Unhappy and Happy Obesity: A Comparative Study on the United States and China

There are studies demonstrating that the impact exerted on SWB varies by culture and by social institutions. Studies that examine the relationship between having children and SWB may be considered as examples of this aspect (Aassve et al. 2012; Margolis and Myrskylä 2011). These research initiatives indicate that the effect of having children on SWB can be positive or negative according to culture and social institution to which the individuals belong.

Buss (2014) pointed out several limitations of BMI. For example, it is difficult for BMI to make a distinction between lean and fat mass. In addition, BMI also has difficulty in identifying fat distribution. The other indicators to measure individual fatness, such as weight-to-height ratio, have begun to be used to deal with these issues. However, this study uses BMI as the measure of individual fatness because of the two reasons. First, as BMI has been widely used in many studies, it is easy to compare the results obtained in this study and those in previous studies. Second, the data used in this study did not survey alternative anthropometric measurements other than BMI.

Concerning the IV, the average fatness of biological relatives and the respondent’s height that were used by Bargain and Zeidan (2019) and by Katsaiti (2012) are promising candidates. Although the heights of the respondents are available in the data used for this study, the F-statistic on the height in the first stage estimation of the IV is under 10 for both China and the United States; thus, it may be considered to suffer from the problem presented by weak instruments. For this reason, the use of IV estimation was reconsidered. Finding a better IV remains an incomplete task for this study.

The yearly family income in China is calculated in dollars. The exchange rate is 1 yuan = $0.14 (this rate is based on data as on October 4, 2019). The average yearly family income is $11,067 in China and $69,143 in the United States, indicating that family income is higher in the United States than in China. Log family income is also higher in the United States than in China.

The health-related variables used in the analysis are mainly subjective indicators since the data did not survey the objective health measures. This is a limitation of this study, and future studies will need to be re-examined using other data.

Estimated results by pooled OLS are also presented in the “Appendix”.

The data of China annually surveyed the three types of consumption expenditures, namely, foods, durable goods, and other than these. Consumption expenditures of foods and others were surveyed for monthly averages. However, since consumption expenditure of durable goods was a 1 year value, it was divided by 12 to calculate the value for 1 month. In the analysis, the total of these three types of consumption expenditures is used.

In this analysis, the number of household members is not used as an independent variable to avoid multicollinearity with the consumption expenditures, which are divided by the square root of the number of household members.

Concerning cultural backgrounds, diversity in race and ethnicity in the United States can cause the varying effect of obesity on happiness. The ratio of obesity among adults aged over 20 is higher at non-Hispanic Blacks than at non-Hispanic Whites and Hispanics (Craig et al. 2020). These differences in the prevalence of obesity can bring divergent perceptions of obesity, resulting in the varying effect of obesity by race and ethnicity on happiness. Separate analyses by race and ethnicity are conducted to check this point. We divided the sample of the United States by non-Hispanic Whites, non-Hispanic Blacks, and Hispanics and analyzed the results by using fixed effect OLS. The estimated results of the non-Hispanic Whites show that coefficients of being underweight are positively significant, especially for women. This result is a similar pattern with the result of the undivided sample. This might be because non-Hispanic Whites make up the majority of the sample. Estimated results of the non-Hispanic Blacks show that coefficients of being obese are negatively significant, particularly for women. This result is somewhat surprising as some previous studies indicate that non-Hispanic Blacks are more tolerant of obesity than non-Hispanic Whites (Flynn and Fitzgibbon 1998; Shafer 2010). Regarding the results of Hispanic, none of the coefficients of obesity are statistically significant. These estimated results show that the effect of obesity on happiness is divergent by race and ethnicity. Particularly, the negative association between being obese and happiness at non-Hispanic Blacks is interesting. Although it might appear puzzling to have such a result, an explanation may be found by looking at the association with marital status. Body weight is vital in finding a good partner in the marriage market, and if the sample size of unmarried people is large, the adverse effects of obesity can be significant. The analysis that split the sample by race, ethnicity, and marital status is conducted to check this point. The estimated results of the non-Hispanic Whites show that the coefficient of being underweight is positively significant only at the unmarried samples. This result indicates that being underweight is valuable, especially for a single person who is searching for a partner in the marriage market. Estimated results of the non-Hispanic Blacks show that coefficients of being obese are negatively significant, particularly for unmarried samples. This result indicates that unmarried individuals who are obese are less happy, and it can be caused by the relation that heavy body weight is stigmatized in the Western standard of beauty and prevents finding a good partner in the marriage market. Moreover, in the case of non-Hispanic Blacks, the ratio of unmarried samples is 60.6%, and it is higher than non-Hispanic Whites (36.7%) and Hispanics (41.7%). These results suggest that a relatively large unmarried sample may be responsible for the negative relationship between obesity and happiness. Regarding the results of Hispanics, none of the coefficients of obesity are statistically significant.