In hotel site acquisition decisions, water is usually filed on the operating-expense side of the business plan as a “monthly variable cost.” But the moment a site is secured, a separate set of line items sends cash out the door in a lump sum, before service even begins: the water supply connection fee (also called the water service connection charge or water supply contribution), the sewerage beneficiary contribution, and the water service installation work cost. This article compares these “initial-side infrastructure costs” in actual yen by municipality, and estimates the total gap for a 100-room hotel together with its impact on acquisition cost, depreciation and yield.

Connection fees are not “operating expenses” — they are intangible depreciable assets amortized over 15 years

The main reason connection fees get overlooked in site acquisition decisions is that where they land in the accounts is not intuitive. The National Tax Agency’s Tax Answer No.5463, “Tax treatment of development contributions paid in connection with residential land development,” divides development contributions into three categories according to their character.

What matters in hotel development practice is that water and sewer contributions are not capitalized into land, but treated as intangible depreciable assets that can be amortized over 15 years. Once folded into the acquisition cost of land, an amount can never be depreciated; as an intangible depreciable asset it is expensed against taxable income period by period. Even among “lump sums paid at the development stage,” a water connection fee and an estate-road contribution work on cash flow in completely different ways. It is worth sorting each contribution into the right category at the point where the up-front cost table of the business plan is built.

In addition, many municipalities state explicitly in their ordinances that amounts already paid are not refunded. Article 29 of Naha City’s water supply ordinance, for example, provides both for payment at the time of application and for non-refund of amounts already paid. If a design change reduces the meter diameter, the fee already paid does not come back. This irreversibility is a practical constraint on when to freeze the design.

Up to ¥5.11 million at a 75 mm meter — zero in Tokyo’s 23 wards, ¥5.11 million in neighbouring Yokohama

Using the water supply ordinances and fee schedules published by each municipality, we compiled the connection fee by meter diameter for 10 major municipalities in actual yen. The comparison covers the 40 mm and above range that a hotel-scale water supply implies.

Source: prepared by MetroEngines Research & Consulting from the water supply ordinances and fee schedules of each municipality

The results split sharply. Tokyo’s 23 wards have no water supply connection fee scheme at all. The Tokyo Metropolitan Waterworks Bureau laid its mains publicly from the outset, and within the 23 wards no connection fee is payable regardless of meter diameter (service-line construction costs are still charged separately). By contrast, Yokohama charges ¥5,115,000 at a 75 mm meter and ¥8,745,000 at 100 mm, collected as a lump sum on installation — for a city within 30 km of central Tokyo serving the same metropolitan market. For sites targeting the same catchment, simply crossing an administrative boundary determines whether more than ¥5 million of up-front cash goes out.

Even among municipalities that do levy a fee, the spread is wide. At 75 mm, Yokohama’s ¥5,115,000 against Atami’s ¥1,738,000 is 2.94x; at 50 mm, Yokohama’s ¥2,145,000 against Atami’s ¥638,000 is 3.36x; and at 100 mm, Sapporo’s ¥10,120,000 against Atami’s ¥3,542,000 is 2.86x. Against the 2.1x inter-municipal gap in running water and sewer tariffs covered in our earlier article, the initial-cost side shows an even wider dispersion.

One step up in meter diameter costs millions of yen — and the shape of the escalation curve differs by municipality

Connection fees do not scale linearly with meter diameter; they escalate steeply. Because the gradient of that escalation differs by municipality, the design decision of “which diameter to plan for” carries different weight depending on where the site is.

Source: prepared by MetroEngines Research & Consulting from the water supply ordinances and fee schedules of each municipality

Take Yokohama: stepping up from 25 mm to 40 mm alone jumps the fee from ¥82,500 to ¥1,402,500. From 40 mm to 50 mm adds another ¥742,500, and from 50 mm to 75 mm adds ¥2,970,000. In Fukuoka City, the single step from 50 mm to 75 mm adds ¥2,068,000; in Kyoto City, the step from 75 mm to 100 mm adds ¥6,798,000. Kyoto City’s 100 mm fee (¥10,076,000) is 11.2 times its own 50 mm fee (¥902,000).

