The gap is real, it favors carbohydrate, and it is worth about 26 calories a day#
At identical calories, a diet lower in fat has a small measurable edge. Pooling 32 controlled feeding studies that swapped carbohydrate for fat while holding energy constant, energy expenditure ran 26 kcal/day higher and fat loss 16 g/day greater on the lower-fat diets, both at P < 0.00011. So the folk belief that a fat calorie is a slightly more expensive calorie to eat is, unusually, pointing the right way.
Now the size. Sixteen grams of body fat a day is about half a kilogram a month if it persisted — my arithmetic on their figure, and the if is doing all the work, because the feeding studies that produced it ran days to weeks, not months. Twenty-six calories a day is inside the noise of any method you could use to detect it. What makes this question worth 1,500 words anyway is that almost every mechanism people offer for the difference is wrong, and the one real asymmetry between fat and carbohydrate has nothing to do with what happens after you swallow.
The storage-cost argument, and why the body rarely pays that bill#
The standard explanation runs like this: dietary fat is already fat, so shelving it as body fat is nearly free, while carbohydrate has to be chemically rebuilt into fat first, and that conversion — de novo lipogenesis — costs energy. Therefore a carbohydrate calorie is partly spent on its own filing. The biochemistry is correct. The inference is not, because it assumes the body routinely takes a route it mostly declines to take.
Reviewing the stable-isotope evidence, Hellerstein reached a conclusion that undercuts the entire argument: "eucaloric replacement of dietary fat by CHO does not induce hepatic DNL to any substantial degree," and adding carbohydrate to a mixed diet "does not increase hepatic DNL to quantitatively important levels, as long as CHO energy intake remains less than total energy expenditure"2. Instead of being converted, the extra carbohydrate is simply burned, and the fat that would otherwise have been burned is stored instead. Body fat still accrues. The expensive pathway is never traversed.
Carbohydrate does not become your body fat. It takes the place of your body fat in the furnace, and your body fat stays put.
That displacement is a mechanism this blog has met before under a different name — it is exactly what ethanol does to fat oxidation, for the same reason: a fuel the body cannot store must be cleared now, and everything storable waits. Hellerstein's summary of what this means for the macro question is worth stating plainly: carbohydrate and fat "are therefore not interconvertible currencies" and have "independent, though interacting, economies."
What it actually takes to convert carbohydrate into fat#
De novo lipogenesis is not imaginary — it is just expensive to provoke, and one study measured the price of admission. Three men had their glycogen stripped by diet and exercise, then were overfed for seven days on a diet of 86% carbohydrate, 11% protein and 3% fat, starting at 15.25 ± 1.10 MJ (3,642 ± 263 kcal) on day one and climbing to 20.64 ± 1.30 MJ (4,930 ± 311 kcal) by the last.
Two numbers came out of it. Glycogen storage capacity in humans is about 15 g per kg of body weight, which can absorb roughly 500 g of carbohydrate before net lipid synthesis begins to add to fat mass. And once those stores are saturated, the disposal route opens: about 150 g of lipid synthesized per day, from roughly 475 g of carbohydrate3.
Read the entry conditions rather than the result. A depleted starting point, nearly five thousand calories a day, 86% of them carbohydrate, sustained for a week. That is what it took to make the pathway quantitatively visible. Nobody reaches those conditions by having rice with dinner, which is why "carbs turn into fat" survives as folklore while failing as a description of an ordinary Tuesday.
Overfeeding is the one place the asymmetry gets large#
If the difference is trivial at matched calories and irrelevant below maintenance, there is one condition where it is neither: a genuine surplus. That was tested head-to-head. Nine lean and seven men with obesity were overfed isoenergetic amounts of fat and of carbohydrate — 50% above their energy requirements, 14 days each — inside a whole-room calorimeter that measured expenditure and nutrient oxidation on days 0, 1, 7 and 14.
| Overfeeding arm | What the body did with the excess | Share of excess energy stored |
|---|---|---|
| Carbohydrate, +50% for 14 days | Carbohydrate oxidation and total energy expenditure rose progressively | 75–85% |
| Fat, +50% for 14 days | Fat oxidation and total energy expenditure barely moved | 90–95% |
Data: Horton et al., 1995. The authors' conclusion is exactly as broad as the design supports: excess dietary fat leads to greater fat accumulation than excess dietary carbohydrate, and the difference was largest early in the overfeeding period.
A ten-to-fifteen-point spread in storage efficiency is not a rounding error. It is also a statement about surplus, not about diets. Carbohydrate defends itself by burning faster when it floods in; fat has no equivalent reflex, so a surplus of it goes almost entirely to storage. In a deficit — which is the state anyone reading this is likely in — the relevant question is which fuel you stop eating, and that reverses the frame entirely.
At matched calories, the difference collapses toward nothing#
Which brings back the number this article opened with. In the isocaloric case, the 32-study pooled analysis found lower-fat diets ahead on both expenditure and fat loss, but by 26 kcal/day and 16 g/day1. Under the tightest short-term control available — a metabolic ward, six days, a 30% energy cut taken out of one macro or the other — cutting fat again beat cutting carbohydrate for fat loss, a result worked through with the trials that followed it.
