It has already been documented that women with a higher Body Mass Index (women who weigh more) make less money and have bleaker marriage prospects. In GENDER, BODY MASS AND ECONOMIC STATUS, Dalton Conley and Rebecca Glauber verify this correlation and tackle the question of causality. Does being overweight cause women to have worse jobs/marriage prospects or does having worse jobs/marriage prospects cause women to become overweight? They study for women and men the relation of BMI to income, spouse’s income, rate of marriage, rate of divorce, and job prestige. For women, the negative impact of increased BMI is significant and appears to partly cause these undesired effects but alone cannot account for them. For men, BMI does not negatively impact these variables. In fact, the only significant effect of men’s body mass appears to be a significant positive association with their wife’s earnings.
Dalton Conley and Rebecca Glauber believe that their work is unique because it is one of the first articles that tackles the question of causality. In theory, having a low quality job or a poorer quality marriage could make someone depressed, and this depressed person could eat more (and become overweight) to ease their depression. With this logic, higher BMIs are correlated with these undesired qualities but it is not the cause of them. Another alternative is that overweight people are less attractive and therefore are less successful in choosing and keeping a spouse. Also, these overweight people may be less effective at their jobs because of health issues or can only find worse jobs because of simple discrimination. In this reasoning, the higher BMI is the cause of these deleterious effects.
The authors used data from the Panel Study of Income Dynamics (PSID). The PSID obtained respondents’ height and weight information during 1986 and then in 1999 and 2001. BMI was compared to three economic status outcomes—occupational prestige, labor earnings, and total family income. The authors also examined marital status outcomes (incidence of marriage and divorce). Their model controlled for years of schooling, age, parental status in 1986, and children.
To address causality, the authors compare siblings of different BMIs and also compare estimates for individuals younger than age 35 and individuals age 35 and older. Siblings provide a useful measure because they have the most identical backgrounds. Therefore, on average, they should have the most similar outcomes in life. For age differences, the authors theorize that individual socioeconomic outcomes are largely stable by age 35. With this reasoning, if differences of body mass after 35 affect socioeconomic outcomes, then it suggests that perhaps it is socioeconomic status causing weight gain in their model. Weight gain is correlated but not the cause of differences in status. If differences of BMI have a greater effect on the on socioeconomic and marital status outcomes of people under 35 before 1986 instead, then this would support the idea that the increased weight is the cause of these undesired effects.
For the results, the authors compare the BMIs of the participants in 1986 to their socioeconomic outcomes in 1999 or 2001. The authors find that for younger and older women a one percent increase in a woman's body mass resulted in a 0.6% decrease in her family income and 0.4% decrease in her occupational prestige 13 to 15 years later. Similarly, increased BMI was associated with decreased likelihood of marriage, her spouse's occupational prestige, and her spouse's earnings. For women married in 1986, a one percentage increase in the BMI in 1986 led to a 0.284% decrease in their husband’s current occupational prestige and a 1.085% decrease in current earnings. The authors also ran random effects models with a race term included and discovered that their coefficients did not significantly change.
The authors found evidence in support of a causal effect of body mass on socioeconomic and marital status outcomes as body mass had more of an effect on younger respondents’ outcomes, especially younger respondents’ occupations. One example of this difference was the effect of BMI on the job prestige between women who are over 35 and under 35 in 1986. For women over 35, the difference was relatively small. For women under 35 the effect was huge: a 1% increase in BMI decreased current occupational prestige by .395%. The authors discovered a rather small correlation in BMI for a woman her sister and a man and his brother. The correlations were 0.332 and 0.371, respectively. Using sister/sister and brother/brother comparisons, the idea that increased BMI hurts socioeconomic and marital status outcomes was supported.
For men, increases in BMI do not reduce their earnings, job prestige or likelihood of marriage. It also does not increase their likelihood of divorce, separation, or widowhood. A one percent increase in husband’s 1986 BMI is positively associated with their wife’s labor market earnings by .628 percent. Because there was no correlation, there also was no significant difference in these variables between men younger than 35 and men older than 35.
The principal flaw of this paper is its use of the BMI. BMI only accounts for weight relative to height. Muscularity and body frame do not enter the equation. The result of these omissions is that somebody like Michael Jordan, who is very lean but quite muscular, has a BMI that is in the overweight range. Of course, the idea that Michael Jordan is overweight is ridiculous. Body fat percentage is a much more accurate statistic. It is difficult and expensive to measure and is likely undoable for this paper.
Although women have different levels of muscularity, the difference is much greater in men. Most professional male athletes and many men who work out and put on some muscles have higher BMIs even though their body fat is low. Therefore, the BMI is less accurate for men than women. Given this, it is not surprising that there is no correlation between marriage prospects and also income/job prestige for men compared to their BMI and there is one between these two variables and BMI for women. In fact, the authors find that men with higher BMIs have spouses who make slightly more. All things being equal, more muscular men can be more selective choosing their wife. These more selective wives will earn more money on average. Also, for women that are muscular enough that it affects their BMI, this is not deemed attractive. Therefore, they cannot be as selective in the spouse they choose. The fact that the BMI can produce poor results is common knowledge. I did not see this idea mentioned in the paper once, which surprises me given the gravity of its effect.
Also, this study was very confusing. The data was not well discussed in the paper, and often I could not determine whether the group being discussed was over or under 35. I found myself relying two much on the appendices because the information was so obfuscated in the actual paper. I am not sure how I feel about the separation being over or under 35 also. I wish the authors could have studied specific professions because I feel like there were so many other variables at work that could be explanatory. The authors do address this by saying that years of education does not separate different types of training, which clearly has a powerful effect on income. This additional data is probably not available so I understand why the authors did not include it. They may have done a good job with the limited data, but I feel the argument lacks power.