Over 400,000 people visited Farnam Street last month to learn how to make better decisions, create new ideas, and avoid stupid errors. With more than 100,000 subscribers to our popular weekly digest, we've become an online intellectual hub. To learn more about we what do, start here.

Gary Taubes on What Really Makes Us Fat

Nutrition, as a scientific field, is reminiscent of psychology in its infancy — before Skinner’s or Pavlov’s ideas had made a dent, before there was Piaget, Kahneman, Tversky, Munger; before we had a field called “evolutionary psychology.” Heck, before biology itself had really come to grips with Darwin’s ideas about evolution through natural selection. (That didn’t happen until the modern synthesis in the 1940s.)

There were theories of mind, theories of self, theories of everything, but nearly all of them were incomplete and contradictory explanations of human behavior. Most were unscientific, in the Popperian sense. They couldn’t be falsified, and they claimed to explain too much, even patently opposite behaviors. It would be a long time before we started to pull the threads together and come up with coherent explanations of why we do what we do.

A similar pattern emerges as we survey the field of nutrition today. The ground is not steady. The most widely accepted and promulgated advice of the last fifty years is under attack: Maybe a calorie is not a calorie. Maybe the simple advice of “Eat less, move more,” nice as it sounds, is too simple, and frankly, wrong. (Reminding one of the second half of Einstein’s dictum: As simple as possible, but not simpler.) Maybe saturated fat is actually good for us. Maybe *gasp* salt is actually good for us.

These aren’t slight modifications of a prior theory but a 180 degree turn. (The visual metaphor is more literal than it seems: It’s not hard to argue, with the new evidence, that the old FDA food pyramid we’re familiar with should literally be turned upside down.)

As of late, this heretical thinking has been promoted most heavily by scientific journalist Gary Taubes, with his many articles, and his books: Good Calories, Bad Calories, and the slimmer and more easily digestible (!) Why We Get Fat (And What to Do About It).

He and his intellectual partner, Peter Attia, call it the Alternative Hypothesis. It goes roughly as follows: We don’t gain weight and get modern diseases like heart disease, obesity, and hypertension because we eat too many calories or consume too much dietary fat, but because we are consuming carbohydrates, especially sugar and easily digestible starches, at a pace that nature never intended. The resulting insulin resistance is the main culprit.

In Why We Get Fat, Taubes lays out the modern epidemic:

Fifty years ago, one in every eight or nine Americans would have been officially considered obese, and today it’s one in every three. Two in three are now considered overweight, which means they’re carrying around more weight than the public-health authorities deem to be healthy.

This, in the face of a rise in recreational exercise, gym-going, health-consciousness, and food pyramids. (Not to mention fast food, increasingly sedentary lifestyles, and so on — all the things you’ve heard before. Taubes thinks those are red herrings, by the way.)

The Causality Problem

The premise of the book is relatively simple: Trying to explain the explosion of obesity, hypertension, and heart disease by focusing on overeating or under-exercising (consuming more calories than we’re using) is missing the boat. It’s explaining something causally by simply describing it. Of course we’re storing more energy than expended. Why?

Step-by-step, detective-like, Taubes explores the current hypotheses and finds them wanting:

…(I will argue) that it is absurd to think about obesity as caused by overeating, because anything that makes people grow–whether in height or weight, in muscle or fat–will make them overeat. Children, for example, don’t grow taller because they eat voraciously and consume more calories than they expend. They eat so much-overeat–because they’re growing. They need to take in more calories than they expend.

We want to know what causes the overeating relative to energy expenditure. Why would any animal or human be driven to store more fat than they needed to function? Wild animals do not carry excess fat unless it has a useful physiological mechanism. (Whales keeping warm, squirrels preparing for winter, etc.)

Many have tried to explain this in psychological terms: Either we’re being manipulated into over-eating or we’re too weak to resist the temptations of the modern age. Today’s food is just too damn yummy.

Taubes thinks this is the wrong way to approach the problem. What we really need to understand is the regulation of our fat tissue itself and why it would go awry. He calls it Adiposity 101:

The message of eighty years of research on obese animals is simple and unconditional and worth restating: obesity does not come about because gluttony and sloth make it so; only a change in the regulation of fat tissue make a lean animal obese.

The regulation of fat tissue is a bit different than we often imagine. Fat tissue is not like a savings account we add to and then pull from at an indeterminate time later. It’s more like the battery in a solar energy system: It stores excess energy when it is available, then releases it later when the energy isn’t available, at regular periodic intervals.

We’re constantly storing and releasing fat: During every meal, the body mobilizes its resources to store energy not immediately needed. Later, when we’re between meals or sleeping, the body breaks down the stored energy and uses it to fuel the cells. This is how a normally functioning body works. The principle of homeostasis tells us that the body wants to stay in balance: We should be storing only as much fat as needed to survive.

The problem comes when the body isn’t functioning normally, and the balance is lost. We tend to think this happens because we’re overeating. But what if that’s the effect, not the cause? It’s counter-intuitive, but so is the fact that the sun doesn’t revolve around the earth.

