Imagine building the same 200-room business hotel, from the same business plan, in Tokyo and in Yamanashi. Construction cost, assumed ADR, the GOP margin assumption — almost every field an investor fills in first is identical. Yet one variable separates the two in the first year of operation: how many of the people you need you can actually hire in that location. This article lays out that “volume and geography of labor supply” across all 47 prefectures — as a distribution of headcount rather than of wage rates — and sets out where it belongs in a business plan.

HotelBank has covered the labor cost of the accommodation industry from three angles so far. The minimum-wage series asked a pricing question — how far an increase in the wage rate can be absorbed through ADR. The employee-housing and workforce-infrastructure series was about recovering Capex. The labor-saving investment series was about replacing people with equipment. This article overlaps with none of them. Its subject is the geography of hireability: the single point that “the same plan is easier or harder to staff depending on the prefecture.” Wage rates are fixed as a common assumption across every prefecture; only the distribution of headcount is allowed to move, and we look at how that shows up in GOP and yield.

The national picture: 2.9x in a year, still only 14.9% of the quota

Start with the overall volume. Table 5 of the Immigration Services Agency’s “Number of Foreign Residents with Specified Skilled Worker Status (end-December 2025)” puts Specified Skilled Worker (i) residents at 382,341 across all fields. Of these, 1,968 are in the accommodation sector. Six months earlier, at end-June 2025, the figure was 1,265; a year earlier, at end-December 2024, it was 671 — roughly a 2.9x increase over twelve months.

A more recent preliminary figure exists. The agency’s “Intake Targets and Resident Numbers for Specified Skilled Worker (i) (as of end-March 2026, preliminary)” puts accommodation-sector residents at 2,207, with an annual increase of 1,365. That release, however, covers sector totals only and carries no prefecture breakdown. Reading the geography prefecture by prefecture requires the semi-annual Table 5 — which is why every prefecture figure in this article is as of end-December 2025.

The denominator is what deserves attention. The same preliminary table sets the accommodation sector’s intake target at 14,800, against which the fill rate is only 14.9%. With the all-field fill rate at 50.2% (404,527 residents against a target of 805,700), accommodation is among the fields with the most headroom left against its own quota. Put the other way round: if the current growth continues, today’s geography is a snapshot of something still being repainted. How far that growth runs also depends on the design of the scheme itself — in particular on the treatment of the “(ii)” status, which removes the cap on length of stay.

Raw counts and per-room density rank differently: Gunma 6.16, Tokyo 0.39

Line the raw headcounts up by prefecture and the top of the list reads Hokkaido 185, Kanagawa 141, Gunma 113, Hyogo 104, Okinawa 100. But since room stock varies more than tenfold between prefectures, that ordering is little more than a ranking of market size. So divide each prefecture’s resident count by its room stock and restate it as persons per 1,000 rooms. Nationally that is 1,968 ÷ 1,324,155 rooms, or 1.49 — roughly one person per 670 rooms.

Switch to that yardstick and the ranking moves sharply. Gunma leads at 6.16, followed by Gifu at 4.80, Wakayama at 3.71 and Yamanashi at 3.26 — all prefectures whose room stock is anchored by onsen ryokan serving dinner and breakfast. Tokyo and Osaka, ninth and tenth on raw counts, fall to 0.39 and 0.55, less than half the national level. Hokkaido, first on raw counts, sits 18th at 2.01; Kanagawa, second on raw counts, sits 8th at 2.97. How far a prefecture moves differs entirely from one to the next. The gap between the highest and lowest density is about 15.8x between Gunma and Tokyo, and widens to 41x once Aomori (0.15) is included.

This reshuffle is not a statistical artifact. The accommodation-sector visa covers person-facing work — front desk, planning and PR, guest service, restaurant service — so business models with more person-facing work per room, such as ryokan and resorts serving dinner and breakfast, need more staff per room. That prefectures with onsen destinations cluster at the top of the density ranking reads as a direct reflection of business-model mix. Where the stock is dominated by urban business hotels, the headcount needed per room is smaller to begin with, so the same number of people produces a lower density.

