Childhood obesity and the income gradient: evidence from Australia

ABSTRACT

This article examines the dynamic nature of human capital formation in the context of childhood obesity and the association of household income and childhood obesity in Australia using the first five waves of the Longitudinal Study of Australian Children. Our results show a strong evidence of dynamic nature of child obesity: the lag obesity is a significant and robust predictor of obesity in the current period. We also found that the main channel for childhood obesity is inter-generational trait: the probability of obesity in children born to an obese mother or father is 15% higher than that of other children. Other important determinants are lifestyle factors, including the consumption of drinks with a high sugar content and the amount of time watching TV. Income becomes an insignificant determinant of childhood obesity once we control for unobserved individual heterogeneity.

KEYWORDS:

• Childhood obesity
• income gradient
• Australia

JEL CLASSIFICATION:

• I10
• I12

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ CI is defined as twice the area between the concentration curves and the 45° line and ranges from $-1$ to 1. A CI of zero (i.e. concentration curve coincides with the 45° line) reflects perfect equality; a negative CI suggests obesity is greater among the poor and a positive CI suggests the rich has higher rate of obesity.

² Since t parameters of probit/logit estimators with different covariates are not comparable (Norton 2012), we do not report the results of this basic specification (but they can be made available from the authors upon request). We focus instead on the significance of income parameters to show that this simple specification can overestimate the effects of income. We thank an anonymous reviewer for reminding us of this issue.

³ We do not present the result of this naive model for brevity and since comparison paramters of probit/logit estimators with different controls are not advisable (Norton 2012). But this result can be obtained from the authors upon request.

⁴ These 4-point answers are (1) Underweight, (2) Normal weight, (3) Somewhat overweight, and (4) Very overweight. We define self-reported obesity if children are in option 4.