Does Ethnicity Affect Your BMR? The Research Behind the Adjustment

Plug your stats into ten different online calculators and you'll get ten different resting metabolic rates, sometimes a few hundred calories apart. Most of that scatter comes from which formula the site happens to use. But one variable that rarely gets discussed, and probably should, is where your ancestors are from. So does ethnicity affect your BMR? The honest answer is: for some populations, yes, measurably, and the size of the effect is smaller than people assume. That's the premise behind MacroMentor's ethnicity adjusted BMR calculator, and it's worth walking through what the studies actually show before trusting any adjustment, including this one.

Where the standard formula comes from

Most calorie calculators, MacroMentor included, lean on the Mifflin-St Jeor equation. A 1990 study by Mifflin and colleagues derived it from 498 healthy adults in the U.S., measured by indirect calorimetry. It's held up well since. A 2005 systematic review by Frankenfield and colleagues, which compared it against older formulas like Harris-Benedict, Owen, and the WHO/FAO/UNU equation, found Mifflin-St Jeor predicted resting metabolic rate within 10% of measured values in more people than any of the alternatives. It also flagged something easy to miss: the review found that US-residing ethnic minorities were underrepresented in the development of these predictive equations, Mifflin-St Jeor included, and in the studies that validated them afterward. A formula built mostly on one population and then applied to everyone will drift when the underlying body composition shifts. That drift is what the ethnicity research is actually measuring.

So does ethnicity affect metabolism, or just the formula?

Both, and it's an important distinction. A 1998 study by Soares and colleagues compared basal metabolic rate in 96 adults from Bangalore against 81 Caucasian Australians. On the surface, there was a gap. Once the researchers adjusted for fat-free mass, the gap disappeared. Their literal conclusion: no evidence for an ethnic influence on basal metabolism once body composition is controlled for. The weight-based formula was biased; the person's own metabolism was never in question.

A 2016 study by Song and colleagues in Singapore dug into why. Comparing Chinese and Asian-Indian men, they found a 59-calorie-a-day gap that survived adjusting for total lean mass, but vanished once they adjusted specifically for trunk lean mass, which includes the liver, kidneys, heart, and spleen. Those organs burn far more energy per pound than skeletal muscle does. Two people can carry identical lean mass and still run different metabolic rates if one of them has proportionally larger organs and less muscle making up that total. That's the mechanism, and it shows up again and again across this literature.

The numbers, population by population

For South Asian ancestry, the studies above put the gap in the range of roughly 2 to 6%, with formulas built on total body weight overestimating BMR more than formulas anchored to fat-free mass. MacroMentor applies a 4% downward correction on the Mifflin-St Jeor path for this group, landing inside that published range.

For East Asian ancestry, a 2016 study by Camps and colleagues measured basal metabolic rate by indirect calorimetry in 232 Singaporean Chinese adults and tested it against six predictive equations, Mifflin-St Jeor included. Mifflin overestimated measured BMR by about 39 calories a day, roughly 2.6% off. MacroMentor applies a 4% downward correction here, a touch above that single-study figure but within the wider range these population comparisons report.

Worth flagging: a 2017 study of 30 elite Indian male weightlifters found the opposite problem. Every formula tested, Mifflin-St Jeor and Cunningham among them, underestimated their measured energy expenditure, because lean body mass, not total body weight, was the real driver of their metabolic rate. That's not a contradiction of the ethnicity research; it's a reminder that individual body composition can override a population average, which is exactly why MacroMentor turns the ethnicity adjustment off the moment you enter a body fat percentage.

Why entering body fat sidesteps the whole question

Once body fat is known, the calculator switches to the Katch-McArdle or Cunningham equation, both built directly on lean mass rather than total weight. Since the organ-and-muscle-mass difference is the actual driver behind most of these ethnicity gaps, an equation that already accounts for lean mass has less bias left to correct. Soares' 1998 comparison found the lean-mass-based Cunningham equation predicted accurately across both his Indian and Australian groups, while the weight-based formula didn't. Song's 2016 organ-mass analysis tells the same story from a different angle. Lean-mass equations aren't bias-free in every population: a 2026 validation study in perimenopausal women found Cunningham overestimating measured resting energy expenditure by nearly 160 calories a day even as Mifflin-St Jeor stayed closest to the measured values. But for the specific ethnicity-driven bias this article is about, they do the job.

What the research doesn't support yet

Some of this evidence base is much thinner than the rest. A 2011 study by Manini and colleagues measured genetic ancestry directly, using ancestry-informative genetic markers rather than relying on self-reported race, in a group of 141 older African American adults. Each percentage point of European ancestry was associated with 1.6 additional calories of daily resting metabolic rate. Within that single self-identified group, measured ancestry ranged from 0.1% to 70.7% European. In other words, a category like "African American" or "South Asian" is a rough proxy for whatever biology actually drives the effect, not a clean biological box. The National Academies' 2023 review of energy requirements reaches a similar conclusion: it treats metabolic data across diverse ethnic groups as too limited to support formal per-group adjustment factors, and it doesn't issue any.

That's a reasonable place to land. The published sample sizes behind most of these findings run somewhere between 70 and 250 people per study, single cities, single cohorts. The direction of the effect is consistent and real. The exact size of it, for any one person, isn't something a spreadsheet can promise you.

If you know your body fat percentage, enter it. It's the more direct fix. If you don't, the ethnicity field gives you a small, literature-grounded nudge in the right direction rather than none at all. Either way, the calculator's at /calculator whenever you want the actual number.