Conclusion — the price gap is about 1.7x, but the margin comes from a change in who carries the major repairs, not from pricing

The conclusion first. Across the properties we track whose names include “Kokumin-Shukusha” and for which continuous OTA sales could be confirmed over the past year — 38 properties — the estimated settled ADR was a median of ¥8,800 (Q1 ¥7,400 / Q3 ¥13,500). Average room count was 28.9, small for a hotel investment target.

By contrast, in Chugoku/Shikoku, where this stock is most densely concentrated, the estimated settled ADR of 280 ryokan and resort hotels with 20 or more rooms was a median of ¥15,100, with Q3 at ¥22,200 and the 90th percentile at ¥31,200. Median to median that is a gap of roughly 1.7x; reaching the upper quartile would make it about 2.5x.

It would be a mistake, however, to read that difference straight across as “room to raise prices.” Public properties price low not through weak management but because Kokumin-Shukusha were established for a policy purpose: to provide sound recreation at low cost. The legal basis is a 1963 directive from the Vice-Minister of Health and Welfare (the Guidelines for the Establishment and Operation of Public Kokumin-Shukusha), and from 1993 onward the equivalent guidelines issued by the Vice-Minister of the Environment. The price level is simply what that institutional design produces. What an investor should therefore examine is not “why are they cheap,” but two things only: how far the price moves the moment the property leaves that framework, and how much capital expenditure that move requires.

Distribution of the surviving stock — 30% of the 43 properties sit in Chugoku/Shikoku; by name, 106 properties and 2,673 rooms nationwide

By the Public Kokumin-Shukusha Association roster, there are 43 public Kokumin-Shukusha. The regional breakdown is Hokkaido/Tohoku 6, Kanto/Chubu 11, Hokuriku/Kansai 9, Chugoku/Shikoku 13, Kyushu 4. Chugoku/Shikoku accounts for 30.2% of the total, followed by Kanto/Chubu at 25.6%. This stock is not skewed toward the three major metropolitan areas; if anything it survives most thickly in the regional areas now losing population.

Searching our database for properties whose names include “Kokumin-Shukusha” — including those whose name records remain after suspension or closure — returns 106 properties and 2,673 rooms in total. Because the set mixes privately run Kokumin-Shukusha and properties that kept only the name after moving to private operation, the denominator is larger than the association roster’s 43. The regional distribution is Chugoku/Shikoku 28, Kanto/Chubu 28, Hokuriku/Kansai 19, Kyushu 17 and Hokkaido/Tohoku 14, broadly consistent with the roster’s pattern. Room counts are small: a median of 22 and an average of 25.2.

That small scale cuts both ways for deal-making. With an investment of only a few hundred million yen per property, a fund investing alone is inefficient; for a regional bank or a local operating company, however, the size is within reach, and there is room to bundle several properties into a portfolio. In practice, a profile of 28.9 rooms on average and an estimated ADR of ¥8,800 means annual room revenue at a single property does not reach ¥100 million.

Distribution of actual rates — a median estimated ADR of ¥8,800 across 38 Kokumin-Shukusha, against an upper quartile of ¥22,200 for regional ryokan

This is the core of the article: using our estimated settled ADR to quantify where municipally owned stock stands today, and how far that sits from the price band private operators achieve in the same trade area.

Start with the municipally owned side. Of the 106 name-matched properties, 38 had an observable estimated settled ADR for six months or more over the past year (August 2025 to August 2026). The distribution: Q1 ¥7,400, median ¥8,800, Q3 ¥13,500, 90th percentile ¥15,400. Fifteen properties sit below ¥8,000 and nine between ¥8,000 and ¥12,000, so a little over 60% fall under ¥12,000. The median guest review score is 4.21 (N=37 properties) — hardly low, which suggests pricing is restrained relative to how guests rate the experience.

For comparison, we took the estimated settled ADR of 280 ryokan and resort hotels with 20 or more rooms across the nine prefectures of Chugoku/Shikoku, where this stock is most densely concentrated. Q1 ¥9,900, median ¥15,100, Q3 ¥22,200, 90th percentile ¥31,200.

What is striking is that property size does not explain the gap. Splitting the 280 Chugoku/Shikoku properties by room-count band gives medians of ¥13,900 for 20-39 rooms, ¥13,500 for 40-59, ¥18,500 for 60-99 and ¥17,600 for 100 or more. Even in the 20-39 room band — the main range for Kokumin-Shukusha — the median is ¥13,900, far above the ¥8,800 of the 38 name-matched properties. They are not cheap because they are small.

