Families of children with disabilities: income poverty, material deprivation, and unpaid care in the UK

The study examines the tripartite relationship between child disability, unpaid (informal) care provided by co-resident family members, and poverty among UK families with dependent children. Prior work shows high child poverty rates in the UK, rising in-work poverty, and links between disability and economic hardship. The authors hypothesize that families supporting a child with disabilities may face greater financial strain due to extra care needs and potential labour market constraints, and that unpaid care provision could further influence poverty risk. The research aims to (1) compare poverty likelihood between families with and without a disabled child; (2) assess whether, among families with a disabled child, unpaid care provision increases poverty risk; (3) identify covariate risk factors for poverty by disability status; and (4) compare risk factors between families with disabled children who do and do not provide unpaid care. Poverty is measured using relative income (after housing costs, AHC) and a child material deprivation index.

In the UK, poverty measurement is contested. While the Child Poverty Act 2010 established income and material deprivation indicators, subsequent legislative changes (Welfare Reform and Work Act 2016) shifted official emphasis away from income-based metrics, despite broad expert support for retaining income and material deprivation measures. AHC income is often preferred as it accounts for housing costs and better reflects living standards. Prior research shows that income poverty and material deprivation do not always align at the individual level, underscoring the importance of multidimensional measures. Families with disabled children have elevated poverty risks, and disability definitions in UK surveys follow the Equality Act (2010) social model emphasizing functional limitations. Disability incurs additional direct costs that benefits may not fully offset. Sociodemographic correlates of poverty and hardship include lower education, less advantaged occupations, non-white ethnicity, lone parenthood, and non-traditional family structures. Unpaid care is widespread and economically significant but can negatively affect caregivers’ employment and well-being. However, there is no universally agreed definition of unpaid care activities, and survey capture varies. The UK literature has seldom directly examined the combined effects of child disability and parental unpaid care on family poverty, motivating this study.

Design and data: Secondary analysis of the UK Family Resources Survey (FRS) 2018/19, linked with Households Below Average Income (HBAI) data. The FRS is a face-to-face, repeated cross-sectional survey with stratified cluster sampling in Great Britain (PAF) and systematic, geographically stratified sampling in Northern Ireland; response rate ~50%. Sample: 5451 benefit units (BUs) with dependent children (ages 0–16, or 16–19 in full-time education). Subsamples: (1) families with no disabled child (N=4716); (2) families with at least one disabled child (N=735), further split into (2a) no unpaid care (N=491) and (2b) unpaid care (N=244). Measures: - Poverty outcomes: (a) Relative income poverty before housing costs (BHC) and after housing costs (AHC), using 60% of the median equivalised income threshold; binary indicators derived (poor/not poor). AHC subtracts housing costs (rent, mortgage interest, insurance) from net income. Although both BHC and AHC were examined, only AHC results are reported due to similar patterns. (b) Child material deprivation: continuous score 0–100 based on 21 items (prevalence-weighted), dichotomised at ≥25 to indicate material deprivation. - Group variables: Child disability per FRS definition aligned to Equality Act 2010 (long-term condition with substantial limitation). Unpaid care is a binary indicator if any co-resident family member provides unpaid care to someone with long-term ill-health/disability or old age; FRS intentionally broad and not prescriptive about care activities. - Covariates: Number of dependent children; number of adults with disability; lone parent status; average self-reported health (higher values worse); highest education in BU (GCSE or lower; A-levels; degree+); ethnicity (White vs Black/Asian/Other); housing tenure (owner vs private rent vs social rent); highest occupational classification (Managers/Professionals/Associate professionals [ref]; Admin/Skilled/Caring; Elementary/None; Sales/Process/Plant operators); employment status incorporated via occupation. Missing data: Education (7.56%) and average health (1.12%) were imputed via multiple imputation (m=5) using predictive mean matching (MICE in R), with all other variables as predictors. Analytical strategy: Descriptive statistics by groups; multivariable logistic regression models for material deprivation and AHC poverty. Models addressed: (i) disability vs no disability (full sample); (ii) unpaid care vs no unpaid care among families with a disabled child; (iii) separate models within no-disability and disability groups; (iv) separate models within disability group by unpaid care provision. Multicollinearity assessed via GVIF (all <2). Analyses conducted in R.

