Identifying the Effects of Income on Children's Development Using Experimental Data

Prior research suggests that poverty can be detrimental to low-income children's development. Is this relation capturing the effects of poverty or the effects of other characteristics of low-income families associated with poverty? Can low-income children benefit from increases in income? In this paper, an instrumental variables estimation strategy is used with data on nearly 900 children from a random assignment evaluation of a pilot welfare reform program in Minnesota in order to answer these questions and to identify the causal effects of income on children's development. There are some suggestions that increased income improves the development of low-income children, at least with regard to their school engagement and positive social behavior. Results are discussed with regard to their implication for analysis, as well as research and policy. Key Words: experimental data, income, instrumental variables estimation, poverty. Between 1975 and 1993 the percentage of children under age 6 living in families with incomes below the poverty line rose from 18% to 23%; although child poverty rates have declined by more than one fourth since 1993, these rates are still very high-nearly 16% of children lived in poverty in 2000 (Song & Lu, 2002). Despite considerable research documenting a negative relation between low income and child functioning, research on the causal and reversible effects of poverty on children is still inconclusive. In this paper, we utilize an analytic approach to estimate the causal effects of increases in income on children, yielding conclusions that can inform poverty research and public policy. Children in poverty are more likely to experience poor health, to score lower on standardized IQ and achievement tests, and to be retained in a grade or drop out of school (Haveman & Wolfe, 1994; Smith, Brooks-Gunn, & Klebanov, 1997). Despite considerable advances in research on the effects of poverty on children's development, two critical questions remain unanswered: Is poverty itself the cause of the greater risk observed among children in low-income families? And do interventions that aim to move families out of poverty benefit children? First, because parents in poor and nonpoor families differ in a variety of unmeasured characteristics in addition to their income level, the effects of poverty on children may reflect these unmeasured differences, rather than their income level. Second, research to date has not attempted to identify the effect of increases in income on child well-being. Even if poverty were causally related to children's outcomes, the relation between income and children's outcomes may not be reversible. That is, increasing income may not remove the previous disadvantage children may have experienced from poverty. Using data from the evaluation of the Minnesota Family Investment Program, a random assignment experiment that resulted in increases in employment and income among long-term welfare recipients, we examine whether increases in income resulting from the program had benefits to children's school performance and social behavior. In so doing, we extend prior work on the effects of poverty on children's development by using an analytic method that allows us to estimate the causal and reversible effects of income on preschool and early school age children in the context of an experimental antipoverty program. The effect of poverty on children in the extant research is typically estimated using techniques that cannot identify causal relations. The technique we employ in this paper combines the use of a random assignment design with an instrumental variables estimation strategy. The key element that makes this technique especially successful is that random assignment of an individual or family to a treatment group or to a control group is uncorrelated with any observable or unobservable factors that may confound estimated effects on children's development. …