More telling is the price difference by review-score band. Stratifying the same 280 properties by guest review score, the 92 properties at 3.99 or below have a median of ¥13,500 and the 68 properties between 4.00 and 4.29 sit at ¥13,300, while the 111 properties at 4.30 or above reach ¥20,700 — about 1.5x. This is not a claim of causation between scores and prices, but the fact that, among regional ryokan and resorts, the top band of guest experience actually sustains a price level around ¥20,000 is an empirical reference point when setting a target ADR after renovation.

The demand side — regional ryokan show a two-layer structure: low actual occupancy, but high sell-through of listed inventory

Before discussing price, it is worth checking the depth of demand on two measures. Macro actual occupancy comes from the Japan Tourism Agency’s “Overnight Travel Statistics Survey.” National room occupancy for 2025 was 61.8%; by property type, business hotels ran at 75.3%, city hotels 74.2% and resort hotels 56.9%, while ryokan came in at 38.4% — far behind. Across the ryokan segment as a whole, that means six out of ten rooms sit empty through the year.

On the other hand, the occupancy we estimate from the sell-through of OTA sales inventory (July 2026, ryokan category) is high everywhere: Aomori 90.3%, Ehime 88.5%, Hokkaido 87.6%, Akita 85.8%, Nagano 85.7%. These are different metrics with different populations and definitions, and cannot be compared directly. Place them side by side, though, and a structure emerges for regional ryokan: “the inventory put on OTAs sells well, but the property’s rooms as a whole are not filling.” Exposure to sales channels, troughs in demand (weekdays, winter), or a deliberate throttling of rooms offered because of staffing constraints — one of these is very likely at work, which makes it one of the most important items to verify in due diligence.

Operating benchmarks — results disclosed by listed hotel REITs for regional and resort portfolios

When testing whether the ADR, occupancy and RevPAR assumed in a deal are plausible against industry levels, the monthly results listed hotel REITs publish are the most reliable reference point. The figures below are quoted directly as each REIT published them (we have not re-aggregated them).

Read the table this way. A portfolio centered on regional business and limited-service hotels runs at ADR ¥9,684 and OCC 84.4% (July 2026); a portfolio including resort and ryokan properties runs at roughly ADR ¥18,500 and OCC 74.2% (June 2026). The level a renovated piece of municipally owned stock should aim for sits between those two bands — realistically ADR ¥13,000-¥22,000 with OCC of 55-70%. That is consistent with our own data: an upper quartile of ¥22,200 for Chugoku/Shikoku ryokan and resorts, and a median of ¥20,700 for the 4.30-and-above review band.

The other thing the table shows is that ADR at resort and ryokan portfolios has turned negative year on year. The 2023-2025 price upcycle driven by urban inbound demand is running its course in regional resorts. For deals from 2026 onward, the financials should be built on moving the property’s own price band through product strength, not on riding a market-wide price rise.

A conversion track record — two years and ten months from transfer, and 16 months from opening, to move the estimated ADR from the ¥13,000 range to ¥20,400

The most direct answer our data can track to the question “how far does price move once a property leaves the framework” is a case in Oshu, Iwate. The former Kokumin-Shukusha Sun Hotel Koromogawa-so was transferred by the City of Oshu to the private sector in July 2022. The buyer was Nagamori Shoji Co., Ltd. (now Libera Hotels & Resorts Co., Ltd.). A renewal opening was initially planned for March 2024, but material-cost inflation stretched the construction schedule, and the property opened on 25 April 2025 as “Oshu-Hiraizumi Onsen Ryokan ITSUMU” with 42 rooms. Two years and ten months passed between transfer and opening.

The estimated settled ADR we observed after opening moved as follows. The opening month, April 2025, was around ¥13,000. It rose in stages to about ¥16,200 in November 2025 and about ¥17,100 in January 2026, reaching about ¥20,400 in August 2026. That is up 56.5% from the opening month and up 40.4% against the same month a year earlier (August 2025, around ¥14,500). The trailing 12-month average is about ¥15,600.

What deserves attention here is less the level reached than the shape of the climb. The property did not come out of the gate at a high rate; it lifted its price band in stages over a year and a half from opening. For a regional onsen ryokan, it takes one or two seasons before accumulated reviews and an occupancy record function as support for price — which means that, in a business plan, the realistic design is to set a conservative ADR for the first year of operation and reach the target level in years two to three. A plan that assumes the upper quartile of regional ryokan from the opening year is, on the evidence of this case, too optimistic.

Note too that this deal took two years and ten months from transfer to opening, with the schedule extended by a surge in construction material costs along the way. Building construction-cost inflation into the plan, discussed below, is not desk-bound conservatism — it is what actually happened.

Where the deals come from — municipalities open up their options first, through market sounding

Deals involving municipally owned lodging stock do not begin with a sale notice. In most cases, a sounding-type market survey comes first: a dialogue-based procedure in which the municipality invites opinions and proposals from the private sector before fixing the scheme. Whether you can take part in it materially shapes your position in the public tender that follows. Sourcing starts not at the notice but at the sounding.

