Where Foreigners Live in Japan: Census Data by City
Where foreigners live in Japan is not evenly spread: five prefectures, out of forty-seven, hold more than half the country's entire foreign national population.1 For a prospective mover or a newly arrived resident choosing a first city, that concentration pattern, and the data behind it, shapes where community networks, language support, and jobs are easiest to find.
Procedures, fees, and requirements can change. Confirm current details at e-Stat, the Statistics Bureau of Japan's official statistics portal.
Overview
This article is built on Japan's 国勢調査 (Kokusei Chōsa, "National Census"), the government's five-yearly headcount of everyone usually resident in the country, Japanese and foreign national alike.23 The most recent cycle with a full nationality, prefecture, and municipality breakdown is 2020; the 2025 cycle was conducted, but as of this writing only preliminary population and household totals have been released, not the nationality detail (as of 2026-07; confirm current figures with e-Stat, the Statistics Bureau of Japan's official statistics portal).3
The data below draws on the "都道府県・市区町村別の主な結果" (Main Results by Prefecture and Municipality) table series, specifically table 44-1, cross-checked against the Statistics Bureau's own published overview reports for each cycle.1456 It answers three questions: which prefectures and cities have the largest foreign populations, why the pattern looks the way it does, and whether it has been intensifying or spreading out since 2000.
Reading the Data: Census vs. Registry Statistics
Japan actually runs three separate government systems that each produce a "how many foreigners live here" number, not two. Most ranking articles and blog posts use one of these without ever saying which, which is the single biggest reason a figure in this article may not match a number a reader has seen elsewhere.
| Japanese term | Romaji | English gloss | Compiled by | Frequency |
|---|---|---|---|---|
| 国勢調査 | Kokusei Chōsa | National Census | Statistics Bureau (総務省統計局) | Every 5 years |
| 在留外国人統計 | Zairyū Gaikokujin Tōkei | Residency/immigration registry statistics | Immigration Services Agency (出入国在留管理庁) | Quarterly |
| 住民基本台帳 | Jūmin Kihon Daichō | Basic Resident Register ("juki") | Municipal (city/ward/town) offices | Continuous |
The census counts everyone at their place of 常住 (jōjū, "usual residence") as of a single fixed date, October 1, defined as living or being scheduled to live at a dwelling for three months or longer. Foreign diplomatic and military personnel and their families are excluded entirely; students are counted at their school lodging address rather than a home-country address (as of 2026-07; confirm current figures with e-Stat's glossary of statistical terms).2
The registry counts two groups: 中長期在留者 (medium-to-long-term residents, meaning anyone approved for a stay exceeding three months) and 特別永住者 (special permanent residents, mostly Zainichi Korean and Chinese residents with pre-war-era historical status). Short-term visitors and diplomatic status holders are excluded from this core figure, though the Immigration Services Agency (ISA) separately publishes a broader total each quarter that adds them back in (as of 2026-07).78 The registry is compiled on a "place of residence" basis tied to the residence card, which can diverge from where a person is actually registered on the Basic Resident Register.7
Foreign residents were folded into the Basic Resident Register system on 2012-07-09, replacing the older Alien Registration Act regime.9 It is updated in near-real-time by individual municipalities, and it is what most "foreign population of my town, right now" figures cited in local news actually come from, not the census.
No census figure exists for 2024 or 2025 yet, since the census only runs in years ending in 0 or 5 and the 2025 nationality breakdown has not been published. Any current-year ranking a reader sees is almost certainly sourced from the quarterly registry or the continuously updated juki register instead.3
The two systems have produced materially different totals for as long as both have existed. In 2000, the census-reported foreign national population was 1,310,545, versus 1,686,444 in the registry (then still the Alien Registration Act) for the same year, a gap of 375,899 people (as of 2000).10 That gap shows the two methodologies are not interchangeable, and it is not a recent phenomenon.
