Site evaluation for hotel development leans heavily on demand-side indicators. Catchment population, station passenger counts, tourist arrivals, the surrounding room stock — all of them measure who will come to stay. But the people who actually run the building once it is built are the people who can work there. Of Japan’s roughly 27,000 lodging facilities, few have been evaluated from the supply side of labor.

This article measures a “labor catchment ratio” — the working-age population within a 5km radius divided by the room count in that same radius — across 20 catchments nationwide, and converts it into the number of staff dormitory units required and the capital expenditure (Capex) per room. The conclusion up front: in catchments where this ratio falls below 1.5 workers per room, staff housing becomes a de facto mandatory facility, and it moves GOP yield by 0.18 to 0.66pt. Put the other way round, a business plan that prices this in from the start can still treat a thin-labor catchment as an area with substantial upside.

Divide working-age population by room count and catchments differ by about 66x

We began by matching the working-age population within a 5km radius against the room count in that same 5km radius, for 20 catchments nationwide. Population comes from MLIT’s National Land Numerical Information 250m mesh future population projections (2025 estimate), aggregated in concentric circles; room counts are built up from the property master maintained by MetroEngines Research & Consulting, covering properties open and not yet closed as of 2026.

The result is shown below: the lowest, Niseko-Hirafu, sits at 0.17 workers per room, and the highest, Nagoya Station, at 11.22 — a spread of roughly 66x. The Niseko-Hirafu catchment has just 1,701 working-age residents against 10,169 rooms. Even if every working-age resident in the catchment worked in lodging, that would still amount to only 0.17 people per room.

The ratio alone, however, does not tell us whether a given level is high or low. The reference point is a unit requirement: how many staff does lodging need per room? According to the 2016 Economic Census for Business Activity (MIC & METI), private-sector “ryokan and hotels” nationwide comprised 18,447 establishments with 476,731 workers and 991,925 rooms. Divide those and you get 0.48 workers per room. That national average includes the labor-intensive ryokan format; lodging-focused and resort hotels run lower.

This article uses 0.30 workers per room as the base assumption to reflect operating efficiency, and tests a 0.25–0.35 range in the sensitivity analysis later. It is worth noting this is conservative relative to the 0.48 national average.

The thinner the catchment, the thinner it gets by 2040

The labor catchment ratio is not a static indicator. Pull 2040 total population projections from the same mesh data and most of the low-ratio catchments are set to lose more people. Against 2025, the Kusatsu Onsen catchment is projected at −32.3%, Nozawa Onsen −32.0%, Zao Onsen −28.4%, Hakuba −16.5% and Hakone-Yumoto −21.3%. For deals with management contracts or hold periods beyond ten years, the number that matters is not the labor catchment ratio at opening but the level at exit.

What is interesting is that the low-ratio catchments split into two futures. Niseko-Hirafu’s projected 2040 total population change is −3.9%, one of the smaller declines among the 20. That suggests the catchment itself is thickening as a destination for relocation and settlement, which is a relative upside on the labor side. Kusatsu Onsen and Nozawa Onsen, by contrast, combine a low ratio with a steep projected decline.

Look at the mesh level and the character of that “thinness” comes into focus. In the Niseko-Hirafu catchment, resident population clusters in a limited set of central meshes, with a wide blank beyond. Its elderly share, at 18.4%, is the lowest of the 20 catchments.

Shade the Kusatsu Onsen catchment by elderly share and most of the inhabited meshes fall into the band above 32%. Across the catchment as a whole there are 2,958 working-age residents against 2,518 elderly residents, an elderly share of 43.0%. That age structure sits behind the labor catchment ratio of 0.74 workers per room.

Treat employment data as a secondary reference — coverage is uneven

The other material for measuring the supply side of labor is establishment and employment data from the Economic Census. Query a 1km radius and Nagoya Station returns 33 districts, 8,965 establishments and 207,026 workers (office-sector share 27.4%), a thick reading, while Niseko-Hirafu returns 1 district, 113 establishments and 1,075 workers (20.7%). As a secondary line for gauging how much existing labor demand has agglomerated, it is useful.

