The Poor Are Not Fat

By Robin Hanson · August 4, 2011 8:15 am · Comments (42) · ·

In both the popular press and academic research, there is the argument that the growth of fast food and energy-dense food has been an important cause of the overweight epidemic in the U.S. and that this has disproportionately affected poor people. [Some] argue that limited economic resources may shift dietary choices toward a diet that provides maximum calories at the least cost. An implication of this line of research is that the poor cannot afford healthy diets. (more; ungated)

In fact, however, the poor are on average not fatter than the rich:

Contrary to conventional wisdom, … the poor have never had a statistically significant higher prevalence of overweight status at any time in the last 35 years. Despite this empirical evidence, the view that the poor are less healthy in terms of excess accumulation of fat persists.

A new paper manages to find a relation between poverty and fat – both the very fattest and the very thinnest people tend to be poor:

Distribution-sensitive measures of overweight … [show] that the severity of overweight has been higher for the poor than the nonpoor throughout the last 35 years. … The strongest relationship between income and BMI is observed at the tails of the distribution. … For those at the tails of the BMI distribution, increases in income are correlated with healthier BMI values.

OK, but this hardly supports a general story that the poor can’t afford healthy diets. It fits much better with there being a small fraction of “broken” folks who tend both to have low abilities to earn money, and to have an unhealthy weight, both too high and too low.

GD Star Rating
loading...
Tagged as: ,
Trackback URL:

8 Trackbacks

  1. [...] This surprised me big time. There is no correlation between being overweight and income. [...]

    By Browsing Catharsis – 08.05.11 « Increasing Marginal Utility August 5, 2011 at 8:02 am

  2. [...] No. HT to Robin Hanson: [...]

    By Are the Poor Getting Fatter than Everyone Else? | John Goodman's Health Policy Blog | NCPA.org August 5, 2011 at 3:27 pm

  3. [...] No. HT to Robin Hanson: [...]

    By Are the Poor Getting Fatter than Everyone Else? « Daniel J. Smith August 5, 2011 at 4:35 pm

  4. [...] Study. HT: Tyler via Robin Hanson. [...]

    By People Judge Future Actions More Harshly Than Past Actions | John Goodman's Health Policy Blog | NCPA.org August 17, 2011 at 12:28 pm

  5. [...] the poor more likely to be fat? Robin Hanson argues against conventional [...]

    By I Think I Believe » Links for the weekend August 19, 2011 at 7:19 pm

  6. [...] evidence, the view that the poor are less healthy in terms of excess accumulation of fat persists. (source) 51.053562 3.720867 Share:ShareFacebookTwitterStumbleUponPrintEmail This entry was posted in [...]

    By Human Rights Facts (158): Poverty and Obesity, Ctd. | P.a.p.-Blog | Human Rights Etc. August 20, 2011 at 8:23 am

  7. [...] while the relationship between income and obesity is perhaps more complicated than we've been led to believe, it's still true that metabolic diseases like diabetes and [...]

    By You shouldn’t be able to buy soda with food stamps | Grist August 23, 2011 at 7:32 pm

  8. [...] evidence, the view that the poor are less healthy in terms of excess accumulation of fat persists. (source) 51.053562 3.720867 Share:ShareFacebookTwitterTumblrStumbleUponRedditPrintEmail This entry was [...]

    By Human Rights Facts (32): Poverty and Obesity | P.a.p.-Blog | Human Rights Etc. December 19, 2011 at 9:53 am

34 Comments

  1. Sister Y
    August 4, 2011 at 11:54 am | Reply

    That’s amazing! Especially since obesity by BMI measure is associated with lower IQ, and poverty is also associated with lower IQ (still true, right?).

    By “broken,” do you mean schizophrenia, mental retardation, and things like that, or simply subclinical poor impulse control/inability to defer gratification, or something else?

  2. Low
    August 4, 2011 at 12:10 pm | Reply

    How does the paper define poor? Is it only focused on the US? (I’m not willing to pay $42 for the paper.)

    If not, since “both the very fattest and the very thinnest people tend to be poor”, another explanation could be:

    1. Rich – can afford “good” food.
    2. Poor – only affords “bad” food.
    3. Really poor – can’t afford enough food.

    • Low
      August 4, 2011 at 12:40 pm | Reply

      The critique is not against the conclusion, but that the data provided is not enough to draw the provided conclusion.

      Some argue that limited economic resources may shift dietary choices toward a diet that provides maximum calories at the least cost. An implication of this line of research is that the poor cannot afford healthy diets.

      If this is the hypothesis they want to try, it feels like the averaging they weight of poor people is a very awkward way to prove anything.

      Q: Is it possible to buy cheap and healthy food?
      Way to solve: Go to the supermarket and try.

