Your BMI equals your weight in kilograms divided by your height in meters squared. A person who weighs 70 kg and stands 1.75 m tall, for example, has a BMI of about 22.9. The calculation takes seconds, which is exactly why it became the world’s default shorthand for body size. But the number it produces is a rough proxy for fatness, not a direct measurement of it, and the gap between the proxy and reality varies depending on who you are.
The Formula in Both Unit Systems
In metric units the math is straightforward: divide your weight in kilograms by your height in meters multiplied by itself. If you weigh 82 kg and stand 1.80 m, you divide 82 by 3.24 (which is 1.80 × 1.80) and get roughly 25.3.
If you work in pounds and inches, multiply your weight in pounds by 703, then divide by your height in inches squared. So someone who weighs 180 pounds and stands 5 feet 10 inches (70 inches) would calculate 180 × 703 = 126,540, then divide by 4,900 (70 × 70), arriving at about 25.8. The factor of 703 is just a unit-conversion constant that makes the two systems line up.
Where the Number Came From
The formula traces back to Adolphe Quetelet, a Belgian mathematician who in 1832 observed that in adults, body weight tends to scale with the square of height rather than height itself.1Nephrology Dialysis Transplantation. Adolphe Quetelet (1796–1874)—the average man and indices of obesity For over a century the ratio was called the Quetelet Index. It was not designed as a health tool; Quetelet was interested in statistical patterns across populations, not in diagnosing individual patients. The name “body mass index” arrived in 1972, when American physiologist Ancel Keys published a landmark paper arguing that this particular ratio correlated better with body fat than other height-weight formulas available at the time.2International Journal of Epidemiology. Commentary: Origins and evolution of body mass index (BMI): continuing saga That paper cemented BMI’s place in clinical practice and public health, where it has remained ever since.
Standard BMI Categories
The World Health Organization groups adult BMI values into four broad bands:
- Underweight: below 18.5
- Normal weight: 18.5 to 24.9
- Overweight: 25.0 to 29.9
- Obese: 30.0 and above
These cutoffs were drawn primarily from mortality and disease-risk data in populations of European descent. They are the thresholds used on most online calculators and in most doctors’ offices, but as the sections below explain, their accuracy varies by sex, ethnicity, age, and body composition.
How Well BMI Actually Tracks Body Fat
At a population level, the correlation between BMI and measured body fat is strong, especially in women. A large analysis using DXA scans (a precise body-composition technology) found that BMI correlated with fat mass at roughly 0.94 in men and 0.98 in women, though the link weakened with age and varied by ethnicity. In Korean participants, for instance, those correlations were lower than in other groups.3Scientific Reports. Different correlation of body mass index with body fatness and obesity-related biomarker according to age, sex and race-ethnicity Studies in Sri Lankan adults and Indian adults have confirmed a moderate-to-strong positive correlation overall, while also noting that the strength of the link can differ substantially between men and women within the same population.4PubMed Central. Relationship between Body Mass Index (BMI) and body fat percentage, estimated by bioelectrical impedance, in a group of Sri Lankan adults: a cross sectional study 5PubMed Central. Relationship Between Body Mass Index and Body Fat Percentage in a Group of Indian Participants: A Cross-Sectional Study From a Tertiary Care Hospital
The takeaway is that for large groups of people, a higher BMI usually does mean more body fat. But “usually” is doing a lot of work in that sentence. At the individual level, two people with the same BMI can carry very different amounts of fat, and the mismatch is large enough to matter clinically.
