Room Stock in the Six Areas — 52 of 53 Hotels (23,741 Rooms) on Level 1 Reclaimed Land

Table 1 breaks down the six areas. The largest is Maihama with 15 hotels and 7,199 rooms, where large properties of 500+ rooms line up around the Tokyo Disney Resort’s directly operated and official hotels. Ariake/Odaiba (12 hotels, 5,810 rooms) is dominated by city hotels supported by exhibition and convention demand, while Kaihin-Makuhari (8 hotels, 3,811 rooms) still has the first generation of the Makuhari New City that opened in the early 1990s. Kobe’s two islands hold 6 hotels with 1,721 rooms, Osaka Nanko 6 hotels with 1,685 rooms, and the Chubu Centrair airport island and Rinku 6 hotels with 3,515 rooms. The three Tokyo Bay areas alone account for 35 hotels and 16,820 rooms, or 70% of the total.

The key axis for reading this is opening year. Of the 53 hotels, 14 (8,051 rooms) opened in 1994 or earlier, i.e. before the 1995 Southern Hyogo Prefecture Earthquake; 16 (7,575 rooms) opened between 1995 and 2010; and 23 (8,115 rooms) opened in 2011 or later, after the 2011 Tohoku earthquake. The 2000 amendment to the Building Standards Act and the MLIT notification of the following year (2001) are understood to have positioned the assessment of liquefaction risk within the ground surveys used to determine allowable bearing capacity, so buildings designed before 1995 and those designed after 2011 may differ in how far liquefaction was assumed at the design stage. This does not mean “older buildings are weaker.” Hilton Tokyo Bay (828 rooms) and Hotel Okura Tokyo Bay (427 rooms), both opened in Maihama in 1988, and the large properties that opened in Kaihin-Makuhari in 1993 came through the 2011 liquefaction in Urayasu and are still operating, examples of pile-supported buildings that suffered no critical damage from liquefaction (Hilton Tokyo Bay’s room count and opening year match the table of 37 hotels in Japan and the history section of Hilton’s “Company Overview,” updated August 25, 2026). Opening year is best used as a pointer for deciding “which parts of the geotechnical survey report and structural calculations to read” at acquisition.

Port Island 1995 and Urayasu 2011 — Lifelines and Access, Not Buildings, Were What Stopped

How liquefaction of reclaimed land actually plays out is recorded in primary sources for two earthquakes. Table 2 extracts only the events with figures from the Kobe City Development Bureau report held in Kobe University’s Earthquake Disaster Archive, MLIT City Bureau materials, and Urayasu City’s reconstruction grant evaluation.

What the two cases have in common is that damage concentrated on surrounding ground, seawalls, lifelines and access routes rather than on pile-supported buildings. On Port Island, quay displacement and settlement of the ground behind made the container terminals unusable, and lifelines such as electricity and water stopped. In Urayasu, although liquefaction hit 86% of the city, only 24 buildings were totally destroyed; the core of the damage was “half destroyed or worse” certifications due to tilting (about 3,700 buildings) and damage to water, sewer and roads. In Maihama, Tokyo Disneyland closed for about a month for liquefaction repairs to its parking lots, and the two hotels on site reopened on the same April 15. For hotels, the impact showed up as “the period when they could not operate” more than as building repair costs.

Differences in fill material also matter. Port Island’s fill is mainly masa soil from the Rokko Mountains, while Rokko Island also used Kobe Group sediment long considered resistant to liquefaction based on grain size, yet the island still experienced liquefaction-induced ground settlement. Even within the same “reclaimed land, Level 1” mesh, outcomes vary with fill material, construction era and whether seawalls were seismically reinforced. The 250m-mesh landform classification does not distinguish at this level. That is exactly why, at acquisition, the fill’s grain size and N-values and the liquefaction assessment (PL value) need to be read individually in the geotechnical survey report.

The countermeasure cost estimate in Urayasu’s materials is also instructive for holders of existing stock. A study-stage figure of ¥4M–10M per house was given for grid-type ground improvement under existing detached-house lots, and MLIT likewise explains that improving ground under an existing building costs considerably more than on a vacant site because of working-space limits and buried utilities. Improving the ground directly beneath an operating hotel is not realistic, so investment goes not toward “changing the ground” but toward “not stopping operations even if liquefaction occurs”: lifeline flexibility, the floor on which equipment is installed, and a contingency for restoring exterior works.

Is the Hazard Priced Into Rates? — Estimated Settled ADR in the Six Municipalities Follows Demand Anchors

If reclaimed-land conditions were reflected in room rates, they could be used as grounds in acquisition price negotiations. Table 3 lists the estimated settled ADR of the municipalities containing the six areas for September 2026 (a settled month), the trailing 12-month average, and the ratio to the prefectural average.