The culprit is resistance to insulin, a natural hormone which (among other things) encourages fat storage and discourages release of fat stores while it’s circulating.

Remember, we depend on fatty acids for fuel in the hours after a meal, as blood sugar levels are dropping to their pre-meal level. But insulin suppresses the flow of fatty acid from the fat cells; it tells the other cells in the body to burn carbohydrates. So, as blood sugar returns to a healthy level, we need a replacement fuel supply.

(But) if insulin remains elevated, the fat isn’t available…as a result, the cells find themselves starved for fuel, and we quite literally feel their hunger. Either we eat sooner than we otherwise would have or we eat more when we do eat, or both. As I said earlier, anything that makes us fatter will make us overeat in the process. That’s what insulin does.

Taubes thinks we’ve gotten the causality backwards: Overeating doesn’t make us fat any more than overeating makes young children get taller; getting fat causes us to overeat. Our cells are literally being starved for energy by the selfish fat tissue, growing on its own accord in a tumor-like fashion driven by an abundance of circulating insulin. A isn’t causing B…B is causing A. This error has been the source of bad dieting advice.

Resistance is Futile 

How do we become insulin resistant in the first place? Simple: We over-consume easily digestible carbohydrates, which causes heavy and constant insulin response in our bloodstream. This level of carbohydrate consumption was not available to our ancestors year-round, making it a modern problem. (One objection that’s been raised here is that our ancestors simply didn’t live long enough to contract modern disease. I hypothesize that there’s an error being made: Many earlier humans lived perfectly long lives, certainly long enough to get obese and hypertensive, but one major reason average life expectancy was low is because child mortality was so high.)

In any case, our cells eventually become resistant to insulin, and as more and more of the hormone is released in response–to keep our blood sugar down–our fat remains stored away tightly in the form of triglycerides (tri = 3 fatty acids, glycerol = a binding molecule; how our body stores fat for later).

The indicators for diseases we now associate with obesity, including hypertension, diabetes, and stroke, have now largely been lumped under the term Metabolic Syndrome. Because these indicators–high triglycerides, high blood pressure, low HDL cholesterol among them–are often associated with an expanding waistline, many have been led to believe that obesity causes the other associated diseases.

Not so, according to Taubes:

The simplest way to look at all these associations, between obesity, heart diseases, type 2 diabetes, metabolic syndrome, cancer, and Alzheimer’s (not to mention the other conditions that also associate with obesity and diabetes, such as gout, asthma, and fatty liver disease), is that what makes us fat–the quality and quantity of carbohydrates we consume–also makes us sick.

The insulin resistance is causing obesity and the other problems. It isn’t that A causes B, it’s that C is causing A and B. Another important logical flaw.

What Now?

The implications are fairly obvious, from a diet perspective. We should cut our consumption of carbohydrates drastically, especially those that digest into the blood stream most easily and cause the most drastic insulin response.

Taubes smartly delimits himself from going beyond that advice, though:

It would be nice if we could improve on the foods to eat, foods to avoid, foods to eat in moderation. Unfortunately, this can’t be done without guessing. The kind of long-term clinical trials have not been undertaken that would tell us more about what constitutes the healthiest variation of a diet in which the fattening carbohydrates have already been removed.

Taubes admits that the right science must be done in order to prove that he’s right:

Our conventional dietary wisdom, as I’ve described in my books, is based on science that was simply not adequate to the task of establishing reliable knowledge — poorly-controlled human experiments, observational studies incapable of establishing cause and effect, and animal studies that may or may not say anything meaningful about what happens in humans.

That’s why he and Peter Attia created the Nutritional Science Initiative, or NuSi, with the idea of pulling together a range of scientists and researchers to perform more precise experiments. The goal is understanding whether a calorie is really a calorie, whether carbohydrate-restrictive diets are more effective (and why) and a host of other pertinent questions. Only good data will convince the public (and the health officials it turns to for advice) that things need to change. Their work is underway as we write.

Given the resistance most people have to admitting they were wrong, it’s an uphill battle.


What makes Taubes an effective thinker, and writer, is that he’s pulling from diverse fields in his quest to solve the obesity problem. While nutritionists typically silo themselves, to a certain extent, in their own field, with some influence from psychology to explain why obese people tend to overeat, Taubes instead pulls from anthropology, biochemistry, endocrinology, epidemiology, and the field of nutrition itself. By asking the right questions of the right people, he gets better answers. He isn’t afraid to step on toes. He also uses careful induction, sorting point by point the problems with competing hypotheses. This is something that reminds us of all great thinkers, from Newton to Darwin to Holmes to Munger.

Nutrition is a field changing in real-time: It needs precise thinkers with a strong penchant for self-criticism, truth-seeking, contrarianism, and boundary-crossing. Whether or not Taubes’ Alternative Hypothesis ends up being the correct one, the job needs to be done right. Observational studies, on which most nutritional recommendations are based, are not science: They are hypothesis-generators. Correlation generators. Figuring out causality is a tougher task.

The good news is that this core problem is solvable, and we suspect, like psychology, the field will come together in time.