- Descriptives (unweighted): Material deprivation was higher among families with a disabled child (26.67%) vs no disability (14.72%). AHC poverty was slightly lower among disability families (24.22%) vs no disability (26.99%). Within disability families, AHC poverty was higher when no unpaid care was reported (29.74%) vs unpaid care (13.11%). Lone parent prevalence was higher with disability (38.37% vs 25.15%). Social renting was more common among disability families (34.29% vs 18.94%). - Disability vs no disability (adjusted, Table 3): Presence of a child with no disability increased odds of AHC poverty versus disability families (OR=2.173, p<0.001); disability indicator not significant for material deprivation. Sociodemographic covariates associated with higher risk included renting privately (material deprivation OR=4.725; AHC OR=3.808) and social renting (material deprivation OR=5.660; AHC OR=3.394), non-white ethnicity (material deprivation OR=2.206; AHC OR=2.078), worse average health (material deprivation OR=1.635; AHC OR=1.120), lone parenthood (material deprivation OR=2.132; AHC OR=1.268), lower occupational classes (e.g., Elementary/None: material deprivation OR=3.132; AHC OR=2.810), more adults with disability (material deprivation OR=1.702; AHC OR=1.260), and more dependent children (material deprivation OR=1.273; AHC OR=1.378). Degree-level education was protective (material deprivation OR=0.585; AHC OR=0.720). - Unpaid care among disability families (Table 3 right): Unpaid care was not associated with material deprivation (OR=0.790, p=0.278) but was associated with lower odds of AHC poverty (OR=0.311, p<0.001). Among disability families, renting remained a strong risk factor (private rent AHC OR=3.203; social rent AHC OR=2.492). Lower occupations increased AHC poverty risk (e.g., Elementary/None OR=2.251; Sales/Process/Plant operator OR=2.333). - Stratified models (Table 4): Effects of key sociodemographics were broadly similar in direction across families with and without disabled children. Notable differences included a stronger protective effect of higher education for material deprivation among disability families (Degree+ OR=0.322) vs no-disability families (OR=0.678), and a larger material deprivation penalty for non-white ethnicity among disability families (OR=3.200) vs no-disability families (OR=2.061). In disability families, number of dependent children and adults with disability increased material deprivation risk, but not AHC poverty. - Within disability families by unpaid care (Table 5): For material deprivation, private renting had a larger association among unpaid care providers (OR=7.326) than non-providers (OR=4.621); social renting was also strongly associated (OR≈6–7). For AHC poverty, private renting increased odds in both groups (non-care OR=3.066; care OR=3.295). Lower occupational class (Elementary/None) raised AHC poverty among unpaid care providers (OR=4.710).

The study set out to analyze whether child disability and unpaid care provision influence family poverty and whether sociodemographic correlates differ across these contexts. Adjusted models indicate that families without a disabled child had higher odds of AHC poverty than those with a disabled child, contrary to expectations. The authors attribute this to the treatment of disability-related benefits within income: benefits are included in AHC income and may lift some disability families just above the poverty threshold despite higher disability-related costs. When disability benefits are conceptually removed, AHC poverty becomes higher among families with disabled children (e.g., 35.10% vs 28.12% without disability), aligning with the higher material deprivation rates observed. Unpaid care provision among disability families was associated with lower AHC poverty odds, which is counterintuitive given evidence that caregiving reduces labor supply. The authors posit selection processes: families with higher baseline incomes may be more able to choose unpaid care without falling into poverty, offsetting negative income effects at the aggregate level. Longitudinal data and qualitative inquiry could unpick decision-making and consequences of unpaid care over time. Across groups, classic sociodemographic disadvantages—renting (especially social or private), non-white ethnicity, lone parenthood, lower occupational class, poorer health, and more dependents—were consistently associated with higher material deprivation and often higher AHC poverty, with some stronger effects among disability families (e.g., ethnicity, education).

This study provides an initial, UK-focused examination of the interplay between child disability, unpaid care, and poverty using FRS 2018/19 and HBAI data. It shows that sociodemographic risk factors for material deprivation are broadly similar across families, with pronounced penalties for renters, lone parents, lower occupational status, and non-white ethnicity, and protective effects of higher education. In adjusted models, families without a disabled child had higher AHC poverty odds than those with a disabled child, likely reflecting inclusion of disability benefits in income rather than lower needs; removing benefits reverses this pattern. Among families with a disabled child, unpaid care provision is associated with lower AHC poverty odds, potentially reflecting selection by higher-income families into caregiving. Future research should (a) refine poverty measurement to account for disability-related costs and benefit structures; (b) use longitudinal designs to disentangle selection into caregiving from its economic consequences; and (c) collect targeted data on families of disabled children, including caregiving intensity, types of disability, and indirect care, to inform policy.

Key limitations include: (1) survey response rate (~50%) with potential non-response bias (e.g., time-constrained carers underrepresented); (2) proxy responses for some items and self-identification of caregiving may cause misclassification; (3) FRS unpaid care questions use a show card listing activities, which may both prompt identification and constrain reporting; (4) likely capture of primarily direct/primary care, potentially missing indirect care; (5) disability measured dichotomously without type/severity differentiation due to low counts; (6) cross-sectional design limits causal inference and conflates selection into unpaid care with its effects; (7) inclusion of disability benefits in income may mask true living standard deficits; (8) modest missingness in education and health addressed via multiple imputation, but residual bias is possible.