In short, no ground-related discount can be identified in any of the six municipalities. Urayasu was ¥22,360 in September 2026, 2.09x the Chiba Prefecture average (¥10,680), and its trailing 12-month average of ¥23,780 runs above the Tokyo average. The -4.4% YoY is a reaction against a busy autumn 2025 base; the monthly trend (Figure 3) shows 2026 tracing almost the same seasonal pattern as 2025, peaking at ¥26.6K in March and sitting in the ¥18K range in summer. Koto Ward was ¥14,290 (+12.4% YoY) and Mihama Ward, Chiba City ¥17,009 (+43.6%), driven respectively by convention/exhibition events and depth of demand including a new opening in June 2026 (note that Mihama Ward has only N=13 properties, so month-to-month swings are large).

Meanwhile, Suminoe Ward’s -31.5% is a reaction to September 2025 falling within Expo 2025 Osaka, Kansai, and the Osaka Prefecture average fell by a similar -32.4%. Tokoname’s +23.0% and the Aichi average’s +45.9% reflect a month overlapping with lodging demand for the Aichi-Nagoya Asian Games, held from late September to early October 2026. Chuo Ward, Kobe was flat at ¥12,120 (+0.2% YoY), but its N=64 includes inland properties such as Sannomiya and is not a rate isolated for the six hotels on the two artificial islands. In every municipality, rates are driven by demand anchors (theme parks, exhibition centers, the Expo, international sporting events, airports), the same picture as our earlier analysis of Chuo Ward, Niigata City (100% of 6,415 rooms on liquefaction-prone ground with no ADR gap).

For holders and acquirers, this has two implications. First, there is little room on the market-price side to negotiate on the basis of the hazard. Second, for the same reason, liquefaction preparedness has to be designed as an investment recovered not through room rates but through shorter closures and improvements in PML and insurance premiums. The following sections lay out the items needed for that design.

Seven Items to Check in Acquisition Due Diligence — Piles, Utility Intakes, Equipment Floors, PML, Insurance

An engineering report (ER) is a survey that quickly identifies a building’s physical risks during real estate due diligence, consisting of building condition assessment, regulatory compliance, repair and replacement costs, replacement cost, environmental risk and seismic risk assessment. When acquiring a hotel on reclaimed land, this standard structure needs to be read with liquefaction in mind. Table 4 sets out what to check, and in which documents, for existing stock like the 53 hotels in this article.

The first thing to look at is the combination of items 1 and 3. If the piles reach the bearing stratum, the building itself is unlikely to settle or tilt from ground liquefaction, and the investment focus shifts to the lifelines and equipment in items 2–4. Conversely, if power-receiving equipment, generators and water pumps are concentrated in the basement or on the ground floor, operations stop as soon as settlement of the surrounding ground damages the intake pipes. Remember that in Urayasu, the main causes of closure were water, sewer and parking lots, not damage to buildings.

Item 5, seismic PML, expresses the physical loss at the 90% non-exceedance probability, for ground motion with a 10% probability of exceedance in the next 50 years, as a percentage of replacement cost; in transaction practice, 15–20% or lower is referenced as a benchmark. J-REITs customarily disclose property-level seismic PML in their securities reports, and 7 of the 53 hotels in the six areas fall within this disclosure framework as investment-corporation holdings (Table 7). Whether liquefaction is included in PML calculation conditions depends on the order specifications, so at acquisition we recommend explicitly ordering it “with liquefaction conditions” and reading it together with the insurance terms.

On item 6, insurance, standard commercial fire insurance does not cover earthquake-caused losses, so the decisions are whether to add an earthquake endorsement and how to set limits and deductibles. Residential earthquake insurance added criteria for liquefaction damage in June 2011, but how commercial building policies treat liquefaction losses needs to be checked insurer by insurer. As Ministry of Finance materials show, earthquake insurance rates have historically been designed not to vary greatly by region, ground or structure, so ground conditions are reflected only to a limited degree in premiums. That makes BCP equipment and shorter closures easier to justify as investments than insurance.

BCP Capex per Room — ¥139K–362K for Retrofitting Existing Buildings

So how much does equipment investment to prepare existing buildings for liquefaction cost? Rather than ground improvement, we estimated four packages aimed at “not stopping operations even if liquefaction occurs,” by property size band, based on published market prices and public-works unit costs (Table 5). The electrical assumptions are the same published market prices used in our article on tsunami inundation along Onna Village to Nago Bay, with liquefaction-specific flexible joints at utility intakes and a contingency for restoring settled exterior works added.

Per-room cost comes to ¥343K–1,000K in the 100-room band, ¥154K–417K in the 300-room band and ¥111K–271K in the 700-room band. Because power-receiving equipment and generator costs do not scale with room count, larger properties carry a lighter per-room burden. Mapping the 53 hotels in this article onto the size bands gives 13 hotels (1,418 rooms) under 200 rooms, 23 hotels (7,782 rooms) with 200–499 rooms and 17 hotels (14,541 rooms) with 500 or more; assuming every hotel took the same measures, the total would be ¥3.30bn–8.60bn, or ¥139K–362K per room. The concentration of 500+ room properties in Maihama pulls the per-room burden down.