Below are cases that have actually been published, organized within the limits of public information.

The pattern on the municipal side is clear from these three. First, deterioration and seismic performance are the direct triggers for suspension or closure. Second, municipalities present more than one scheme rather than narrowing to a single option (designated manager / sale / Park-PFI / concession / lease). Third, there are cases where continuing the lodging business is not a mandatory condition (Boyo-so is asking about uses excluding lodging). On the question of whether to extend the life of pre-1981 seismic stock through renovation or commit to rebuilding, we set out the tipping point across 588 buildings in Japan’s Pre-1981 Hotel Stock: 236k Rooms, 19.1% — Retrofit or Rebuild.

The second and third points mean considerable freedom for the proposing side. Keep the frame and recompose the use; cut room count and raise the rate; scale back lodging and pivot to day-use bathing and restaurants — all of these are on the table as far as the rules are concerned. Put the other way round, the municipality’s decision turns not on “which scheme” but on “who takes on the repair burden and the jobs.” That is where a proposal is won or lost.

Issues by contract structure — what you acquire, and what you carry

The institutional base is broad. According to a Ministry of Internal Affairs and Communications survey, as of 1 April 2021 the designated manager system had been introduced at 77,537 public facilities, of which 43.5%, or 33,708 facilities, had a “private company or similar” — a joint-stock company, an NPO and so on — as designated manager, up 3.1 points from the previous survey. At the prefectural level it covered 6,721 of 11,301 public facilities, an introduction rate of 59.5%. Lodging and recreation facilities are among the categories covered by that survey, so room for private-sector participation is already open as a matter of policy.

Deterioration and major repairs — ¥930 million over ten years, or about ¥16 million per room

The biggest variable in structuring a deal around municipally owned stock is neither price nor occupancy but capital expenditure. And here there are concrete numbers to be read off public documents.

Kokumin-Shukusha “U-no-Misaki” in Ibaraki opened in 1971; its main building, constructed in 1997, is a steel-framed reinforced concrete structure of eight storeys with a total floor area of 10,458 m², and the Country Plaza, built in 1989, is a single-storey reinforced concrete building of 741.57 m². The long-term repair plan drawn up for the property commits about ¥930 million over the next ten years, of which the prefecture bears roughly ¥650 million for frame work such as exterior wall repairs, and the designated manager, the Ibaraki Prefectural Development Corporation, bears roughly ¥280 million for building and equipment upgrades. The fiscal 2025 program is ¥41.237 million, covering roof and exterior wall repairs to the No. 3 women’s dormitory building, exterior wall repairs to the Country Plaza, and roof repairs to the gas store.

Divided simply by the 58 rooms our database records for the property, that is about ¥16 million per room over ten years, or about ¥1.6 million per room per year. And this is the cost of “maintaining the property at its current level” for a building around 30 years old; renovation spending to improve the product is on top. Anyone considering a transfer must assume that capital expenditure of this magnitude comes with the property, even where the transfer price itself is nominal. For reference, capital expenditure disclosed by listed hotel REITs for their holdings runs between ¥130,000 and ¥620,000 per room per year (Hotel Capex: ¥134k-¥624k per Room a Year — 5 REITs’ IR Disclosures), so the roughly ¥1.6 million a year required by public stock of about 30 years’ age sits well above the top of that range.

Construction-cost inflation — lodging buildings nationwide are up 62.5% per tsubo in the five years from 2019

The comparison point, the cost of building new, has itself moved sharply over these five years. Calculating cost per tsubo (3.3 m²) from lodging-use buildings in the Ministry of Land, Infrastructure, Transport and Tourism’s “Construction Starts Statistics Survey” (estimated construction cost divided by floor area), the national average went from ¥1.203 million in 2019 through ¥1.386 million in 2022 to ¥1.955 million in 2024 — up 62.5% on 2019. The number of buildings has also recovered, from 2,300 in 2019 to 2,836 in 2024.

Regional differences in that unit cost are large, however. In 2024, against ¥2.174 million per tsubo in the three major metropolitan areas, Hokkaido/Tohoku ran at ¥2.003 million, Chugoku/Shikoku at ¥1.569 million and Kyushu/Okinawa at ¥1.353 million. In the regional areas where municipally owned stock is concentrated, building new costs 20-40% less than in the three major metropolitan areas. That works against the relative advantage of reusing an existing frame, and means one cannot automatically say “it is regional, so renovating is cheaper.”

