At the 3rd Asia Hotel Investment Conference Japan & Far East Asia (AHICE), held in September 2026, Seth Sulkin of Pacific Hotels GK said construction costs are running “50–100% over budget,” and that projects are being cancelled, postponed or redesigned one after another (as reported by Travel Voice on October 23, 2026). This article treats that remark as a single piece of testimony rather than a statistic, and uses our in-house model of a 100-room new-build hotel to convert “the budget overrun itself” into a required ADR. The baseline of ¥13,454 rises to ¥18,836 if construction costs come in at 1.5x, and to ¥24,217 at 2x. Measured against the 12-month average estimated settled ADR of ten major cities and wards, only Tokyo’s Chuo Ward clears the 1.5x level. Using the same formula, we also work through the land carry cost when the build period stretches from three years to six, the break-even point for switching to acquiring and renovating an existing building, and how much contingency lenders should build in.

*All estimates in this article are approximations from our simplified model; investment decisions on individual projects require a detailed feasibility study. Remarks by AHICE speakers are quoted from the Travel Voice article and are not industry-wide statistics.

“50–100% Over Budget” at AHICE — Testimony, and the Step-Up in Unit Costs in Construction Statistics

First, the remarks themselves. According to the AHICE report Travel Voice published on October 23, 2026, Sulkin, speaking on a panel about the investment environment, described construction costs as running “50–100% over budget,” and said that in regional areas it is hard even to secure a contractor, leaving no room to negotiate on price. Choe Peng Sum of Pan Pacific Hotels Group named soaring construction costs as the biggest barrier to growth outside Tokyo. After noting that labor is also scarce, he pointed out that a new build takes “more than twice the usual time, five to six years” from planning to completion, and that holding costs during that period squeeze yields. Andrew Langdon of Accor’s remark that “70% of the rooms signed this year are conversions of existing hotels” came in the same context.

These are testimonies based on each speaker’s own project experience, not statistics representative of all projects in Japan. Public statistics, however, point the same way. In the Ministry of Land, Infrastructure, Transport and Tourism (MLIT) Building Construction Starts Survey, the cost per tsubo for accommodation buildings (nationwide, all structures), calculated as planned construction cost divided by floor area, rose from about ¥1.11M in 2017 to about ¥1.39M in 2022 and about ¥1.95M in 2024. A 41% jump over the two years from 2022 to 2024 is most likely to surface as an overrun “against budget” in projects with a gap of a year or more between estimate and groundbreaking. The 2024 figure is partly inflated by large SRC projects, so it is more sensible to read the slope than the level of any single year.

Our earlier article, “Hotel Construction Cost ¥8.7M–71.4M per Room”, multiplied the 2024 RC unit cost from the same statistics by segment-specific per-room floor area, giving construction costs per room of about ¥8.7M for limited-service hotels, about ¥24.8M for city/upper-midscale hotels and about ¥71.4M for resort/upscale hotels. The ¥20M per room assumed in this article’s 100-room model sits slightly below the city/upper-midscale tier of those three.

The 100-Room Model — Every 50% Overrun Adds About ¥5,400 to the Required ADR

The model’s assumptions are kept simple: 100 rooms, construction cost of ¥20M per room and land of ¥5M per room, for a total project cost of ¥25M per room (¥2.5bn for 100 rooms). With an investor-required NOI yield of 5.5%, full-year occupancy of 80% and an NOI-to-revenue ratio of 35%, the required ADR is “total project cost × 5.5% ÷ 365 days ÷ 0.8 ÷ 0.35.” The baseline comes to ¥13,454. Table 1 moves only the construction cost, by +25%, +50%, +75% and +100% against budget; the land cost is held constant.

The result is linear. If construction costs reach 1.5x (¥30M per room), the total project cost becomes ¥35M and the required ADR ¥18,836; at 2x (¥40M), it becomes ¥45M and ¥24,217. The required ADR rises by ¥5,382 for every 50% overrun, which means Sulkin’s “50–100%” range maps directly onto a range of ¥18,836 to ¥24,217. The overrun described in the AHICE testimony can be read as lifting the baseline ADR by 40% to 80%.

Table 2 varies the yield and occupancy assumptions. On the optimistic side, a 5.0% yield with 85% occupancy (assumed, full-year) brings the baseline required ADR down to ¥11,511; on the conservative side, 6.5% and 70% lifts it to ¥18,172. The right-hand column shows the +50% overrun case at 80% occupancy: even at a 5.0% yield it requires ¥17,123, higher than every cell in the on-budget table except the toughest one, 6.5% and 70% at ¥18,172. Put another way, a project whose construction costs have reached 1.5x still needs about ¥16,100 even if investors accept a 0.5pt lower yield and occupancy is lifted by 5pt — close to the ¥16,774 required by an on-budget project assuming a 6.0% yield and 70% occupancy. The overrun is too large to be absorbed by adjusting assumptions. For expected yield levels by city, our article on hotel expected yields from the 54th Real Estate Investor Survey lists eight cities.

