Executive summary — the 20-year ceiling picks the structure

Start with the conclusion. For hotel investment, the Park-PFI 20-year ceiling bites not as a constraint on location but as a constraint on structure. Under the National Tax Agency’s statutory useful lives for buildings used as inns and hotels, reinforced concrete runs 39 years and wood frame 17 years. In a framework where the business period ends at 20 years, reinforced concrete reaches expiry with 48.7% of construction cost still undepreciated, while wood-frame and light-gauge steel structures are written off in full. The intuition that low-capex formats suit parks is underwritten not merely by cheapness but by an accounting fit: the depreciation schedule fits inside the permit period.

And there is a way out of the constraint. Article 5, Paragraph 4 of the same Urban Park Act allows an establishment and management period of up to 30 years, but only where the work is carried out as a project selected under the PFI Act. A 100-room reinforced concrete hotel that only pencils out at Tokyo- or Kanagawa-level pricing under a 20-year window sees its required ADR fall to around ¥7,500 under a 30-year window. The question of putting a hotel in an urban park is therefore not whether to use Park-PFI, but which permit period to go after.

Reading the scheme as an investment period — three exemptions and the 10-year renewal in practice

Park-PFI grants three exemptions as a package. Organised by statutory provision, they look as follows. All derive from the Urban Park Act and its Enforcement Order, and note that local government discretion (the ratio set by municipal ordinance) enters at several points.

In practice, the internal structure of that period is easily overlooked. The 7th-edition operating guidelines state explicitly that “within the validity period of the approval, an establishment and management permit may be granted for up to 20 years. However, since the establishment and management permit period is 10 years at most, the approved operator must apply to renew the permit before it expires.” In other words, a 20-year permit is not issued in one piece: you reach 20 years by renewing a 10-year permit once. Because the park administrator is obliged to grant permits applied for under an approved plan, renewal is highly probable — but a financing structure must be built on the fact that an administrative act is interposed in year 10.

A second point: the building coverage uplift is not cumulative with other exemptions. The guidelines state that “where recreation facilities, sports facilities and solicited park facilities are present, the building coverage ratio for these buildings as a whole falls within the ratio set by ordinance with reference to 2/100, plus the ratio set by ordinance with reference to 10/100.” If existing management buildings, amenity facilities or a gymnasium already consume floor area, only the remainder can be allocated to a hotel.

Adoptions peaked at 34 parks in FY2021 and have since run at 20–30 parks a year. Parks are classified by the fiscal year in which the solicitation guidelines were published, and unsuccessful or withdrawn projects are not listed as of the end of FY2025. MLIT reports a further 187 parks where use of the scheme is under consideration. For the origins of the scheme itself and movement on individual projects, see our earlier Park-PFI hotel investment analysis of Japan’s public park hotel market.

Where the 203 parks are — Tokyo 18, Fukuoka 17, six prefectures at zero

Tabulated by prefecture from published material, the 203 parks are spread across 41 prefectures; Akita, Niigata, Shimane, Tokushima, Ehime and Kumamoto have no adoptions at all. The leaders are Tokyo with 18, Fukuoka 17, Kanagawa 14, Aichi 13 and Osaka 13 — the top five alone account for 75 parks, or 36.9%. Adoption is skewed towards metropolitan areas where designated cities and special wards act as park administrators, so the map of the scheme overlaps with the map of deep lodging demand.

20 years vs statutory useful life — what structure leaves behind at expiry

This is the core of the analysis. Under the National Tax Agency’s table of useful lives for depreciable assets, buildings used as inns and hotels carry statutory useful lives ranging from 15 to 39 years depending on structure — a 2.6-fold spread. The Park-PFI business period, by contrast, is capped at 20 years regardless of structure. Overlay the two and the question of which structures fit inside the 20-year vessel answers itself mechanically.

Cost per m² rose sharply in 2024 from the prior year. SRC in particular rests on a small base of just 17 construction starts, where large, high-end projects push the unit cost up. Read the figures as material for relative comparison between structures and treat the absolute values as ranges.

Here is how to read it. Building a 100-room reinforced concrete hotel (3,500 m² gross) costs roughly ¥1.84bn, and at year 20 about ¥896m remains undepreciated. Whether the business period ends with demolition of the facility, transfer to the park administrator, or succession by another operator through re-solicitation, the investment does not work unless the project contract settles in advance where that ¥896m lands. Conversely, with wood frame or light-gauge steel the book value reaches zero at 20 years, and all that remains at expiry is a choice: demolish, or win the re-solicitation and keep operating. That difference shapes what lodging facilities in parks look like.