Dividing this added cost by the estimated settled ADR of each municipality lets us compare the weight of the investment as “how many days of room revenue” (Table 6). The ¥154K–417K per room for the 300-room band equals 8.1–21.9 days at Urayasu’s trailing 12-month average ADR of ¥23,780 and 80% occupancy (a full-year assumption used for conversion), or 23.1–62.6 days at Tokoname’s ¥8,333. Given that the 2011 closure in Maihama lasted about a month, the room revenue forgone in a one-month closure at Urayasu’s level is about ¥571K per room, roughly the same scale as the upper end of the added cost. The argument that equipment investment is recovered through shorter closures can be used as-is in credit and investment committee reviews.

Table 6 assumes 80% occupancy, but even large bay-front properties see annual occupancy swing with seasons and events. Table 6-2 holds Urayasu’s trailing 12-month average ADR of ¥23,780 fixed and varies days of room revenue along two axes: occupancy (full-year average assumption) and equipment cost per room. Even at 60% occupancy, the 53-hotel upper bound of ¥362K is only 25.4 days, below the room revenue lost in a 30-day closure (¥428K). Lower occupancy lengthens the payback in days, but it also shrinks the opportunity loss of a one-month closure, so the rule of thumb “equipment cost ≈ one month of room revenue” broadly holds across 60–90% occupancy.

Note that the amounts in Table 5 assume retrofitting an operating existing building, not the incremental cost of building measures into a new design, and structural reinforcement and temporary works for construction during operation may be extra. As in the tsunami inundation article, generators are not liquefaction-only equipment: they also support business continuity during annual typhoons and outages, and can be depreciated on that basis.

2026 Openings and Planned Supply — 301 Rooms in Makuhari, 381 in Kobe, a 1,000-Room Plan in Mihama Ward

New hotels and rebrands continue on the bay front and artificial islands. On June 18, 2026, KOKO HOTEL Premier Tokyo Bay Makuhari (301 rooms, Polaris Holdings) opened in front of Kaihin-Makuhari Station, and on October 1, Courtyard by Marriott Kobe (381 rooms) opened on Port Island as a rebrand of a fully renovated existing hotel. Both are included in the 53 operating hotels in this article. Among planned projects, building-permit-application data show a 1,000-room mixed-use hotel/cinema/hot-bath complex in Hibino 1-chome, Mihama Ward, Chiba City (scheduled for completion in December 2027), and a 126-room plan in Chidori, Urayasu City. The 750-room project in the same Chidori district (Mitsubishi Estate) already opened on February 13, 2026 as “The Royal Park Hotel Maihama Resort Tokyo Bay,” about 1.6km from our Maihama tally point and just outside this article’s 1,500m radius.

New builds can incorporate longer piles or ground improvement at the design stage. That is the difference from acquiring existing stock, and why this article limits itself to the “acquire and hold” side. For analysis from the new-build cost side, see our separate article on construction cost differences from pile length and bearing-stratum depth on Tokyo Bay hotel sites.

Investment Takeaways — Three Steps to Keep Holding Hotels on Reclaimed Land

Frequently Asked Questions

Q. Will a hotel on a “very prone to liquefaction” mesh be damaged in an earthquake?

A. Not necessarily. MLIT’s tendency map shows a relative tendency based on the landform class of 250m meshes and does not reflect site-specific ground improvement or piles. In the 1995 and 2011 cases, damage concentrated on lifelines, exterior works and access rather than on pile-supported buildings. Check individually with the geotechnical survey report and PML at acquisition.

Q. How much do liquefaction measures cost for an existing hotel?

A. Our estimate for four packages (flexible joints at utility intakes, moving power-receiving equipment upstairs, emergency generators and an exterior-works restoration contingency) is ¥154K–417K per room for a 300-room property and ¥111K–271K for a 700-room property (retrofit of existing buildings, editorial estimate). Improving the ground itself is not realistic for an operating building, so costs center on equipment and lifelines.

Q. Is being on reclaimed land reflected in room rates?

A. It cannot be confirmed in the estimated settled ADR of the six municipalities. Urayasu was ¥22,360 in September 2026, 2.09x the Chiba Prefecture average, and Koto Ward and Mihama Ward (Chiba City) rose year on year; rates move with demand anchors such as theme parks, exhibition centers and international sporting events.

Q. What should be added to the engineering report order at acquisition?

A. Three things: include liquefaction in the seismic PML calculation conditions; confirm on drawings the floor level of equipment and whether utility intakes have flexible joints; and obtain from the seller the settlement and repair records from 2011 and 1995. For insurance, check in the policy wording whether commercial property insurance covers earthquake risk and how liquefaction damage is assessed.