Build Periods Stretching from 3 Years to 5–6 — Land Carry Cost Rises from ¥0.94M to ¥2.05M per Room

To apply Choe’s “five to six years,” we only need to add the cost of holding land acquired in advance to the project cost. Our earlier article “Hotel Land Carry Cost 5.9%/yr” set the annual carry cost at 5.9%, combining an effective 0.83% for fixed asset tax and city planning tax, 1.89% in interest (60% LTV, 3.15% rate), 3.20% in opportunity cost of equity (8% required return) and land management costs. This article applies that rate, compounded, to land of ¥5M per room. From planning to completion, the cumulative cost is ¥0.94M over three years, ¥1.66M over five and ¥2.05M over six. For an on-budget project the required ADR rises from ¥13,959 to ¥14,559; combined with a +50% overrun it reaches ¥19,729 (five years) to ¥19,940 (six years).

Table 3 shows that a longer build period has a considerably smaller effect than a construction overrun. A six-year build adds +¥1,105 to the required ADR, only about 40% of the +¥2,691 from a 25% construction overrun. That, however, is the carry cost of the land alone. Applying the roughly ¥22.6M per month (100 rooms; the sum of borrowing interest, pre-opening staff costs and deferred GOP) shown in our earlier article “Hotel Construction: 24 Months Standalone, 49 Mixed-Use” to three years gives an economic cost of about ¥8.14M per room, equivalent to a 40% construction overrun. When Choe said “holding costs squeeze yields,” he was probably closer to this right-hand view including deferred revenue than to the left-hand columns of Table 3. Build period and construction cost are not independent either: the longer the period, the higher the unit cost at groundbreaking, which the same earlier article estimated at about ¥8.9M per month of delayed start.

Comparing with Market Rates in 10 Major Cities and Wards — Only Tokyo’s Chuo Ward Reaches 1.5x

Next, we set the required ADR against market rates, using our estimated settled ADR (an estimate applying segment adjustments to each property’s lowest-plan rate, tax-exclusive equivalent), taking a simple 12-month average from November 2025 to October 2026 for each city or ward. Tokyo’s Chuo Ward stands at ¥19,200 (N=151), Shinjuku Ward ¥16,200 (N=102), Fukuoka’s Hakata Ward ¥14,700 (N=152) and Sapporo’s Chuo Ward ¥14,700 (N=139), all above the ¥13,454 baseline; the six areas from Osaka’s Chuo Ward (¥11,600, N=193) down do not even reach the baseline. Only Tokyo’s Chuo Ward exceeds the 1.5x level of ¥18,836, and no area reaches the 2x level of ¥24,217.

The right-hand column restates this from a lender’s perspective. The total project cost supported by Tokyo Chuo Ward’s average ADR is ¥35.75M per room, so even if construction costs swell by up to 54%, the overrun can be absorbed at an average rate. The ceiling is +26% for Shinjuku Ward and around +12% for Hakata Ward and Sapporo’s Chuo Ward; only one of the ten areas can absorb Sulkin’s 50% overrun at its average rate. The six areas from Osaka’s Chuo Ward down already require a rate above their average at the baseline.

Note that an area’s average ADR reflects the typical level of existing properties, not the rate of a new hotel. New builds normally command a premium for being new, and the “premium needed” figures in Table 4 show what percentage premium is required. Reaching the baseline requires +16% over the average in Osaka’s Chuo Ward and +26% in Nagoya’s Naka Ward, within the range a new build can achieve. By contrast, the +63% required in Osaka’s Chuo Ward and +88% in Sendai’s Aoba Ward under a 1.5x overrun would leapfrog the entire upper tier of the same market, and are hard to reach without changing the product design itself.

The Break-Even with Acquiring and Renovating — At a 1.5x New Build, an Existing Building Is Equivalent up to ¥29.5M per Room

Several AHICE speakers leaned toward “acquiring and renovating existing buildings” because it is a choice that carries no overrun risk. Compared using the same required-ADR formula, the break-even is simple: if the acquisition price of an existing building plus renovation cost comes in below the total project cost of a new build, the existing building wins. For renovation cost, we used the three conversion CapEx cases of ¥600,000, ¥800,000 and ¥1M per tsubo (about ¥4.1M, ¥5.5M and ¥6.8M per room, based on 18 m² of guestroom floor area) shown in our earlier article “Office-to-Hotel Break-Even ADR: ¥5,500–¥10,700”.

If a new build comes in on budget at ¥25M per room, a buyer allowing ¥5.5M for renovation can pay only up to ¥19.5M per room for an existing building. But if the new build lands at 1.5x, an acquisition price of up to ¥29.5M is acceptable, and up to ¥39.5M at 2x. To compare these figures with the market, we divided the disclosed acquisition prices of domestic properties held by seven J-REITs (Ichigo Hotel REIT Investment Corporation, Invincible Investment Corporation, Nippon Hotel & Residential Investment Corporation, Japan Hotel REIT Investment Corporation, Hoshino Resorts REIT, Inc., Mori Trust Reit, Inc. and Kasumigaseki Hotel REIT) by their room counts: the median across 298 properties was ¥24.25M per room, with an interquartile range of ¥12.14M to ¥41.16M. Candeo Hotels Osaka Namba (カンデオホテルズ大阪なんば; completed 2017, 496 rooms), held by Japan Hotel REIT, was acquired for ¥14.32bn, or ¥28.87M per room, and is classified as a limited-service hotel in the REIT’s portfolio list.