There is a third, more subtle wrinkle worth flagging before the first data table below: even within the 2020 census alone, two different official totals exist for the same nationwide figure. The Statistics Bureau publishes both a raw tabulated count (原数値, gensūchi) that leaves "not stated" nationality responses in their own bucket, and a 不詳補完値 (fushō hokan-chi, "unknown-imputed value") reference table that statistically apportions those "not stated" responses into known categories (as of 2020-10).11 The next section explains what that means for the numbers you are about to read.
The Concentration Pattern
By Prefecture
In the 2020 census, the top five prefectures by foreign national population, in order, were Tokyo, Aichi, Osaka, Kanagawa, and Saitama, together holding 53.9% of Japan's entire foreign national population (as of 2020-10; confirm current figures with e-Stat, the Statistics Bureau of Japan's official statistics portal).1
The table below uses the Statistics Bureau's own headline overview figures, which are 不詳補完値 (imputed) values; the Bureau's overview report labels this table with that footnote directly. The same report separately discloses the pre-imputation national total (原数値, raw): 2,402,460, versus 2,747,137 imputed, a gap of roughly 14%.
A secondary aggregator publishing prefecture-level figures it labels as raw table 44-1 values reports a national total of 2,402,460, an exact match to the Bureau's own disclosed raw figure, with Tokyo 483,372, Aichi 231,369, Osaka 208,681, Kanagawa 195,535, and Saitama 161,439 (as of 2020-10).112 The prefecture ordering is identical either way.
This article uses the Bureau's own headline (imputed) set as the primary reference, since it is the figure the Bureau itself foregrounds. Both totals are genuine official numbers, depending on whether "not stated" nationality responses are left raw or statistically imputed, which is exactly the kind of methodology distinction this article exists to make visible.
2020 census: top 5 prefectures
| Rank | Prefecture | Foreign national population | Share of national foreign population | Share of prefecture's own population |
|---|---|---|---|---|
| 1 | Tokyo | 564,000 | 20.5% | 4.0% |
| 2 | Aichi | 259,000 | 9.4% | 3.4% |
| 3 | Osaka | 242,000 | 8.8% | 2.7% |
| 4 | Kanagawa | 231,000 | 8.4% | 2.5% |
| 5 | Saitama | 186,000 | 6.8% | 2.5% |
Figures are the Statistics Bureau's headline (不詳補完値-based) 2020 census values; see the note above for why a secondary aggregator's alternate figures run lower.112
The national total foreign-national population in 2020 was 2,747,000, up 43.6% from 2015, against a 0.7% national population decline over the same period (as of 2020-10).1 Three of the top five prefectures, Tokyo, Kanagawa, and Saitama, sit within the Greater Tokyo metropolitan area, reinforcing that region as a single large cluster rather than three separate ones.
The Aichi-Osaka reversal
The Aichi-versus-Osaka ordering that different sources disagree on is a genuine historical reversal, not just an artifact of sources mixing 2020 census data with newer registry data. Osaka led narrowly through 2005 and 2010; Aichi overtook Osaka by 2015 and held the lead through 2020.
| Census year | Osaka | Aichi | Which is ahead |
|---|---|---|---|
| 2005 | 175,766 | 150,115 | Osaka |
| 2010 | 164,704 | 160,228 | Osaka (narrowing) |
| 2015 | 150,890 | 166,150 | Aichi (reversed) |
| 2020 | 242,000 | 259,000 | Aichi |
Each figure is dated to its own census cycle; the 2020 row uses the Bureau's headline figures, consistent with the table above.4513
By City
The census table series supports municipality-level detail, not just the prefecture level: table 44-1 is explicitly structured as "national, prefecture, municipality."6 That granularity is what makes small manufacturing towns with unusually high foreign-resident shares visible in the data at all.
Ōizumi Town (大泉町), Gunma, a small manufacturing town with a longstanding industrial base, is a genuine, census-confirmed high-foreign-share outlier. A Gunma prefectural government statistics report puts Ōizumi's foreign-resident share at 17.6% in the 2020 census, the highest of any municipality in the prefecture, well ahead of the runner-up, Shōwa Village at 6.3% (as of 2020-10).14 Within Ōizumi, Brazilian nationals are the largest single group (4,248 people, 57.5% of the town's foreign residents), followed by Peru (952 people, 12.9%) and Nepal (5.1%) (as of 2020-10).1415
The 17.6% figure is the raw tabulation; the Statistics Bureau's own commentary on the 2020 results puts Ōizumi's imputed foreign-resident share at 18.4%, up from 14.7% in 2015.15 That gap between 17.6% and 18.4% is the same raw-versus-imputed split covered above, playing out at the municipal level.