      Q: Is the cheap, healthy food less tasty?
      Way to solve: Interview people. (eww…)

      People will probably try to maximize (Healthiness * Taste) / Price.

      Food: Healthy, Tasty, Cheap. Pick two.

      • Floccina
        August 7, 2011 at 10:52 pm |

        I agree that is way to address he question. Carrots and bananas are delicious and cheap. Cabbage, Collard and mustard greens are very cheap, I like them but not everyone does. It seems to me that their are plenty of cheap fruits and vegetables.

        I think that causation might run the other way that is very fat and very skinny people have trouble find and holding good jobs, that could be because of their appearance or there could be a common cause.

      • daedalus2u
        August 8, 2011 at 10:21 am |

        That is not correct. You need to consider the price per calorie, not the price per unit weight.

        http://www.mymoneyblog.com/what-does-200-calories-cost-the-economics-of-obesity.html

        Baby carrots cost $2.50 per 200 calories. Broccoli costs $1.93.

        For what you would spend on 200 calories of carrots you could get 2,000 calories of glazed donuts.

        People need to eat a certain number of calories per day, not a certain weight of food. If you need to eat 2,000 calories per day on a minimum budget, you are going to try and maximize consumption of things with low cost per calorie. Of the low cost sources of calories that do not require lots of preparation, cooking, refrigerated storage are all junk food.

    • daedalus2u
      August 4, 2011 at 1:30 pm | Reply

      They define “poor” as having an income less than 130% of the Federal poverty limit. In 2009, the poverty limit for a single person is $10,830.

      So if you make more than $14,080 per year you are considered “non-poor”. That is plenty of money to buy food. The USDA estimats that a single male eating a thrifty diet will need $40.60 per week, or $2,112.20 per year.

    • Sigivald
      August 4, 2011 at 4:04 pm | Reply

      “Good” food can be afforded by pretty much anyone, anywhere.

      If poor folks don’t eat “healthy”/”good” food, it’ll be because of factors other than affordability, in almost all cases.

      (Such as: Not being a good or even competent cook.

      Not desiring to spend the time and effort required.

      Not valuing nutrition enough to eat the “good” food instead of some delicious thing from a bag.)

      • Constant
        August 4, 2011 at 8:56 pm |

        Affordability is still a factor. If you are not a competent cook and do not want to make the effort and are rich, it’s easier for you to get good food anyway than if you are not competent and do not want to make the effort and are poor.

      • Psychohistorian
        August 5, 2011 at 4:55 pm |

        …or not having time to cook, because you’ve got too many obligations. Having no or very limited kitchen facilities. Not being able to afford tasty healthy food, and not wanting to live a life eating food that doesn’t taste good.

        It’s fun to pick all the morally blameworthy reasons, but not all the actual reasons are morally blameworthy.

  4. Gabriel Rossman
    August 4, 2011 at 1:35 pm | Reply

    So basically, there’s no difference in the mean but there is a difference in kurtosis. This strikes me as a very interesting finding but as we all learned from l’affaire du Larry Summers it’s not as if kurtosis is meaningless.

  5. Amos Zeeberg
    August 4, 2011 at 1:49 pm | Reply

    There’s a strong inverse correlation between women’s income and obesity.

    Interestingly, among Blacks and Latinos, men with higher income have have *higher* rates of obesity. [No strong correlation between income and obesity in white men.]

    http://timewellness.files.wordpress.com/2010/12/db50_fig1.gif

    • Brian
      August 5, 2011 at 12:45 am | Reply

      And the fact that this correlation exists only for women pretty strongly suggests discrimination as the main cause. Unless someone wants to suggest a mechanism where obesity affects a woman’s intelligence/persistence/health but not a man’s…

      Poverty causes obesity about as much as it causes blackness.

      • Miley Cyrax
        August 5, 2011 at 4:07 am |

        Hiring discrimination could be possible, but I have an additional hypothesis. Women who make less income tend to be around and interact with men who make less income, and men who make less income tend to black or latino, who are less discriminating against female obesity when it comes to mate choice, especially when the female is white. Thus, on average, women who make less have lower incentive to stay thin.

      • Brian
        August 5, 2011 at 9:00 am |

        “Women who make less income tend to be around and interact with men who make less income, and men who make less income tend to black or latino, who are less discriminating against female obesity when it comes to mate choice, especially when the female is white. Thus, on average, women who make less have lower incentive to stay thin.”

        Mate discrimination is still a form of discrimination. In this explanation, culture causes obesity, and the culture we’re talking about is correlated with poverty.