And in free-living trials lasting a year or two, nothing survives. That is not a contradiction; it is what a small effect looks like once it has to compete with whether people ate what they were assigned. Two things are simultaneously true: composition does move physiology by a measurable amount, and the amount is smaller than the week-to-week variance in anyone's actual eating. A difference you can only detect in a calorimeter is not a difference you can diet on.
The processing-cost story people reach for next — that fat is cheaper to digest than carbohydrate, so its calories land heavier — is real in direction and even smaller in size, and it belongs to the thermic effect of food. It cannot rescue a 26-calorie gap into a reason to fear olive oil.
The asymmetry that costs you is at the table#
Here is where fat genuinely is different, and it has nothing to do with metabolism. Six men each spent three separate weeks under whole-body indirect calorimetry with unlimited access to a covertly manipulated diet — 20%, 40% or 60% of energy as fat, with the fat content disguised. They were not told which diet they were on and were free to eat as much as they wanted.
Energy intake climbed with fat content (P < 0.001), producing average daily energy balances of −0.27, 0.77 and 2.58 MJ/day across the low-, medium- and high-fat diets5. In calories — my conversion at 4.184 kJ per kcal — that is a 65-calorie daily deficit on the low-fat diet against a 617-calorie daily surplus on the high-fat one, in the same six men, eating to appetite, with no idea anything had changed. The paper's own summary of the mechanism is that protein and carbohydrate appeared able to reduce subsequent intake while "there was no apparent reductive effect due to fat."
Nothing signalled to them that they had crossed from a small deficit into a 600-calorie surplus. That is the whole of fat's disadvantage, and it happens before digestion, not after.
One caveat is essential and is arguably the real finding. Those three diets also differed in energy density — 4.80, 5.59 and 7.04 kJ/g — because raising fat raises density, and the design cannot separate the two. That is not a flaw so much as a description of food: outside a laboratory, dietary fat and energy density arrive together. Whether the culprit is the macronutrient or the calories-per-gram it drags with it, the practical consequence is identical, and it is the reason a fat floor rather than a fat ceiling is the useful way to think about the number.
So: does a fat calorie count more?#
Not on the label, and not in any way the conversion factors behind 4/4/9 fail to capture. In a surplus, fat is stored more efficiently than carbohydrate by 10 to 15 percentage points of the excess. At matched calories, the gap is 26 calories a day. And eating to appetite, a high-fat diet moved six men 680 calories a day without their noticing — which is roughly 25 times the isocaloric effect and the only version of this question with practical stakes.
So the useful reframing is that dietary fat is not metabolically penalized; it is behaviorally expensive, because it delivers more energy per bite and per bite is how people actually decide when to stop. That points at portion control on dense foods rather than at macro percentages, and it leaves the size of the deficit exactly where it was: the variable that decides the outcome.
FAQ#
Do carbohydrates get converted into body fat?#
Hardly ever, in the direct sense. Replacing dietary fat with carbohydrate at matched calories does not meaningfully switch on the fat-synthesis pathway, and neither does adding carbohydrate to a mixed diet while total intake stays below expenditure2. Extra carbohydrate is burned, and the dietary fat it displaces is stored instead — so body fat rises without any conversion happening. The pathway does open under extreme carbohydrate overfeeding: about 150 g of lipid a day, once glycogen stores of roughly 15 g/kg are saturated3.
Is a calorie of fat stored more efficiently than a calorie of carbohydrate?#
In a surplus, yes, and measurably. Overfed by 50% for 14 days inside a whole-room calorimeter, men stored 90–95% of the excess energy on the fat arm against 75–85% on the carbohydrate arm, because carbohydrate overfeeding raised oxidation and total expenditure while fat overfeeding did not4. At or below maintenance the difference shrinks to about 26 kcal/day1.
Does a high-fat diet make you eat more without noticing?#
That is what the controlled test found. Six men eating freely from covertly manipulated diets of 20%, 40% and 60% fat ran daily energy balances of −0.27, 0.77 and 2.58 MJ — about a 65-calorie deficit against a 617-calorie surplus — without being told the diets differed5. The fat content and the energy density moved together, as they do in real food, so the study cannot separate them.
Sources#
- Hall KD, Guo J. Obesity Energetics: Body Weight Regulation and the Effects of Diet Composition. Gastroenterology. 2017;152(7):1718-1727.e3.
- Hellerstein MK. De novo lipogenesis in humans: metabolic and regulatory aspects. Eur J Clin Nutr. 1999;53 Suppl 1:S53-S65.
- Acheson KJ, Schutz Y, Bessard T, Anantharaman K, Flatt JP, Jequier E. Glycogen storage capacity and de novo lipogenesis during massive carbohydrate overfeeding in man. Am J Clin Nutr. 1988;48(2):240-247.
- Horton TJ, Drougas H, Brachey A, Reed GW, Peters JC, Hill JO. Fat and carbohydrate overfeeding in humans: different effects on energy storage. Am J Clin Nutr. 1995;62(1):19-29.
- Stubbs RJ, Harbron CG, Murgatroyd PR, Prentice AM. Covert manipulation of dietary fat and energy density: effect on substrate flux and food intake in men eating ad libitum. Am J Clin Nutr. 1995;62(2):316-329.
- Hall KD, Bemis T, Brychta R, et al. Calorie for Calorie, Dietary Fat Restriction Results in More Body Fat Loss than Carbohydrate Restriction in People with Obesity. Cell Metab. 2015;22(3):427-436.
