“Visitor arrivals” (訪日外客数) is the number of foreign nationals who entered Japan, published monthly by the Japan National Tourism Organization (JNTO). It covers temporary visitors such as tourists, business travelers and people visiting relatives, and excludes permanent residents and similar categories. It is estimated from the Ministry of Justice’s immigration statistics and revised in three stages: estimate → provisional → final. As of writing (October 18, 2026), the latest published month is August 2026, with an estimate of 3,098,900 arrivals, down 9.6% YoY (JNTO, released September 16, 2026). China was 418,000, down 59.0% YoY. The September estimate is scheduled for release at 16:15 on Wednesday, October 21, 2026, and had not been published at the time of writing. This article explains the definition and scope of the statistic, the size of revisions across its three stages, the release schedule, and how to read the breakdowns by nationality and purpose of visit, and then sets out the steps hotel investors and operators can use to convert “headcount” into “lodging demand (guest nights)”.

What Are Visitor Arrivals? Who Is Counted and Who Is Excluded

JNTO’s press releases define visitor arrivals as figures “calculated based on the immigration statistics compiled by the Ministry of Justice.” The starting point is foreign nationals admitted through regular entry. From that number, foreign permanent residents and others whose main country of residence is Japan are subtracted, and foreign nationals granted shore passes (port-of-call landings, transit landings and cruise-ship tourist landings) are added. In other words, the concept is close to “the number of times foreigners who do not live in Japan entered the country.”

The point readers most often misread is how expatriate staff, their families and international students are treated. Even though these people live in Japan, they do not fall under permanent residents and similar categories, so they are counted as visitor arrivals each time they enter or re-enter. Aircraft and ship crews, on the other hand, are not included. And because JNTO counts “one entry procedure as one person,” the same person entering twice in a month is counted as two. It is more accurate to read visitor arrivals as “the number of entries” rather than “the number of individual people.”

Source: Compiled by HotelBank Editorial Team from notes in the Japan National Tourism Organization (JNTO) press release “Visitor Arrivals (August 2026 Estimate)” (September 16, 2026) and JNTO “FAQ (Statistics and Data)”

Note also that the “number of foreign entrants” published by the Immigration Services Agency does not match JNTO’s visitor arrivals. JNTO’s own statistics page warns readers about the difference: definitional differences such as excluding permanent residents and adding shore-pass passengers translate directly into different numbers. When a news report or document cites a “number of foreign visitors to Japan,” always check the source line to see which statistic it is.

Estimate, Provisional and Final Figures: What Changes Across the Three Stages

Visitor arrivals for a given month are published three times. According to JNTO’s FAQ, the estimate is an approximate figure released in the month after the target month and given to the hundreds place; the provisional figure is a near-final number released two months after the estimate and given to the single unit; and the final figure is the definitive value calculated once the Ministry of Justice finalizes its annual totals, from July of the following year onward. In practice, what matters is not only that precision improves with each release but that the granularity of the published breakdowns changes as well.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrivals and Japanese Departures: Press Release Dates (Scheduled)” and JNTO “FAQ (Statistics and Data)”

So how much do the estimate and the final figure differ? Lining up JNTO’s past estimates against the provisional and final figures in later releases, revisions to the total amount to only a few hundred people. For the 2025 annual total, the estimate released on January 21, 2026 was 42,683,600, the provisional figure in the July 15 release was 42,683,837, and the final figure in the September 16 release was 42,683,301. The difference from estimate to final was −299 people, less than 0.001%.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) press releases (October 15, 2025; January 21, July 15 and September 16, 2026). Differences are between published values

Revisions to the total are this small presumably because most of the Ministry of Justice’s entry data is already reflected at the estimate stage. The YoY change in the estimate (−9.6% for August 2026) can therefore be expected to settle at almost the same level in the provisional and final figures. At the estimate stage, however, nationality data covers only rounded figures for some markets, and the breakdown by purpose has to wait until the provisional figure. For documents that discuss market composition in detail, it is safer to use the provisional figures rather than the estimates.

