Do Carbs Make You Fat? The Insulin Hypothesis, Honestly

Carbs don't make you fat. Caloric surplus makes you fat. The reason this confuses people is that low-carb diets often work well for fat loss, which gets read as evidence that carbs are the cause. The actual mechanism is simpler and more boring: cutting carbs cuts a major source of easy-to-overeat hyperpalatable food, kills several common appetite triggers, and drops glycogen-bound water in the first week. None of that requires insulin to be the villain. The metabolic-ward data — the gold standard for testing this question — has been clear for a decade.

This is the carb-and-insulin question, handled without the tribal signaling. The underlying frame — that fat loss is about caloric deficit, food form, and adherence — is in the energy density explainer and the PE diet guide. The short version: at matched calories and matched protein, low-carb and high-carb diets produce nearly identical fat-loss results. Low-carb often feels easier. That's not the same as carbs being uniquely fattening.

The Insulin Hypothesis, Stated Fairly

The carbohydrate-insulin model of obesity, as articulated by Gary Taubes, David Ludwig, and others, runs roughly like this:

Eating carbohydrates raises blood glucose, which raises insulin.
Insulin promotes fat storage and inhibits fat release from adipose tissue.
Therefore, chronic high-carb eating drives fat accumulation independent of total calorie intake.
The corollary: cutting carbs lowers insulin, releases stored fat, and produces weight loss even without consciously eating less.

This is internally coherent and has the virtue of being testable. The first three claims are partially true at the biochemistry level — insulin does rise with carb intake (it also rises with protein, less famously), and insulin does facilitate fat storage. The fourth claim — that this matters for net body fat over weeks — is what the ward studies have addressed directly, and it doesn't hold up.

Kevin Hall's Metabolic Ward Studies

The cleanest tests of the carbohydrate-insulin model come from Hall et al. at the NIH. The setup: subjects live in a metabolic ward for weeks, every calorie measured, every output measured, no self-report. The diet macros change; everything else is held constant.

The headline findings:

2015 ward study (Hall et al., Cell Metabolism). Matched-calorie low-fat vs. low-carb diets. Low-fat group lost slightly more body fat over six days (statistically significant, clinically tiny). Insulin levels obviously differed; body fat loss did not favor the low-carb side.
2016 ward study (Hall et al., AJCN). Four weeks of ketogenic vs. baseline high-carb diet at matched calories and protein. Subjects lost fat at near-identical rates. The keto group did not gain a metabolic advantage. The carbohydrate-insulin model predicted ~0.5 lb/week additional fat loss on keto; observed value was essentially zero.
2021 ward study (Hall et al., Nature Medicine). Plant-based low-fat vs. animal-based low-carb diets, ad libitum (eat as much as you want). The low-fat group spontaneously ate ~500 fewer calories per day and lost more weight. Insulin was higher on the low-fat diet. The body fat went the other direction from what the carbohydrate-insulin model predicts.

The ward studies are not the only evidence, but they are the strongest because they remove self-report and ad-libitum variability. The verdict from this line of work, stated by Hall himself, is that the carbohydrate-insulin model's quantitative predictions are not supported. Calories are the primary lever; carbs and insulin are not a separate fat-storage pathway that bypasses caloric balance.

Why Low-Carb Often Works Anyway

If carbs aren't uniquely fattening, why do so many people lose weight on Atkins, keto, and "lazy keto"? Three real mechanisms, none of which require the insulin model:

Appetite suppression from high protein and fat. Cutting out the bread-and-pasta-and-cereal-and-rice quadrant of the diet usually means the remaining food is meat, eggs, vegetables, and dairy — all higher in protein and fat per calorie than what got removed. Protein and fat are more satiating than refined carbohydrate. People spontaneously eat less. The 2021 Hall study found the opposite direction for low-fat, but in the broader real-world data, keto wins the ad-libitum appetite battle for most people.
Elimination of hyperpalatable food vehicles. Cookies, ice cream, chips, pizza, pastries, soda, and most fast food are carb-dominant by design. "No carbs" effectively means "none of the foods that are engineered to override your satiety signals." The hyperpalatable foods writeup covers why this matters more than the carb count itself.
First-week water loss. Every gram of stored glycogen holds ~3g of water. A typical adult has 300–500g of glycogen. Cutting carbs to keto levels depletes most of it within a week, releasing 1–2 kg of water. This shows up on the scale immediately and looks like rapid fat loss. It isn't, but it's a real motivator.

These mechanisms are all calorie-driven. They make the deficit easier to achieve and easier to sustain. They do not make carbs themselves fattening — they make carb-rich processed food easy to overeat.

The Long-Run Equivalence

The clearest summary of decades of comparison trials is the 2018 DIETFITS study (Gardner et al., JAMA): 609 adults randomized to either a healthy low-fat or a healthy low-carb diet for 12 months. Both groups lost weight at near-identical rates (~6 kg average), with no statistically significant difference between arms. Genetic and insulin-secretion subgroups also didn't predict differential response — the "find your right macro split" intuition failed.

The honest summary: at matched protein and matched adherence, the macro split mostly doesn't matter for long-run fat loss. What matters is:

Caloric deficit. The TDEE calculator and macro calculator give you a baseline.
Protein intake. ~1.6–2.2 g/kg bodyweight per day, regardless of carb level, preserves muscle and increases satiety. The PE diet writeup is the case for protein as the lever.
Food form and SPC. Whole foods, less processing, more volume per calorie. The satiety-per-calorie ranking tells you which foods to favor.
Adherence. Whatever diet you'll actually do for months.

If you find low-carb easier — fine. If you find a moderate-carb diet with high satiety per calorie easier — also fine. The macro split is a personal-fit question, not a metabolic-truth question.

What Actually Drives Body Fat

The mechanism is unromantic. Adipose tissue grows when more calories enter the body than leave it, sustained. It shrinks when the opposite is true, sustained. Insulin is the hormone that signals fed-state storage, but storage requires the calories to be there in the first place. You cannot store fat from a meal you didn't eat regardless of how high your insulin goes.

What actually pushes people into chronic surplus, the boring catalog:

Liquid calories (soda, juice, alcohol, sweetened coffee).
Hyperpalatable processed foods that override satiety signals.
Low-protein meals that leave you hungry two hours later.
Plate sizes and portion creep at restaurants and at home.
Snacking out of stress, boredom, or social context rather than hunger.

Carbs appear in some of these (soda, pastries, chips), but so do fats (ice cream, fried food, nut butters by the spoonful) and so does alcohol. The pattern is energy density, palatability, and ease of overconsumption — not the macronutrient label.

The Verdict

Do carbs make you fat? No. Calorie surplus makes you fat. The carbohydrate-insulin model is a tidy story that the metabolic-ward data does not support at the quantitative level it would need to in order to be the explanation.

Are low-carb diets a legitimate fat-loss approach? Yes — for appetite-suppression, palatability-control, and adherence reasons. Not for special metabolic ones.

The practical rule: pick the macro split you'll actually stick to. Hit a protein target around 1.6–2.2 g/kg. Build the plate around high-SPC foods. Let the macro war happen on Twitter without you. The macro calculator gives you a starting deficit; the PE diet guide and the top-50 foods ranking cover the food-choice side of the equation that actually moves the needle.