagement (e.g., supportive leave policies, “visit parents” facilitation), prioritize support for rural and central/western elderly, expand and tier social welfare and health insurance coverage, and establish systems for identifying and monitoring multidimensional poverty among older adults. Future research should disaggregate by urban–rural (hukou) and regional contexts and further explore causal mechanisms across support types and poverty dimensions.

The analysis pools urban and rural elderly without fully exploring heterogeneity by household registration (hukou) or spatial disparities. Given marked urban–rural and regional differences in China, future studies should stratify samples to compare how intergenerational support types affect multidimensional poverty across these contexts and investigate additional sources of endogeneity.