How Many Koreans Visit Japan? Why the Numbers Differ
JNTO's final 2025 count of South Korean visitors to Japan is 9,459,711; Japan's Justice Ministry counts 9,230,881 new entries. Why they differ and what to check before citing.
In 2025, 9,459,711 South Koreans visited Japan, according to final figures from the Japan National Tourism Organization (JNTO). That's the value JNTO lists in its visitor statistics by nationality and month (Excel). For the same year, the Immigration Services Agency of Japan, under the Ministry of Justice, counted 9,230,881 new entries by South Korean nationals in its 2025 preliminary report.
Both numbers come from Japanese agencies, and both count the same nationality in the same year. They just cover different groups of people. Table 1 (第1表) of the Justice Ministry's preliminary report splits foreign entrants into "new entries" and "re-entries," and the ministry figure above counts new entries only. So where does the number Korean media report as "Koreans who went to Japan" come from?
This explainer follows the footnotes in the source documents to see who publishes Japan visitor numbers, under what definitions, and when. It answers three questions in order: how many Koreans visit Japan, why the agencies' numbers differ, and which number to cite and how. The short answer: the "Korean visitors to Japan" figure used in Korea comes from JNTO. When JNTO revised its estimates to the next stage, the numbers barely moved (in the two cases PRISM checked). Swapping in the Justice Ministry's count of new entries for the same year produced a gap in the 2% range (in 2024 and 2025). That's why, when you cite a source, you should match the agency and the definition before the release stage.
Who counts "Korean visitors to Japan"?
JNTO's numbers start from Justice Ministry statistics. A footnote in JNTO's monthly press release says its foreign visitor count is "calculated based on immigration control statistics compiled by the Ministry of Justice" (法務省集計による出入国管理統計に基づき、算出したもの). JNTO takes the ministry's data and adds or subtracts groups to fit its own definition.
Korea's own departure statistics don't break down by country. According to the Korea Tourism Organization's (KTO) statistical description for Korea Tourism Statistics, the outbound tourist series (국민 해외관광객) has three breakdowns: by sex, age and mode of transport. Its history section reads: "As filling out a 'departure card' when nationals leave the country was abolished (July 1, 2006), outbound tourist statistics by destination and by purpose were discontinued." Korea's departure data stopped tracking destinations in July 2006.
Other countries' figures fill the gap. In the same description, the KTO says it "collects and provides the Korean arrival statistics compiled and published by major countries, for the convenience of statistics users." Open the Excel attachment to its outbound tourist statistics by major destination, as of September 2026, posted November 5, and the "Source" sheet (출처) names JNTO as the source for the Japan column. In that table, the 12 monthly figures for Japan in 2025 and the annual total of 9,459,711 match JNTO's final figures down to the last digit.
Trace it back, and the Japan visitor figure cited as "Korea Tourism Organization statistics" is also JNTO's.
Who's in JNTO's count and who's left out
JNTO's whole definition fits into one paragraph, footnote 3 (注3). Here it is, unpacked.
- It starts with nationality. JNTO counts foreign regular entrants to Japan (外国人正規入国者) by nationality.
- Foreigners living in Japan are excluded. Permanent residents and others whose main country of residence is Japan (日本を主たる居住国とする永住者等の外国人) are taken out. For the Japan Tourism Agency's published "foreign travelers to Japan" figure (訪日外国人旅行者数), which follows the same structure, a footnote in the Justice Ministry report lists who's excluded: entrants under the residence statuses of permanent resident (永住者), spouse of a Japanese national (日本人の配偶者等), spouse of a permanent resident (永住者の配偶者等) and long-term resident (定住者), plus special permanent residents.
- Cruise and transit landings are added. Foreigners granted port-of-call landings, transit landings and cruise-ship tourist landings (寄港地上陸、通過上陸、船舶観光上陸) are included. Justice Ministry tables count these landing permits separately from entrants, as "special landings" (特例上陸).
- Expats and international students are included. The original reads: "Entrants and re-entrants such as expatriate staff and their families and international students are included in foreign visitors" (駐在員やその家族、留学生等の入国者・再入国者は訪日外客に含まれる). A Korean expat flying back to Tokyo after a business trip to Seoul is in this number, too.
