Causal analysis of the effect of weight status on education and wage outcomes

Abstract

While the study of the effects of weight status on socioeconomic indicators has been an active research area, most of the evidence is based on the US population, with little understanding of the mechanisms through which weight status affects the hourly wage in the UK. As such, this thesis aims to: i) synthesise the available evidence on the relationship between weight status and education, occupation, and hourly wage in high income countries; ii) examine the long-term effects of weight status on educational outcomes; and iii) estimate the pathways through which weight status affects hourly wage. Questions (ii) and (iii) are answered empirically using two UK representative datasets: the 1958 National Child Development Study (NCDS) and the 1970 British Cohort Study (BCS). In Chapter 1, background information on overweight and the specific aims and objectives as well as the significance of this thesis are presented. Chapter 2 presents the economic theories that attempt to explicate the link between weight status, education, and hourly wage. Chapter 3 systematically reviews 54 studies on the effect of weight status on hourly wage (N=27), employment (N=21), annual earnings (N=14) and other related outcomes (N=5). A total of 37 studies were summarised in a narrative synthesis and 17 studies were analysed in a meta-analysis. The results from the meta-analyses indicate that BMI and wages are inversely related among White women, Black women, and Black men, while for White men the results were not statistically significant. The results for employment status suggest that women with obesity are less likely to be employed than women without obesity while the pooled effects for men indicate no evidence of any effect. There was no evidence of any impact of BMI on log annual earnings as well. The narrative synthesis indicates that the impact of weight status on employment outcomes differs according to the geographical region, estimation method, and study population. A description and a critique of the econometric methods that allow to estimate the causal effects of weight status on employment related outcomes in the previous chapter is presented in Chapter 4. Chapter 5 presents a literature review on the relationship between weight status and educational outcomes, and then, in Chapter 6, using OLS, multinomial logit and Two-Stage Least Squares (2SLS) regressions, analyses for the educational outcomes using the NCDS, and the BCS datasets are presented. The findings from this analysis indicates that there is a positive relationship between overweight and test scores at ages 7, 11, and 16 among females while a negative effect was obtained for males. Furthermore, females with overweight at age 16 have reduced likelihood of obtaining a degree or a higher qualification compared to non-overweight counterparts. Chapter 7 assesses the effects of weight status the role of education and occupation as potential mediators in the weight status and hourly wage relationship using the BCS dataset. The estimation of the multiple mediator model is carried out using the inverse odds weighting (IOW) estimation. The mediation analyses show that although there is no direct or indirect effect of weight status at age 10 on hourly wages at age 38; the effect of weight status at age 26 has both direct and indirect effects on wage. In Chapter 8, the discussion and conclusions from this study are presented.