When hotel running costs come up for discussion, the conversation almost always turns to electricity and gas. But the lodging industry buys another resource in enormous volume — water. According to a survey of roughly 100 properties nationwide by the Eco Subcommittee of the Japan Association of Tourism Facilities, water consumption per guest is 630 liters. For a single overnight stay, that is more than twice what one person uses in an entire day at home (282 liters).

What this article examines is what price that 630 liters is bought at. And the crucial point is that the price is set not by the national government but by the water utility of the municipality where the property stands. Building the same 100-room business hotel in seven different municipalities and stacking up its water and sewer bill produces an annual figure ranging from ¥6.99mn to ¥14.85mn — a 2.12x spread. The ¥7.85mn difference compounds to ¥78.5mn over ten years, and at a 5.0% cap rate translates into ¥157mn of asset value.

Where the 630 liters comes from — of ¥1,300 in utility costs, how much is water

Start with the primary source for the consumption intensity. At the “Symposium on Improving Productivity in Lodging Facilities” held in 2017 by the Japan Association of Tourism Facilities and the Japan Ryokan & Hotel Association, the facilities association reported that a survey of approximately 100 properties nationwide found utility consumption per guest of 630 liters of water, 30.4 kilowatt-hours of electricity, and 4.8 liters of oil. The association set a target of ¥900 for utility costs per guest and stated that achieving it would require holding water down to 600 liters (Kanko Keizai Shimbun, December 5, 2017).

One caveat about the 630-liter figure: it is the average of a sample that includes many ryokan with large communal baths. The association noted that “ryokan with hot springs use large volumes of water, electricity, and heavy oil, and there is considerable room for improvement in the operation of large communal baths,” so the figure should be understood as swinging widely by property type. Indeed, a 26-room hot-spring ryokan on the Sea of Japan coast presented at the same symposium halved its utility cost per guest from ¥1,956 to ¥992 through equipment renovation.

As a guide on the cost side, the Japan Ryokan & Hotel Association’s FY2015 statistics put utility costs per guest at ¥1,259 for large ryokan (100+ rooms), ¥1,307 for mid-size ryokan (31–99 rooms), and ¥1,346 for small ryokan (30 rooms or fewer) — clustered around ¥1,300 regardless of room count. That is a level from more than a decade ago, however, and cannot be used as a current value. The chair of the association’s Eco Subcommittee wrote in a 2023 column that “before COVID, ryokan spent an average of about ¥1,350 per guest on utilities including water; with soaring energy prices I would think it now exceeds ¥2,000” — a practitioner’s estimate rather than a statistic.

Hotels buy water on a different tariff table than households — the structure of meter size and use category

Here is the core of the matter. A water bill consists of a base charge plus a volumetric charge, but the tariff table applied to large commercial users is a different instrument from the residential one. How it is set breaks into three broad systems that vary considerably by municipality.

First, by meter size. The larger the meter diameter, the more sharply the base charge jumps. In Tokyo’s 23 wards (Tokyo Metropolitan Bureau of Waterworks), the base charge for a 13 mm meter is ¥860 per month (before tax), against ¥20,720 for 50 mm, ¥45,623 for 75 mm, and ¥94,568 for 100 mm. Moreover, in Tokyo’s case the cheaper volumetric tiers disappear entirely on large-meter contracts. Households on 13–25 mm meters pay ¥0 for the first 5 cubic meters and ¥22 for 6–10 cubic meters, but on a 50 mm contract ¥372 per cubic meter applies from the very first cubic meter. The same water costs a household ¥22 and a hotel ¥372.

Second, by use category. Sapporo separates “domestic use” from “non-domestic use” and levies a higher base charge on the latter. Fukuoka likewise splits “domestic” and “non-domestic,” with hotels falling into the latter. And here lies a practical trap — Osaka’s “business use” category is not the one that covers hotels. According to the application criteria published by the Osaka City Waterworks Bureau, business use is limited to entertainment establishments such as bars and cabarets, ornamental facilities such as fountains, ponds, and waterfalls, public bathhouses other than bathhouse use, and temporary business activity. Hotels do not fall under it and are instead assigned “general use.” Inferring the rate from the name of the category alone will badly misstate the unit price in Osaka.

Third, base-volume systems. Sapporo’s 50 mm “non-domestic use” base charge of ¥53,900 per month (tax included) is among the highest in the country, but it includes up to 100 cubic meters of consumption per month. Kyoto’s 50 mm meter likewise carries a base volume within its base charge of ¥40,260 (tax included, two-month billing). A comparison that lines up base charges alone misses this structure.

And the sewerage charge must always be shown alongside. Most municipalities bill wastewater discharge as equal to water consumption, so using water automatically generates sewerage. In the 100-room model below, the combined water and sewer burden comes to 1.5–2.2x the water-supply figure alone. Compare sites on water supply only and you overlook 30–50% of the burden.

Building up from consumption intensity — annual water and sewer costs for three scale models

Next, build actual annual consumption and cost up from the consumption intensity. The formula is rooms × occupancy × 365 days × guests per room × consumption intensity. The assumptions are as follows, all of them this article’s own settings.

