In the Japan Tourism Agency’s Inbound Consumption Trend Survey for January–March 2026 (first preliminary release, published 15 April 2026), the top spot for travel spending by country and region changed hands from China to Taiwan. Taiwan posted ¥388.4bn (+22.5% year on year) against China’s ¥271.5bn (-50.4%), roughly a halving. For accommodation operators, however, the question is not who ranks first. It is where, and at what rate, the lodging demand China gave up was picked up by the increases from Taiwan and Korea.

This article cross-references the Japan Tourism Agency’s consumption survey and overnight accommodation statistics with MetroEngines Research’s estimated settled ADR to examine how the shift in the nationality mix shows up in room rates by region. The conclusion up front: nationally, the gains from Taiwan and Korea almost entirely offset the decline from China, but that offset does not hold geographically — Kyushu, Hokkaido and Okinawa are moving in the opposite direction from Osaka.

Inside the handover — China’s -¥276.3bn absorbed by Taiwan’s +¥71.3bn and Korea’s +¥35.9bn

Inbound travel spending in Japan for January–March 2026 came to ¥2.3378tn, up +2.5% year on year — a record high for a first quarter. The reason people still say “inbound has broken down” despite a rising total is that the composition beneath it has shifted violently.

China fell from ¥547.8bn to ¥271.5bn, a drop of ¥276.3bn. That decline was absorbed in shares by Taiwan (¥317.1bn to ¥388.4bn, +¥71.3bn), Korea (¥282.3bn to ¥318.2bn, +¥35.9bn), the United States (¥222.3bn to ¥259.2bn, +¥36.9bn) and Australia (¥115.1bn to ¥138.9bn, +¥23.8bn), among others. By visitor count, China fell to 1.045 million (-50.3% year on year), while Korea reached 3.055 million (+22.4%) and Taiwan 1.981 million (+24.6%), both posting double-digit growth.

Source: Japan Tourism Agency, Inbound Consumption Trend Survey, January–March 2026 (first preliminary release); compiled by the HotelBank Editorial Team

What matters for accommodation operators is that within that ¥2.3378tn, lodging was the largest category at ¥857.1bn (36.7% of the total), up +12.1% from ¥764.5bn (33.5%) a year earlier. With total spending up only +2.5%, lodging alone grew in double digits. That contrasts with shopping, which fell -12.2% from ¥671.3bn to ¥589.5bn — a quarter in which the centre of gravity of spending moved from goods toward staying and eating. Lodging remained the largest category into April–June as well, and the distance between that total and real-world room rates is examined in Lodging Is 37% of Japan’s ¥2.51T Inbound Spend: ¥10,446 vs Real ADR.

On a per-person, per-night basis, Taiwan, Korea and China sit at nearly the same lodging rate

Chinese visitors are often discussed on the premise that they spend more per head, but for lodging specifically the picture is a little different. On the Japan Tourism Agency’s estimates, lodging spend per person from China was ¥82,845, slightly above the all-nationality average of ¥81,255, while average nights stayed stood out at 16.6 (up 8.2 nights year on year). As a result, lodging spend per person per night comes to ¥4,994 — the lowest of the five major markets that can be compared.

This reflects the relatively high share of long-stay purposes among arrivals from China — study, work and visiting relatives. Restrict the view to tourism and leisure and the picture changes. On that basis, lodging spend per person per night is ¥10,285 for Taiwan, ¥10,207 for China and ¥9,916 for Korea — three markets essentially level with one another. Hong Kong at ¥12,777 and the United States at ¥13,567 sit a step above.

Source: Japan Tourism Agency, Inbound Consumption Trend Survey, January–March 2026 (first preliminary release), Reference Tables 1 and 2; compiled by the HotelBank Editorial Team

In other words, judged solely on the lodging revenue received per night, swapping Chinese guests for Taiwanese and Korean ones creates almost no downward pressure on rate. The substance of the effect is not the rate itself, but the change in nights stayed and in where those nights are spent. Organised by nationality, the breakdown looks like this.