Overlaying the domestic labor market: 2.30x for service staff, 1.4pt off the peak

The visa scheme is a supplementary channel; most accommodation-industry staff are still hired from the domestic labor market. So overlay the occupational data from the Ministry of Health, Labour and Welfare’s “Employment Referrals for General Workers.” Nationally, in July 2026, the openings-to-applicants ratio stood at 2.30x for service and waiting staff and 2.18x for food preparation workers — more than double the 1.05x across all occupations. More openings per job seeker means, in plain terms, occupations where employers compete hard for candidates.

The direction of travel, however, is toward easing. The ratio for service and waiting staff peaked at 3.70x in December 2022 (the current occupational classification series starts in April 2022), then fell steadily to 2.30x by July 2026 — a decline of 1.40 points in about three years and seven months. Food preparation fell from the low threes to 2.18x over the same period. In the sense that applications now gather more readily per opening, the time and effort a search requires can be assumed to be looser than a few years ago. When writing “how many months to hire” into a business plan, there is no need to extrapolate the felt tightness of the post-pandemic years.

One caution is needed here. The ministry publishes occupational and prefectural series separately, but no occupation-by-prefecture cross-tabulation is published. A value such as “the openings ratio for service staff in Yamanashi” therefore does not exist. This article treats the occupational level (national, 2.30x for service staff) and the regional level (prefecture, all occupations) as separate axes, and reads regional differences on the all-occupations prefecture basis.

On that prefecture basis (place of work, July 2026), the highest readings are Yamanashi at 1.68x, Toyama at 1.67x, Fukui at 1.64x and Shimane at 1.59x. The lowest are Osaka at 0.92x, Fukuoka at 0.94x, Kanagawa at 1.00x, Wakayama at 1.00x and Hokkaido at 1.01x. In the major metropolitan labor markets, openings fall below — or roughly match — the number of job seekers, an environment in which applications gather readily. Rather than the simple picture of “the more rural, the harder to hire,” the statistics show that on the openings ratio it is the prefectures of Chubu, Hokuriku and San’in that are tight, while the big metropolitan areas are comparatively loose.

Sorting the two supply channels into quadrants: a labor supply map of all 47 prefectures

Take the two axes established so far — the thickness of the visa channel (residents per 1,000 rooms) and the looseness of the domestic labor market (the openings ratio) — and cut each at its median (1.36 and 1.26x) to form quadrants. Bubble size is room stock.

In correlation terms, the larger a prefecture’s room stock, the lower its openings ratio (−0.49) and, at the same time, the thinner its residents per 1,000 rooms (−0.20). The two supply channels broadly work in opposite directions, so in volume terms they tend to offset each other. That is precisely why judging “easy or hard to hire” from a single indicator misreads the situation — and why the quadrant view earns its place.

The “labor catchment” works below prefecture level: walk-in employment differs by more than 100x

The prefecture is a coarse unit for an investment decision. Recruiting actually happens around the property, and the size of the commutable area is the applicant pool. Taking establishment counts and employee numbers within a 1.5 km radius for four representative areas from the Economic Census for Business Activity (Ministry of Internal Affairs and Communications / Ministry of Economy, Trade and Industry), the orders of magnitude differ sharply even within one country.

Within 1.5 km of Shinjuku, 483,032 people work, 35.8% of them in office-type jobs — meaning talent moves back and forth with other industries inside the same catchment. Within the same radius of Kusatsu Onsen, the figures are 3,754 and 4.3%: the local employment structure itself leans toward accommodation and food service. The difference should be treated not as “which is better” but as a difference in how wide a recruiting footprint the plan assumes. Where the walk-in labor market is thin, shuttle transport to widen the commutable area — or live-in dormitories and staff housing — enters the business plan as a mandatory component from the outset.

Translating the staffing fill rate into GOP and yield: a 200-room business hotel and a 120-room resort

From here we convert the thickness or thinness of labor supply into business-plan figures. The key is to hold the wage rate constant across every case and move only the fill rate of headcount. The rate discussion belongs to the minimum-wage work; this article isolates how volume alone bites.