Working back from these unit costs gives a per-room cost of building new. Assuming 50 m² per room including common areas (about 15.1 tsubo), the benchmark for new construction is about ¥23.7 million per room in Chugoku/Shikoku, about ¥20.5 million in Kyushu/Okinawa and about ¥32.9 million in the three major metropolitan areas. A renovation that reuses the existing frame and comes in at ¥15 million per room is 63% of the Chugoku/Shikoku cost of building new; a light renovation at ¥5 million per room is 21%. That ratio is the clearest yardstick for the economics of acquiring existing stock. A per-room construction cost built up from room-level unit areas in actual development plans is derived in Hotel Construction Cost ¥20.67M/Room: 35.0m² Floor Area in 31 Plans, which can be read alongside the per-tsubo benchmark used here.

On land, the 2026 official land price survey showed a nationwide average gain of 2.8% across all uses (residential +2.1%, commercial +4.3%), a fifth consecutive year of increases, with gains continuing in regional areas too. But much of the municipally owned stock this article covers sits inside parks, along coastlines or in hot-spring districts, where absolute land values are low. In these deals, value comes not from unrealized gains on land but from the building frame, hot-spring rights, permits and the operating record — a fundamentally different valuation build-up from urban properties.

Development scenarios — three ways to structure a 30-room onsen property

Using the figures above, we model three scenarios for a typical municipally owned property: 30 guest rooms, hot spring, roughly 30 years old. ADR is set from the observed distribution of our estimated settled ADR, and the GOP ratio from general levels by property type. This is a simplified model only; an actual investment decision requires a detailed feasibility study.

Scenario C is recommended not for the absolute level of its yield but for its resilience to downside. With ADR cut 10% and OCC cut 5 points, Scenario C still holds a 9.4% yield (payback 10.7 years), while Scenario B falls to 7.9% (payback 12.7 years). It is more resilient despite costing more than twice as much to renovate because reducing the room count raises the product quality per room and moves the property into a price band with a higher GOP ratio.

Rather than a single sensitivity case, flexing assumed ADR and assumed OCC together on a two-axis grid sharpens the outline of the conditions under which Scenario C works. Even if ADR falls to ¥19,980 (10% below base), securing 55% occupancy holds the yield at 9.4%; conversely, if occupancy stalls at 50%, the yield drops to 9.5% even at an ADR of ¥22,200. What governs the profitability of this scenario is less the ADR level itself than how far the 24 remaining rooms can be filled across the year.

Scenario C does, however, need the time axis shown by the conversion case above built into it: two years and ten months from transfer to opening, and another year and a half from opening to the target ADR. The practical assumption is around four years from acquisition before revenue reaches planned levels, and the structure must account for cash flow and funding costs over that period.

Due diligence checklist — 12 items to add to private-property DD

Due diligence on municipally owned stock layers public-sector-specific issues on top of ordinary hotel acquisition DD. The following is limited to what should be added to a private-property DD list.

A practical sourcing sequence

In summary — operating public stock carries a different risk-return structure from a vacant site

The 43 public Kokumin-Shukusha (Hokkaido/Tohoku 6, Kanto/Chubu 11, Hokuriku/Kansai 9, Chugoku/Shikoku 13, Kyushu 4) are what survives after a fall to roughly one-eighth of the 345-property peak of fiscal 1979-1981. The profile our database returns for name-matched properties — 106 properties, 2,673 rooms, an average of 25.2 rooms — shows what that means: small property by property, but an inventory too large to ignore in aggregate, and concentrated in regional Japan.

The figures established in this article come down to three. First, the distance in price. Against a median estimated settled ADR of ¥8,800 across the 38 properties named Kokumin-Shukusha, the 280 ryokan and resorts with 20+ rooms in Chugoku/Shikoku run at a median of about ¥15,100 and an upper quartile of about ¥22,200. The band scoring 4.30 or above in guest reviews has a median of ¥20,700, an empirical reference point for a target. Second, the proof that the move happens. At a property converted after transfer to the private sector, the rate moved from the ¥13,000 range in the opening month to ¥20,400 in 16 months (up 40.4% year on year). Third, the cost of that move. One public property’s long-term repair plan runs to about ¥930 million over ten years, or about ¥16 million per room, against a cost of building new in regional Japan of ¥20-24 million per room.

The decisive difference from vacant-site and closed-school deals is that an operating record, permits, hot-spring rights and a workforce already exist. That is upside — start-up risk is lower — and at the same time a burden to be assumed: repairs, employment, designated use. The basic discipline for turning operating public stock into a deal is to assess that duality not by acquisition price but by the total of “acquisition price plus ten years of capital expenditure,” and to judge it as a ratio against the regional cost of building new.

And the deal starts at the sounding, not the notice. Notices of suspension or closure citing deterioration and seismic performance, revisions to comprehensive public facility management plans, and published sounding-type market surveys: monitoring those three continuously is the highest-return initial investment available to anyone sourcing regional deals.