A local newspaper reported Ōizumi's foreign-resident share at approximately 20.02% (8,306 of 41,495 residents) as of late November 2023, published January 2024 and described as the town's first time crossing 20%.16 The report does not state whether this is a juki or registry figure, and it should not be read as an update to the 17.6% (or 18.4%) census figure above; they measure different things at different times, which is exactly the kind of mismatch the "Reading the Data" section warns about.
Very small nationality-by-municipality counts are subject to a documented small-cell suppression convention: where a result is "notably small," the Statistics Bureau replaces the published number with "x" rather than a real figure.17 The exact numeric threshold for table 44-1 specifically is not confirmed; the "Nationality Detail Exists at the City Level" entry under Good to know below explains what this means in practice.
What Drives the Concentration
Employment
Foreign workers are not distributed evenly across industries either, and the industry mix differs sharply between Tokyo and Aichi in ways that track the population concentration described above. Nationwide, manufacturing is the single largest industry employing foreign workers: 635,075 people, or 24.7% of all 2,571,037 foreign workers on the employer-notification register (as of 2025-10-31; confirm current figures with the Ministry of Health, Labour and Welfare's regional labour bureau reports).18
Tokyo's foreign workforce skews sharply away from manufacturing and toward office, tech, and service industries. Of Tokyo's 652,251 foreign workers, only 35,821 (5.5%) work in manufacturing, well under a quarter of the national 24.7% rate; information and communications work, dominated by IT, accounts for 77,924 (11.9%), roughly three times the 3.8% national rate. Wholesale and retail (108,878, 16.7%) and accommodation and food services (131,080, 20.1%, versus 12.4% nationally) are also disproportionately large in Tokyo (as of 2025-10-31).18
Aichi's foreign workforce skews just as sharply the other way. Of Aichi's 249,076 foreign workers, 96,288 (38.7%) work in manufacturing, well above the 24.7% national rate, and manufacturing is the largest single industry in every sub-region of the prefecture, Nagoya, Owari, West Mikawa, and East Mikawa alike (as of 2025-10-31).18
Within that manufacturing employment, transport-equipment manufacturing, meaning automotive assembly and parts, is the single largest subsector: 32,205 people, or 12.9% of all of Aichi's foreign workers, not just its manufacturing workers. That concentration is sharper still in the West Mikawa sub-region, which includes Toyota City: manufacturing there employs 37,853 people, 52.2% of the sub-region's entire foreign workforce (as of 2025-10-31).18
An independent labor-market research note on the same government dataset describes the same pattern in its own words: fewer foreign workers in manufacturing and more in IT, wholesale and retail, and accommodation and food service in Tokyo than in other regions, with foundation-materials, industrial-machinery, and electric and electronic manufacturing employment "exceptionally high" in Aichi and Osaka by comparison.19
Tokyo's and Aichi's foreign workforces work in very different industries, but that difference is a correlation, not a demonstrated cause. No source available for this article draws an explicit causal line from a specific industry to where foreign residents choose to live; the honest claim is that the industry mix tracks the population pattern, not that it explains it.
