American Health Fail: What's Making Us Fat? A Decline in Smoking

Americans are fat. The latest obesity estimates reach as high as 30% of the population; and the future looks worse. There’s been much hand wringing over the years, with a new television show sprouting up every season imploring the obese to lose weight.

A new paper by researchers Charles Baum and Shin-Yi Chou provides a detailed look at the leading indicators of weight, using the National Longitudinal Survey of Youth from 1979 and 1997 to compare the habits, similarities and differences between people of the same age – just a quarter century apart.

The results aren’t pleasant: the largest effect on our recent weight gain? The decline in cigarette smoking.

Here’s the abstract:

We simultaneously estimate the effects of the various socio-economic factors on weight status, considering in our analysis many of the socio-economic factors that have been identified by other researchers as important influences on caloric imbalance: employment, physical activity at work, food prices, the prevalence of restaurants, cigarette smoking, cigarette prices and taxes, food stamp receipt, and urbanization. We use 1979- and 1997-cohort National Longitudinal Survey of Youth (NLSY) data, which allows us to compare the prevalence of obesity between cohorts surveyed roughly 25 years apart. Using the traditional Blinder-Oaxaca decomposition technique, we find that cigarette smoking has the largest effect: the decline in cigarette smoking explains about 2% of the increase in the weight measures. The other significant factors explain less.

The researchers examined differences in employment, physical activity at work, food prices, prevalence of restaurants, cigarette smoking, cigarette pricing, food stamp receipts and the prevalence of urban sprawl.

They found that:  “(i) occupational fitness and strength demands significantly affect weight, (ii) cigarette smoking (and being a cigarette smoker) significantly decreases weight, (iii) food stamp receipt significantly increases weight, and (iv) urban sprawl significantly increases weight.” The decline in cigarette smoking, however, only makes up about 2% of weight increase, and the authors saw their criteria show only small percentage changes in weight measures as a whole, as opposed to other studies with much higher numbers.

The authors write: “Ultimately, the socio-economic and demographic factors examined in this analysis, whether considered individually or collectively, explain a minority of the increase in BMI, overweight, and obesity.”


To further this assessment. I'd recommend reading Why We Get Fat by Gary Taubes. It takes this study to the next level and suggests a different sociology Eco favor that is very plausible.


Swapping crutches doesn't make the absence of one a cause of the other. The cause is cultural insanity.

Mike B

I think we just have a society that is just overwhelmingly unhappy. If we aren't smoking or overeating we are finding ways yo make ourselves feel better by pushing others down.

Mike B

Cigarette smoking also makes Medicare and Social Security much more sustainable over the long run. When both those programs were conceived massive numbers of Americans would reach age 50-65 and then drop head from smoking related illnesses. Therefore the real effect of SS and Medicare was to care for the widows of those breadwinners who had killed themselves from smoking (or on the job carcinogen exposure). Take away the population trimming effect of smoking and the benefits that few people were intended to collect now apply to huge amounts of the population. What's even more ironic is that this and other studies show that not only were Americans left to get their dopamine hit from other sources (ie unhealthy food) they also lost their most popular appetite suppressant in the face of plummeting food prices and activity levels.

The result is that we have traded a population that died quickly at the end of their productive years for a population that becomes increasingly sickly and then stays that way for decades at public expense. This is not meant as a call to return people to smoking, just an acknowledgement that the conditions that out social safety net was enacted under no longer apply. It is also pointing out that when unhealthy behavior is a coping mechanism to bring temporary relief to an unhappy existence, taking away that mechanism doesn't solve the problem of unhappiness and will only cause people to find substitutes.


caleb b

I'm sure the Blue Traveler guy is really happy you decided to use a picture of him when talking about fat people.

...given the title of the article, is the picture supposed to beg the question, how much fatter would he be without smoking?


The picture is worth a thousand words. The fat guy is smoking. How can the correlation be so high when so many obese people are smokers?

Chris Sampson

Congratulations, Freakonomics, on your grossly misleading and erroneous headline.


Not too hard to invalidate this hypothesis.. Were people fat before smoking was common?

Causes of obesity are right in front of us. Look at the shopping carts of regular folks and ask yourself "when did we stop eating real food?" Over-engineered, preserved, processed, pulverized and reconstituted food took off in the 60s. Line up those graphs with weight or BMI or waist size or whatever you measure you prefer.

