Our register data for the birth cohorts of 1968 to 1970 showed that, overall, receiving disability benefits was significantly associated with reduced chances of having children. We also found a complex relationship receiving disability and having a first child. On one hand, getting disability at a very early age was associated with extremely low chances of having children over the life course, yet on the other hand, people that had a first child early (13─20 years of age) were likely to receive disability benefits during their 30 s and 40 s. We found that men with disabilities were less likely to have children compared to women with disabilities and to men and women without disabilities.
Our finding that receiving disability benefits was associated with significant reduction in chances of having children supports findings from previous studies 15–17. The association between disability and low chances of having children could very well be explained by several factors including economic barriers as Oppenheimer’s theory suggests 6–8. In addition, assortative mating 10, biological barriers and negative societal attitude could negatively influence the likelihood of becoming a parent for people with disabilities 20–23.
Early use of disability benefits was associated with low chances of becoming parents. An early start on disability benefits might contribute to childlessness as it implies prolonged ill health, and socio-economic constraints, factors that influence parental status. Moreover, the majority of individuals who start on disability benefits continue to receive them on a long-term basis, with an outflow rate at just about three percent 30. We also noted that becoming a parent very early was strongly associated with later use of disability benefits. Mental health problems were shown to independently predict early parenthood in Sweden 32. Mental health problems constitute a common diagnostic group for the award of disability benefits among young people 33–34. It is possible that early parenthood increases stress which leads to chronic mental health problems. Other research have reported that young parents tend to come from socially disadvantaged background 35. A combination of mental health problems and prolonged socioeconomic disadvantage might increase the odds that those who have a first child early end up using disability benefits later in life.
Surprisingly, we observed that once women started to receive disability benefits their chances of having children increased, as represented by a diagonal pattern observed in Fig. 4. This might imply that for women, the period before getting a diagnosis is characterised with even more financial uncertainty leading to postponement of child bearing. However, starting to receive disability benefits might offer some financial security making it possible for women to consider having a first child. This finding is reflected in the reports that show women in the Nordic countries postponing childbearing until they were firmly established in the labour market and could be assured of financial security via parental leave insurance 36–37.
Our study also identified major sex differences in the chances of having children. These chances were much lower for men with disabilities compared to women with disabilities or to men and women without disabilities. Literature suggest that entry to a stable partnership union requires a strong financial underpinning and more occupational stability for men compared to women 6–7. Limited male entry into partnership formation will subsequently result in low number of children for men. However, it is also important to note that our study population consisted of a higher proportion of men than women, implying a surplus of men on the partnership market, which might further constrain the chances of men with disabilities.
We noted that the odds of having children reduced in the younger birth cohort of 1969 and 1970 compared to people born in 1968. We find no known social economic changes during this time to explain these parental status differences by birth cohort. It is therefore possible that this association is due to random variation.
Strengths And Limitations Of The Study
The main strength of this study is the access to a large longitudinal data set that gave us an opportunity to observe a total population for a long duration. Moreover, using register data ensured minimal loss to follow-up, as data on the use of disability benefits and having children was available for all cohorts. Even though our models excluded some people due to missing data, there is no reason to suspect that missing data differed by disability status.
We measured disability based on different disability benefits. However, it is possible that the diagnoses leading to the award of sickness compensation is different from the diagnoses leading to activity compensation. We speculate that such differences could have implications on the possibilities of family formation and, subsequently, parental status. For example, having a chronic developmental disability as opposed to having a musculoskeletal problem could have different social and cognitive effects on average capabilities, and as such, might impact entry into partnership and, subsequently, having children.
Though we had access to longitudinal data, we chose an analytical approach that offers more flexibility to measure the complex interplay between disability and parental status. We chose our regression methods based on the study aim and nature of the variable “receiving disability benefits.” We considered the following: 1) Receiving disability benefits can be a process that takes several years so the exact date of when one gets a disability cannot be determined, 2) the probability of receiving disability benefits is highly dependent on the individual’s own socioeconomic background with largely unmeasured factors, and 3) disability benefits recipients might have a wide spectrum of health circumstances that greatly varies in frequency depending on the age of the individuals, 4) we could also not rule out the fact that parental status might influence the probability of getting disability benefits. Many longitudinal statistical methods, such as survival analysis, which could be a potential candidate here require that the “states” (here disabled/non-disabled) are well-defined both in character and timing, which is not the case here. Based on the above reasons, we preferred using descriptive and regression methods that measure the association, rather than attempting to establish causality between disability and parental status.