One caveat matters, though: this employment data has uneven coverage. Of the six locations queried here, Miyakojima-Hirara returned zero districts and zero establishments. That does not mean employment is zero; it means the boundary data for that location is not included. In reality the Hirara district contains a large number of establishments. Where coverage returns as “none,” the working rule has to be not to use employment data as evidence for demand or labor supply. Nor can the gap itself be generalized as an economic phenomenon.

Sizing the dormitory requirement — the threshold is 1.5 workers per room

From here we model the number of dormitory units required. We narrowed the assumptions to three.

Rearrange the formula and the condition for needing no dormitory reduces to “labor catchment ratio > 0.30 ÷ 0.20” — a single threshold of 1.5 workers per room. Set the assumptions at 0.25 workers/room and a 30% ceiling and the threshold moves to 0.83; at 0.35 workers/room and a 10% ceiling it moves to 3.5. But on the base assumptions, 1.5 workers per room is the dividing line. Seven of the 20 catchments fall below it: Niseko-Hirafu, Zao Onsen, Hakuba, Nozawa Onsen, Kusatsu Onsen, Kinugawa Onsen and Karuizawa.

Assume a 100-room hotel and the staffing requirement is 30 people. In the Niseko-Hirafu catchment only 11% of that can be filled locally, leaving 27 units of housing to provide. Zao Onsen follows at 26 units, Hakuba at 21, Nozawa Onsen and Kusatsu Onsen at 15 each, Kinugawa Onsen at 12 and Karuizawa at 7. That unit count is the starting point for the Capex, or fixed cost, that belongs in the business plan.

Cost per unit — about ¥7.0 million to build, ¥415,000–¥438,000 a year to lease

Next we set a cost per unit. The build-and-own case assumes a wood-frame apartment format at 25m² per unit (20m² of living space plus 5m² of allocated common area).

The construction unit cost is grounded in MLIT’s Construction Starts Statistics. In 2024, “residential-only buildings” nationwide totaled 59,571,227m² of floor area against ¥14.86 trillion in planned construction cost, which works out to ¥249,428/m² (¥825,000 per tsubo). That is +29.6% above the ¥192,481/m² recorded in 2019.

Build 25m² at that national average and one unit costs ¥6.24 million. Apply market tsubo rates for wood-frame apartments (roughly ¥840,000 in regional cities, around ¥1.1 million in urban areas) and a unit runs ¥6.35–¥8.32 million. Given that these sites are resort locations where materials and labor carry transport costs, this article adopts ¥7.0 million per unit as the base assumption. Land cost is excluded from the model on the assumption that construction takes place within the existing site.

Annual cost is set at ¥318,182 in depreciation, using the 22-year statutory useful life for wood-frame structures, plus ¥200,000 a year for maintenance, utilities and repairs — ¥518,182 per unit per year in total.

For leased corporate housing we use the prefectural monthly rent for privately rented housing from the Statistics Bureau’s 2023 Housing and Land Survey (excluding zero-rent units): Hokkaido ¥51,199, Yamagata ¥49,805, Tochigi ¥51,179, Gunma ¥49,447 and Nagano ¥52,170 (nationwide ¥65,496). Set the employer’s share at 70% and the annual cost is ¥415,000–¥438,000 per unit.

The GOP yield impact is 0.18–0.66pt — and higher rates absorb it

The revenue assumptions for the 100-room model are an ADR of ¥30,000 (tax-exclusive equivalent), OCC of 65% (a modeling assumption, not measured occupancy; full-year), non-room revenue at 35% of room revenue, a GOP margin of 30% and total investment of ¥4.0 billion. That gives room revenue of ¥712 million, total revenue of ¥961 million and GOP of ¥288 million, for a GOP yield of 7.21% before staff housing. The 30% GOP margin is a conservative setting for a first year of operation, against the 38.9% GOP actually recorded across the 91 managed properties disclosed by Invincible Investment Corporation (fiscal period ended December 2024).

