Abstract
Body mass index (BMI) is an imperfect measure of body–fat and recent studies provides evidence in favour of replacing BMI with waist circumference (WC). Hence, I investigate how WC and BMI are associated with employment status in England. I use fifteen rounds of the Health Survey for England (1998-2013), which has measures of employment status in addition to measured height, weight and WC. WC and BMI were entered as continuous variables and obesity as binary variables defined using both WC and BMI. I used multivariate models controlling for a set of covariates. I found significant and negative associations between both WC and BMI with employment in women. In men, the coefficient of WC was significant, but not BMI. In both men and women my findings suggest that the association of WC with employment was of greater magnitude, than the association between BMI and employment. To explore whether any difference in the association was mitigated by conventional instrumental variables methods I reran the analysis using the “workhorse” in analyses of BMI and employment. I.e. instrumenting for WC and BMI based on WC and BMI of a biological family member. Significant endogeneity was detected for both WC and BMI. The finding, that the impact of WC on employment was larger than BMI, did not persist in the IV-models. Although the IV models found significant impacts of obesity on employment, they were not more pronounced when WC was used to measure obesity, compare to BMI.
Keywords: waist circumference; body mass index; obesity; employment status
JEL classifications: I00, I12, J24
ISBN 978-82-7756-253-7