By contrast, it is the JTA’s Overnight Travel Statistics that see large revisions. Foreign guest nights for July 2026 were 13.58 million in the first preliminary release on August 31 and 13.31 million (13,314,610) in the second preliminary release on September 30, a downward revision of about 270,000 guest nights, or −2.0%. That is two orders of magnitude larger than the roughly 0.01% revisions to visitor arrivals. When matching headcounts against guest nights, assume that the preliminary guest-night figures are the ones more likely to move. How the first preliminary, second preliminary and final figures work in the overnight travel statistics, with real examples of preliminary-versus-final differences, is covered in our Overnight Travel Statistics Survey Guide.

Release Schedule: Wednesdays at 16:15, Mid to Late Each Month

In recent years, JNTO has released visitor arrivals at 16:15 on a Wednesday in the middle to latter part of each month. The last three releases were July 15, 2026 (June estimate), August 19 (July estimate) and September 16 (August estimate and June provisional). According to JNTO’s published schedule, the next releases are October 21, November 18 and December 16. Remembering that each release date brings both “the estimate for the latest month” and “the provisional figure for two months earlier” makes it easy to know which month’s numbers are being updated.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrivals and Japanese Departures: Press Release Dates (Scheduled)” (September–December 2026) and monthly press releases. Target months of the provisional figures for the July and August releases follow the rule of “two months after the estimate.” Schedules are subject to change

For anyone matching arrivals against lodging demand, it helps to line up the JTA’s release dates as well. According to the JTA’s release schedule, the Overnight Travel Statistics Survey will publish the second preliminary figures for August 2026 and the first preliminary figures for September on October 30, and the second preliminary figures for September on November 30. The first preliminary release of the July–September quarter of the International Visitor Survey, which covers foreign visitors’ average length of stay and spending, is scheduled for October 21, the same day as JNTO’s September estimate. Headcount (JNTO) comes out first, and guest nights (JTA) follow one to two weeks later. With that lag in mind, simply remember: September headcount on October 21, September guest nights on October 30.

Latest Figures: August 2026 at 3,098,900, Down 9.6% YoY

Visitor arrivals in August 2026 were 3,098,900, 9.6% below the 3,428,406 (final) recorded in the same month a year earlier. JNTO explained that while East Asia, Europe/the Americas/Australia and the Middle East saw stronger travel demand around summer holidays and public holidays, this is a quieter season mainly for Southeast Asia, and flight reductions and typhoon-related cancellations weighed on some markets. The January–August cumulative total was 27,626,300, down 2.7% YoY.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) press releases. 2024 and 2025 are final figures; January–June 2026 are provisional; July and August are estimates

Overlaying the monthly trend year by year, 2026 exceeded the prior year in February, March and July, but ran below the 2025 line in January, April–June and August, with the gap widening in August. The level itself, however, was above 2024 in every month from January to August. The figures include a rebound effect from a record prior year, so it would be premature to read market contraction from the YoY decline alone.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrivals (August 2026 Estimate)” (September 16, 2026). Selected major markets

By market, South Korea ranked first at 850,500 (+28.7%), followed by Taiwan at 666,000 (+7.3%), China at 418,000 (−59.0%) and Hong Kong at 247,300 (+9.4%). Fourteen markets set records for August: South Korea, Taiwan, Hong Kong, Malaysia, Indonesia, India, the United States, Mexico, France, Italy, Spain, Russia, the Nordic region and the Middle East. Italy and Spain also set all-time single-month records. For China, JNTO cited the Chinese government’s advisory urging citizens to avoid travel to Japan and the impact of flight reductions. The decline from China alone (about 600,000) exceeded the drop in the total (about 330,000), which means markets other than China grew by about 270,000 combined.

Reading the Nationality and Purpose Breakdowns: What the Estimate Shows and What Must Wait for the Provisional Figure

Two premises matter when reading figures by nationality or region. First, the tally is based on passport nationality, not on the country or region of departure. Travelers of other nationalities living in Europe, or travelers who arrived via a third country, are counted under their nationality. Second, the nationality breakdown published with the estimate is rounded (to the nearest 100); nationality figures to the single person, and the breakdown by purpose, appear only with the provisional figure. JNTO’s releases list 23 markets/regions plus “Others,” but detailed market comparisons are best redone with the provisional figures.

