Olga Archangelidi1, Emmanouil Mentzakis2. 1. National Heart and Lung Institute, Imperial College London, London, UK. 2. Economics Department, University of Southampton, Southampton, UK.
Abstract
Background: While the consequences of body weight for physical health are well explored, the evidence for psychological well-being is less straightforward. An instrumental variable approach is used to address the endogenous relationship between body weight and well-being in the UK general population. Methods: Data from the Health Survey for England (2003, 2004 and 2006) are used to fit linear and ordered probit instrument variable models for a sample of 13 862 individuals, with frequent white meat consumption instrumenting for body-weight. Non-linearities in the relationship, robustness to weak instruments and relaxation of strict exogeneity assumption are further examined. Results: Accounting for endogeneity and conditional on health a protective effect on well-being is observed. A unit increase in body mass index (BMI) improves General Health Questionnaire (GHQ) by 0.17 (95% CI: 0.02-0.31) points and reduces the probability of reporting very low GHQ by 2.5% (95% CI: 0.01-0.05). Empirical testing showed that the instrument performs well, with increased meat consumption adding 0.58 points (95% CI: 0.42-0.74) to ones' BMI. Conclusions: We present support for the jolly-fat hypothesis, however, caution is recommended in drawing inferences. Further research needs to resolve the mixed findings in the literature.
Background: While the consequences of body weight for physical health are well explored, the evidence for psychological well-being is less straightforward. An instrumental variable approach is used to address the endogenous relationship between body weight and well-being in the UK general population. Methods: Data from the Health Survey for England (2003, 2004 and 2006) are used to fit linear and ordered probit instrument variable models for a sample of 13 862 individuals, with frequent white meat consumption instrumenting for body-weight. Non-linearities in the relationship, robustness to weak instruments and relaxation of strict exogeneity assumption are further examined. Results: Accounting for endogeneity and conditional on health a protective effect on well-being is observed. A unit increase in body mass index (BMI) improves General Health Questionnaire (GHQ) by 0.17 (95% CI: 0.02-0.31) points and reduces the probability of reporting very low GHQ by 2.5% (95% CI: 0.01-0.05). Empirical testing showed that the instrument performs well, with increased meat consumption adding 0.58 points (95% CI: 0.42-0.74) to ones' BMI. Conclusions: We present support for the jolly-fat hypothesis, however, caution is recommended in drawing inferences. Further research needs to resolve the mixed findings in the literature.