Existing community and chain migration
The commonly cited historical driver for Aichi's, and specifically the Toyota area's, Brazilian and Nikkei population is the 1990 revision to the Immigration Control and Refugee Recognition Act (入管法). Effective 1990-06, it created the 定住者 (teijūsha, "long-term resident") status, letting Japanese-descent (Nikkei) individuals down to the third generation, and their spouses, live and work in Japan without the occupational restrictions attached to other visa statuses.20
Hamamatsu City, Shizuoka, self-reports as the single Japanese municipality with the largest Brazilian-national population of any city nationwide: 9,462 Brazilian residents out of 24,932 total foreign residents, about 38% (as of 2022-04-01).21 The city's own account ties this to the same mechanism already described for Aichi: a late-1980s Japanese labor shortage combined with economic and political instability in Brazil drove Nikkei 出稼ぎ (dekasegi, "working away from home") labor migration, and the 1990 law then removed the occupational restrictions that had limited what second- and third-generation Nikkei and their families could do for work.21
Japan's registered Brazilian-national population grew through the 1990s and peaked at the end of 2007, then fell sharply during the 2008 global financial crisis. By the end of 2011 it stood at 210,032, a decline of 106,935 people, or 33.7%, from the 2007 peak.22
In 2001, Hamamatsu proposed and founded the 外国人集住都市会議 (Gaikokujin Shūjū Toshi Kaigi, "Council of Cities with Concentrated Foreign Populations"), a voluntary council of municipal governments that shares knowledge and lobbies the national government on issues facing cities with large foreign populations. The city's own stated rationale for founding the council generalizes the same pattern beyond Aichi: South American Nikkei residents surged and concentrated around manufacturing clusters, particularly transport-equipment, meaning automotive, manufacturing.21
Gunma's Ōizumi Town, covered above for its 2020 census share, fits the same pattern. The Statistics Bureau's own commentary on the 2020 census results attributes Ōizumi's high Brazilian-national share to nearby automobile-manufacturer factories, and separately notes that Brazilian immigration nationally rose "in particular from the relatively recent 1990s," tying the town's demographic profile directly to the 1990 law's timing.15
The specific channel connecting community networks to jobs shows up in a peer-reviewed study of Nikkei Brazilian residents at the 保見団地 (Homi danchi), a public-housing complex in Toyota City. As residents' length of stay increased, they built social networks centered on fellow Brazilian residents, and those kin and friend networks were the specific channel connecting people to regular employment at Toyota-affiliated companies.23
The same study found real limits to that mechanism. Despite the network-facilitated access to regular employment, workers still faced limited upward mobility, workplace discrimination, and Japanese-language barriers, remaining in what the study describes as a semi-peripheral occupational position; chain migration eased entry into the job market, but it did not erase the disadvantages foreign workers faced once there.23
Historical settlement patterns
Osaka's Korean-resident community, historically concentrated in Ikuno Ward (生野区), in what was once the Ikaino (猪飼野) district, predates the current census-era concentration pattern by roughly a century. Settlement began in the late Meiji and early Taishō era, and grew with laborers brought in for Hirano River channel-straightening construction that ran from 1919 to 1939.24
This pre-existing community is one reason later Korean, and broader East Asian, arrivals have continued to cluster in Osaka and the wider Kansai region even as the overall composition of Japan's foreign population has diversified in recent decades, with the Korean and Chinese shares of the national total declining and the Vietnamese share rising sharply between 2005 and 2020.15
The Trend Across Five Census Cycles (2000-2020)
Japan's census-counted foreign national population has grown in every published cycle since 2000, with growth accelerating sharply in the most recent 2015-to-2020 interval.
National foreign-national population by census cycle
| Census year | National total (foreign nationals) | Top-5-prefecture share of total |
|---|---|---|
| 2000 | approx. 1,310,545 | 51.6% |
| 2005 | 1,555,505 | 49.9% |
| 2010 | 1,648,037 | 52.1% |
| 2015 | 1,752,368 | 53.9% |
| 2020 | 2,747,000 (headline, imputed) / 2,402,460 (raw table) | 53.9% |
Each total is dated to its own census cycle (as of 2000-2020, per cycle); confirm current figures with e-Stat, the Statistics Bureau of Japan's official statistics portal. The 2020 row again shows both totals discussed above, for the same raw-versus-imputed reason.10145132512
Top-5-prefecture concentration dipped slightly between 2000 and 2005, then climbed steadily from 2005 through 2015, holding flat at 53.9% into 2020 (as of 2000-2020, per cycle).101451325 Read together, that is mild intensification of concentration since the mid-2000s, not a dramatic multi-decade climb and not a spreading-out trend.