Source: Japan Tourism Agency, Inbound Consumption Trend Survey, January–March 2026 (first preliminary release), Figure 5 and Reference Tables 1 and 2; compiled by the HotelBank Editorial Team. Lodging spend per person per night is on an all-purpose, general-visitor basis, and is published for only six segments: all nationalities, Korea, Taiwan, Hong Kong, China and the United States

The unusually high lodging spend per person from Europe, the US and Australia reflects length of stay as much as rate. France’s 24.4 nights and Australia’s 14.5 nights, converted to a per-person-per-night basis, are unlikely to look very different from Taiwan or Hong Kong. Read the other way, Korea’s 4.8 nights and Taiwan’s 6.2 nights make the amount spent per person look small, but the structure earns its guest nights through visitor volume.

Nationally, 95.7% of the guest nights are offset — Taiwan +695k and Korea +422k against China -1.167m

Measured not in yen but in total guest nights — the figure that maps directly onto hotel supply — the scale of the swap becomes clearer. In the nationality tabulation of the Japan Tourism Agency’s Overnight Accommodation Statistics Survey for February 2026 (second preliminary release), Taiwan took the top position for the first time at 2,373,260 guest nights (+41.4% year on year), followed by Korea at 2,081,950 (+25.4%) and China at 1,721,030 (-40.4%).

Working back to prior-year levels from the year-on-year rates, Taiwan added roughly 695,000 guest nights and Korea roughly 422,000. Against that combined increase of 1.117 million, China’s decline was about 1.167 million. The gains from Taiwan and Korea alone offset 95.7% of China’s decline. With additional contributions from the United States (+10.2%), Hong Kong (+11.6%), Canada (+49.6%) and the United Kingdom (+39.5%), national foreign guest nights held on to growth of +5.8% as of February.

This near-complete offset is the structure sitting behind the national aggregates of +2.5% in total spending and +12.1% in lodging spend. But the offset holds only in the national total; on a map, it does not hold at all. An organised, single-map view of which prefectures are hosting which nationalities is set out in US Now Tops Japan Inbound at 13.5% — Prefecture Host Map 2026.

Three major metro areas -1.7%, regional areas +7.8% — the offset is not happening geographically

The same statistics publish foreign guest nights split between the three major metropolitan areas (the eight prefectures of Saitama, Chiba, Tokyo, Kanagawa, Aichi, Kyoto, Osaka and Hyogo) and regional areas. For February 2026, the three major metropolitan areas were -1.7% year on year against +7.8% for regional areas. The regional share rose 5.0 points in two years, from 36.4% in February 2024 to 39.2% in February 2025 and 41.4% in February 2026.

Broken out by prefecture, the gap is sharper still. Tokyo at -0.3%, Osaka -0.5%, Kyoto -9.8%, Aichi -15.5% and Hokkaido -3.7% all came in below the prior year, while receiving regions in Kyushu and western Japan grew strongly: Saga +76.8%, Kagoshima +77.4%, Yamaguchi +52.2%, Nagasaki +22.2%, Oita +21.3% and Fukuoka +15.8%. Taiwan and Korea have dense direct-flight service into Kyushu’s airports, Kansai and Hokkaido, and their itineraries tend to be completed within a single region rather than touring several cities in a short stay. That itinerary characteristic shows up directly as the geographic skew in guest nights.

So how do those swings in guest nights translate into actual room rates? We compared MetroEngines Research’s estimated settled ADR for January–March 2026 against January–March 2025.

Source: MetroEngines Research; compiled by the HotelBank Editorial Team (N=155–1,145 properties per prefecture, as of March 2026)

Okinawa +22.6%, Hokkaido +15.6% and Fukuoka +14.2% led the table, while Osaka -4.3% and Saga -2.5% sat at the bottom. Sixteen of the 18 prefectures covered were positive. In some prefectures the growth in guest nights and the growth in ADR line up; in others they clearly diverge. Set side by side, they look like this.