A 12% coverage ratio decides which parks qualify — 80% of Japan’s 115,917 parks physically cannot host a hotel

After period comes area. According to MLIT’s Urban Park Database, at the end of FY2024 (as of 31 March 2025) Japan had 115,917 urban parks and equivalent sites covering 131,242.58 hectares. The composition is extremely skewed: neighbourhood-block parks account for 92,677 sites, or 80.0% of the total. Their average area is 1,591 m², so even using the full 12% building coverage allowance, the building footprint caps out at 191 m².

A 100-room reinforced concrete hotel with 3,500 m² of gross floor area needs a 700 m² footprint at five storeys, implying a park of 5,833 m² (0.58 ha). Build to eight storeys and 3,646 m² (0.36 ha) suffices. Mapped onto the categories, anything from a neighbourhood park (1.8 ha average) upwards works on paper, while a neighbourhood-block park does not. Summing the categories at neighbourhood-park scale and above gives roughly 11,500 sites, or 9.9% of the total. Eight in ten urban parks drop out on area before the regulatory discussion even begins.

Conversely, even 191 m² in a neighbourhood-block park is enough for four or five cottages of 40 m² each. Low-capex formats suit parks not only because unit costs are low. It is because the small footprint per unit allows multiple units to be arranged within the 12% coverage vessel.

There is also room to fall outside the coverage denominator entirely. Article 4, Paragraph 1 of the Urban Park Act applies to “buildings” as defined in Article 2, Item 1 of the Building Standards Act. For structures built on vehicles, such as trailer houses, an established administrative practice since Ministry of Construction Notices No. 419 (1 December 1987) and No. 170 (31 March 1997) holds that they are not treated as buildings so long as they remain capable of being moved at any time and at will. Remove the wheels, fix the unit to the ground or leave no access route for relocation, and it will be judged a building — so this applies only where the operating requirements are genuinely met. Within those limits, choosing mobile units carries real meaning for building coverage.

Working backwards to the required ADR — ¥11,300 on a 20-year window, ¥7,500 on 30 years

Now translate these constraints into revenue terms. The assumptions are set out below. All figures are estimates and do not guarantee the viability of any individual project.

Set these required ADRs against market ADR in the prefectures where Park-PFI is most widely adopted. Trailing 12-month average estimated settled ADR compiled by MetroEngines Research (August 2025 to July 2026) is as follows.

The reading is unambiguous. Making a 100-room reinforced concrete hotel work under a 20-year Park-PFI window requires a market where a typical property sells at ¥11,300 or more. That is limited to the Tokyo, Kanagawa, Hyogo and Fukuoka class; in ¥8,000-level markets such as Aichi and Saitama, you have to come down to wood frame or light structures to get inside the payback window. Yet the same 100-room reinforced concrete building, if a 30-year window can be secured as a project selected under the PFI Act, needs only ¥7,500 — and all eight prefectures examined here move into viable territory. For hotel development in urban parks, choosing between the Park-PFI route and going after a 30-year period under the PFI Act shifts the class of asset you can put on the site by a full step.

Sensitivity analysis — the conditions for fitting 100 concrete rooms into 20 years are narrow

A required ADR is only a single point estimate. Flexing ADR and OCC to see how the simple payback period moves shows just how thin the region around the 20-year line is.

At ¥12,000 and 80% OCC, payback takes 18.9 years — leaving just 1.1 years of headroom against a 20-year permit. Payback completes exactly as the period expires, and almost nothing is left for the investor. Making reinforced concrete work on a 20-year window realistically means targeting a ¥14,000-plus price band, or building the plan around continuation through renewal and re-solicitation after the approval period ends.

For reference, consider recent actuals from a listed hotel REIT. Invincible Investment Corporation reported monthly operating results for July 2026 covering 101 domestic hotels (of the 112 domestic hotels held at the start of the fiscal period ending December 2026, those operated by Iconia Hospitality Co., Ltd. and its subsidiaries): occupancy 86.3%, ADR ¥15,100 and RevPAR ¥13,039 (year on year, occupancy +1.9pt, ADR −1.0%, RevPAR +1.2%). That a portfolio centred on limited-service hotels reaches this level suggests that in a favourable urban location, 20-year payback on reinforced concrete is within realistic reach. These are results for stabilised properties, however, and a more conservative view is warranted for a newly built asset in its ramp-up phase.

