Term bites harder than rate — the supportable price moves 42.6% on identical NOI

Start with the conclusion. Even with identical NOI (net operating income), what a buyer can pay changes substantially when the tenor changes. The reason is simple: under level-payment amortization, the annual debt service is divided back out across the term. With annual debt service DS, interest rate i and term n years, the borrowable amount L is determined as follows.

DS = NOI ÷ DSCR / L = DS ÷ [ i ÷ ( 1 − (1+i)⁻ⁿ) ] / Supportable price P = L ÷ LTV

The rate i enters linearly in the numerator, but the term n sits inside the exponent. The shorter n becomes, the faster the annuity present-value factor shrinks, and the smaller the principal that the same DS can carry. The chart below puts the price impact of moving the rate and the price impact of moving the tenor on the same vertical axis.

Doubling the rate from 1.5% to 3.0% takes the supportable price only from ¥4,064 million to ¥3,521 million — a 13.4% decline — so long as a 20-year tenor holds. But keep the rate at 2.0% and move the tenor from 30 years to 15, and ¥5,301 million becomes ¥3,041 million, a fall of 42.6%. The contrast is sharper still at the extremes. A 1.5% rate with a 15-year tenor supports ¥3,158 million; a 3.0% rate with a 30-year tenor supports ¥4,639 million. Even at double the rate, double the term lets the buyer pay 46.9% more. Grinding the rate assumption down in 0.25% increments has its place in an investment memo, but there are situations where negotiating one more year of tenor moves the price by an order of magnitude more.

Why 39 years is the starting point — statutory useful life and the simplified method set the remaining years

So what determines the tenor? The practical benchmark is the building’s remaining useful life. Under Appended Table 1 of the Ministerial Ordinance on Useful Lives of Depreciable Assets, the statutory useful life of hotel- and ryokan-use buildings of steel-framed reinforced concrete or reinforced concrete construction is set at 39 years (within the same hotel-use category, brick, stone and block construction is 36 years; metal-frame construction is 29, 24 or 17 years depending on the thickness of the frame members; and wooden and synthetic-resin construction is 17 years).

In lending practice, financial institutions widely use the “statutory remaining life” — 39 years less elapsed years — as the ceiling benchmark for loan tenor. There is also a second yardstick, the simplified method used to determine the useful life of used assets for tax purposes. An asset that has passed its full statutory useful life is assigned “statutory useful life × 20%”; one that has passed part of it is assigned “(statutory useful life − elapsed years) + elapsed years × 20%”. Some regional banks and shinkin banks will extend loan terms out to the simplified-method figure, or apply their own standard (for RC construction, “70 − building age”, for example), so tenor is not uniform. The direction of travel is common to all of them, however: for every additional year of building age, the term you can borrow for shortens by one year.

The threshold to watch here is age 25. At that point the statutory remaining life drops below 14 years. Assuming lenders for whom regional-bank and shinkin credit practice makes it hard to lend beyond the statutory remaining life, properties older than 25 years effectively enter a world of 14-year tenors or less, and those older than 30 years, 9 years or less. On the sensitivity above, a tenor of around 14 years sits in the zone where the supportable price falls by more than 40% versus 30 years.

Supportable price by tenor — ¥5,301 million at 30 years, ¥2,126 million at 10

Reading the table across, three consequences of a shorter tenor happen at once. First, the supportable price falls. Second, because the price falls, the effective cap rate jumps from 3.77% to 13.06% — which shows that the main driver of the “older assets look high-yielding” phenomenon is not only property risk premium, but the fact that financing constraints have pushed the price down in the first place. Third, levered cash-on-cash improves in the opposite direction (2.49% → 8.61%). A shorter repayment term amortizes principal faster, so cash flow during the hold is thinner but equity turns over more quickly. In other words, a short-tenor property is an investment that is “cheap to buy, thin on interim cash flow, and fast on principal repayment”. That is a different kind of equity story, and an investor seeking interim income and one targeting equity build-up through principal amortization will pay different amounts for the very same property.

Up to this point NOI was held constant and only the tenor moved, but in practice the two move together. Placing a 25-year-old property into the three scenarios above puts the supportable price at ¥1,690 million to ¥4,082 million — the optimistic case is 42.5% above the base case (¥2,865 million) and the pessimistic case 41.0% below it, a 2.41-fold spread from bottom to top. What is worth noting is that even in the pessimistic scenario the effective cap rate looks high, at 10.35%. Looking high-yielding and being a good investment are two different things.

