Skip to main content

Inter-generational effects of disability benefits: evidence from Canadian social assistance programs


Individuals with disabilities face greater challenges in the labor market than able-bodied individuals, and a growing body of research is finding that their children also tend to have more developmental problems than the children of able-bodied parents. Can transfer payments help reduce this gap? In this paper, we present the first evidence on how parental disability benefits affect the well-being of children. Using changes in real benefits under ten disability benefit programs in Canada as an identification strategy and Statistics Canada’s National Longitudinal Survey of Children and Youth (NLSCY) as the data source on child outcomes, we find strong evidence that higher benefits lead to improvements in children’s cognitive and non-cognitive development, as measured by math scores in standardized tests, hyperactive symptoms, and emotional anxiety behavior. The effect is larger on children with a disabled mother than on those with a disabled father.

This is a preview of subscription content, access via your institution.


  1. For example, Statistics Canada’s Participation and Activity Limitation Survey (PALS) identifies persons with a disability by using two filter questions on the Census. The first question asks the respondents whether they have any difficulties in hearing, seeing, communicating, walking, climbing stairs, bending, learning, or doing any similar activities. The second question asks whether there are any physical/mental conditions or health problems that reduce the amount or kind of activity the person can do at home, in leisure activities, at work, or at school. A yes to any of these questions will result in a respondent being selected into the PALS survey pool.

  2. Documentation on the prevalence of parents with disabilities has been extremely scarce in the literature. A main reason may be data availability—most national surveys provide information on the number of people with disabilities or the number of parents but not the combination of these two characteristics (Preston 2012).

  3. For example, Mont and Cuong (2013) find in the 2006 Vietnam Household Living Standards Survey that children of parents with a disability have a lower enrollment rate in primary and secondary school. Bratti and Mendola (2014) presents evidence from the Bosnia and Herzegovina (BiH) Living Standards Measurement Survey (LSMS) that children of mothers with severe limitations in activity of daily living are less likely to be in school at ages 15–24. Morefield, Mühlenweg, and Westermaier (2011) find from the German Socio-Economic Panel (GSOEP) that work-limiting disabilities of either parent significantly increase children’s problem behaviors and negatively affect their personality traits.

  4. All dollar values expressed in 2008 dollars.

  5. For example, Milligan and Stabile (2011) using data from the Canadian NLSCY conclude that the provincial child tax benefit policies initiated in 1998 have a significant positive effect on educational outcomes, physical health, and mental health for the general population of children.

  6. In this paper, they measure the hours of parental time available for children as total weekly hours less “sleep time”, less weekly hours of paid work. They argue that weekly rather than annual hours are more relevant for children, since care cannot be deferred until a later point in the year. See also Phipps, Burton, and Osberg (1996) who find it is weekly hours which generate most time stress for adults.

  7. According to this study, most mothers do not place their adolescent children in structured care after school as they did with their younger children. Adolescents may have difficulties if left alone after school and into the evening hours as mothers take on off-hours and shift work. They also find that adolescent children may be asked to take on greater household responsibilities and may be encouraged to engage in employment themselves when their single mothers move into employment. While there is limited research on the effects of household chores on children, a high level of employment during adolescence (particularly more than 20 h of employment) has been linked with children’s difficulties in school and increased drug and alcohol use.

  8. The disability support program in PEI provides income support to persons with disabilities on a case-by-case base. In this study, we use data on PEI’s social assistance program. Alberta also has a distinct program for persons with disabilities: the Assured Income for the Severely Handicapped (AISH) program. Different from other provincial disability benefit programs, the AISH clients are provided with a flat rate living allowance benefit which is not contingent on family size. We replicate our analysis later with children from these two provinces excluded. As are shown, our main results are not substantially affected.

  9. In 2001, of the 3.42 million adults with disabilities in Canada, 10 % received income support from provincial disability benefit programs, about the same proportion as those receiving a Canadian/Quebec Pension Plan (Prince 2008).