  6. Peter St Onge
    August 4, 2011 at 5:39 pm | Reply

    Gary Taubes and Kurt Harris believe that diets high in grains lead to both obesity and malnutrition. Obesity where gluten intolerance is high, malnutrition (but not obesity) where gluten intolerance is low.

    Which gives you low-willpower correlating with both obesity and over-thinness, with degree of gluten intolerance pushing in one or the other extreme.

    If this is the mechanism, it’s not that weight is uncorrelated, it’s that it strongly correlates, but with an independent variable (gluten intolerance) changing the sign of the correlation.

    • Brian Moore
      August 5, 2011 at 10:02 am | Reply

      Exactly. All the people here talking about low or high willpower are missing the point: if you eat foods that actually give your body what it wants, you feel full and can compliment yourself on your excellent willpower.

      If you don’t (see: “I feel hungry 30 minutes after eating that rice-laden Chinese food”) then your body says “we need more” and it looks like you have bad willpower.

  7. Erika
    August 4, 2011 at 10:13 pm | Reply

    I think depends a lot not only on what and you are eating but the amount also.

    It is not the same eating a 1 dollar chocolate bar to a dollar yogurt. Although both are the same price, the later is healthier and makes you feel fuller. However the chocolate is high in calories, and probably you could do much more than only eating a yogurt; but because your stomach is feeling empty, you end up eating (unnecessarily) more calories ..

    On the other hand, take a hamburger, you could have the same amount of calories for the same price in a cup of plain boiled rice; the same applies to fried potatoes, for the same price, you could have even the double amount of healthy boil potatoes, if you just choose to buy them in the supermarket.

    Eating healthy, does not anything to see with being rich or poor, but with “knowing” what to eat.

  8. Jack
    August 5, 2011 at 12:34 am | Reply

    Perhaps naively, it would seem that only including ‘poor’ and ‘non-poor’ categories is a bit of a drawback of the study.

    I would expect very poor people to have lower BMI (due to a variety of factors – harsher lifestyle, possible lack of food), and (according to conventional wisdom*) expect poor – lower middle class people to have higher than average BMI (less free time to make healthier choices, less leisure time for exercise, etc.).

    Perhaps a study that included a larger range of different income types would be more suitable.

    * This is the common assumption that this study seems to be addressing.

  9. jqhart
    August 5, 2011 at 2:10 am | Reply

    Amos suggests the reality of the situation. The causation is obviously obesity causes lower relative income, not the reverse, and for the obvious reasons. For the same kinds of reasons, being abnormally skinny causes lower relative income. If we could measure beauty, we would also find that beauty causes higher relative income.

    Since this is so obvious to normal people, it doesn’t send enough “I’m so clever” signals to merit publication in a scientific journal. Thus the myriad of papers containing quite clever but quite false theories that assume that lower income causes obesity.

  10. Miley Cyrax
    August 5, 2011 at 4:11 am | Reply

    If such a relationship exists between obesity and poverty, I suspect obesity among the poor is not only caused by the poor’s decision to go for the best bang for the buck when it comes to calories/$$, but also their low impulse control and low future time orientation, which will cause both poorness and poverty. Foods rich in sugar, fat, and salt taste good to most human palates, and it takes restraint to refrain from them, knowing the effects they’ll have on your body-shape.

    • daedalus2u
      August 5, 2011 at 3:09 pm | Reply

      Interesting idea that the poor are fat because they have low impulse-control.

      However the data that Robin is talking about doesn’t show that the poor are fatter than the rich, Therefore there is no evidence of obesity that can be used to justify the belief that the poor have low impulse-control.

      So where did the idea that the poor have low impulse-control come from? I suspect that the impulse-control of the poor is not that different than the impulse-control of the non-poor, however impulsive behavior has more detrimental effects on the poor than on the non-poor (as does everything).

  11. Nancy Lebovitz
    August 5, 2011 at 2:29 pm | Reply

    Not everyone has convenient access to supermarkets or time to cook. For the former see “food deserts” and consider needing to use inconvenient mass transit. For the latter, consider people working long hours at low wages, not to mention how it would feel to risk missing meals if your experiments in learning to cook don’t work out.

  12. Xander Miller
    August 6, 2011 at 1:48 pm | Reply

    There is another problem with the cheap food, poor people hypothesis: Food price does not correlate with how healthy a food is. A more likely culprit is stress. Poor people tend to be more stressed than the wealthy and they can make unhealthy food selections in attempts to self-medicate through comfort foods or stress and depression can also lead to loss of appetite.

  13. Robbie
    August 6, 2011 at 2:35 pm | Reply

    I may be reading it incorrectly but that (very poor) graph seems to show poor and non-poor populations having very similar BMI distributions, sure the poor curve is slightly flatter but the difference is minimal.

    Anyone care to point out where I’m incorrect?