Start by building up the required headcount. For housekeeping we use the operational rule of thumb of 25-35 minutes per room for checkout cleaning and roughly six productive hours (360 minutes) per cleaner per day. For the front desk we take a staffing example of three on day shift and two to three on night shift at 200-room scale, then multiply by a factor of 1.4 — the two-days-off-per-week equivalent — to cover all 365 days and convert to FTE. On that build-up, the 200-room business hotel comes to 34.0 FTE (0.17 per room), and the 120-room resort, with food and beverage added, to 55.0 FTE (0.46 per room).

Next, model what happens when the fill rate falls short of 100%. Responses on the ground broadly split in two: topping up with agency or spot workers, or restricting the number of rooms offered for sale. This article adopts an intermediate assumption — half the shortfall is covered by agency staff at a 1.5x rate, and the other half is absorbed by restricting rooms offered. Payroll falls only slightly, because the direct cost of the unfilled positions disappears while the agency premium is added on top, whereas revenue falls further as occupancy declines.

The result: a 10-point drop in the staffing fill rate erodes GOP by 6-7%, and a 20-point drop by 12-14%. In yield terms, the 200-room business hotel moves from 6.23% to 5.47% (−0.76pt) and the 120-room resort from 5.02% to 4.32% (−0.70pt). The more labor-intensive resort is the more sensitive in percentage terms (−13.9% versus −12.2%). A yield difference of roughly 0.7 points is about the same magnitude as a 10% move in construction cost, which makes this a variable that deserves the same weight in a business plan as the precision of the construction estimate.

Return to the quadrants. In a Quadrant D prefecture, where both supply channels are still ahead of the market, assuming a 100% fill rate in the first year of operation is somewhat bullish. In Quadrant A or B, where an intake track record and living-support infrastructure already exist nearby, the ramp-up under the same plan is easier to call. That the same business plan carries a different probability of working out depending on the prefecture is, in the end, a question of where these 0.7 points are placed.

How investors should build it in: four practical points

When the geography of labor supply is written into a business plan, there are broadly four variables to move. In practice all four are better treated not as “how to avoid a disadvantage” but as design questions about how to open up room for the plan to work in a thin area.

One point on the visa channel deserves attention on an investment time horizon. Against the accommodation sector’s 14,800-person intake target, residents numbered 2,207 in the end-March 2026 preliminary release, a fill rate of 14.9%. Were that target reached, apportioning it across the 1,324,155 rooms of stock used here would put national density at 11.18 per 1,000 rooms — 7.5 times today’s 1.49. Even Gunma, the densest prefecture at 6.16, sits at barely half that level. Today’s geography is a way station, and for an investment decision spanning three to five years of operation, not treating current density as a fixed value matters.

Summary

Specified Skilled Worker residents in the accommodation sector numbered 1,968 at end-December 2025, roughly 2.9x higher than a year earlier. Ranked by raw count, however, the list only reveals the ordering of market size and is of no use for an investment decision. Restated per 1,000 rooms, the ranking reshuffles: from Gunma at 6.16 to Tokyo at 0.39 is 15.8x, and 41x once Aomori is included. The prefectures at the top are those with onsen ryokan serving dinner and breakfast — business-model mix showing through directly.

Overlay the domestic labor market indicator and the two supply channels turn out to work broadly in opposite directions. Prefectures with large room stock have low openings ratios (correlation −0.49) and, at the same time, thin visa-channel density. Judging hireability from either alone therefore misreads it. Of the four quadrants cut at the medians, the 11 prefectures (150,908 rooms) where both the visa channel and the domestic market are still ahead of them are where designed countermeasures do the most work.

And the value of those countermeasures can be put in figures. A 10-point drop in the staffing fill rate erodes GOP by 6-7%, a 20-point drop by 12-14%, and moves GOP yield by roughly 0.7 points. That is on par with a 10% move in construction cost, and deserves the same weight in a business plan as the precision of the construction estimate. Where labor supply is thick, operations are easier to assemble; where it is thin, staff housing, labor-saving investment and operator selection open up room for the plan to work. Geography is not a constraint — it is a variable to be built into the plan.

Related reading

• Japan Hotels: 1,968 Skilled Foreign Workers, 1.55 per 1,000 Rooms
• Japan Ends the ¥1.06M Wage Threshold Oct 2026: Hotel Shift Design
• Face Recognition Check-In: Survival Strategy for Mid-Size Business Hotels (80 vs 150 Room GOP Simulation)