  10. All provincial programs exempt most fixed assets, such as principal residence, vehicles (up to a certain limit), the value of prepaid funerals and property/equipment required for employment, while liquid assets are only partially exempt.

  11. All provincial programs exempt a portion of employment income although using slightly different formulae. For example, Nova Scotia allows its client families to keep the first $200 of earned total income and one fourth of earnings exceeding $200 per month.

  12. PEI increased its basic earnings exemption level from $600 to 900 per month in 2001.

  13. Along with the medical certification requirement, an applicant must be of certain age (between 18 and 65) and be resident of particular province to be eligible for the benefits.

  14. We use the maximum benefit in the empirical analysis because we do not know the disability payments that are actually received by individuals and because these would be endogenous with child outcomes.

  15. The National Council of Welfare computes the disability benefits as the sum of the basic assistance rate (i.e., amounts for food, clothing, shelter and utilities, personal and household needs), additional benefits (i.e., supplementary allowances that were automatically provided to persons with a disability), and the provincial tax credit and GST credit that are intended for the disabled. These estimates assume a single disabled person who (1) qualifies for long-term rates of assistance, (2) lives in the largest urban area in the province or territory, (3) goes on disability benefits on January 1 of each year and remains on benefits for the entire calendar year, and (4) is a tenant in the private rental market rather than a homeowner or social housing tenant and who also does not share accommodation.

  16. In some provinces, actual entitlement to disability benefits may vary according to the circumstances of each individual family, including household size, composition, and the children’s age. We do not differentiate these family types because of data limitation, i.e., the National Council of Welfare did not produce benefit schedules for couple-families with a disability. Doing so also avoids potential endogeneity in fertility decisions and living arrangements to the generosity of needs-tested benefits (Moffitt 1990; Milligan 2005). We are unaware of any systematic legislation change that affected benefit schedules for single persons differently than for couple-families with a disability during the study period. This paper exploits within-province variations in benefit levels for the single disabled over time, which will reflect changes in benefit levels for other family types, unless there was such a change.

  17. We subtract the sample mean from each respective benefit level.

  18. Since the difficulty of the math test in the NLSCY varies with the school grade of the child, the standardized scores increase as a child grows older (Lefebvre, Merrigan, and Verstraete 2008). We thus include parent-report school grade (K through ten) for the child in our regression. Inclusion of this variable causes less than 2 % of 4 to 5 year olds who did not attend kindergarten to drop out of our estimation sample. In a robustness check that is not reported in this paper, inclusion of these children does not affect our results qualitatively (available upon request). In another study that uses the same data source (i.e., the NLSCY), we find only a very small number of children in grade K through four who do not comply with the school entry regulations in their year/province (Chen et al. 2015). Our data is consistent with this finding.

  19. For example, less patient parents may be more likely to engage in risky behavior, experience disability, and at the same time invest less time and money in children.

  20. The sample of children living in lone mothers where the lone mother had a disability was unfortunately too small to produce reliable estimates. To the extent that such families may be most affected by higher disability benefits, we may be underestimating effect size.

  21. Data on the take-up rate of means-tested disability benefits in Canada are unfortunately difficult to obtain. Based on our only source of information, Prince (2008), using the 2001 Participation and Activity Limitation Survey (PALS) reports that, of the 3.42 million of Canadian adults with disabilities, 10 % received income support from provincial disability benefit programs. If the vast majority of benefit recipients do not have a university degree, the benefit take-up rate can be around 15 % among the population under analysis, since about 70 % of children reported to live with parents neither of whom has a university degree during our study period. If true, the ATT effect of disability benefits should be 6–7 times as large as the size of the ITT effect estimated in current paper. However, since the unit of analysis in the NLSCY is the child instead of the adult, above extrapolation is rather crude and cautions should be taken in interpretation.

  22. Given our main coefficient of interest is the benefit effect on children; we choose to cluster standard errors at the highest level with the smallest number of clusters, which is a relatively conservative strategy. Clustering at the household level produces qualitatively similar results.