In other words, when a new build comes in on budget, existing assets sit in a band where they cost “more than a new build once renovation is included”; the moment a new build swells to 1.5x, that flips to “equal to or less than a new build even after renovation.” A nine-year-old property at ¥28.87M per room is at almost the same level as the ¥29.5M ceiling under a +50% overrun (with ¥5.5M renovation), and a newer building needing no further renovation would have the edge over a 1.5x new build. Acquiring an existing asset has one more advantage: the land carry cost and deferred revenue in Table 3 barely arise. Choe’s citing of the five-to-six-year build period as a reason to acquire existing assets can be read as a judgment that includes this time value. That said, the ceilings in Table 5 assume renovation brings the product up to new-build standard; whether the post-renovation rate can exceed the market average depends on each property’s design.

For Lenders — Contingency Levels Translated into Required ADR

Finally, we use the same formula to organize how to set contingency in loan underwriting. Table 6 lists total project cost and required ADR for each contingency ratio against construction cost, along with the number of areas whose average ADR reaches that required ADR. With +10% contingency the required ADR is ¥14,530, +20% gives ¥15,607 and +30% gives ¥16,683, narrowing the number of the ten areas reached at average ADR to four, two and one respectively. Setting +50%, as the AHICE testimony suggests, gives ¥18,836 with only one area; +100% gives ¥24,217 with none.

Three practical points follow from this table. First, the contingency level should be backed out from “the market’s rate headroom,” not from “estimate accuracy.” In a market like Tokyo’s Chuo Ward, where the average ADR absorbs up to a 54% construction overrun, a project can work even with 30–50% contingency; in markets like Hakata Ward or Sapporo’s Chuo Ward, where the absorbable range is around +12%, any contingency beyond about 12% presupposes a premium over the average rate. Second, for projects that cannot carry contingency, it is effective to keep open early on the option of absorbing an overrun through “design” rather than “build period.” The wave of redesigns Sulkin mentions likely reflects how many projects are forced to change room counts or specifications after groundbreaking. Third, noting the room to switch to an existing acquisition at the underwriting stage adds exit options once an overrun is confirmed. The ceilings in Table 5 can serve as a guide for this. For the contractual question of whether the owner or the contractor bears higher construction costs, Who Pays When Hotel Build Costs Rise? sets out the issues around the price-adjustment clauses of the revised Construction Business Act.

Seen from another angle, persistently high construction costs also raise the relative value of existing stock. As JLL points out, Japan’s new supply rate of 1.5–1.7% is far below the Asia Pacific average of 6.6%, and the harder it is for new builds to pencil out in a market, the better protected existing properties’ occupancy and rates tend to be. Overrun risk is a constraint for developers, but for owners of existing assets and their lenders it is also an upside in the form of restrained supply. The share of hotel supply coming from routes other than new construction is broken down from construction-start statistics in Hotel Supply: 16.7% of 5,707 Starts Are Not New Builds.

Summary — Underwrite Overruns Against the Market’s Absorbable Range

The formula in this article is simple, which is precisely why it can be reused by swapping in each project’s assumptions. Before adding contingency to a construction estimate, working out first how large a total project cost the market’s average ADR can support will speed up the decision — “change the design, switch to an existing asset, or recover it through rate” — once an overrun becomes clear.

Frequently Asked Questions

Q. How is the required ADR calculated?
A. We multiply the total project cost per room by the 5.5% NOI yield to get the annual required NOI, then divide by 365 days, 80% occupancy (assumed, full-year) and a 35% NOI-to-revenue ratio. For the ¥25M baseline, this gives ¥13,454. If you change the assumptions, see the sensitivity table in Table 2.

Q. Is “50–100% over budget” an industry-wide figure?
A. No. It is a remark by an AHICE speaker reported by Travel Voice, testimony based on each speaker’s project experience. As public statistics, MLIT’s Building Construction Starts Survey confirms that construction cost per tsubo for accommodation buildings rose 41% from 2022 to 2024.

Q. Can a new build not work in markets where the average ADR is below the required ADR?
A. The average reflects the typical level of existing properties, and new builds normally sell at a premium to it. The “premium needed” in Table 4 is the required uplift; around +20–30% is within a new build’s reach, but levels above +60% should be seen as requiring a change in product design.

Related Articles

• Hotel Construction Cost ¥8.7M–71.4M per Room: 3 Tiers, 12 Markets
• Hotel Construction: 24 Months Standalone, 49 Mixed-Use; 2029 Deadlines
• Hotel Land Carry Cost 5.9%/yr: ¥2.42M per Room Before Opening
• Office-to-Hotel Break-Even ADR: ¥5,500–¥10,700 Across 7 Districts
• Hotel GOP Explained: USALI Tiers, Formula, ADR ±5% = GOP ±10.7% (2026)
• Japan Hotel Forward Commitments: 5,243 Rooms Due 2027-28, 20% Penalty