The largest impact is the Niseko-Hirafu catchment’s −0.66pt under the build-and-own case. Divide the ¥13.99 million annual cost by 23,725 room-nights sold a year and it comes to ¥590 per occupied room-night, or 1.97% of a ¥30,000 ADR. In other words, if you can add about 2% to the rate, the dormitory Capex is neutralized in earnings terms. In Karuizawa the figure is just ¥153 (0.51%).

For context on actual rates, median estimated settled ADR over the past 12 months runs ¥18,805 in Kutchan Town (N=21 properties), ¥12,519 in Hakuba Village (N=42), ¥17,298 in Kusatsu Town (N=85) and ¥22,206 in Karuizawa Town (N=52). Kutchan, though, swings to ¥63,522 in its peak winter month — the seasonality is extreme. A business plan that assumes ¥30,000 as a full-year average depends on how far peak-season rates can be pushed. We covered how snowfall works on both rate and cost, snow-clearing Opex included, in Snow as Cost or Asset: Winter RevPAR 53% vs Winter ADR 2.66x.

The other relationship worth holding onto is that the higher the rate, the smaller the yield impact of dormitory Capex. For the Niseko-Hirafu catchment’s 27 units under the build-and-own case, moving ADR and total investment together produces the following.

Because the annual cost of staff housing does not depend on room rate, its weight against the denominator falls as the rate rises. A location with a thin labor catchment only works as a business when paired with a high-rate positioning — a relationship worth settling at the stage when the development concept is decided.

Flex the assumptions and the range stays at 0.49–0.80pt — how robust the conclusion is

We tested sensitivity across two axes: staffing requirement (0.25–0.35 workers/room) and local hiring ceiling (10–30%). Results for the Niseko-Hirafu catchment, 100 rooms, build-and-own, are as follows.

All nine combinations land within a 0.49–0.80pt range. The unit count swings between 20 and 33 depending on how the assumptions are set, but the conclusion itself — a fixed assumption that moves yield by roughly half a point — does not. Put the other way round, build that range into the business plan from the outset and it will never show up later as a surprise.

Translating this into investment decisions and management contracts

According to the Japan Tourism Agency’s White Paper on Tourism 2026, a survey of 522 lodging facilities conducted between December 2025 and January 2026 found that 72.2% were short of staff. Labor productivity and pay levels in lodging are both said to run at around 70% of the all-industry average, and the white paper points to the need for improved conditions through capital investment. Providing housing is one component of those conditions.

Limitations of this analysis

Conclusion

Divide the working-age population within a 5km radius by the room count and the 20 catchments spread from 0.17 to 11.22 workers per room, a range of roughly 66x. Assume 0.30 staff per room and a 20% local hiring ceiling and the threshold at which dormitory Capex becomes necessary is 1.5 workers per room — with 7 of the 20 catchments below it.

In the 100-room model the impact is 0.18–0.66pt of GOP yield, or ¥153–¥590 per occupied room-night. Flex the assumptions and it stays within 0.49–0.80pt, so the conclusion is robust. And the higher the rate, the smaller that impact becomes.

That half a point or so is a design variable if it is priced in beforehand, and a gap that appears afterwards if it is not. Alongside land price, floor-area ratio and disaster risk, the labor catchment ratio is worth adding as a line item in site evaluation.

Related reading

• Tokyo Bay 2040: Dual-Growth Zones for Hotel Investment (+23.8%)
• Japan Hotels: 1,968 Skilled Foreign Workers, 1.55 per 1,000 Rooms
• Snow as Cost or Asset: Winter RevPAR 53% vs Winter ADR 2.66x