  23. Children living in lone-parent families make up only 3 % of all children aged between 4 and 15 years in the NLSCY. To the extent that lower benefits may increase the chance of parental separation/divorce and/or such families may be most affected by changes in benefit levels, we may be underestimating effect size.

  24. Earlier cycles contained an additional question for each scale. We re-constructed these scales by dropping this question so they can be compared consistently across cycles.

  25. The survey question that we used to identify parental disability in the NLSCY is as follows: “Because of a long-term physical or mental condition or a health problem, are/is … … … limited in the kind or amount of activity you/he/she can do: 1) At home? 2) At school? 3) At work? 4) In other activities such as transportation to or from work or leisure time activities? 5) In caring for children?”. If the parent answered “yes” to any of the above items for herself or her spouse, parental disability is coded one indicating disabled, otherwise zero for non-disabled.

  26. In assessing the impact of parental disability on parenting, we conduct a simple descriptive analysis on the association between disabilities of different functional domains and child outcomes (see Appendix Table 3). In general, all types except for limitations at transportation/leisure/other are associated with significantly worse outcomes of children even after adjusting for child gender, age, and school grade. There is, however, no clear pattern suggesting one matters more than another, possibly due to the high correlation among the limitations reported in different functional domains.

  27. A chi-squared test for group difference suggests a statistically significant difference for the latter.

  28. In this study, we use the Luxembourg Income Study (LIS) scale to calculate “equivalent income.” The LIS scale is calculated as the square root of family size.

  29. For the sake of completeness, we also report coefficients on other covariates in the model. These results indicate that: (1) conditional on school grade, older children score higher on standardized math test, but tend to exhibit more hyperactive symptoms and anxiety behavior; (2) boys do better at math tests but suffer more hyperactive symptoms than girls; (3) compared with children of high school dropouts, children of parents with a high school degree or post-secondary education have better outcomes in all three cases.

  30. Consumer Price Index (CPI), 2011 basket content is from Statistics Canada’s CANSIM (database) Table 326-0021: Access on April 4, 2013.

  31. A fully interacted model that includes interaction between parental disability dummy and every single covariate in the regression produces highly similar results (available upon request).

  32. In our regression analysis, the disability benefit variable is measured in 1000 dollars.

  33. Cognitive achievement is more likely to be linked to the cumulative process of human capital acquisition (Todd and Wolpin 2007) as opposed to external shocks, such as variations in family income, or parental stress associated with employment.

  34. We do not attach any fundamental meaning to these estimates since in the presence of interaction terms, their magnitude and significance merely reflect the group difference at the average benefit level and whether or not this difference is statistically different from zero at this point.

  35. Data on the provincial expenditure of education (Table 478-0014) and Consumer Price Index (Table 326-0021) were retrieved on April 4, 2013 from Statistics Canada’s CANSIM (database): // We extrapolate the total number of school-age children by province by year using data from cycles 1–8 of the NLSCY.

  36. As part of Early Child Development (ECD) initiative, the NLSCY dropped many young children aged 6–10 from its cross-sectional sample in later cycles. Standardized math test score thus is the only outcome for which there is sufficient sample in each year to enable comparisons over time.

  37. Less than 2 % of children in the estimation samples moved inter-provincially during our study period.

  38. Apart from published statistics, we cannot find any existing studies that estimate or extrapolate the rate of take-up for a means-tested disability benefit program in Canada.

  39. In addition, most existing research in Canadian context seems to agree that the rise and fall of the number of people on welfare do not coincide with benefit levels. As pointed out by a survey by the National Council of Welfare (1998), every significant study has shown that welfare caseload growth tends to coincide with periods of recession and rise of unemployment. It is not the meager benefit levels that attract people to welfare or discourage them from leaving to find a job.

  40. Around 200 children, representing 10 % of all children with at least one disabled parents are excluded from our estimation samples.