Source: Japan Tourism Agency, Overnight Accommodation Statistics Survey, February 2026 (second preliminary release); MetroEngines Research; compiled by the HotelBank Editorial Team. The Japan Tourism Agency advises caution on Fukuoka and Oita, where the standard error rate exceeds 20%

Three patterns can be read from this table. First, prefectures where guest nights and ADR both rose: Fukuoka, Oita and Hyogo. Second, prefectures where guest nights fell below the prior year yet ADR rose sharply: Okinawa, Hokkaido and Tokyo. Third, prefectures where guest nights rose substantially but ADR stayed flat or lower: Saga and Kumamoto. The same phrase — “inbound went up” — reaches rate or fails to depending on which guest segments are being received and on the existing supply-demand balance.

Fukuoka and Osaka — two prefectures that traced opposite monthly curves in the same year

Comparing like-for-like across the seasonal year, Fukuoka and Osaka moved in opposite directions. Fukuoka climbed to roughly ¥12,000 in January 2026, +25.7% against January 2025, and stayed above the prior year consistently through July. Osaka, by contrast, has run below the prior year every month from February 2026 onward.

Source: MetroEngines Research; compiled by the HotelBank Editorial Team

Osaka’s drop from April onward, however, must not be explained by the change in nationality mix alone. Expo 2025 Osaka, Kansai ran from 13 April to 13 October 2025, so ADR over April–October 2025 was lifted by Expo demand. The large year-on-year declines — -22.1% in April 2026 and -32.3% in June — look as steep as they do largely because the 2025 comparison base was exceptional. The -4.3% reading confined to January–March, which the Expo does not touch, is the realistic figure for assessing the nationality swap.

Fukuoka’s curve shows January–March, which had nearly overlapped in 2024 and 2025, shifting a step higher in parallel in 2026. That shape indicates a lift in the base level from the start of the year rather than a single-month event, consistent with a year-round demand uplift supported by direct flights from Korea and Taiwan. That said, the pace of growth itself has begun to moderate recently, and how to rebuild the rate ladder is covered in detail in Fukuoka Settled ADR Slows +14.4% to +3.6% — Breakfast Premium Next?.

At ward level the gap widens further — Naha +31.7%, Naniwa-ku Osaka -5.9%

Cutting the entertainment districts long considered China-dependent down to city and ward level reveals differences that prefectural averages hide. Comparing estimated settled ADR for January–March 2026 with the same quarter a year earlier, the leaders were Naha +31.7%, Kita-ku Sapporo +19.0%, Chuo-ku Sapporo +16.1%, Hakata-ku Fukuoka +15.4% and Chuo-ku Fukuoka +14.8%.

Source: MetroEngines Research; compiled by the HotelBank Editorial Team (N=22–192 properties per ward, as of March 2026)

The Tokyo–Osaka contrast is what deserves attention. Chuo City in Tokyo, which contains Ginza, rose +13.3%, the strongest in the capital. It is an area where Chinese visitors loomed large in duty-free spending, yet on the room-rate metric it is comfortably above the prior year. Shinjuku City also gained +11.4%. Against that, Chuo-ku in Osaka, which contains Shinsaibashi and Dotonbori, fell -5.8%, and Naniwa-ku, which contains Namba, fell -5.9% — moving opposite to their Tokyo counterparts.

The factor that separates them can be framed as the substitutability of demand. Chuo City in Tokyo also functions as a destination for high lodging-spend markets such as the United States, Australia and Europe, leaving room for higher-spending segments to replace the decline in Chinese guests. The Japan Tourism Agency’s estimates of ¥13,567 for the United States and ¥12,777 for Hong Kong in per-night lodging spend on tourism trips run 20% to just over 30% above Taiwan and Korea. Chuo-ku and Naniwa-ku in Osaka, meanwhile, had in part set both supply and pricing aggressively the prior year on pre-Expo demand expectations, and an adjustment down from that level may have compounded the effect.

That Kita-ku in Osaka (Umeda) held roughly flat at +0.7% is consistent with this reading. Kita-ku carries a heavier weighting of business demand and is less exposed to shifts in the nationality mix of leisure travellers. Even within the same city, the demand portfolio determines how resilient rate is.

