VoxEU Column Health Economics

Obesity and wages: using body composition as an alternative to BMI

Most social science research on obesity, such as its impact on wages, uses the body mass index (BMI). BMI is imperfect since it fails to distinguish fat from lean body mass. Here is evidence that both men’s and women’s wages are lowered by body fat when a direct measure is used instead of the BMI.

In the US, the proportion of the adult population which is obese has risen from 15% in the mid-1970s to around 33% in the early 2000s.1 This dramatic increase in the prevalence of obesity is causing tremendous concern among public health officials because of the well-known links between obesity and overweight and the risk of developing health problems such as hypertension, dyslipidemia, type 2 diabetes, coronary heart disease, strokes, and some cancers. Health problems associated with overweight and obesity also impose a substantial economic burden on the US health care system, which have been estimated at about 117 billion dollars in 2000.2

In addition to the adverse health consequences, obesity is also known to be associated with negative social and economic outcomes, such as social stigmatisation, discrimination, lowered self-esteem, and marriage problems. Recently, these social and health consequences have motivated economists to investigate the potential relationship between obesity and labour market outcomes, such as wages. These economic studies usually point to a negative association between obesity and wages for white females, but no clear evidence of a wage penalty has emerged for males or other female groups from these studies. One common characteristic of the economic research on obesity is that all studies use a measure of obesity that is based on Body Mass Index (BMI), which is defined as the ratio of weight in kilograms and height in meters squared. The World Health Organization sets the universally accepted cut-off points for classification of overweight and obesity as having a BMI over 25 and 30, respectively.

While BMI is the widely accepted measure of obesity by social scientists, it is increasingly acknowledged by clinical researchers that it is at best an imperfect measure of obesity (or excessive body fatness) because it does not distinguish body fat from lean body mass. A systematic review of the medical literature on the association between BMI-based measures of obesity and total mortality for patients with coronary artery disease between 1996 and 2005 suggests that overweight patients actually have a better survival rate and fewer cardiovascular events than underweight or obese patients.3

This phenomenon is known as the obesity-paradox. An increasing number of studies point to the inability of BMI to properly distinguish between body fat and lean body mass as a possible explanation for the obesity paradox and stress the need for developing alternative measures of obesity, which would better characterise individuals who truly have excess body fat. Another study reports that BMI cut-off points for overweight and obesity may not represent the same levels of body fat in various ethnic groups due to differences in body build, fat patterning, and muscularity that alters the relationship between BMI and body fat.4 Some studies conclude that the reliability of BMI as a tool for measuring body fat is questionable, and that direct measurements of body fat would provide a significant improvement towards detecting and diagnosing obesity in individuals.

Although relatively unknown to economists, body composition is a widely-used method for detecting body fat by epidemiologists, nutritionists, and physiologists for studying nutritional health, physical growth, and physical performance. In the two-compartment model of body composition, the total body weight is divided into body fat and fat-free mass. Body fat accounts for about 20% to 40% of body weight. It basically consists of adipose tissue whose main role is to store energy in the form of fat, while providing a measure of insulation. Sometimes referred to as lean body mass, fat-free mass is the larger component that includes everything else.

Models based on body composition have several advantages over models based on BMI. First, body composition better reflects the biological condition of human body in which, as documented by the medical literature, body fat is responsible for inferior health outcomes, while fat-free mass is closely associated with improved health. Unlike BMI, whose marginal effect is not subject to a direct interpretation, the marginal effect of body fat or fat-free mass has a biological meaning that can be traced to a physical increase in one of the body components. Second, through their combined but opposite effects on health and physical performance, body fat and fat-free mass can exert a complex influence on the economic and social outcomes that cannot necessarily be captured by a measure that fails to distinguish one from the other. Because their expected effects are opposite from each other, there could be instances when body fat and fat-free mass cancel each other out. In such a scenario, a single index such as BMI will not be able to uncover the true effect of fatness on socio-economic outcomes.

Body fat and wages

One problem in estimating the effect of body fat and fat free mass on wages is the difficulty of finding data sets that contain both these measures and economic outcomes such as wages and employment. One way to solve this problem is to combine information from two data sets – one being rich in economic outcomes and the other one containing enough information to construct the body composition measures. In the United States, two such data sets are the National Longitudinal Survey of Youth (NLSY), which contains rich data on economic variables, and the National Health and Nutrition Examination Survey III, which contains data on body fat as measured by observed electrical resistance (so-called BIA readings, which are converted into measures of body fat and fat-free mass by entering it into one of the predetermined prediction equation developed by clinical scientists.).

In a recent paper, Roy Wada and I examine the effect of body composition on wages using measures of body fat and fat-free mass.5 The wage regressions on body composition measures strongly indicate that increased levels of body fat decreases wages of both males and females. The effects are very clear for white males, white females, black females, and to a lesser extent for Hispanic males and females. The effects of body composition on the wages of black males are usually much smaller than other groups in magnitude and they are not statistically significant.

These findings are in contrast to the previous studies that found strong evidence of a negative effect on white females but not for other population groups. We believe that these studies might have missed the effect on other groups, especially on males, possibly due to the inability of BMI to distinguish body fat from lean body mass. Given that a higher proportion of women’s bodies consists of fat than do men's bodies (due to demands for childbearing and other hormonal functions), BMI may serve as a better measure of excessive body fatness for women than men. Such gender-dependent correlation could particularly explain the previously mixed and unstable findings for men. The regressions also indicate that individuals with high levels of fat-free mass or lean body mass earn a wage premium. In summary, these results suggest that public health officials should pay particular attention to the opposing effects of body composition components on health and labour market outcomes when designing nutrition intervention programs aimed at reducing the incidence of obesity.

These results are not only a significant expansion on the economics of obesity literature, but they also expand our understanding of the role of non-cognitive factors on wage determination. We know in labour economics literature that most of the variation in wages across individuals remains unexplained even after extensive controls for human capital variables like education and experience. This has motivated economists to increasingly focus on the role of non-cognitive factors like beauty, leadership, height, and sociability on wage determination. Evidence discussed above also contributes to the development of a better understanding of wage determination.




1 Centers for Disease Control and Prevention (2007a). Overweight and Obesity. http://www.cdc.gov/nccdphp/dnpa/obesity/
2 Centers for Disease Control and Prevention (2007b. Diseases and Risk Factors. http://www.cdc.gov/steps/disease_risk/index.htm.
3 Romero-Corral, Abel, Virend K. Somers, Justo Sierra-Johnson, Michael D. Jensen, Randal J. Thomas, Ray W. Squires, Thomas G. Allison, Josef Korinek, and Francisco Lopez-Jimenez (2007). “Diagnostic Performance of Body Mass index to Detect Obesity with Coronary Artery Disease,” European Heart Journal, July 17.
4 Rush, Elaine, Lindsay Plank, Vishnu Chandu, Manaia Laulu, David Simmons, Boyd Swinburn, and Chittaranjan Yajnik (2004). “Body Size, Body composition, and Fat Distribution: A Comparison of Young New Zealand Men of European, Pacific Island, and Asian Indian Ethnicities,” The New Zealand Medical Journal, vol. 117, no. 1207, December, pp. 1-9.
5 Wada, Roy, and Erdal Tekin (2007). “Body Composition and Wages” National Bureau of Economic Research Working Paper No. 13595, November. See the paper for details on how the data sets are integrated.



210 Reads