Feminization of poverty: an analysis of multidimensional poverty among rural women in China

The study addresses the persistent issue of the feminization of poverty, focusing on rural women in China who face compounded disadvantages by gender and geography. Traditional income-based poverty metrics miss important non-monetary deprivations. Guided by Sen's capability approach and OPHI's recommendations for country-specific multidimensional poverty indices (MPI), the research develops and applies an individual-level, six-dimensional MPI to rural women using 2010–2020 CFPS data. It asks: (1) What is the current status of rural women's multidimensional poverty (MP)? (2) What are the spatial and temporal characteristics? (3) How can the risk of MP among rural women be reduced in the new era? The work is important for informing targeted, gender-sensitive poverty governance and for understanding dynamic changes during China's poverty alleviation period.

The literature has evolved from unidimensional, income-based measures (e.g., Rowntree) to multidimensional conceptions of poverty aligned with Sen's capability approach, recognizing deprivations in opportunities, freedoms, and non-market goods such as health, education, and wellbeing. Various multidimensional methods exist; the Alkire-Foster (A-F) approach is widely used due to its axiomatic rigor, flexibility, decomposability, and communicability. Gendered analyses show women often face higher MP, with rural women especially disadvantaged due to social norms, lower access to assets and services, and time poverty. Prior studies frequently measure at the household level, which can obscure intra-household gender disparities; individual-level measures better capture gendered deprivations. Country- and context-specific MPIs are recommended by OPHI. Existing female-focused MPIs differ across contexts and over time, increasingly including non-material dimensions such as subjective wellbeing. In China, evidence suggests higher MP in rural areas and among women, but comprehensive, individual-level, dynamic analyses specific to rural women have been limited.

Data: The study uses China Family Panel Studies (CFPS) biennial panel microdata (2010, 2012, 2014, 2018, 2020). Household and adult datasets were merged; missing data were dropped casewise. Due to >50% missing BMI in 2016, that wave was excluded. Final pooled sample: 133,557 individuals over five waves; rural women observations: 33,381 (age 16–97, M=44.84, SD=16.32). For panel analyses, 8,950 rural women participated in all five surveys (age 16–87, M=46.64, SD=12.92). Subgroups: Comparisons by gender and rural/urban; spatially by Eastern, Central, Western Regions; temporally 2010–2020. MPI construction (Alkire-Foster): Six dimensions with 12 binary indicators, dimension-equal weighting. Dimensions and indicators: (1) Economic: per capita net household income below official poverty line (year-specific). (2) Education: years of schooling less than 9. (3) Health: BMI<18.5 (undernutrition), chronic disease in last six months, poor or worse self-rated health. (4) Social welfare: lacks any medical insurance; lacks any pension insurance. (5) Living standards: does not own housing; lacks access to clean drinking water; lacks access to clean cooking fuel. (6) Subjective wellbeing: low life satisfaction; depression (CES-D 6 score >10). Indicators are coded 1 if deprived, else 0. Weights: Dimensional equal weighting; within-dimension indicator weights as specified (e.g., 1/6 for single-indicator dimensions like income, education; 1/18 for each of three health and living standard indicators; 1/12 for each of two indicators in social welfare and subjective wellbeing). Correlations among indicators were modest (all <0.3) and VIFs <4, supporting equal weighting. Poverty identification: Dual cutoffs—indicator-specific deprivation cutoffs, and overall deprivation cutoff k=0.3 (validated via sensitivity; clear threshold). Compute headcount H (share multidimensionally poor), average deprivation share A among the poor, and MPI M0=H×A. Decomposition by subgroup and time; contributions by dimension/indicator. Dynamics: Classify poverty spells over five waves with time threshold t=2: never poor (0 waves), temporary poor (1–2 waves), chronic poor (>2 waves). Sensitivity analyses varied k and t. Statistical reporting includes absolute and relative differences with CIs; robustness checks estimated a three-dimension MPI (education, health, living standards) following OPHI baseline.

