An empirical analysis on the determinants of overweight and obesity in China

ABSTRACT

Overweight and obesity in adult populations is considered to be a growing epidemic worldwide, and appears to be rapidly increasing in China. From 1992 to 2002, the incidence of overweight in adults increased by 39.0%, while that of obesity doubled. To identify the determinants of adult overweight and obesity in China, micro-level data from a questionnaire survey entitled the ‘Preference Parameters Study,’ which was conducted by the Global Centers of Excellence programme at Osaka University, were analysed. In addition to the entire sample, data from urban and rural subsamples were also analysed in order to investigate whether the determinants of overweight and obesity differed. The results suggested that body mass index (BMI) is correlated with subjective well-being, gender, age, labour intensity and drinking and eating habits among urban respondents, and with age, monthly income, number of siblings and eating habits among rural respondents.

KEYWORDS:

• Body mass index (BMI)
• overweight and obesity
• urban residents
• rural residents
• China

JEL CLASSIFICATION:

• C21
• D12
• I12

Acknowledgements

We wish to thank the Global COE programme at Osaka University for providing the data we used in this study. Financial support for this study was provided by the Japanese Ministry of Education, Culture, Sports, Science and Technology through Grants-in-aid for Scientific Research (C) 25380234 and 15K03353. All views expressed in this article and any errors are the sole responsibility of the authors.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ Time preference refers to the consumer’s inclination towards current consumption over future consumption. Having a high (resp. low) time preference means that the consumer prefers current (resp. future) consumption to future (resp. current) consumption.

² The excludability requirement for the instruments employed means that these three variables affect someone’s BMI only through their effect on the overall subjective well-being. It is true that thinking about them as excludable might be a doubt because each of these variables is a specific dimension of overall well-being; however, the result that all these variables are not statistically significant when regressing them directly on BMI implies that this manipulation should not be a problem.

³ One anonymous referee points out that investigating the correlation among various characteristics is vital to understand their effects on BMI. Most of the correlation coefficients in absolute value are found to be less than 0.30; and only three (i.e. 0. between age and number of children, the one between marital status and number of children, and the one between monthly income and educational level) are between 0.30 and 0.55. We also run other regressions by using these highly correlated variables separately; however, we did not find large differences between these results and the current ones presented in Tables 3 and 4.

Additional information

Funding

Financial support for this study was provided by the Japanese Ministry of Education, Culture, Sports, Science and Technology through Grants-in-aid for Scientific Research (C) 25380234 and 15K03353.