Enter your name...

Except that your story doesn't account for the fact that so many middle-class and upper-class people were overweight and obese in previous centuries, when "real food" was the only possible form of food.

Our "average" weight was low in previous centuries only because poor people were malnourished (this was back when people actually starved to death in "developed" countries), but take a look at the pictures of the people with modern levels of food security: For example, Queen Victoria had a BMI of 30, which puts her into the obese category, and she was not unusual.


Nor does your story explain much. We have, for instance the stereotypical image of the obese, cigar-smoking plutocrat, c.a. the late 19th century, and his equally obese non-smoking spouse. We can likewise find many examples of middle/upper-class people from those times who were emphatically not obese, for instance:

Could it be that both smoking and food are red herrings, and that the primary determinant of obesity is, in fact, simply exercise?

Eric M. Jones.

But is the increase in obesity correlated with the decline in smoking of the population of people who smoked?...or merely the overall population?....

These are very different issues.

Joe Dokes

The study confuses correlation with causation. The summary itself clearly states that only 2% of the rise in obesity can be linked to a decrease in smoking.

Thus, blaming the decline in smoking is just wrong, the real reason people are fat is that over the past 30 years we've been lied to, we've been told that FAT makes us fat and that we should eat carbs. Wrong, carbs make us fat. As the poster above stated read Gary Taubes Good Calories Bad Calories, or Why We Get Fat, for the real causes of American obesity.

The study itself shows this, who is most likely to be obese? People on food stamps. Why, because the have to eat cheap, what's cheap? Food loaded with cheap processed carbohydrates.


Joe Dokes

Mike B

Calories make you fat, doesn't really matter where they come from. There are a lot of issues that come together here, but the truth is that cigarettes suppress appetite (wonder while ballerinas all smoke?) and cigarettes provide a narcotic hit to the brain. Eating sugar, fat and salt also provides a narcotic hit to the brain and with all the current price incentives (cigarettes = expensive, unhealthy food = cheap) it's no surprise that people are eating themselves to death the same way they used to smoke themselves to death. The only difference is the time between getting sick and dropping dead.

Joe Dokes

I used to believe that as well. I always thought of it as a simple equation, calories in vs. calories out. Turns out I was wrong, to sum up 600+ pages in Taubes and others books it boils down to carbohydrates ability to quickly and dramatically raise blood sugar that induces both additional hunger and fat storage.

For me personally since I cut way back on my carbs, particularly refined carbs, I've dropped 35 Lbs. without ever being hungry. Prior to that I tried Jenny Craig and a couple of other diets in which I was simply STARVING all the time with minimal weight loss.


Joe Dokes


It is interesting how you cover one study after other with different fallacies, this one is a classic case of narrative fallacy.

"“Our need to fit a story or pattern to a series of connected or disconnected facts”


every smoker knows to expect weight gain if and when they quit smoking, this concept alone is not really news.


How do you explain the dramatic rise in childhood obesity? I doubt smoking has much to do with that.


The smoking rates for adult men peaked at about 60% sometime in the 1950s.

Today about 66% of adults are overweight or obese.

67% of the population drinks alcohol while 33% consider themselves total abstainers.

Two-thirds of the adult Americans population placed some kind of bet last year.

Smoking, over-eating, drinking and gambling are all examples of maladaptive coping mechanisms.

Walter Mishcel's Stamford Marshmallow study on the psychology of self-control found that about 66% of the kids in the study were unable to self-impose delayed gratification.

That 2/3rds number keeps popping up.


Great Blog. I will raise one issue however:

Have you guys (the Freakonomics team) considered making a small tab with an explanation of econometrics? I know you explain it in accessible terms in the books, but I just keep seeing the words "correlation" and "how can you tell from correlations" and "what about this factor, it explains everything" in the comments here.

The whole point of Freakonomics is to get at causality using robust econometric techniques, but some of the comments seem to reflect a serious lack of understanding on the part of the readers about the methodologies used in the studies.

People: there are indeed multiple causes for shifts in social phenomena. Econometrics (e.g. here on panel data) (1) shows what % of the variance is explained by a particular variable, such as the change in smoking, and (2) controls for hosts of other causes, which ultimately tries to simulate a controlled experiment. I keep seeing anecdotes here from posters - but again, the point of Freakonomics is that using econometric techniques on lots of data often turns our anecdotal notion on its head.