  41. The NLSCY contains retrospective information on labor market activities for both parents such as paid work participation, usual weekly hours, and family income received from all sources (before taxes), 12 months prior to the survey. Parents reported their hours of work in six categories: less than 10 h, 10–19 h, 20–29 h, 30–39 h, 40–49 h, and 50 h or more. We create a pseudo continuous variable coded at the mid-point of each category to capture the non-linear nature of parental hours, and another indicator variable that identifies a parent’s full-time work status (i.e., equal to one if a parent works 30 h or more), to test the effect of hours of work.

  42. As mentioned before, a reduction in disability benefits will both directly decrease poor families’ disposable incomes and induce poor parents to participate more in paid work. Although a negative association between benefit level and parental paid work for both the disabled parent and his or her non-disabled spouse is to be expected, the benefit effect on family income is ambiguous.

  43. Considering that in our sample mothers’ average weekly hours is lower than fathers’, the size of benefit effect for own-labor supply is slightly larger for disabled mothers than disabled fathers.

  44. Overall, there seems to be an asymmetric incentive effects on spousal labor supply—as the benefit level declines, fathers of disabled spouse increase their full-time employment and hours of work. By contrast, wives of disabled spouse do not behave differently. The pattern is consistent with previous research on the effect of spousal ill health on labor supply in particular, husband’s health on wife’s labor supply in North America (Berger and Fleisher 1984; Berger 1983; Gallipoli and Turner 2009), and studies related to spill-over effects of public transfer payments in the U.S. (e.g., Colie 2004). However, cautions should be taken in interpretation since the unit of analysis in the NLSCY is the child instead of the adult. Even though the results can be replicated when one parent per child is randomly selected (results available upon request), further analysis using alternative data sources is needed.

  45. Log of family income measure is used in column (5).


  • Akerlof GA, Kranton RE (2000) Economics and identity. Q J Econ 115:715–753

    Article  Google Scholar 

  • Alessandri SM (1992) Effects of maternal work status in single-parent families on children's perception of self and family and school achievement. J Exp Child Psychol 54(3):417–433

    Article  Google Scholar 

  • Avery RC, Hogan DP (2006) Family Configurations of Disability in the 2000 Census. Conference presentation at the annual meeting of the Population Association of America

  • Beauregard L, and Noreau Luc (2009) Spouses of persons with spinal cord injury: impact and coping. Brit J Soc Work: bcp140

  • Becker GS, Tomes N (1986) Human capital and the rise and fall of families. J Labor Econ: S1-S39

  • Berger MC (1983) Labor supply and Spouse’s health: the effects of illness, disability, and mortality. Soc Sci Quart 64(3):494–509

    Google Scholar 

  • Berger MC, Fleisher BM (1984) Husband's health and wife’s labor supply. J Health Econ 3(1):63–75

    Article  Google Scholar 

  • Blau DM (1999) The effect of income on child development. Rev Econ Stat 81(2):261–276

    Article  Google Scholar 

  • Bound J, Burkhauser RV (1999) Economic analysis of transfer programs targeted on people with disabilities. Handb Labor Econ 3:3417–3528

    Article  Google Scholar 

  • Bratti M, Mendola M (2014) Parental health and child schooling. J Health Econ 35:94–108

    Article  Google Scholar 

  • Brooks-Gunn J, Duncan GJ (1997) The effects of poverty on children. Future Child 7:55–71

    Article  Google Scholar 

  • Burton P, Lethbridge L, Phipps S (2008) Children with disabilities and chronic conditions and longer-term parental health. J Socio Econ 37(3):1168–1186

    Article  Google Scholar 

  • Burton P, Chen K, Lethbridge L, Phipps S (2014) Child health and parental paid work. Rev Econ Househ 5:1–24

    Google Scholar 

  • Cameron SV, Heckman JJ (1998) Life cycle schooling and dynamic selection bias: models and evidence for five cohorts of American males. J Polit Econ 106(2):262–333

    Article  Google Scholar 

  • Campolieti M (2004) Disability insurance benefits and labor supply: some additional evidence. J Labor Econ 22(4):863–889

    Article  Google Scholar 

  • Campolieti M, Gomez R, Gunderson M (2009) Volunteering, income support programs and persons with disabilities. Ind Relat 64(2):189–208