- Crew are excluded. "The above foreign visitors do not include crew" (上記の訪日外客には乗員は含まれない).
So calling the entire JNTO figure "tourists to Japan" doesn't match the definition, since expats and students are in it as well. To look at tourism alone, use the figures by purpose of visit. According to JNTO's release schedule, those come out at the provisional stage.
Each month's number comes out three times
JNTO publishes each month's figure in three stages: estimate (推計値), provisional (暫定値) and final (確定値). According to its release schedule, estimates give only the total and rounded figures (概数) for some nationalities, while provisional figures give each nationality down to the last digit. Final figures arrive the following year. A footnote in the October 21, 2026, release says "the 2025 values are final" (2025年の値は確定値である), and the tables mark estimates as italic figures with an asterisk (*印の斜体部分).
The revisions between stages were small. In the May 2026 estimate, Korea stood at 951,300. The provisional figure was 951,328, a difference of 28 (PRISM calculation). For Korea's 2025 annual total, the estimate released January 21 was 9,459,600 and the final figure was 9,459,711, up 111 (PRISM calculation).
When you look at year-on-year changes, check the stage of the comparison figure, too. The May 2026 estimate used the 2025 provisional value for the same month a year earlier (上記の2025年の数値は暫定値、2026年の数値は推計値である).
The KTO's destination spreadsheet mixes stages as well. Its 2025 Japan figures are JNTO's final values. August 2026's 850,500 and September's 842,300 are JNTO estimates carried over unchanged, rounded to the hundreds.
Korea, Taiwan and China each count departures their own way
Japanese agencies count people coming in. Countries of departure count their own people going out. And each country counts differently.
Korea's "outbound tourists" figure is defined on e-Narajipyo (Statistics Korea's indicator portal) as "outbound tourists = total departures of nationals − nationals residing abroad + crew." The page says the formula follows recommendations from the World Tourism Organization (UNWTO). The KTO's statistical description also says it "excludes departures by nationals residing abroad (overseas Koreans)." It adds crew and leaves out overseas Koreans. The formula for the opposite direction, inbound foreign tourists, includes both crew and arriving overseas Koreans.
Taiwan assigns a destination based on a flight's first point of arrival. The introduction to the tourism statistics database of the Taiwan Tourism Administration (Ministry of Transportation and Communications) puts it this way: "Outbound data for nationals is compiled on the principle of the first stop a flight arrives at (首站). For data on nationals' travel to individual countries, refer to the official arrival figures published by each destination country" (國人出境數據係以飛航到達首站為統計原則,如需國人赴各國旅遊數據,請以各旅遊目的地國家官方公布入境數字為準). The raw data comes from departure records kept by the National Immigration Agency (Ministry of the Interior, Taiwan). The unit is person-trips (人次), which counts the number of departures.
The National Bureau of Statistics of China splits outbound travel just three ways in its 2025 statistical communiqué: total outbound trips by mainland residents, private trips within that total, and trips to Hong Kong, Macao and Taiwan. There's no breakdown by destination country. A footnote says all figures are "preliminary statistics" (初步统计数).
Same Koreans, different numbers: a definitions table
For 2025, JNTO's final count of Korean visitors is 228,830 (PRISM calculation) higher than the Justice Ministry's count of new entries (9,230,881). That gap equals 2.42% of JNTO's final figure (PRISM calculation). The two cover the same year at different stages: JNTO's number is final, and the ministry's is preliminary. In 2024, the gap was also in the 2% range. JNTO's 8,817,765 (from the October 2026 Excel file) and the ministry's 8,629,184 new entries (the 2024 value in the 2025 preliminary report) differ by 188,581, or 2.14% of JNTO's figure (both PRISM calculations).
Put the two definitions side by side, and you can see where gaps could come from. JNTO includes re-entrants and special landings, which aren't in the ministry's new-entry count, and it excludes permanent residents and others living in Japan. For all foreign nationals in 2025, the ministry's Table 1 shows 39,184,546 new entries, 3,246,369 re-entries and 4,048,112 special landings. A reference column in the same table lists the Japan Tourism Agency's published "foreign travelers to Japan" figure at 42,683,600. That's the same value as JNTO's 2025 annual estimate for all visitors.
