Surprising New Findings on Obesity

One of the first Freakonomics Radio podcasts we made was an episode about the (surprisingly tenuous) link between obesity and health problems. A new study in The Journal of the American Medical Association finds that “Grade 1 obesity overall was not associated with higher mortality, and overweight was associated with significantly lower all-cause mortality.”  Writing for The Daily Beast, Kent Sepkowitz explains:

Compared to people with a normal weight (a BMI less than 25), the overweight (BMI between 25 to 30) had a 6 percent lower mortality rate—and both groups had a rate about 15 percent lower than the obese, especially the very obese (BMI above 35).

The explanation for the finding is uncertain. Perhaps the pleasantly plump but not obese have an extra reserve—a literal spare tire—that confers a survival advantage should they become seriously ill, whereas the lean-iacs do not. Or maybe the thin ones were thin because of a serious illness that, in the course the various studies, killed them. Or maybe the thin ones were thin because they were chain smokers living off Scotch and potato chips. Or just maybe the occasional pig-out does soothe the soul and make for a happier, healthier individual.

(HT: Andrew Sullivan)

Scott Templeman (@tallbonez)

Real findings should be that BMI is a garbage metric


I couldn't agree more about BMI being a grabge metric.

I am a triathlete, and in a recent medical at my company, the nurse exprerssed concern about my low BMI, after remarking on my powerful sounding heart and spectacular peak flow measurement. I am built like a whippet, am uber-fit and extremely healthy.

The next guy in after me was a 230 pound bodybuilder, really fit, ripped as anything but only 5'10", guess what?...


We have the same problem at my office. Except every other person is morbidly overweight and many are struggling with diabetes.

If you bring up any simple measure of obesity (like BMI), they all point out how it's a junk metric and it doesn't address their muscle mass. I wonder if they could do a single push-up.

BMI is pointless. Fit people know that at best it's a weak guideline. Fat people dismiss it because they don't want to change to a healthier lifestyle. Rarely do you meet a person who actually wants to know what their healthiest weight would be and then change their life to achieve it.

As a person who is constantly trying to live healthier, I'd love to have an ideal weight for my height and build. I'm 5'11" and 167 lbs, and I meet the BMI guidelines, but I'd love to know whether another 10 lbs either way would extend my life (and it's quality) or reduce it.


Or maybe the study included people with terminal cancer and other diseases which cause them to lose weight. This would be a very poor study indeed. And a source of incorrect information about the health risks of being overweight.

Enter your name...

The combined studies covered nearly 3 million people long enough that almost 10% of them died during the tracked time period. I think it was probably a pretty good study overall.

I understand that they excluded anyone in hospice care or in the hospital at the start of the study. It probably does include people with eating disorders like anorexia nervosa (known to have a high mortality rate among young adults), though; on the other hand, it probably includes a lot more people with binge-eating and other disorders that cause weight gain.

It is also true that most of the excess mortality in "healthy weight" people comes from the people whose BMI is in the skinnier half of that category. I think it is probably a good basis for us to consider whether the BMI categories need to be adjusted slightly, so that "barely heavy enough to be normal" is now considered "underweight".


Is it also possible that the use of BMI as the gauge for obesity is outdated? It doesn't measure muscle mass or fitness level. I understand it was used in this particular study for comparisons to other studies, but it seems to me that BMI is misleading at best.

Tomasz T.

It's true that BMI cannot distinguish between a plump couch potato and muscle-packed bodybuilder, but I would guess the number of these latter guys is so minuscule compared to the former, that they cannot disturb the results too much. BMI may be a very good proxy for overweight/obesity at a population level, even though it can be wildly wrong for a particular individual.


Using BMI to determine "weight" is like using horsepower to determine how fast a vehicle can go. I'm sure there are some relatively close relationships, but when it comes down to it, there are so many other aspects that should be taken into consideration before drawing a conclusion.


How long until insurance companies start dropping premiums for higher BMI's?

Enter your name...

"Lives longer" and "incurs fewer health costs" are not the same thing.


Life Insurance, not health.


Perhaps BMI is a terrible way to measure healthy weight. This has been well insinuated by a number of beyer weight and health professionals.


Yup, there are a lot of very fit people who are in the 25-28 BMI range. BMI doesn't take muscle mass vs. fat mass into account.

When you have a selection of fit, muscular people being called overweight and have likely not weeded out the very sick, alcoholics, anorexics, etc. from your below 25 BMI, that may well explain the "advantage."

It may still indicate that some extra weight isn't actually harmful though.


Not just the use of BMI, basically this whole "study" was full of garbage methodology. We are so desperate to continue living our crappy lifestyle that we'll allow anything to pass for science if it gives us hope.


Even if we decide that the method for calculating BMI is reasonable (which I'm not sure it is), the ranges for what is considered "normal" vs "overweight" seem a little skewed to me. I have friends who I don't think for a minute need to be concerned about weight loss, who live healthy lifestyles and everything, they're just not model skinny, but they are technically in the "overweight" range. Maybe this study is just showing that those ranges are a little off, and the upper limit for "normal" should be raised a little bit, because those weights are actually still healthy.


It's not so much that the BMI ranges are off, as that it was designed to measure a mostly sedentary population. As a consequence, there's a part of the range from high "normal" to moderately overweight in which the excess weight can come from either fat or muscle. BMI can't really distinguish between the two groups, so you'd have to apply some other measure - say chest to waist ratio for men.