    Google Scholar 

  • Chen K, Fortin N, Phipps S (2015) Young in class: implications for inattentive/hyperactive behavior of Canadian boys and girls. Can J Econ: Forthcoming

  • Coile C (2004) Retirement incentives and couples’ retirement decisions. Top Econ Anal Policy 4(1):1–30

    Google Scholar 

  • Cunha F, Heckman JJ, Schennach SM (2010) Estimating the technology of cognitive and noncognitive skill formation. Econometrica 78(3):883–931

    Article  Google Scholar 

  • Currie J (1998) The effect of welfare on child outcomes: what we know and what we need to know. Northwestern University/University of Chicago Joint Center for Poverty Research. Accessed 1 April, 2014

  • Currie J (2008) Healthy, wealthy, and wise: socioeconomic status, poor health in childhood, and human capital development. National Bureau of Economic Research Working Paper No. 13987. Accessed 1 April, 2014

  • Currie J (2011) Inequality at birth: some causes and consequences. No. w16798. National Bureau of Economic Research. Accessed 1 April, 2014

  • Currie J, Almond D (2011) Human capital development before age five. Handb Labor Econ 4:1315–1486

    Article  Google Scholar 

  • Currie J, Stabile M (2003) Socioeconomic status and child health: why is the relationship stronger for older children? Am Econ Rev 93:1813–1823

    Article  Google Scholar 

  • Curtis L, Phipps S (2000) Economic resources and children's health and success at school: an analysis with the National Longitudinal Survey of Children and Youth. Applied Research Branch, Human Resources Development Canada No. W-01-1-4E. Accessed 1 April, 2014

  • Curtis LJ, Dooley MD, Lipman EL, Feeny DH (2001) The role of permanent income and family structure in the determination of child health in Canada. Health Econ 10(4):287–302

    Article  Google Scholar 

  • Dooley M, Stewart J (2004) Family income and child outcomes in Canada. Can J Econ 37(4):898–917

    Article  Google Scholar 

  • Fortin B, Lacroix G, Drolet S (2004) Welfare benefits and the duration of welfare spells: evidence from a natural experiment in Canada. J Public Econ 88(7):1495–1520

  • Galarneau D, Radulescu M (2009) Employment among the disabled. Perspect Labour Income 10(5):5–15

    Google Scholar 

  • Gallipoli G, Turner L (2009) Household responses to individual shocks: disability and labor supply. Nota di lavoro//Fondazione Eni Enrico Mattei: Global challenges No. 97.2009. Accessed 1 April, 2014

  • Gordon PA, Perrone KM (2004) When spouses become caregivers: counseling implications for younger couples. J Rehabil 70(2):27–32

    Google Scholar 

  • Haveman R, Wolfe B (1995) The determinants of children's attainments: a review of methods and findings. J Econ Lit XXXIII:1829–1878

    Google Scholar 

  • Haveman R, Wolfe B (2000) The economics of disability and disability policy. Handb Health Econ 1:995–1051

    Article  Google Scholar 

  • Hogan DP, Shandra CL, Msall ME (2007) Family developmental risk factors among adolescents with disabilities and children of parents with disabilities. J Adolescence 30(6):1001–1019

    Article  Google Scholar 

  • Lethbridge L, Phipps S (2006) Income and the outcomes of Cchildren. Statistics Canada Analytical Studies Branch Research Paper Series 2006281e. Accessed 1 April, 2014

  • Lundberg SJ, Pollak RA, Wales TJ (1997) Do husbands and wives pool their resources? Evidence from the united kingdom child benefit. J Hum Resour 32:463–480

    Article  Google Scholar 

  • Mann C, Dieppe P (2006) Different patterns of illness‐related interaction in couples coping with rheumatoid arthritis. Arthrit Care Res 55(2):279–286

    Article  Google Scholar 

  • Mayer SE (1997) What Money Can't Buy: Family Income and Children's Life Chances. Harvard University Press

