HEALTH ECONOMICS Health Econ. 16: 145–162 (2007) Published online 6 October 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hec.1142 FAMILY INCOME, PARENTING STYLES AND CHILD BEHAVIOURAL–EMOTIONAL OUTCOMES MARTIN DOOLEY a and JENNIFER STEWART b, * a Department of Economics, McMaster University, Canada b School of Public Policy and Administration, Carleton University, Canada SUMMARY A positive relationship between income and child outcomes has been observed in data from numerous countries. A key question concerns the extent to which this association represents a causal relationship as opposed to unobserved heterogeneity. We use data from the National Longitudinal Survey of Children and Youth to implement a series of empirical strategies for estimating the existence and size of the effect of income on behavioural–emotional outcomes. We also examine the role of parenting style. Our results indicate that there is little evidence of an effect of income on behavioural–emotional scores. The exclusion of parenting style from the models was found to not bias the estimated income effect, but parenting style was found to have a consistent impact on child outcomes. Copyright # 2006 John Wiley & Sons, Ltd. Received 21 December 2004; Accepted 4 May 2006 JEL classification: I10 KEY WORDS: child development; income INTRODUCTION A positive correlation between income and child outcomes has been observed in data from numerous countries and surveys. Canadian data also indicate that children from lower income families have worse behavioural, emotional and cognitive outcomes than do children from higher income families (Dooley et al., 1998; Phipps, 1999). Governments have used this observation as one of the motivations for income transfers to families with children. For income transfers to be effective, this empirical link must represent a causal relationship. Specifically, can governments expect to improve child behavioural and emotional outcomes substantially by increasing cash transfers to low income families? Or should the government focus on the provision of services, such as early childhood education or parenting training? It is difficult to conclude that a causal relationship exists between household income and child outcomes using survey data. Unmeasured variables, such as parental health and skill may be correlated with both child outcomes and income. Estimated effects of income on child outcomes will be biased if methods are not employed to account for these unobserved variables. One solution for heterogeneity is to use instrumental variables, but this approach is not commonly encountered in the child outcome literature due to the scarcity of good instruments in most surveys. We believe there does not exist one clearly superior and feasible solution for heterogeneity given the available data. In this paper, we first provide estimates of a conventional model of the effects of income and other socio-economic variables *Correspondence to: School of Public Policy and Administration, Carleton University, 1125 Colonel By Drive, Ottawa, ON K1S 5B6, Canada. E-mail: jennifer stewart@carleton.ca Copyright # 2006 John Wiley & Sons, Ltd.