PRISM doesn't know how much of the 228,830 gap for Koreans comes from re-entries and how much from special landings. The Justice Ministry's preliminary report has no tables of re-entries or special landings by nationality. PRISM couldn't check the effect of excluding permanent residents and others by nationality, either. Ministry data on re-entrants and special landings by nationality would make it possible to split the gap. Until then, the explanation that the gap comes from definitions is an interpretation resting on the two agencies' original definitions.
For Taiwan, the gap is small. According to the Executive Yuan's national overview (the Executive Yuan is Taiwan's cabinet), Taiwanese nationals made 16,849,683 outbound person-trips in 2024. Of those, 6,006,116 person-trips (35.65%) had Japan as their first stop. JNTO counted 6,044,316 visitors of Taiwanese nationality that year. The two numbers use different units (people and person-trips) and different definitions (nationality and first stop). The difference between them is 38,200, or 0.63% of JNTO's figure (both PRISM calculations). Because it subtracts person-trips from people, we've left it without a unit.
The "Korean visitors to Japan" figure used in Korea is JNTO's count, carried over. For the same 2025 Korean visitors, the Justice Ministry's "new entries" figure is 228,830 (2.42%) lower than JNTO's final count. In the two cases PRISM checked, JNTO's revisions to the Korean figure from one stage to the next were +111 for the 2025 annual total (estimate to final) and +28 for May 2026 (estimate to provisional). Taking the single 2025 annual pair for Korea (JNTO final vs. ministry preliminary), the gap between agencies and definitions (2.42%) is roughly 2,000 times the estimate-to-final revision (0.0012%), by PRISM's calculation. With only two revision cases checked, we can't go as far as saying estimates are always this close. Within what we checked, replacing an estimate with a provisional or final figure barely moved the number, and switching from JNTO's figure to the ministry's new-entry count opened a gap in the 2% range. That's the basis for matching the agency and the definition before the stage when you carry a number into a report or an article.
The table below brings together the definitions of six of the statistics in this article. "Not found in source" marks cells where PRISM couldn't find the relevant wording in the source documents it opened.
| Statistic | Who counts | Basis | Crew | Cruise/transit landings | Residents and overseas nationals | Destination | Release cycle |
|---|---|---|---|---|---|---|---|
| JNTO foreign visitors (訪日外客数) | Japan (arrival) | Foreign regular entrants by nationality (built from Justice Ministry statistics) | Excluded | Included | Permanent residents etc. living in Japan excluded; re-entries by expats and students included | Japan | Monthly estimate → provisional → final the following year |
| Japan Justice Ministry new entries | Japan (arrival) | Foreign new entrants by nationality | Not found in source | Counted separately as special landings | Re-entrants counted separately | Japan | Not found in source (this article uses the 2025 preliminary report) |
| Japan Tourism Agency Inbound Consumption Survey | Japan (arrival) | Spending per person × number of travelers = total spending (general-visitor table; a monetary amount) | Not found in source | Not found in source | Not found in source | Japan | Quarterly (first preliminary report) |
| Korea outbound tourists | Korea (departure) | Total departures of nationals − nationals residing abroad + crew | Included | Not found in source | Nationals residing abroad (overseas Koreans) excluded | None since July 2006 (KTO compiles other countries' statistics) | End of each month (previous month) |
| Taiwan Tourism Administration outbound nationals | Taiwan (departure) | Outbound person-trips by nationals (National Immigration Agency departure records) | Not found in source | Not found in source | Not found in source | First arrival point of the flight (首站) | Not found in source |
| National Bureau of Statistics of China communiqué | China (departure) | Outbound person-trips by mainland residents | Not found in source | Not found in source | Mainland residents only | No country breakdown (Hong Kong, Macao and Taiwan only) | Annual (preliminary figures) |
Does "demand" mean people or money?
The Japan Tourism Agency's Inbound Consumption Survey, first preliminary report for July–September 2026 measures how much foreign visitors spent in Japan. For that quarter, it estimated foreign visitors' travel spending at ¥2.2076 trillion. By country and region, Taiwan ranked first at ¥381.8 billion (a 17.3% share). Korea came in at ¥293.8 billion (13.3%).