The Athlete Problem
BMI cannot tell the difference between a kilogram of muscle and a kilogram of fat. This creates obvious problems for anyone who carries above-average muscle mass. In a study of adolescent athletes, only about 38 percent of those flagged as obese by BMI were also obese by their measured body fat. The remaining 62 percent were false positives: muscular young people who looked overweight on paper but were not by any meaningful body-fat standard.6PubMed Central. Body Mass Index and Percentage of Body Fat as Indicators for Obesity in an Adolescent Athletic Population
Research in adult male athletes tells a similar story. Using WHO cutoffs, about a quarter of athletes with normal body-fat levels were misclassified as overweight or obese. Agreement between the BMI classification and actual body-fat status was poor.7PubMed Central. Revising BMI Cut-Off Points for Overweight and Obesity in Male Athletes: An Analysis Based on Multivariable Model-Building If you lift weights regularly, play a sport that builds lean mass, or just happen to be naturally muscular, your BMI can overstate your health risk. No one seriously thinks a rugby forward with a BMI of 29 and 15 percent body fat faces the same metabolic profile as a sedentary person at the same BMI carrying twice as much fat.
Normal-Weight Obesity and the Problem BMI Misses in the Other Direction
BMI’s blind spot cuts both ways. While it can label fit people as overweight, it can also give a clean bill of health to people whose weight falls in the “normal” range but whose body-fat percentage is dangerously high. This condition, sometimes called normal-weight obesity, is more common than most people assume. A study in young adults found that roughly 29 percent qualified as normal-weight obese, and these individuals had about three times the odds of cardiometabolic risk compared to lean people at the same BMI, along with more than seven times the odds of high abdominal fat.8PubMed Central. Normal-Weight Obesity Is Associated with Increased Cardiometabolic Risk in Young Adults A normal BMI can be falsely reassuring if you carry excess fat around your organs but not enough total mass to push the number past 25.
Why BMI Cutoffs Do Not Apply Equally Across Ethnicities
The standard thresholds were developed largely in European-descent populations, and body composition at a given BMI differs meaningfully across ethnic groups. At the same BMI, Asian populations tend to carry a higher percentage of body fat, especially at younger ages, than white or Black populations. Black individuals, on the other hand, tend to have a lower percentage of body fat at the same BMI, particularly as they age.9The American Journal of Clinical Nutrition. How to Calculate Your BMI: Formula and Limitations
These differences are not trivial. The optimal BMI threshold for identifying cardiometabolic risk in African-American women was found to be about 32.9, roughly 3 points higher than the threshold for white women. For men, the ethnic gap was smaller and not statistically meaningful.10PubMed Central. Ethnic-specific BMI and waist circumference thresholds Several Asian countries have already lowered their official overweight threshold to a BMI of 23 rather than 25, reflecting the higher metabolic risk at lower body weights. If you are using a BMI calculator and your ancestry is not European, the standard cutoffs may be telling you the wrong story.
Sex Differences in What BMI Means
Women naturally carry more body fat than men, so at the same BMI, a woman and a man have different body compositions. Research has found that at a given BMI, women carry about 10 percent more fat mass than men.11PubMed Central. Obesity: a gender-view The relationship between BMI and body fat also follows a different curve depending on sex. In women the relationship tends to be more curved (quadratic), while in men it is somewhat more linear, meaning that each additional BMI point corresponds to a slightly different amount of fat depending on whether you are male or female, and where on the scale you fall.12PubMed. What kind of Relationship is Between Body Mass Index and Body Fat Percentage?
Yet the WHO categories are the same for both sexes. A BMI of 27 flags a man and a woman identically, despite their likely differences in fat mass, muscle mass, and where on the body that fat sits. This is one reason some clinicians look at waist circumference alongside BMI when assessing women’s cardiovascular risk.
BMI in Older Adults and the “Obesity Paradox”
One of the stranger patterns in BMI research is that among older adults, being slightly overweight by the standard cutoffs appears to be associated with lower mortality than being at a so-called normal weight. This has been called the obesity paradox, and it has generated years of debate.
The paradox is probably not what it seems. As people age, they tend to lose muscle and gain abdominal fat, but total weight may stay the same or even drop. BMI cannot detect this redistribution. A study of more than 130,000 UK Biobank participants found that much of the supposed protective effect of overweight in older adults disappeared once researchers accounted for central fat accumulation, smoking, and disease-related weight loss.13PubMed Central. Central adiposity and the overweight risk paradox in aging: follow-up of 130,473 UK Biobank participants Systematic reviews have noted that most studies reporting the paradox evaluated BMI alone without measuring body composition or visceral fat, and that central fat and loss of lean mass may matter more than BMI for health risk in older people.14PubMed. Obesity paradox and aging 15PubMed Central. The Obesity Paradox and Mortality in Older Adults: A Systematic Review
The practical implication: if you are over 65, a BMI in the normal range does not guarantee a healthy body composition, and a BMI in the overweight range does not guarantee a risky one. The number becomes less informative the older you get.
