Why this question does not have a published answer
No US government agency publishes a count of full names. The Census Bureau publishes surname frequency; the Social Security Administration publishes given-name frequency; nobody publishes the joint frequency of "John" and "Kim" occurring together, because doing so at the level of specific, identifiable name combinations would raise the same privacy concerns that keep the Census Bureau from publishing names attached to individual records at all. Anyone who states a specific number for "how many John Kims there are" is not quoting a government figure. They are either guessing, or performing a calculation on public rate data — and the calculation, done the obvious way, produces a number that is wrong for a specific, explainable reason. This page works through what the two halves of that calculation actually say, why multiplying them together does not answer the question, and what can honestly be said about the size of the collision instead.
What the surname side says: Kim in the United States
The Census Bureau's decennial surname tabulations are the standard source for US surname frequency. Independent compilations of the Bureau's 2010 surname file consistently place Kim at approximately the 77th most common surname in the country, with roughly 262,000 people reporting it — a large jump from its position roughly two decades earlier, when 1990 census-derived tabulations put Kim closer to 233rd most common, with well under half that count. Later tabulations drawing on the Bureau's 2020 surname release put Kim slightly higher still, in the neighborhood of the low-to-mid 70s by rank with a count approaching 275,000. These figures come from secondary compilations of the Bureau's published surname files rather than a figure this site pulled directly from the Bureau's own table, so they should be read as "approximately" rather than exact — different compilations round or process the underlying file slightly differently — but multiple independent tabulations converge closely enough to trust the order of magnitude: roughly a quarter million people in the United States carry the surname Kim, and the number has grown substantially since 1990 alongside the growth of Korean immigration to the United States over the same period.
One more figure from the Bureau's surname data matters for what follows: Kim's demographic breakdown is unusually concentrated by race and ethnicity compared with most surnames in the top 100. Compilations of the Bureau's data put the share of people with the surname Kim who self-identify as Asian or Pacific Islander in the low-to-mid 90th percentile range — an overwhelming majority, in sharp contrast to a surname like Smith or Johnson, whose bearers are spread far more evenly across the Bureau's racial and ethnic categories. That concentration is the fact the rest of this page turns on.
What the given-name side says: John nationally
The Social Security Administration's baby-name data is the standard source for US given-name frequency, built from Social Security card applications going back to 1880. Over the hundred-year period the SSA highlights in its own published rankings (1926–2025), John is the third most common name given to boys in the United States, behind only James and Michael, with just over 4.2 million total recorded occurrences across that century. That is an enormous number — John has been one of the handful of default American boys' names for generations — but it is also, critically, a cumulative figure spanning a hundred years of US births overall, not a figure specific to any one community, generation, or era. John's popularity as an annually assigned name has also declined substantially from its mid-twentieth-century peak; it no longer ranks among the most frequently chosen names for boys born in a given recent year, even though the accumulated stock of men named John, built up across a century in which it was consistently common, remains very large.
Why simply multiplying the two rates is wrong
The tempting shortcut is to treat surname frequency and given-name frequency as independent probabilities and multiply them: take the share of Americans named Kim, multiply by the share of American men named John, and treat the product as the expected share of Americans who are John Kim. This calculation assumes that a given name is distributed identically across every surname group in the population — that being named Kim tells you nothing about how likely someone is to be named John, and being named John tells you nothing about how likely someone is to be named Kim. That assumption is false, and it is false in a way that is directly documented by the data already discussed on this page and in a companion guide on this hub.
Given names are strongly correlated with ethnicity, generation, religion, and immigration history — none of which are randomly distributed across surname groups. "John" is not evenly spread across every surname in America; it is disproportionately common among people whose family history runs through Christian Europe or its diaspora, for centuries of largely independent reasons documented in a companion guide on this hub. "Kim" is not evenly spread across every ethnic and religious group in America either — the Census Bureau's own demographic breakdown puts the overwhelming majority of Kims in the Asian/Pacific Islander category, and within that population, a companion guide on this hub documents why Korean-Americans specifically, for their own separate set of historical reasons — Korea's comparatively large Christian population and a documented, decades-long practice of adopting English given names weighted toward biblical choices — are considerably more likely to carry an English given name like John than an unrelated cross-section of the general population would be. Both "John" and "Kim" are elevated, for related historical reasons, within the population that intersects: Christian-influenced, English-speaking Korean-Americans. Naive multiplication, which assumes no such relationship exists, does not merely introduce a small margin of error here — it gets the direction of the correlation backward. The true joint rate is very likely higher, potentially substantially higher, than the product of the two raw national rates would suggest, precisely because the two rates are not independent of each other in the way the multiplication assumes; they are correlated through the same underlying population.
What would actually be needed to answer the question properly
A defensible estimate would require given-name frequency data broken out specifically within the population that shares the surname Kim, or equivalently, surname frequency broken out specifically within the Korean-American population — not the national given-name rate applied uniformly to everyone. That kind of cross-tabulated, name-by-ethnicity data is exactly the sort of information the Census Bureau and Social Security Administration do not publish at the individual-name level, for the same privacy reasons that prevent publication of the joint frequency directly. Academic researchers in fields such as demography and onomastics have occasionally used approaches along these lines — for instance, treating a distinctive surname like Kim as a proxy for Korean ethnicity in population studies — but that research is aimed at estimating ethnic population sizes from surname lists, not at producing a public count of a single specific full name, and it is not designed to, and does not, yield a defensible "how many John Kims" figure either.
What can be said with real confidence, without inventing a number that does not exist in any published source, is directional and qualitative rather than a count: John Kim is very likely one of the more frequently occurring full names in the United States relative to how uncommon either "John" or "Kim" would be if drawn from unrelated, independent populations — precisely because the historical forces that made each half common are not independent of each other. That is a meaningfully different claim from a number, and it is the honest one.
What the uncertainty means in practice
For anyone actually trying to identify a specific person named John Kim — a patient's doctor, a business contact, a public figure in the news — the practical lesson is not the missing number itself but what the missing number implies: the name provides very little narrowing power on its own, considerably less than an equally common-sounding name would provide in a population where given names really were distributed independently of surname. A name search, a directory listing, or an unqualified mention of "John Kim" should be treated as identifying a large, genuinely indeterminate set of unrelated people rather than a single likely individual, and any attempt to disambiguate has to rely on identifiers beyond the name itself — an employer, a professional license, a specific institution, a documented credential — rather than on the assumption that the name is rare enough to narrow things down by itself. This site's other silos exist specifically because the name alone cannot do that work.