  • Mayer SE, Jencks C (1993) Recent trends in economic inequality in the United States: income vs. expenditures vs. material well-being. In: Popademitrious D, Wolff E (eds) Poverty and prosperity in the USA in the late twentieth century. Popademitrious Press, London

    Google Scholar 

  • Miller AR, Zhang L (2009) The effects of welfare reform on the academic performance of children in low‐income households. J Policy Anal Manag 28(4):577–599

    Article  Google Scholar 

  • Milligan K (2005) Subsidizing the stork: new evidence on tax incentives and fertility. Rev Econ Stat 87(3):539–555

    Article  Google Scholar 

  • Milligan K, Stabile M (2011) Do child Tax benefits affect the well-being of children? Evidence from Canadian child benefit expansions. Am Econ J: Econ Policy 3(3):175–205

    Google Scholar 

  • Moffitt R (1990) The effect of the US welfare system on marital status. J Public Econ 41(1):101–124

    Article  Google Scholar 

  • Moffitt R (1992) Incentive effects of the US welfare system: a review. J Econ Lit: 30:1–61

    Google Scholar 

  • Mont D, Cuong N (2013) Does parental disability matter to child education? Evidence from Vietnam. World Dev 48:88–107

    Article  Google Scholar 

  • Moore KA, Driscoll AK (1997) Low-wage maternal employment and outcomes for children: a study. Future Child 7:122–127

    Article  Google Scholar 

  • Morefield B (2010) Parental health events and children's skill development. University of North Carolina at Greensboro Department of Economics Working Papers 10-11. Accessed 1 April, 2014

  • Morefield B, Mühlenweg AM, Westermaier F (2011) Impacts of parental health on children's development of personality traits and problem behavior: evidence from parental health shocks. ZEW Discussion Papers No. 11-049. Accessed 1 April, 2014

  • Morris P, Michalopoulos C (2000) The self-sufficiency project at 36 months: effects on children of a program that increased parental employment and income. Applied Research Branch, Human Resources Development Canada. Accessed 2 May, 2015

  • Phipps S, Burton P (1998) What’s mine is yours? The influence of male and female incomes on patterns of household expenditure. Economica 65(260):599–613

    Article  Google Scholar 

  • Preston P (2012) Parents with disabilities. International Encyclopedia of Rehabilitation. Accessed 1 Nov. 2014

  • Prince M (2008) Bold feasibilities: a new policy social architecture for Canadians with disabilities. Accessed 25 April, 2013

  • Ruhm CJ (2004) Parental employment and child cognitive development. J Hum Resour 39(1):155–192

    Article  Google Scholar 

  • Human Resources and Social Development Canada (2009) Advancing the inclusion of people with disabilities 2009. Accessed 1 Nov. 2014

  • Todd PE, Wolpin KI (2007) The production of cognitive achievement in children: home, school, and racial test score gaps. J Hum Capital 1(1):91–136

    Article  Google Scholar 

  • Woolley F (2004) Why Pay child benefits to mothers? Can Public Pol 30:47–69

    Article  Google Scholar 

Download references


We gratefully acknowledge many insightful comments and suggestions by the anonymous reviewers, and scholarship support from Dalhousie University and the Canadian Labour Market and Skills Researcher Network (CLSRN). NSLCY data were accessed through the Atlantic Research Data Centre; we thank Heather Hobson for vetting our output.

Conflict of interest

The authors declare that they have no conflict of interest.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Kelly Chen.

Additional information

Responsible editor: Erdal Tekin



Table 13 Disability designations in ten disability benefit programs
Table 14 Number of recipients, cases, and disability benefit level (Ontario Disability Support Program)
Table 15 Associations between parental disability and child outcomes by functional domains

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chen, K., Osberg, L. & Phipps, S. Inter-generational effects of disability benefits: evidence from Canadian social assistance programs. J Popul Econ 28, 873–910 (2015).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:


  • Disability benefits
  • Child well-being
  • Welfare
  • Inter-generational transmission

JEL codes

  • J1
  • J6
  • I38