Demand that fills rooms and demand that reaches rate are different things — the Saga case

What the explanation “Taiwan and Korea filled in for China” most easily overlooks is that a recovery in volume does not necessarily bring a recovery in rate. Saga posted one of the strongest gains in the country, with foreign guest nights up +76.8% year on year in February 2026, yet its estimated settled ADR for January–March was ¥12,800, down -2.5% from the same quarter a year earlier. Guest nights rose more than 70% and the room rate still did not reach the prior year.

This is not unrelated to Korea’s lodging spend per person per night of ¥7,465 (all purposes), which sits below the all-nationality average of ¥7,851. Korean stays average a short 4.8 nights and include a relatively high share of price-sensitive travellers. The market delivers headcount and guest nights, but it does not lift the rate band as strongly as Taiwan, Hong Kong or the Western markets. In Kyushu’s Korean gateway prefectures, occupancy improvements tend to come first, with rate following later.

Okinawa is the reverse: guest nights were essentially flat at +1.3%, yet ADR rose +22.6%, the largest gain in the country. What worked there was likely not an increase in visitor numbers but a rising share of higher-priced plans in the mix. In other words, “how many came” and “which price band sold” need to be looked at separately.

Translated into operational practice, the question of how to fill the gap left by Chinese guests breaks into two. One is filling guest nights (who to reach, and through which channel); the other is holding rate (which price band of product to prepare for the segments that fill those nights). The growth from Taiwan and Korea solves the first with high confidence; it does not automatically solve the second. That the Japan Tourism Agency estimates per-night lodging spend on tourism trips at ¥10,285 for Taiwan and ¥9,916 for Korea, against ¥10,207 for China, shows that the room to hold the price band exists structurally. Downward pressure on rate can be treated less as a question of what demand can afford and more as a question of product design and pricing latitude.

Summary — offset nationally, redistributed on the map

Here is what the January–March 2026 data confirms. First, against a -50.4% fall in Chinese travel spending, total inbound lodging spend rose +12.1%, from ¥764.5bn to ¥857.1bn. In aggregate terms, lodging demand has not contracted.

Second, on a guest-night basis, increases of roughly 695,000 from Taiwan and 422,000 from Korea offset 95.7% of China’s decline of about 1.167 million. Nationally, the offset holds.

Third, that offset does not hold geographically. The three major metropolitan areas were -1.7% against +7.8% for regional areas, and by prefecture Kyoto -9.8% and Aichi -15.5% sit alongside Saga +76.8% and Kagoshima +77.4%. Estimated settled ADR split by direction too, with Okinawa +22.6%, Hokkaido +15.6% and Fukuoka +14.2% against Osaka -4.3%.

Fourth, on a per-person, per-night basis, and restricted to tourism and leisure trips, lodging rates are near-identical across the Taiwan, Korea and China markets. The change in nationality mix does not in itself push rates down. What pushes them down is the regional supply-demand gap created by demand moving elsewhere, and how closely pricing follows it.

Into the second half of 2026, the pace of recovery from China is the largest variable. What can be said at this point, though, is that even in areas that were heavily China-dependent, wards such as Chuo City in Tokyo have grown their rates. Taking the change in nationality mix as given, the design choice of which price band to offer which guest segments is what is producing the regional differences on the ground.

Frequently Asked Questions

Related Reading

• US Now Tops Japan Inbound at 13.5% — Prefecture Host Map 2026
• Lodging Is 37% of Japan’s ¥2.51T Inbound Spend: ¥10,446 vs Real ADR
• Did Non-China Inbound Fill China’s -58% Gap? April 2026 Stats & Regional ADR
• Fukuoka Settled ADR Slows +14.4% to +3.6% — Breakfast Premium Next?
• City Hotels: 45.9% of Guest Nights Are Foreign — Ryokan +3.7%
• Japan Inbound July 2026: 3.44M Arrivals, 41.7-42.8M Full-Year Range
• China National Day 2026: 21pt Weekday OCC Gap Across 26 Prefectures
• 60% of Municipalities Sit Below Their Prefectural ADR: 946 Checked
• JTA 2025 Annual Stats Deep Dive: Japanese -3.8%, Foreign +8.2% Structural Shift