The purpose breakdown appears in the provisional-figure table as “of which, tourists.” Summing the provisional figures for January–June 2026, tourists accounted for 19,169,708 of 21,085,262 visitor arrivals, or 90.9%. The composition was almost the same as a year earlier (19,635,967 of 21,518,575, or 91.3%); tourists fell 2.4% YoY versus 2.0% for the total, so tourists declined slightly more. Non-tourist arrivals include business travelers and “Others,” which covers study, technical intern training and the like. JNTO notes for Vietnam that “it should be noted that this is a market with many ‘other’ visitors, including students and technical interns,” which shows that in some markets growth in visitor arrivals does not translate directly into growth in lodging demand.

Source: Compiled by HotelBank Editorial Team from the 2026 visitor arrivals and Japanese departures table (provisional figures) in Japan National Tourism Organization (JNTO) “Visitor Arrivals (August 2026 Estimate)”. Shares and growth rates are calculated from published values

Visitor Arrivals vs Foreign Guest Nights (Japan Tourism Agency)

In hotel demand analysis, JNTO’s visitor arrivals and the JTA’s foreign guest nights are easily confused. The former is “the number of people who entered the country”; the latter is “the cumulative number of guests who stayed at lodging facilities.” They differ in both counting unit and population. And only the latter is available by prefecture. JNTO’s FAQ states that it does not provide visitor arrival statistics by prefecture and directs users to the JTA’s Overnight Travel Statistics Survey and International Visitor Survey.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) press releases and FAQ, and Japan Tourism Agency “Overnight Travel Statistics Survey” and “Tourism Statistics Release Schedule”

The practical implication is clear. To know how many inbound visitors are staying in your area, you need to look not at visitor arrivals but at the JTA’s foreign guest nights by prefecture. The basic two-step approach is to use visitor arrivals as a leading indicator that quickly shows the national “inflow at the entrance,” and then confirm regional lodging demand in the overnight travel statistics released a month later. Differences in nights per visitor by prefecture are also covered in Record Inbound Arrivals, Fewer Guest Nights: 3.96 Nights Per Visitor.

How to Convert Visitor Arrivals into Lodging Demand

The basic formula for converting visitor arrivals into hotel demand is “headcount × average nights = guest nights.” Bringing in an average length of stay from outside adds assumptions, however, so the robust approach is to first take ratios between published values and interpret from there. Proceed in the following four steps.

Source: Compiled by HotelBank Editorial Team

Applying these steps to January–August 2026 makes the relationship between headcount and guest nights clear. Visitor arrivals were down 2.7% cumulatively for January–August, while foreign guest nights were 109.2 million, down 8.7% (August is the first preliminary figure), so guest nights fell about 6 points more. Month by month, the gap was wide at 10.9 points in January, 7.2 points in June and 6.6 points in July, but in August visitor arrivals were −9.6% and foreign guest nights −10.3%, narrowing the gap to 0.7 points.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrival Statistics” and Japan Tourism Agency “Overnight Travel Statistics Survey”. Visitor arrivals for July and August 2026 are estimates; guest nights for January–July 2026 are second preliminary figures and August is the first preliminary figure; 2025 figures are final for both. The JTA changed its stratification criteria starting with the January 2026 survey

Guest nights per visitor, dividing figures for the same months, fell from 4.21 in January–August 2025 to 3.95 in the same period of 2026. Had the ratio stayed at 4.21 in 2026, foreign guest nights for January–August would have been about 116.4 million, roughly 7.2 million more than the actual figure. That number shows the size of the “extra” decline: guest nights fell by more than the drop in headcount alone would explain.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrival Statistics” and Japan Tourism Agency “Overnight Travel Statistics Survey”. Guest nights per visitor is a simple ratio of the two statistics’ published values and does not equal the actual average length of stay

Because the denominator of this ratio also includes entrants who do not use lodging facilities (cruise passengers on port calls, people staying with relatives or friends, and people staying at facilities outside the survey), it is not the actual average length of stay itself. Even so, comparing it with the prior year on the same definition tracks changes in how many nights each inbound visitor spends at hotels and ryokan. In monthly terms, August 2026 was 4.08, nearly level with 4.11 a year earlier, recovering from 3.87 in July (4.14 a year earlier).