BMI and Mortality Risk
Despite its limitations, BMI does track meaningfully with death risk across large populations. A UK cohort study of 3.6 million adults found a J-shaped curve: mortality was lowest around a BMI of 25, rose by roughly 21 percent for each 5-point increase above 25, and also rose (by about 19 percent per 5-point decrease) below 25.16The Lancet. Association of BMI with overall and cause-specific mortality: a population-based cohort study of 3·6 million adults in the UK Another analysis showed the same general shape across different baseline health profiles: the very lowest and very highest BMI categories carried the most risk, while a wide middle range showed comparably low mortality.17PLOS ONE. Association of BMI, comorbidities and all-cause mortality by using a baseline mortality risk model
The J-shape is worth understanding because it means being underweight can be just as risky as being moderately obese, something that gets far less attention in public conversation about BMI. The nadir of the curve sitting near 25 also suggests that the boundary between “normal” and “overweight” may be drawn a bit too sharply for mortality purposes.
Waist-to-Height Ratio as a Better Screening Tool
If BMI has trouble capturing where fat sits on the body, waist-based measurements tackle that problem more directly. Your waist-to-height ratio is exactly what it sounds like: divide your waist circumference by your height. A ratio above 0.5 is a commonly used flag for elevated risk. The measurement takes a tape measure and about 10 seconds.
A systematic review and meta-analysis of more than 300,000 adults found that waist-to-height ratio had significantly better discriminatory power than both BMI and waist circumference alone for detecting diabetes, hypertension, and cardiovascular disease in men and women across multiple ethnic groups.18PubMed. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: systematic review and meta-analysis For people with diabetes specifically, the advantage was confirmed in a separate systematic review of prospective cohort studies, which found that waist-to-height ratio consistently outperformed BMI in predicting adverse cardiovascular outcomes.19PubMed Central. Comparative Evaluation of Waist-to-Height Ratio and BMI in Predicting Adverse Cardiovascular Outcome in People With Diabetes: A Systematic Review
That said, the practical advantage over BMI is modest. A large cohort study with replication found that the population-attributable fractions for multiple long-term conditions were less than 1 percentage point higher for waist-based measures than for BMI. The real gains showed up primarily among people in the overweight BMI range (25 to 29.9), where BMI is least informative and waist measurements add the most clarity.20PubMed Central. Waist, waist-height-ratio vs body mass index and the risks of multiple diseases: a cohort study with replication If your BMI is clearly in the normal range or clearly in the obese range, the two measures will usually agree. It is in the gray zone around 25 to 30 that a tape measure around your waist adds the most useful information.
DXA Scans and Other Direct Measurements
If you want to know your actual body composition rather than an estimate based on your height and weight, the gold standard is a DXA scan (dual-energy X-ray absorptiometry). DXA directly measures bone, lean tissue, and fat tissue, and can even show where fat is distributed regionally. A study comparing BMI-based obesity classification to DXA-based measurement found a dramatic gap: BMI classified about 51 percent of participants as obese, while DXA scans put the figure at 92 percent. The mismatch was especially large in men, where BMI labeled only about 41 percent as obese while DXA found 86 percent had excess body fat.21PubMed Central. Accuracy of body mass index compared to whole-body dual energy X-ray absorptiometry in diagnosing obesity in adults in the Eastern Province of Saudi Arabia: A cross-sectional study
Bioelectrical impedance analysis (BIA), the technology behind most consumer smart scales, is more accessible than DXA but less accurate. A comparison of over 3,600 measurements found that BIA overestimated lean mass and underestimated fat mass at most BMI levels, with the disagreement reaching as much as 8 kg of lean mass for people in the higher BMI ranges. The limits of agreement between the two methods were wide regardless of BMI, meaning any single BIA reading can be off by a fair amount.22PLOS ONE. Comparison of body composition assessment by DXA and BIA according to the body mass index: A retrospective study on 3655 measures A smart scale is better than nothing for tracking trends over time, but it should not be treated as a precise measurement.