The low-capex market today — the formats among 1,315 openings in 2026 that work on unit rates

The other viable answer is the low-capex format. Within MetroEngines Research’s coverage, 1,315 lodging facilities were confirmed to have opened in 2026, averaging 26 rooms. The small-scale, low-rise unit formats that could sit in a park break down as follows.

Cottages, glamping and lodges together come to just 67 openings, but their significance as an investment structure is large. A single wood-frame cottage of 40 m² costs roughly ¥17.17m to build including designated park facility costs. To reach simple payback in 20 years, the required nightly rate per unit is ¥13,400 at 55% OCC, or ¥11,300 at 65% OCC, on a full-year business-plan basis. Whole-unit rentals for four to six guests routinely price at ¥25,000 to ¥40,000 per unit, so the payback range is reachable even at half occupancy — in sharp contrast to a reinforced concrete hotel that cannot get there without assuming 80% OCC.

Wood frame also reaches the end of its statutory useful life at 17 years, and wood frame with mortar finish at 15. Depreciation finishes inside the 20-year business period, leaving no undepreciated balance at expiry. The advantage of the low-capex format lies less in being cheap than in having depreciation, payback and permit period all fit into the same 20-year vessel.

Financing — what it means to hold neither freehold nor leasehold

What an operator obtains under Park-PFI is not a right in land. It is a standing, based on an administrative disposition in the form of an establishment and management permit, to install and manage park facilities. Three practical issues follow.

First, only the building can be pledged as security. Ownership of the solicited park facilities vests in the operator, so a mortgage over the building itself is conceivable, but the site belongs to the park administrator — in most cases a local government — and cannot be taken as collateral. In ordinary hotel development, land accounts for the bulk of collateral value, so leverage has to be structured differently at the same project scale.

Second, succession to the operator’s standing requires the park administrator’s approval. Article 5-8 of the Urban Park Act provides that a general successor to the party that submitted the approved plan, and a person who acquires ownership or other necessary title to the solicited park facilities established and managed under the approved plan, may succeed to the approved standing with the park administrator’s approval. Put the other way round, a third party who acquires the building through enforcement of a security interest cannot necessarily continue the business: an administrative judgment in the form of approval is interposed. In a project finance structure, how that approval process is reflected in the project contract and the direct agreement becomes the practical focus.

Third, measures for business failure are required in advance. The operating guidelines list, among the matters desirable to include in solicitation guidelines, “where operation of the solicited park facilities by the approved operator becomes difficult, succeeding the business to another private operator with the park administrator’s approval under Article 5-8, or taking a deposit at the start of the business to be applied to the cost of removing the solicited park facilities”. Where the design calls for a removal-cost deposit up front, that amount has to be booked as a cash outflow at commencement.

These are constraints, but they should be read together with the benefits on the other side. Zero land acquisition cost — or no land rent, replaced by a park usage fee — is substantial upside in a market where land prices keep rising. This connects directly to the fixed-term business leasehold structure examined in our analysis of leasehold hotel development and supportable ground rent across nine markets, where working backwards to the maximum supportable ground rent showed viability up to 1.38 times the land price that the freehold approach could bear. Park-PFI sits on the same continuum, positioned as a scheme that compresses land cost further in exchange for surrendering business period and collateral structure.

Summary of viability conditions

Park-PFI is not a scheme that lets you build anything you like in a park. It is a scheme in which a 20-year period ceiling and a 12% building coverage ratio sharply narrow the structure, scale and price band of the asset you can put on the site. That is precisely why the room for design is so clearly defined. Choose a low-capex, short-useful-life format and fit it cleanly into the 20-year vessel; push reinforced concrete payback into 20 years in a high-ADR market; or go after the 30-year period itself as a project selected under the PFI Act. Deciding between those three branches at the outset is the starting point for anyone approaching urban parks as a third category of development site.

Frequently asked questions

Related reading

• Park-PFI Hotel Investment: Inside Japan’s 165-Site Public Park Hotel Market
• Leasehold Hotel Development: 1.38x Land Price Reach Across 9 Markets
• Department Store Sites: 6 of 14 Go Hotel, ADR Needs 1.4-3.1x Market
• Japan’s Pre-1981 Hotel Stock: 236k Rooms, 19.1% — Retrofit or Rebuild