Ned Cody

There's something dodgy about the mathematics, I think. How can mortality rates which differ by 6% both be equally below another mortality rate? And the cut-off values must be deceptive: how can a grouping of 29.8 bmi-ers be 15% less likely to die than a group of 30.0 bmi-ers.
I'm only saying...

Eric M. Jones.

Fat jiggles to attract attention--signaling breeding fitness.


Smokers usually weigh less. Reexamine the data excluding them and see what is revealed.


I have a theory. It comes from a lifetime of being the "big guy." Very simply, I believe that some people are "designed" to carry additional weight. We've all seen those folks who, with just a few extra pounds, are huffing and puffing. Or even those folks who, morbidly obese, can barely get around.

And then we've seen folks who, even 50 lbs. overweight, are active, vibrant, etc.

For many years, I didn't understand why some people suffer heart attacks before 30, while I, much, much larger than them, enjoyed relatively good health.

I have concluded that some people have a body, metabolism, or something that simply permits them to successfully carry more weight. I'm not saying that's the best thing. I'm simply saying that IF they have more weight, it doesn't negatively affect them like it might another person. Perhaps later in life, but at least year-for-year, pound-for-pound, they do much better. No, they won't be going to the Olympics, but they will be able to enjoy a larger plate of food and desserts than others, all with relatively minimal impact to their health (at least until much later).

I see some morbidly obese who die in their 50s...then others who are big until true old age. Maybe they don't live as long, but for those special ones that seemed designed to handle lots of extra weight, it sure seems that they enjoyed life without worrying too much about calories.


Phoebus Rising

Sounds like Mr. Sullivan might be one of the "pleasantly plump".

Matthew V

I also remember seeing a study on BMI that showed various football players. Linemen and running backs (or some other position, don't like the sport) that had the same BMI, but completely different fitness levels had similar life expectancy. In other words, the fat guys and the fit guys lived to similar ages. Compared apples to oranges and got the same shelf life.

Rob S

Surprised this hasn't come up on Freakanomics, but isn't the mortality rate 100% for everyone?


Not yet. It's only 100% for those who have already died. Saying that it will be 100% (or indeed, any particular number) is predicting the future, which the Freakonomics guys keep telling us is impossible :-)

Eric M. Jones.


Predicting the future for the INDIVIDUAL is impossible. Predicting the certain future for a population is merely statistics.

That's why I know that you and everyone else will perish whilst I shall live forever.


Did the study take into account how the subjects died? Fit people may be more fit in general because they are more active. But going out and doing things puts you at a higher risk for accidental death than sitting on your couch watching TV...

Also agree with most here that BMI is far from an ideal way to measure health/fitness. Body fat percentage would probably be a better indicator overall.


Not only are there lots of superfit people with higher BMIs but there are lots of metabolically healthy and moderately fit people with higher BMIs. If someone is thin, no one assumes that they have to also be a bodybuilder to be basically healthy. Many studies through the years have also found that the same is true for fat people. This study of 3mil people also bares that out. Fat people do not fall into only two categories: coach potato or body-builder. Fat people lead the same lives as thin people: some are coach potatoes, some go to the gym 2-3 times per week, some walk to work and use the stairs, some play intramural sports, and some are in peak physical condition as professional athletes (bodybuilders, football players, Olympic judo or discus throwers, etc.).


"Fat people do not fall into only two categories: coach potato or body-builder."

Perhaps this is a quibble over your phrasing, but bodybuilders are seldom if ever fat. Competitive (male) bodybuilders typically aim for a body fat percentage under 6% (14-17% is considered normal for a fit male.)

That's really what most of the comments about the inadequacy of BMI as the sole measure of fitness have tried to address: the fact many who are moderately overweight by BMI can be carrying the excess as muscle rather than fat. Bodybuilders of course take this to an extreme.

Likewise, I'm sure we all have met a few slender couch potatos of the skin, bones, and small pot belly variety. BMI may say that these folks are at a normal, healthy weight, but are they really? Would they not be much healthier if they started working out and put on some muscle mass?


The results may be quite different if body fat % was used rather than bmi. Bmi is a notoriously poor method of gauging if a person is truly overweight. Tim Ferris wrote a great critique of bmi and is worth reading (I think it can be found on his blog). Bmi was originally developed as a statistical measure that makes generalizations about a population but was not intended for evaluation of an individual's body fat %.

Henry Lahore

Vitamin D is a possible reason. The extra weight stores vitamin D as well as calories. Many studies have shown that as a person gets sick and loses weight, the vitamin D levels in their bodies increases. The higher level of vitamin D protects the lower weight body against a wide variety of diseases.


My explanation: some of us are genetically programmed to be plumper, and it's "normal" for us. Unlike the morbidly obese, who clearly suffer from a syndrome we'd rather blame them for then explore and treat, we "plumps" are "as we should be." I have been in the neighborhood of 200 pounds my entire life. I'm now 66. I have no more health problems than the average person. It's foolish to think that human variation is so narrow that only "thin" is healthy. To puncture other myths, I don't live on chocolate cake and "choco covered sugar bombs" cereal. I eat a lot of vegetables and fruits, try to stick to whole grains, and prefer plain meats to "smothered in" kinds of dishes. I am as my grandmother was, and she died at 87.