Laid out on two axes, the difference in impact is immediately visible. Moving horizontally (shortening the tenor from 30 years to 10), the price change at each rate level is −56.5% to −63.3%, while moving vertically (raising the rate from 1.0% to 3.0%, a tripling), the price change at each tenor is no more than −9.9% to −24.1%. In the most extreme diagonal comparison, a 3.0% rate with a 30-year tenor (¥4,639 million) supports 2.07 times what a 1.0% rate with a 10-year tenor (¥2,242 million) does. Even at one-third the rate, one-third the term halves the price and more. And because the rate bites harder the longer the tenor is (−24.1% at 30 years, −9.9% at 10), the ordering also tells you that negotiating the rate only matters on deals where long-term funding is available in the first place.

Real data: building age is not pushing room rates down monotonically

Everything so far has been a worked example. So do hotels that have aged actually generate less cash flow? From here we test that against real data — our own lodging market data and the property-level results disclosed by listed hotel REITs.

Start with our own data. We extracted properties with a known opening year and 30 or more rooms into building-age bands (taking the top 400 properties by room count in each band as the population, and limiting it to properties with at least three recent months of results), then aggregated the property-level median of estimated settled ADR for March–August 2026 by band (N=1,767 properties). The result is not a monotonic decline.

The highest band is under 10 years old (¥13,900), but that reflects new openings being concentrated in high-occupancy urban locations. The trough is the 20–24-year band (¥8,700), and from there the figures actually rise. The 40+ band is ¥12,100, 39.1% above the 20–24 band. The reason is category mix. The 40+ band contains a heavy weighting of hot-spring ryokan and resorts that serve two meals and command high rates, while the 20–24 band is concentrated in limited-service properties supplied in bulk in the early 2000s. Building age is a weak variable for explaining room rates.

The same pattern holds in the property-level results disclosed by listed hotel REITs. Of the 299 properties held by seven investment corporations (our ownership master, snapshot as of 31 August 2026, domestic properties), the 171 for which property-level RevPAR was available in the most recent disclosure month line up by building-age band as follows.

Median RevPAR is ¥10,600 in the 30–39 band and ¥10,900 in the 40+ band, both above the ¥7,700 of the under-10 band.On occupancy, the 30–39 band is the highest of the well-populated bands at 89.6% (the outright maximum is the thin 20–24 band at 91.7%, N=5).What these properties actually earn does not fall monotonically with building age.

This is the crux of the article. Earnings are not falling, yet the tenor shortens mechanically in line with the regime. The result is a financing-driven price gap between two properties generating identical NOI, attributable to building age alone. Back in the worked example, a 10-year-old property (29-year tenor) at ¥5,170 million and a 30-year-old property (9-year tenor) at ¥1,932 million are both generating exactly the same ¥200 million of NOI.

Properties aged 25+ hold 42.3% of room stock — 956,200 rooms with a narrowed buyer pool

How large is the asset base this applies to? Breaking down the 38,390 properties and 2,261,074 rooms we track that have a known opening year and 10 or more rooms by building-age band, properties aged 25 and over account for 19,169 properties and 956,200 rooms, or 42.3% of the total on a room basis.

The 956,200 rooms aged 25 and over sit, on standard lender credit criteria, in a world of tenors of 14 years or less. That is selective about buyers. Core investors chasing thin cap rates with long-dated non-recourse loans cannot even enter the bidding, because the repayment term cannot be arranged. What remains are three types: (1) cash buyers able to raise the equity ratio substantially, (2) funds that come in on short-term bridge or value-add money and turn the asset in three to five years after renovation and rebranding, and (3) operating companies that can absorb the short repayment term with unrealized gains on land they hold or with cash flow from their core business. Behind the complaint that trades in older hotels “price inconsistently” or “fail to attract bidders” lies not a demand problem, but a contraction of the buyer pool caused by tenor. How far the option of converting to a non-lodging use supports the downside is a separate question, and one that turns on the breadth of available exits.