Reading Years When Headcount and Guest Nights Diverge: Market Mix Changes Length of Stay

The divergence of headcount and guest nights was already visible in summer 2026. Our article July Foreign Guest Nights −4.6% in Japan: China Is 93% of the Drop showed that July visitor arrivals reached 3,442,100, a record for July, while foreign guest nights in the first preliminary release fell 4.6% YoY (revised to −6.5% in the second preliminary release). Guest nights can fall even when headcount holds steady because nights per visitor vary widely depending on which markets the travelers come from.

Source: Compiled by HotelBank Editorial Team from Japan National Tourism Organization (JNTO) “Visitor Arrivals (August 2026 Estimate)” and Japan Tourism Agency “Overnight Travel Statistics Survey” August 2026 first preliminary release, Table 2 (foreign guest nights by nationality, facilities with 20+ rooms, total 11,151,890 guest nights). The two statistics have different populations and the guest-night side covers only facilities with 20+ rooms, so the ratios are for reference

Taking market-by-market ratios from the August 2026 published values, short-haul markets have smaller ratios (South Korea 1.90, Hong Kong 2.53, Taiwan 3.05) and long-haul markets larger ones (United States 4.98, Australia 5.01, France 7.25, Italy 8.05, Spain 8.50). The guest-night side is limited to facilities with 20 or more rooms, so on an all-facility basis the ratios would be larger, but the direction of the differences between markets is clear. The same increase of 100,000 visitors would equal about 190,000 guest nights of lodging demand from South Korea, but more than 700,000 guest nights from the major European countries.

Most of the divergence between headcount and guest nights in 2026 can be explained by this shift in mix. China, at around four nights per visitor, fell by nearly 60% YoY, and South Korea, Taiwan and Hong Kong, at one to three nights, filled the gap in headcount. Total headcount approaches the prior-year level, but because arrivals are being replaced by shorter-stay markets, guest nights do not recover as much as headcount. The narrowing of the gap to 0.7 points in August can be read as the result of long-stay markets such as Europe and the Middle East setting August records, which, together with growth in short-haul markets, offset the erosion in nights.

Another point that cannot be overlooked is a change in the JTA’s survey design. The JTA says it changed its stratification criteria from “number of employees” to “number of rooms” starting with the January 2026 survey, and notes that YoY comparisons may include the effect of this change. Also, the latest month’s guest nights are first preliminary figures and, as July showed, can move by a few percent in the following month’s second preliminary release. When discussing the gap between headcount and guest nights, treat a gap of a few points with some latitude, as it may also reflect factors on the statistics side.

For operators and investors, three practical steps stand out. First, break down foreign stays in your own area by nationality and understand the mix of long-stay and short-stay markets. Second, use the visitor arrival release (mid to late each month) to anticipate changes in the national market mix, and confirm the spillover to your area one to two weeks later in the JTA’s guest nights by prefecture. Third, when the share of short-haul markets rises, design sales around plans that encourage multi-night stays and around short-stay weekend demand, making up for shorter stays with higher room turnover.

Frequently Asked Questions

Summary

Visitor arrivals are “the number of entries by foreigners who do not live in Japan,” calculated by JNTO from the Ministry of Justice’s immigration statistics and revised in three stages: estimate, provisional and final. Revisions to the total have been small in practice, within ±0.02%, so the overall direction can be read at the estimate stage. The latest figure, for August 2026, was 3,098,900, down 9.6% YoY, and the September figure is scheduled for release on October 21.

What drives hotel demand, however, is guest nights rather than headcount. In January–August 2026 the gap widened, with headcount down 2.7% and guest nights down 8.7%, but most of it can be explained by a shift in mix from long-stay to short-stay markets, and in August the gap narrowed to 0.7 points. Check visitor arrivals every month as a leading indicator, then confirm the spillover to your area with the JTA’s guest nights by prefecture and nationality. Combining these two statistics lets you reflect changes in inbound demand in sales plans sooner and more accurately.