Obesity and employment: Causal evidence from Genetic Instruments in the UK Biobank

This study explores the causal impact of obesity on employment using data from the UK Biobank, a comprehensive dataset combining genetic, health, and socioeconomic information. We apply a Mendelian Randomisation (MR) approach, using a genetic risk score for a higher body mass index (BMI) as an instrumental variable. This method leverages genetic predispositions determined at conception, enabling us to isolate the exogenous variation in obesity while addressing key endogeneity concerns, such as omitted variable bias and reverse causation. To account for the heterogeneity in obesity's impact across diverse demographic and socioeconomic factors, we apply a combination of instrumental variable techniques and the person-centred treatment effects (PeT) framework. This methodological approach extends beyond traditional average treatment effect estimations, uncovering nuanced individual-specific impacts that reflect the complex relationship between obesity and employment probabilities. Our findings reveal a significant and negative causal effect of obesity on employment across all estimation methods. Nonlinear approaches, including marginal treatment and PeT effects, yield the same substantive results with differences in marginal effects. Gender-specific subsample analyses indicate that the employment penalty associated with obesity is more pronounced among females in the linear models, whereas the PeT effects approach reveals the opposite pattern. Overall, these results highlight the wider economic burden of obesity, particularly its adverse impact on employment.