That said, listed REITs are not avoiding older assets. The median building age across the 299 properties held by the seven investment corporations is 19 years; properties aged 25 and over account for 44.8% (134 properties), and those aged 40 and over for 16.1% (48 properties). The median building age at the time of acquisition is 8 years, but 30.3% of acquisitions were of properties older than 25 years. Because REITs have unsecured corporate bonds, long-dated syndicated loans and follow-on unit issuance available to them, their loan tenor is less tightly bound to any single property’s remaining life. One way to read this is that a difference in financing structure translates directly into a difference in which buyer the asset suits.

Tenor decay — a seven-year hold can erode the exit price by up to 40.9%

The tenor problem does not end at acquisition. It bites at the exit too. The tenor a buyer can arrange at the point of sale is shorter again by the length of the holding period. Even if NOI does not fall by a single yen, the sale price falls purely from the decay in remaining years — what this article calls tenor decay.

The table below applies the same assumptions as the worked example (NOI ¥200 million, rate 2.0%, LTV 65%, DSCR 1.30), sets a holding period of seven years, and estimates the exit price by building age at acquisition. Tenor follows the statutory remaining life, with a floor of 7.8 years from the simplified method.

The most painful case is buying at age 25 and selling at age 32. Against a 14-year tenor at acquisition, the exit tenor falls to the 7.8-year floor, and on identical NOI the exit price erodes by 40.9%. Offsetting that erosion through price would require NOI to grow over the seven-year hold at 7.80% a year (69.1% cumulatively). Sustaining 7.8% a year for seven years through ADR gains and occupancy improvement is not easy. Put the other way round, an acquisition at around age 25 calls for a design that places the return on interim cash flow and principal amortization rather than on sale proceeds. As shown above, CoC in this band is a healthy 4%-plus, and over a seven-year hold principal falls appreciably. The erosion in exit price is absorbed by the equity recovered through the principal repaid during the hold — that is the shape of the structure.

By contrast, buying at age 10 and selling at age 17 erodes the price by 19.2%, and the NOI growth required is only 3.09% a year. In the recent ADR upcycle that is comfortably within range, producing an asymmetry in which newer assets can “build an exit through NOI growth” while older assets cannot. Whether tenor remains is what governs the choice of exit strategy.

Implications for investment memos, credit underwriting and exit design

Finally, the practical implications by role.

For investment professionals writing an acquisition memo. We would suggest making tenor a mandatory axis in sensitivity analysis, alongside rate and cap rate. As shown above, a five-year difference in tenor has a larger price impact than a 1.5-percentage-point difference in rate. Confirming up front how many years your expected lender can go, before setting the bid price, avoids having to redo the price assumptions later in the process. And when you see a deal with a high effective cap rate, it is worth separating whether that reflects a property-specific risk premium or a price pushed down by a tenor constraint. If it is the latter, simply bringing longer-dated funding captures the upside on the same asset.

For lenders setting collateral values and terms. What the real data in this article shows is thatmedian occupancy in the 30–39-year band, at 89.6%, is the highest of the well-populated bands, and RevPAR there also exceeds the newer bands.Statutory remaining life is a tax yardstick, not a measure of how long cash flow will last. Read alongside the history of major repairs, the record of renewal investment and recent occupancy and rate trends, there is room for term-setting more precise than a mechanical application of statutory remaining life. As our earlier work on hotel capital expenditure showed — annual capex runs ¥130,000 to ¥620,000 per room — properties receiving continuous capital expenditure have extended their physical remaining life.

For asset managers reading the exit on older assets. The fact that tenor falls by one year for every year of the hold feeds directly into how the holding period is designed. If the holding period passes through the band where tenor decay is steepest (ages 20 to 32), you need either a design that absorbs the fall in exit price with interim cash flow and principal amortization, or a decision to sell out before leaving that band. The premise that “growing NOI grows the exit too” holds only for properties that still have tenor left.

Related reading

• A 1.00% Policy Rate and Hotel Development — Working Back to the Break-Even ADR Across 8 Major Markets
• 236,000 Pre-1981 Seismic-Code Hotel Rooms, 19.1% of All Rooms — Retrofit or Rebuild, the Tipping Point Across 588 Buildings
• A Structural Analysis of Regional Bank-Led Small-Scale Continuous Development and Renovation Funds
• Hotel Acquisition Costs Run to 7.1% of Price — Building Up How a 6.5% Gross Yield Becomes 4.65% Net
• Hotel Capital Allocation 2026 — One Deal Took 27% of J-REIT Quarterly Acquisitions, an Annual Breakdown of ¥941.9 Billion Actually Acquired