Run these volumes through the seven municipalities’ tariff tables from the previous section. The result is the table below. The property specification is completely identical; the only variable changed is location.

What stands out is the variation in the share taken by sewerage. In Matsumoto, sewerage at ¥3.81mn exceeds water supply at ¥3.18mn. Sapporo is the reverse, with ¥6.40mn of water against ¥2.93mn of sewerage — less than half. Choose a site on cheap water supply and you can be overtaken on sewerage, as in Matsumoto. There is no alternative to viewing water and sewer together.

Changing scale does not change the picture. For the 200-room city hotel (4,783 m³/month), Fukuoka’s ¥59.27mn against Matsumoto’s ¥25.24mn is a 2.35x gap and ¥34.03mn a year; for the 80-room hot-spring ryokan (2,944 m³/month), Fukuoka’s ¥35.49mn against Matsumoto’s ¥15.30mn is 2.32x and ¥20.19mn a year. The more water consumed, the more the escalating volumetric structure bites, and the gap widens rather than narrows.

This curve is highly instructive for investment decisions. In Matsumoto and Shimonoseki the average unit price keeps falling as volume rises, whereas Fukuoka, Tokyo’s 23 wards, and Osaka bottom out around 200 m³ per month and then turn upward. Municipalities that lean heavily on escalating tariffs are designed to penalize properties with large communal baths and larger properties generally. At 1,314 m³ per month, the marginal cost of one additional cubic meter is ¥425 in Matsumoto against ¥1,055 in Fukuoka — a 2.5x spread.

Translating into NOI — 2.24 points of revenue ratio, ¥157mn of asset value

So far these are absolute costs. To use them in an investment decision they must be translated into the language of revenue and NOI.

First, the structure. The volumetric portion of the water and sewer bill accrues in proportion to guest-nights. Revenue is also proportional to guest-nights (more precisely, to rooms sold). Therefore the volumetric portion’s ratio to revenue is barely affected by either occupancy or room count. What affects it is only four things: guests per room, consumption intensity, the volumetric unit rate, and ADR. The base charge, by contrast, is a fixed cost and dilutes as scale and occupancy rise. The intuition that “water costs thin out if you build bigger” applies only to the base-charge portion.

On that basis, look at the revenue ratio when ADR is fixed at ¥12,000 and only the location changes. This measures how many points apart two hotels selling the same product at the same price end up, purely on water charges.

Another angle: what happens when each city’s actual ADR level is used as the denominator. Using MetroEngines Research & Consulting’s estimated settled ADR (12-month average from August 2025 to July 2026), the room-revenue ratio for the 100-room business hotel model comes out as follows.

The ranking reverses. Even where the absolute cost is mid-range, a market with relatively low ADR carries a heavier revenue ratio. Conversely, Tokyo’s 23 wards, expensive in absolute terms, land mid-table because ADR is also high. In site evaluation the correct sequence is not to hunt for municipalities with cheap water, but to ask what ratio the water bill represents against the ADR achievable in that market. Lodging taxes follow the same structure as a cost set by municipal ordinance, and the spread of adopting municipalities and the weight of the burden are a separate subject in their own right.

As for the impact on the yield on total investment, assuming a 100-room business hotel (18 m² rooms, guest-room ratio of 68%, giving 2,647 m² of gross floor area or 801 tsubo, and construction cost of approximately ¥1.93bn at the 2025 steel-frame unit rate of ¥2.405mn per tsubo from the Ministry of Land, Infrastructure, Transport and Tourism’s construction starts statistics), the effect is 0.39 pt at ¥2.0bn of total investment, 0.31 pt at ¥2.5bn, and 0.26 pt at ¥3.0bn. On the upward pressure on construction cost itself — the denominator here — Construction +41% and Labor Crunch Erase Japan’s 2027–2029 Hotel Supply goes into detail. On its own the effect is small, but what matters is that it is a permanent difference that neither design nor operations can move.

Sensitivity analysis — how far does the conclusion move when consumption intensity is flexed

The calculations so far place the 100-room business hotel’s consumption intensity at 400 L/guest-night. As noted, this is this article’s assumption, not a survey value. The verified survey average is 630 L/guest-night (from a sample containing many ryokan with large communal baths), while a property with water-saving equipment and no communal bath could fall below it. Consumption intensity is the assumption that most strongly drives this article’s conclusion, so it is flexed across 300–630 L/guest-night to test robustness.

Two things can be read from this. First, the gap between municipalities does not disappear wherever consumption intensity is set. Even at the most conservative 300 L/guest-night the Fukuoka–Matsumoto gap is 2.01x, and at 630 L/guest-night it widens to 2.25x. If anything, a higher intensity increases the weight of the volumetric portion, so the gap expands slightly. Cost differences arising from location are not the kind of thing a property’s own water-saving efforts can cancel out.

Second, uncertainty on the absolute figure is large. The same Fukuoka property costs ¥10.70mn a year at 300 L and ¥24.41mn at 630 L — a 2.3x range. For a specific project’s business plan, therefore, one should not simply adopt this article’s 400 L but build consumption up from the equipment specification: communal baths, food and beverage, and whether laundry is done in-house. What this article asserts is not the absolute amount but the size of the relative difference between municipalities.