The general-visitor table (図表4) shows how the math works. For Korea, per-person travel spending of ¥113,959 multiplied by 2.576 million travelers gives total spending of ¥293.6 billion. That ¥113,959 per person is the lowest of any country or region in 図表4. The next lowest, the Philippines, is ¥160,522, and per-person spending across all general visitors is ¥234,000.
The same Korean spending figure differs between two tables: ¥293.8 billion in the totals table (図表3) and ¥293.6 billion in the general-visitor table (図表4). Name the table when you cite it. The Japan Tourism Agency has run the survey under its current name since the April–June 2024 quarter. It says it hasn't changed the survey items or estimation methods, so the data can still be compared with earlier releases (調査項目や推計方法等は変更していないため、2024年1-3月期以前のデータとの比較は可能である).
Five things to check when you see a number
When you see a headline like "X million travelers to Japan" or "Japan's X% share of overseas travel," or copy such a figure into a report, check these five things.
- Which country's agency produced it? Japanese agencies (JNTO, the Justice Ministry) count people arriving in Japan. Korean, Taiwanese and Chinese agencies count people leaving their own countries.
- Is it based on nationality or residence? Check the footnotes to see whether overseas Koreans and permanent residents are in or out. That's the "Residents and overseas nationals" column in the definitions table.
- Are crew, cruise passengers and transit landings included? Agencies split on the same item: included, excluded or counted separately.
- Which stage is it: estimate, provisional or final? A value rounded to the hundreds may be a JNTO estimate (概数). A value down to the last digit is provisional or final.
- What's the unit? Check whether it counts people or person-trips, and whether it's people or money.
Which number to use, and when the next ones come out
Choose the number by purpose. Each one is accurate within its own agency's definition.
- To match Korean media reports or KTO tables, use JNTO's figure.
- To compare two numbers, compare values from the same agency at the same stage.
- For annual figures, use JNTO's 2025 final count.
In the source line, include the agency, the stage and the release date. For example: "JNTO 2025 final figures (October 21, 2026, release)." If someone in a meeting asks, "Where's this number from?", that one line leads straight back to the original document.
If you're writing a report in December, check the stage of each monthly value first. Based on JNTO's release schedule, from December 1 to 15, figures through August 2026 are provisional and September and October are estimates. After the December 16 release, figures through September are provisional, and October and November are estimates. The 2025 annual total is final.
Here's when the next numbers come out. JNTO release times in the table are Japan time, which is the same as Korea time.
| Date | What comes out | Source |
|---|---|---|
| Wednesday, November 18, 2026, 4:15 p.m. | JNTO October 2026 estimate, August provisional | JNTO release schedule |
| Wednesday, December 16, 2026, 4:15 p.m. | JNTO November 2026 estimate, September provisional | JNTO release schedule |
| After December 16 | JNTO schedule not yet published ("to be announced later," 後日ご案内いたします) | JNTO release schedule |
| Around the 6th of each month | KTO outbound tourist statistics by major destination (data from two months earlier) | KTO statistical description |
| End of each month | Korea Tourism Statistics (previous month) | KTO statistical description |
JNTO adds that the schedule "is tentative and may change" (予定のため、変更となる場合があります).
This article is based on source documents published through November 11, 2026. We update the numbers and definitions quarterly, and the next update will follow JNTO's 2026 annual estimate in January 2027. How much visitor numbers to Japan rose or fell, and why, is covered in our September 21 article and the November 18 update. All English translations of Japanese, Chinese and Korean source text in this article are PRISM translations.
This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.
Related Articles
Korea's CSAT checks every test-taker's glasses. Using smart glasses means a one-year suspension by law; 10 years needs a law change. Japan, Taiwan compared.
Chiikawa the Movie took 29.8% of Korea's box office revenue in week 44 (Oct. 26–Nov. 1), above 116 Korean films combined (28.7%). Admissions went the other way.
Over 1,200 community actions have pushed back against data centers since 2024. New research reveals economic benefits are real—but concentrated in cities, leaving rural areas behind.
Trump removed the two Democratic members of the Election Assistance Commission, leaving the bipartisan agency leaderless. A Supreme Court ruling made it legal. Here's why a non-crisis can still be a warning sign.
Thoughts
Share your thoughts on this article
Sign in to join the conversation