- Overall status: Rural women are more likely to be multidimensionally poor than other subgroups. The six-dimensional MPI for rural women shows substantial breadth, depth, and intensity of deprivation relative to men, urban women, and rural men. - Demographic heterogeneity: Older rural women are markedly more deprived than younger ones. Specifically, those aged 60+ had H=0.764, A=0.437, M0=0.334 versus those under 60 with H=0.447, A=0.411, M0=0.184 (relative risk +81.52%). Rural women with spouses are less deprived than those without; economic deprivation intensity is higher among those without spouses (economic contribution 0.115). Confidence matters: women without confidence in the future show higher deprivation, with life satisfaction and depression deprivation rates reaching 44.38% and 57.27% respectively; subjective wellbeing contributes most to MPI in this group (life satisfaction 0.086; depression 0.112). - Spatial disparities: Western Region has the highest poverty (H=0.596, A=0.426, M0=0.254), followed by Central (H=0.476, A=0.416, M0=0.198), and Eastern (H=0.430, A=0.409, M0=0.176). Western exceeds Eastern by 44.32% in M0. Deprivations are particularly high in cooking fuel and drinking water in the West; the absolute difference in cooking fuel deprivation between West and East is 25.896 percentage points, and relative difference in water deprivation is 395%. Eastern rural women are more deprived in medical insurance (6.39%), pension insurance (70.80%), and housing asset (9.37%). - Temporal trend (2010–2020): Rural women's poverty declined substantially. H fell from 0.588 (2010) to 0.401 (2020) (−34.8% relative); A decreased from 0.425 to 0.407. M0 declined from 0.250 (2010) to 0.163 (2018/2020). Improvements were notable in social welfare and living standards: pension insurance deprivation fell from 100% to 44.96% (−55.04%); medical insurance deprivation fell from 12.58% to 0.49% (−96.15%); cooking fuel deprivation fell from 64.05% to 34.77% (−45.71%). Some indicators worsened: self-rated health deprivation rose from 19.06% to 35.01% (+83.74%); depression from 28.10% to 42.11% (+49.89%); chronic disease and housing asset deprivation also increased; education deprivation remained high and roughly unchanged (~57%). - Contributions: Non-material dimensions gained importance over time. From 2010 to 2020, contribution shares rose for education (+14.96%), health (+67.86%), and subjective wellbeing (+32.43%), while economic status and social welfare/living standards contributions declined (e.g., income contribution 0.115→0.079, −31.30%; medical insurance −38.57%; pension insurance −18.25%). Spatially, education (>35%) dominates across regions; social welfare contributes more in the East (19.4%), while living standards contribute more in the West (+32.35% relative). - Dynamics: Over the panel (t=2), 26.21% of rural women were never poor, 26.02% temporary poor, and 47.77% chronically poor. The never-poor share was highest in the Eastern Region and lowest in the Western Region; chronic poverty share was highest in the West and lowest in the East. - Sensitivity: Increasing k reduces MPI sharply; for k>0.6, MP incidence approaches zero. Western Region consistently exhibits the highest incidence and intensity at any k. Time-threshold sensitivity indicates t=2 is a clear cutoff; for t>2, chronic poverty share declines toward zero and temporary poverty rises toward ~0.8. - Robustness (3-dimension MPI): Patterns persist—women and rural residents have higher MP; rural women have the largest MPI (M0=0.32, H=0.641, A=0.499). The 3D measure tends to overestimate breadth and depth relative to the 6D measure, but rankings and trends are consistent.

Findings confirm the feminization of multidimensional poverty in rural China and highlight that rural women experience the greatest breadth and depth of deprivation. Cultural norms and patriarchal structures limit women’s opportunities and access to resources, contributing to persistent educational deficits and social exclusion. Ageing amplifies vulnerabilities through deteriorating health and reduced earning capacity. Marital support can buffer risks via emotional, financial, and caregiving resources; conversely, lacking a spouse raises economic vulnerability. Confidence in the future functions as a psychological capability that can catalyze proactive behaviors and improve subjective wellbeing; lack of confidence is associated with substantial deprivations in life satisfaction and mental health. Spatially, the Western Region’s harsher natural environments, fragile ecosystems, and weaker infrastructure are associated with higher deprivation, particularly in basic services (clean water, clean fuels). Eastern areas, despite better infrastructure, exhibit crowding-out in social welfare and asset ownership for rural women. Temporally, China’s targeted poverty alleviation and expanded social welfare coverage contributed to declines in income poverty and improvements in living standards and social insurance. As material deprivations eased, non-material factors—education, health, and subjective wellbeing—emerged as increasingly salient determinants of rural women’s MP. Rising depression and deteriorating self-rated health underscore growing mental health and health-system challenges amidst social and technological change. These results address the research questions by quantifying current MP levels, revealing pronounced spatial and temporal heterogeneity, and identifying key levers—education, health, and subjective wellbeing—for further reducing rural women’s MP.

This study constructs and applies a six-dimension, individual-level MPI for rural women in China using CFPS 2010–2020 and the A-F method, capturing both static and dynamic aspects of deprivation. It demonstrates that rural women face higher MP than other groups; older age, lack of a spouse, and lack of confidence heighten risk. Spatially, MP is highest in the Western Region and lowest in the Eastern; temporally, MP decreases markedly from 2010 to 2020. Material deprivations (economic status, social welfare, living standards) have fallen, while the contributions of education, health, and subjective wellbeing have grown. Policy implications include: (1) adopting a gender lens in poverty governance, recognizing and valuing women’s labor and strengthening rights protections in education, employment, marriage, and pensions; (2) tailoring interventions to regional contexts—direct assistance and infrastructure/ecological improvements in higher-risk Western and Central regions, with attention to adaptation support in relocations; (3) instituting dynamic monitoring systems to track evolving deprivations, elevating the weight and refinement of indicators in education, health, and subjective wellbeing, and establishing prevention/monitoring mechanisms for transitions into and persistence in poverty. Future research should collect primary, timely data; extend analyses beyond China for cross-national comparability; and develop models that elucidate causal mechanisms underlying rural women’s MP.

- Use of secondary CFPS data limits timeliness and variable scope; first-hand data would enhance precision and relevance. - Results are China-specific due to data design and may not generalize internationally; cross-country datasets are needed for broader conclusions. - The study focuses on measurement and decomposition; causal mechanisms are not fully modeled. More comprehensive causal frameworks are needed to unpack pathways into and out of multidimensional poverty among rural women.