BMI for Children and Teenagers
The standard adult BMI categories do not apply to anyone under 18. Children’s body composition changes rapidly as they grow, so a raw BMI number means different things at different ages. Instead, pediatric BMI is compared against age- and sex-specific growth charts. The WHO provides reference curves for children from birth to age 5 based on length or height, weight, and BMI-for-age.23PubMed. WHO Child Growth Standards based on length/height, weight and age For older children and adolescents, the CDC uses percentile charts based on U.S. reference data. A child at the 85th percentile is classified as overweight; at the 95th percentile, obese. These percentiles shift with age, which is why your pediatrician plots your child’s BMI on a curve over time rather than comparing it to a single static threshold.
Proposed Fixes to the Formula Itself
Some researchers have questioned whether squaring height is even the right mathematical approach. University of Oxford mathematician Nick Trefethen proposed raising height to the power of 2.5 rather than 2, arguing that the traditional formula leads shorter people to underestimate their BMI and taller people to overestimate it. His modification uses a multiplicative constant of 1.3, calibrated so that the new formula matches the traditional one at a height of 1.69 meters (roughly 5 feet 7 inches). If you are shorter than that, the modified formula gives you a slightly higher BMI; taller, and it gives you a slightly lower one.24PubMed Central. Assessment of a proposed BMI formula in predicting body fat percentage among Filipino young adults
The idea makes geometric sense: human bodies do not scale uniformly in all directions as height increases, so a simple square may overcorrect for tall people and undercorrect for short people. However, there is no epidemiological evidence that using the 2.5 exponent improves health-outcome predictions, something Trefethen himself acknowledged. The modification has not been adopted by any major health organization. Similarly, a measure called the triponderal mass index (weight divided by height cubed) has been explored in pediatric populations as a potentially more stable alternative across ages, but it has not replaced BMI-for-age percentiles in clinical practice.25PubMed. Triponderal mass index rather than body mass index: An indicator of high adiposity in Italian children and adolescents
BMI in Insurance and Policy Decisions
Outside the doctor’s office, BMI shows up in places most people do not expect. Life insurance underwriting relies heavily on it. A study examining the predictive value of various medical tests in insurance applicants found that BMI, combined with a physician’s assessment of the health declaration and basic demographics like age and sex, was nearly as powerful at predicting mortality risk as a full panel of additional lab tests. A model using just BMI, the health declaration assessment, and gender achieved an area-under-the-curve of about 0.71, while throwing in every available test only pushed that to about 0.74.26PubMed Central. The Added Value of Medical Testing in Underwriting Life Insurance In other words, BMI carries outsized weight in determining your premiums because, at a population level, it does predict mortality tolerably well, even if it mislabels plenty of individuals along the way.
This creates a real tension. BMI is cheap, universal, and reasonably predictive in aggregate, which makes it attractive for policy and insurance applications. But the same individual-level inaccuracies described throughout this article mean that some people pay higher premiums or face employment screening consequences for a number that misrepresents their actual health.
Psychological Effects of BMI-Based Screening
There is another cost that rarely makes it into clinical discussions. When BMI is used as a screening tool, especially in school-age populations, the number itself can cause harm. A systematic review of weight-screening programs for children found negative psychological consequences including decreased weight satisfaction, increased preoccupation with weight, more frequent peer weight talk, heightened sensitivity about body size, and emotional distress.27PubMed. Psychosocial consequences of weight screening of school-age children – a systematic review For a measurement that cannot distinguish muscle from fat and does not account for growth stages, the confidence with which it gets handed to families can be disproportionate to its accuracy. This does not mean screening should not happen, but it suggests that BMI numbers need to come with context rather than a label stamped in bold.