Overview

Abstract (English)

Since the federal government introduced the first family allowance program in 1945, income transfers to families with children have been a basic social policy in Canada. Over the years, both Ottawa and the provinces have spent important sums of money in these kinds of programs. In 2011 only, federal transfers to families amounted to more than 12 billion dollars. Though prompted by a range of goals, a central expectation motivating these kinds of transfers is that higher family income will lead to better child outcomes, especially in low-income families. Using Canadian evidence, including from longitudinal surveys, this synthesis critically examines this assumption. It provides a comprehensive but critical overview of this body of literature and draws implications for policy and future research: What can we really expect from public policies, such as child benefits, that increase household income? Do they really improve child outcomes? Overall, we identified 34 relevant Canadian studies, five of which used research designs or techniques to identify the causal impact of income. The paper also discusses key theoretical and methodological issues in this area of research. In essence, what does this research tell us? All other things being equal, for most measures, it suggests that higher income improves a range of child outcomes (cognitive, behavioural, social and emotional) but that the improvement is small in magnitude. Much of the correlation between lower income and worse child outcomes is explained by other factors that often go along with low income; income itself has a relatively small influence on child outcomes when researchers account for other factors. The evidence thus suggests that the causal effect of income on child outcomes is, in and of itself, quite small. From a policy perspective, this finding suggests that we cannot expect income transfers to low-income families to vastly improve child outcomes, and that focusing exclusively on income to close the gap between children from low-and higher-income families, via transfers or otherwise, is unlikely to be effective. While our focus is on research attempting to identify the role of income in causing child outcomes, we also included non-causal research: these studies offer a more nuanced portrait and provides insight into the nature of the relationship between income and child outcomes, aside from statistical significance and magnitude of impact. In terms of public policy, these studies suggest that income transfers could be more effective if they were targeted: Income’s effect on child outcomes is non-linear – an additional dollar of income has a larger effect on lower-income children – If the goal is to improve child outcomes, families with lower income should be given significantly more benefits than those with more income. Income’s effect on child outcomes displays diminishing marginal returns – each additional dollar of income has a smaller impact on child outcomes than the previous dollar. The evidence suggests that income transfers will have no significant effect on child outcomes beyond a certain threshold (C$60,000 year). Currently, many child benefits are designed progressively and decrease as income increases, but many of these extend well past this threshold. Income matters more in the early years of life. Since early outcomes persist through childhood, child benefit policies should be differentiated based on age, with higher transfers for younger children. Income also affects the movement between positive and negative trajectories: poor outcomes are more persistent and have more deleterious long-term effects for lower-income children. Middle- and high-income children are not immune from poor outcomes, but they are much more likely to be on a trajectory toward positive outcomes than lower-income children. Canadian research on the effect of income on child outcomes has come a long way in the past 20 years. Better data, better data availability, and advancing techniques have allowed for more sophisticated research. However, untangling the complex relationship between income and child outcomes remains a challenge and requires better data and more analysis. Notably, we need to do more to address the following knowledge gaps: Living in low income for a longer period of time is more strongly associated with worse child outcomes. We need to further explore how longer periods of low-income and how income dynamics throughout childhood affect children. Research so far has mostly focused on early childhood; we need to also examine older children outcomes in middle school and high school. Research so far has mostly focused on cognitive outcomes; we need to also collect data and analyze a wider range of non-cognitive outcomes, such as mental health and self-control. Some studies suggest that income has a different effect on some subsets of the population: for instance, income’s effect is not gender-neutral; there are likely other caveats to income’s effect when looking at other subsets of the population. Unfortunately, there is no dataset currently following Canadian children through from birth to adulthood. The National Longitudinal Survey of Children and Youth (NLSCY) – which followed the same children from 1994 to 2008 and proved essential to examine child development in Canada as this synthesis illustrates – is now inactive. While useful as part of a broader evidence base, findings from other national contexts can only be used with caution in the Canadian context given the importance of domestic institutions and policy arrangements such as public health care and public education; foreign experience is at best a last resort option. If informing Canadian policy to help improve child outcomes and support intergenerational mobility is a key goal, then analyzing the experience of children within Canada must be a priority. And to do that, developing a high-quality national longitudinal survey that will follow children well into adulthood must be a priority.

Abstract (French)

Please note that abstracts only appear in the language of the publication and might not have a translation.