Social determination, health selection or indirect selection? Examining the causal directions between socioeconomic status and obesity in the Chinese adult population.

Abstract

Using nationally representative longitudinal data from 2010 to 2016 in China, this paper aims to systematically investigate the dynamic relationship between socioeconomic status (SES) and obesity in the Chinese adult population. We apply maximum likelihood estimations of dynamic panel models with fixed effects to evaluate the causal directions between SES and obesity, thus providing empirical evidence of the relative importance of the "social determination" hypothesis (SES determines obesity), the "health selection" hypothesis (obesity influences SES) and the "indirect selection" hypothesis (some third factors cause obesity and SES). All analyses were conducted for women and men separately. Our analyses mostly support the contention that after two years, there are no significant causal effects of either lagged SES on Body Mass Index (BMI) or BMI on SES, after controlling for time-invariant unobserved confounders. While there is weak evidence of social determination among men, non-manual workers tend to have a higher BMI than those who are unemployed. We conclude that the indirect selection hypothesis may be the most important explanation in understanding the relationship between SES and obesity in China.