The practical implication is straightforward. The choice of supply method (receiving-tank system versus direct booster connection) and the sizing of the required meter diameter are, simultaneously, an engineering decision and an investment decision worth millions of yen. The steeper a municipality’s escalation, the greater the economic value of engineering the diameter down one step at the schematic design stage. Conversely, in zones where the fee is zero — Tokyo’s 23 wards, or Osaka City at 50 mm and below — there is almost no financial incentive to do that work at all. The same engineer running the same exercise recovers a completely different amount depending on location.

Connection fee map for 10 major municipalities (75 mm meter)

Circle size corresponds to the 75 mm connection fee. Source: prepared by MetroEngines Research & Consulting from the water supply ordinances and fee schedules of each municipality

Plotted on a map, it becomes clear that fee levels cannot be explained by regional bloc. Sapporo in Hokkaido (¥4,950,000) and Yokohama in the Tokyo metropolitan area (¥5,115,000) sit close together, while Tokyo’s 23 wards — in that same metropolitan area — are at zero; and within the Kyoto–Osaka–Kobe region, Kyoto City (¥3,278,000), Osaka City (¥2,530,000) and Kizugawa (¥4,752,000) all diverge. Connection fees are not a regional market rate but the outcome of a financial judgment each water utility sets individually by ordinance, and they are not continuous even between adjacent municipalities. When comparing sites, prefecture-level or region-level estimates will not do; the ordinance of the specific municipality must be checked case by case.

Sewerage beneficiary contributions scale with site area — a 4.9x spread in unit rates

The other initial-cost line item alongside the connection fee is the sewerage beneficiary contribution. Rather than meter diameter, it is calculated by multiplying the area of the subject land by a unit rate in most cases, so it bites harder the larger the hotel site. Unit rates for major municipalities, and the resulting charge on an assumed 2,000 m² site, are as follows.

Source: prepared by MetroEngines Research & Consulting from materials published by each municipality (2,000 m² site assumed)

Unit rates run from ¥200/m² in Kumamoto to ¥700/m² in Matsudo, a spread of 3.5x; including Sapporo’s Chubu contribution district at ¥144/m², the spread widens to 4.9x. On a 2,000 m² hotel site that is a range of ¥288,000 to ¥1,400,000 — smaller in absolute terms than the connection fee, but because it scales with site area, it works very differently on large resort developments. A 10,000 m² resort project at Matsudo’s rate comes to ¥7 million, putting it on par with the connection fee.

Note too that some municipalities levy an area-proportional lump sum on the water side as well. Atami charges a water development contribution of ¥200 per m² (tax-inclusive) for residential land development and ¥600 per m² (tax-inclusive) for multi-unit residential construction, separate from both the connection fee and the sewerage beneficiary contribution. Because line-item names differ from one municipality to the next, checking only for a “connection fee” will miss charges. In site due diligence, the reliable approach is to ask both the water utility and the sewerage authority for every lump-sum charge triggered by a new installation. Municipal ordinances drive project costs through more line items than these: our analysis of hotel parking ordinance requirements, which vary 2.5x across cities, covers the same structure, in which city-by-city additional regulation feeds through to the business plan.

Assumed meter diameter for a 100-room hotel, and the total initial infrastructure cost

Which meter diameter a 100-room hotel needs depends on the supply method, whether it has a large communal bath, and the scale of its kitchens. As covered in our earlier article, water use runs at roughly 630 L per guest; at 100 rooms, double occupancy and 80% occupancy (a modelling assumption as of August 2026, full-year average basis, not an actual result), daily demand comes to around 100 m³. A limited-service hotel on a receiving-tank system typically falls in the 50–75 mm range, while a full-service property with a large communal bath or major banquet kitchens falls in the 75–100 mm range. Here we take a 100-room limited-service hotel, a 75 mm meter and a 2,000 m² site as the base case for the total.