    • Daublin
      August 9, 2011 at 4:23 pm | Reply

      That is my impression as well, Robbie. The data suggest there is nothing here. There are plenty of fat rich people, and also plenty of fat poor people. As well, not just the summary statistics but the overall distributions look very similar.

      The data suggest that BMI is predominantly driven by biology, by aspects of humanity that we are born with.

  14. Yvain
    August 7, 2011 at 7:00 pm | Reply

    I think the paper is trying to say the opposite of how you’re trying to interpret it.

    The paper brings up that the poor (defined by 130% of poverty line) are not necessarily overweight (defined by BMI > 25), but then rejects that as a useful statistical measure and says it’s too simplistic. It then proposes its own better statistical measures which show the expected correlation, and concludes:

    “The regression analysis provides further evidence that the relationship between BMI and income is more closely linked to conventional wisdom than is suggested by the cross-tabulations…this paper provides evidence that the cross tabulation of overweight and poor provides a very incomplete picture of the relationship between income and BMI by indicating essentially no association between poverty and overweight and obesity status. Making policy decisions based on this though would make for poor choices. The UQR estimates further suggest that there…is a negative income gradient for the obese (much larger in magnitude than that estimated by OLS) which matches the standard health and wealth gradient, and conventional wisdom. ”

    So the conclusion of the paper says several times that the conventional wisdom of poor people being fatter is correct. As such, I find your title “The Poor Are Not Fat” to be either misleading or dishonest, without even getting into the other data that does find a simple discrete relationship between the two categories (eg here: “More than 33 percent of adults who earn less than $15,000 per year were obese, compared with 24.6 percent of those who earn at least $50,000 per year.”)

    Whether the poor are fatter or not probably depends on how you define “fat” – the study showed no difference ever in > 25 BMI, historical but not current difference in > 30 BMI, and I am guessing that you would currently find a difference if you looked at > 35 BMI. It may be that middle class society has “caught up” to the poor in the former two groups, but if you want to look at people who are fat, not just by doctors’ standards but by the standards of society, the poor will still be disproportionately represented.

  15. Sister Y
    August 16, 2011 at 12:04 pm | Reply

    “After controlling for all other attributes that contribute to wages (an individual’s job experience, firm size, region in which they reside, occupation, sector, full or part-time employment, health, education, age, and whether or not they have young children) the author finds that married men and single women both have a wage rate that is positively related to the their Body Mass Index (BMI) – the heavier they are, the higher the wage they are paid. Single men and married women have the opposite experience – they are penalized for their weight — the heavier they are, the lower the wage they are paid.<"

    (Emphasis mine. Source)

  16. Chris
    August 19, 2011 at 10:25 am | Reply

    It is still widely accepted that the poor are lazy and that’s all I care about.

  17. George Jochnowitz
    August 19, 2011 at 10:38 am | Reply

    Believing in Hell, and therefore believing in the eternal justice of punishment, leads parents to mercilessly force their children to finish what’s on their plates.
    http://www.jochnowitz.net/Essays/ObesityAndHell.html

  18. OneBias
    August 19, 2011 at 10:52 am | Reply

    Next up, can we test the ridiculous bias that the poor are poor because they don’t want to work?

  19. Matt C
    August 19, 2011 at 11:39 am | Reply

    I am unswayed by your interpretation of the data, or the data as laid out.
    First, BMI is a terrible metric for measuring who is “fat.” Only body fat percentage can truly assess whether a person is fat or not. I am 6’1″ and 190 pounds. My BMI is 25.1. By definition in this study and many other studies on obesity, I am overweight. I also have 7% body fat. I am definitionally not overweight. By aggregating datapoints such as myself into studies like this, there is inherent bias in the measured variable. This is a fault of BMI as a metric.

    Second, I need a explanation as to why Mississippi and West Virginia are both 1) the poorest states in the US, and 2) the hightest BMI states in the US (thought as previously noted BMI is a terrible measure of fat) before I reject that there is no correlation between poverty and obesity, since there are barely any demographic similarities between the two states.

    Second, part A – why is Type II diabetes more prevalent amongst the poor?

    Third – the argument from a public health standpoint is not that the poor can’t afford healthy diets, its that the caloric content of many food products has an inverse relationship to price. Per calorie costs of an apple are much higher than that of Cocoa Puffs. If you are shopping for food on a limited budget, you get more bang for your buck in calories (and, theoretically, satiation) by purchasing these inverse relationship products.

  20. Pusser
    August 19, 2011 at 3:52 pm | Reply

    I think the generalization comes from the fact that the more obese states are those with lower incomes and higher percentages of minority population, therefore the conclusion that the poor and racial minorities are more likely to be obese.

Post a Comment

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>