The fifth prefecture in the top five has not always been Saitama. In 2005, Hyogo (85,943) held the fifth spot; Saitama overtook Hyogo for fifth place sometime between the 2005 and 2010 cycles (as of 2005 vs. 2010).513
Only Tokyo's 2000 figure (213,000, a 16.3% share) and the full 2005 top-5 ranking are confirmed at prefecture level from this research. Other individual prefecture figures for 2000 are not included here, since the only available source for them could not be reliably read as data, only as an image; only the 2000 national total and Tokyo's figure are carried forward with confidence.1025
Good to know
The 2025 census results are not out yet
The 2025 census (令和7年国勢調査) was conducted with a survey date of 2025-10-01. As of this writing, only the preliminary population and household count has a confirmed publication date, 2026-05-29 (as of 2026-07; confirm current figures with e-Stat, the Statistics Bureau of Japan's official statistics portal).3 That preliminary release covers total population and household counts only, not a nationality breakdown, so readers should not expect an updated foreign-population-by-prefecture table yet.
The Statistics Bureau's own published schedule puts the next release, 人口等基本集計 (Population Etc. Basic Aggregation), at "by September 2026" (令和8年9月まで). That release is scheduled to cover population, households, and housing together with the foreign-national breakdown in the same batch, the 2025 equivalent of table 44-1, rather than as a separate, later release (as of 2026-07).3 For comparison, the 2020 cycle's equivalent tabulation was published 2021-11-30, roughly 13 to 14 months after that cycle's census date.
Census figures and "foreigners in Japan" rankings you see elsewhere may not match
The most common reason a reader sees a different "foreign population by prefecture" number elsewhere is that the other source is using registry or juki figures, not the census, without saying so. Registry-based totals are generally larger than the corresponding census total for the same period, partly because the census undercounts through survey non-response and partly because of differing scope and timing.710
Even within the 2020 census alone, as covered above, two official totals exist depending on whether "not stated" nationality responses are left raw or statistically imputed, a second and more subtle source of numbers not matching that most rankings do not explain (as of 2020).11
Shimukappu Village (占冠村), Hokkaido, is a good example of a widely circulated figure that is not a census number at all. The often-quoted "36.6% foreign share" is a January 2025 figure from the municipal population register (juki), reported by a regional newspaper as part of a story on 27 municipalities nationwide exceeding a 10% foreign-resident share, tied to labor shortages in the village's ski-resort and tourism industry.26 The actual 2020 census figure was far lower: 11.1% (up from 5.0% in 2015), per the Statistics Bureau's own commentary on the results.15 The two numbers describe the same village using different data sources roughly four years apart, not a single trend line.
A low foreign-population area is not automatically a bad choice
A prefecture or city with a small foreign population usually means fewer expat-facing services and less English-language support nearby. It does not mean the place is a bad place to live.
Community access is a real trade-off worth weighing against other factors: cost of living, job availability, proximity to family, or simply preference for a quieter area. A low foreign-population area is a different kind of place to settle, not a worse one.
Nationality Detail Exists at the City Level, With Small-Cell Suppression
A common assumption is that named-nationality data (which specific countries a city's foreign residents come from) is only published at the prefecture level. That assumption does not hold up: the 2020 census table 44-1 is explicitly structured with named-nationality categories, Korean, Chinese, Filipino, Thai, Indonesian, Vietnamese, Indian, Nepalese, British, American, Brazilian, Peruvian, and an "Other" bucket, cross-tabulated all the way down to the municipal level.6
The real, narrower limitation is small-cell suppression, not a blanket exclusion. Specific nationality-by-municipality cells with very small counts are liable to be masked with "x" rather than withheld category-wide, a documented convention across the census's small-area products, though the exact numeric suppression threshold as applied to table 44-1 specifically is not confirmed.17 In practice, that means a reader can generally find, for example, how many Vietnamese or Filipino residents live in a given city, except where the count in that specific city-nationality combination is small enough to be masked.