Home Insights

2026年8月11日

Why hotels in Japan are not priced per tsubo

Anyone looking at Japanese real estate meets the tsubo quickly. It is the customary unit of area — about 3.3 square metres — and price per tsubo is how offices, land and residential stock are compared here. It travels badly to hotels. The moment a hotel is reduced to a figure per tsubo, the valuation has already gone wrong.

The same building, two different values

Take two hotels on the same street: the same room count, the same year of construction, the same specification. Their earnings will not match, and usually will not come close. Rate strategy, how occupancy is built, staffing structure, the mix of distribution channels, whether breakfast and banqueting are run as a business or as an obligation — change the operating design and what reaches the owner changes with it.

A hotel is real estate that contains a trading business. The building is the container; the earnings come from what moves inside it. Measuring the container tells you very little about the contents.

Why gross yield misleads here

For an office or a residential block, rent is fixed by contract and moves slowly. Annual rent divided by price therefore produces a gross yield with reasonable predictive value.

Hotel revenue is not rent. It moves daily, by season and by day of week, and it absorbs demand shocks directly. Sitting on top of it is a cost base that moves too — payroll, utilities, distribution commissions. A twenty per cent rise in revenue does not deliver a twenty per cent rise in profit, and the same asymmetry applies on the way down.

Compare hotels on gross yield while ignoring that structure and one of two things happens: an unusually strong year is treated as the underlying run rate, or a temporarily depressed year is taken at face value and the asset is written down to a number it does not deserve.

What we look at instead

The sequence we work through when building a price is broadly this.

  • ADR and occupancy — monthly, over several years. Whether rate was pushed at the expense of occupancy, or the reverse, and what happened next.
  • RevPAR — rate and occupancy combined, as the per-room measure of what the hotel actually achieves.
  • Departmental performance — what rooms, food and beverage and other departments each earn, and what each consumes.
  • GOP — what survives at the operating level.
  • NOI — what reaches the investor after the FF&E reserve and operator fees.

Then the harder question: whether those earnings persist. Is competing supply due to open in the catchment? Is demand concentrated in one source market, one channel, one corporate account? Has major capital expenditure been deferred to flatter recent results? Earnings that do not persist do not command a price.

When price per tsubo does mean something

There is a case where it applies. If the hotel is being sold as a site — for redevelopment, or for conversion to another use — the price is founded on the land, and comparable land transactions in the area become the right instrument. The building is then a cost to be demolished rather than an asset to be underwritten.

So the question is whether the asset is being sold as a business or as a container, and the answer determines which yardstick applies. That question cannot be answered by inspection. Both figures have to be calculated — the earnings-based value and the land-based value — and compared. On assets in central locations with weak trading performance, the second is regularly the higher of the two, and owners are often surprised by which one wins.

Assumptions before figures

What we issue in an indicative valuation is a range, not a single number, and every assumption behind it is written down: on these assumptions the price is here, and if this assumption moves, the price moves to there.

The reason is straightforward. A hotel price is a function of its assumptions. Handing over a number with the assumptions stripped out looks more precise and conveys less — and it is the form in which most owners have been given a valuation before.

There is no need to have decided to sell. Establishing where the asset stands today is a reasonable place to begin, and it is work we do at no cost to owners.

All insights