The practical implication is this. If a site selection is comparing Fukuoka with Matsumoto, then even if the intensity assumption is wrong, the direction of the conclusion — a difference of ¥184 to ¥465 in ADR — does not change. The range is wide but the sign is stable, and that is enough to act on. Conversely, if the aim is to narrow that range, what is needed is not closer scrutiny of tariff tables but a settled equipment specification.

The 2026 revision rush — the headline revision rate is not the property’s revision rate

2026 is a year in which water and sewer tariff revisions are concentrated. The principal cases, confirmed against official announcements, are as follows.

Here is a point of first-order importance for investment decisions. The published “average revision rate” does not match the revision rate as a hotel experiences it. The average revision rate is the growth rate of the utility’s total water revenue, and it is pulled by a distribution in which small residential users predominate. Applying each revision to this article’s 100-room business hotel (1,314 m³/month, 50 mm meter) produces effective increases that scatter as follows.

Matsumoto’s headline 20.11% comes out at an effective 4.7%. The reason is clear: Matsumoto’s revision covers the standby (base) charge only, leaving the volumetric rate of ¥176 per cubic meter untouched. For a heavy user the great majority of the bill is volumetric, so a rise in fixed charges alone is diluted. Sapporo’s sewerage revision, by contrast, raises the unit rate on every tier and so lands at an effective 19.9%, close to the headline. The Chiba Prefectural Waterworks also revises both base and volumetric charges, giving a high effective 16.8%.

The Shimonoseki case raises a further point. The city is using the national priority support grant to temporarily discount the increase, and it extended the freeze period by two months, from meter readings through September 2026 to readings through November. But this is time-limited, and it covers water charges only — sewerage is excluded. From the February 2027 reading the new tariff applies in full. Look at recent invoices during acquisition due diligence and what you will see is the post-discount amount. Project future cash flow straight off recent actuals and you will understate the cost — a point that warrants care.

The seven municipalities mapped, and the geography of the burden

Plot the municipalities covered here on a map and no geographic regularity in the size of the burden appears. It is not a simple matter of big cities being expensive and rural areas cheap; each utility’s asset renewal burden, the trajectory of its served population, and the design philosophy of its tariff structure work individually.

Site comparison checklist — five items verifiable before signing

Water and sewer is an expense line that cannot be moved after acquisition or completion. That is precisely why it is most worth checking at the site selection and design stage. The list below is confined to items that can be completed through published materials and pre-application consultation.

Conclusion — as the sixth axis of location evaluation

To summarize what this article has verified.

First, the lodging industry buys water in bulk. The 630 liters per guest (Japan Association of Tourism Facilities, survey of approximately 100 properties nationwide) is more than twice what one person uses in an entire day at home. That comes to 15,768 cubic meters a year for a 100-room business hotel and 57,396 cubic meters a year for a 200-room city hotel.

Second, the price differs sharply by municipality. The annual water and sewer cost for an identically specified 100-room business hotel ranged from ¥6.99mn to ¥14.85mn — a 2.12x spread — across the seven municipalities examined. Increase the scale and the escalating structure bites, widening the gap: for the 200-room city hotel it reaches 2.35x and ¥34.03mn a year.

Third, looking at water supply alone misses 30–50%. Sewerage charges accrue in step with water consumption, and in municipalities such as Matsumoto sewerage exceeds water supply. There is no alternative to comparing the two together.

Fourth, this difference cannot be closed by operational effort. There is genuine room to cut consumption through water-saving equipment and better management of communal baths (the 26-room hot-spring ryokan cited above halved utility costs per guest from ¥1,956 to ¥992). But the unit price itself is set by the utility through ordinance, and the customer cannot choose it. There is no option to switch retailers as there is with electricity.

Fifth, and consequently, it bites at the site selection stage. For a 100-room business hotel, a property in the most expensive municipality needs ¥269 more per room-night in ADR to produce the same NOI as an identical property in the cheapest. That is 2.2% of a ¥12,000 ADR. Whether the location carries demand able to absorb that price difference is the decision input. To repeat: this is not an argument that particular municipalities are unsuited to investment. At the same ADR, the conditions for viability change — that is all. And once the conditions are known, they can be built into both ADR design and equipment specification.

Alongside station distance, land price, climate, and labor cost, water and sewer is worth adding to the investment decision’s variable table as one more immovable location cost. Hazard maps and disaster-resilience capex belong on the same framework as a further axis, and together with this sixth axis they let a full variable table for a site be assembled.

Related Reading

• 2026 Electricity Price Surge: Supply-Side Cost Impact Analysis for Hotel Operations
• Cleaning Dispatch Rate Surge & Mid-Size Hotel Pressure: 2026 H2 Polarization
• Snow as Cost or Asset: Winter RevPAR 53% vs Winter ADR 2.66x
• Leasehold Hotel Development: 1.38x Land Price Reach Across 9 Markets
• Construction +41% & Labor Crunch Erase Japan’s 2027-2029 Hotel Supply
• Hotel vs Condo Land Rivalry: ADR-Implied Feasible Land Price Map