Even under identical conditions — 100 rooms and a 75 mm meter — total initial infrastructure cost is distributed across a range from ¥0 to roughly ¥6.52 million. Restated per room, that is ¥0 to ¥65,210 — comparable to the FF&E cost of a single guest room. How the per-room investment figure itself should be set is reverse-calculated from planning data in our reverse-calculation of per-room construction cost at ¥20.67 million, derived from a 35.0 m² per-room floor area across 31 planned projects,. Substitute a 100 mm design on a large site and a Sapporo-level connection fee of ¥10,120,000 plus the beneficiary contribution puts the lump sum for a single property in the ¥10 million range.

The yield impact is 1–3 bp — but three reasons not to leave it at “small”

The NOI yield used here is annual net operating income (NOI, i.e. operating profit before depreciation) divided by total project cost. The unit of change, bp (basis point), is 0.01 points, so 100 bp equals 1.00 point. To gauge how this initial-cost gap feeds into investment yield, we model it on the actual results of a listed hotel REIT. As a limited-service benchmark, Ichigo Hotel REIT (いちごホテルリート投資法人) reported July 2026 full-portfolio results of 84.4% occupancy (as of July 2026, full portfolio, N=27 properties), ADR of ¥9,684 and RevPAR of ¥8,175 (RevPAR −1.6% year on year, 27 properties). Applying that level to 100 rooms and assuming a new development with a total project cost of ¥2.2 billion, the sensitivity is as follows.

Source: MetroEngines Research & Consulting (modelled figures)

For a single property, the impact on NOI yield stays within 0.7 to 2.6 bp. That is the natural result of a few million yen set against a ¥2.2 billion project cost. Looking at that number alone, it is tempting to conclude it is within the margin of error. In the practice of site acquisition, however, three points make it impossible to ignore.

First, as pre-opening cash it bites in absolute terms. A Yokohama-level 75 mm connection fee of ¥5,115,000 plus a sewerage beneficiary contribution of ¥1,400,000 is cash leaving the business before completion, with zero revenue coming in. As a ratio it is 1.3 bp; on the cash-flow schedule during the development period it stands as a ¥6.52 million line in actual yen. That is large enough to affect additional equity contributions or the sizing of a credit facility.

Second, across a multi-property rollout the order of magnitude changes. Stack a Yokohama-level fee of ¥5.11 million across ten properties and it becomes ¥51.15 million. For a regionally expanding chain comparing dozens of candidate sites, simply prioritising municipalities with lower fees moves initial investment by tens of millions of yen. What is noise for one property is decision-grade information across the pipeline.

Third, it flows through the P&L for 15 years via depreciation. Booking water and sewer contributions as intangible depreciable assets amortized over 15 years, per the National Tax Agency’s treatment, a ¥5.11 million connection fee is recognised as roughly ¥340,000 of depreciation expense each year for 15 years. Unlike other development contributions that get capitalized into land, it is expensed against taxable income — which, on a cash-flow basis, works in the project’s favour. Bury the connection fee in the land account of the up-front cost table and that depreciation benefit never makes it into the business plan.

Against market ADR — the initial-cost gap is absorbed by a few nights of revenue

Estimated settled ADR compiled by MetroEngines Research as of July 2026 stands at ¥12,784 in Kanagawa, ¥12,642 in Tokyo, ¥10,873 in Fukuoka and ¥8,761 in Osaka. On a most-recent-13-month median basis (July 2025 to July 2026), the figures are ¥14,164 in Kanagawa, ¥13,855 in Tokyo, ¥14,305 in Kyoto, ¥11,708 in Fukuoka and ¥10,252 in Osaka.

Divide Yokohama’s 75 mm connection fee of ¥5,115,000 by Kanagawa’s estimated settled ADR of ¥12,784 and it comes to 400 room-nights. For a 100-room hotel running at 80% occupancy (a modelling assumption as of August 2026, full-year average basis), that is five days of rooms revenue. Once the property is in its operating phase the amount is absorbable, which is a further reason it is not a “factor that determines yield.” What matters is the timing — it goes out as a non-refundable lump sum before opening, when revenue is zero.

Sensitivity compared — against 3 bp from connection fees, ADR and occupancy move 290 bp

The previous section concluded that differences in initial infrastructure cost move NOI yield by only 0.7 to 2.6 bp. That figure only means something once set against a comparison. Using the same 100-room, ¥2.2 billion model, we lay out on two axes how yield swings when the numerator variables — ADR and occupancy — are flexed. The blue-outlined cell is the same base case as the previous section (ADR ¥9,684; occupancy 84.4%, as of July 2026, full portfolio, N=27 properties).

The bottom-left of the grid (ADR ¥8,500, occupancy 76.0% — the lower bound of the modelling range as of August 2026, full-year average basis) gives 4.54%; the top-right (ADR ¥11,500, occupancy 92.0% — the upper bound of the modelling range as of August 2026, full-year average basis) gives 7.43%. The swing reaches 290 bp. Since the connection fee effect seen in the previous section was at most about 3 bp, market-side variation moves yield by roughly two orders of magnitude more.

What this contrast shows is not that connection fees can be ignored. It shows that connection fees belong to a different decision layer. ADR and occupancy are variables moved by operations after opening, and in advance they can only be set as ranges. Connection fees and beneficiary contributions, by contrast, are among the few up-front line items whose amount is fixed by ordinance the moment a site is chosen, readable to the yen at the comparison stage. Debating only the wide-range variables while leaving the readable ones unread inverts the proper order of a feasibility study.

There are two practical implications. First, ranking sites by the size of the connection fee is poor practice. The demand-generating power of a location that moves yield by 290 bp is not offset by an initial cost worth 3 bp. Second, when candidates are level on locational merit, the connection fee is a clear tiebreaker. Across a multi-property pipeline in particular, as seen above, it accumulates into a gap of some ¥50 million over ten properties.

Building this into site acquisition decisions and credit review

On that basis, here are the practical points to observe when treating initial infrastructure cost as a location evaluation axis.

Conclusion

The water supply connection fee, payable as a lump sum before a hotel opens, ranges at a 75 mm meter from ¥0 in Tokyo’s 23 wards to ¥5,115,000 in Yokohama, with a 2.94x spread even among the municipalities that levy it. At 100 mm, Sapporo’s ¥10,120,000 against Atami’s ¥3,542,000 is a 2.86x spread. Sewerage beneficiary contribution rates likewise span 4.9x, from ¥144/m² to ¥700/m². Total initial infrastructure cost for a 100-room hotel with a 75 mm meter on a 2,000 m² site ranges from ¥0 to roughly ¥6.52 million, or ¥0 to ¥65,210 per room.

For a single property the impact on NOI yield stays within 0.7 to 2.6 bp, and on its own it rarely overturns an investment decision. But on three counts — it leaves as non-refundable cash before opening with zero revenue; it accumulates into tens of millions of yen across a multi-property rollout; and under the National Tax Agency’s treatment it is expensed against taxable income as an intangible depreciable asset over 15 years — it is a line item that belongs, correctly placed, in the up-front cost table of the business plan.

The most practical implication is that the level of a connection fee cannot be estimated from a regional market rate. Within the same Tokyo metropolitan area, the 23 wards are at zero and Yokohama at ¥5.11 million. Within the same Kyoto–Osaka–Kobe region, Osaka City is ¥2.53 million, Kyoto City ¥3.28 million and Kizugawa ¥4.75 million. The moment an administrative boundary is crossed, the figure changes discontinuously. When comparing candidate sites, go to the ordinance of the specific municipality. That is the only reliable method.