The Underwriting Files: A 10% Yield in Tokyo Wasn't the Answer — It Was the Question

This article is part of The Underwriting Files, where we walk through real listings the way we actually evaluate them. Identifying details — exact address, chōme, station, listing portal, and broker — have been removed or generalized to avoid identifying a property that may still be listed. No wrongdoing by any party is alleged. This is general information, not investment or legal advice.

A 10.05% gross yield on a building inside Tokyo's 23 wards, an eight-minute walk from a station with a direct line into central Tokyo, is the kind of number that makes you stop scrolling. In a market where prime central-ward condos trade at 3–4% gross, double digits in the capital looks like a mistake in your favor.

We stopped scrolling. We did not say yes. This is the walk-through of why — the reasoning chain that turned "10.05%" from an answer into a question. The value here is not the number. It is the order in which we distrusted it.

The Listing at a Glance

Here is the property as the broker's summary presented it, translated and stripped of anything that would identify it.

Two figures in that table are doing all the work, and they point in opposite directions. The 10.05% is the reason to look. The word vacant — applied to every unit — is the reason the 10.05% is not real yet. Everything below follows from holding those two facts next to each other.

Red Flag #1 — Two Listings, Two Build Years

The single largest discrepancy is the build year. The summary we were shown listed the building as constructed in May 2021. A separate listing for what appears to be the same property — same location, same price, same land and building areas, same assumed rent — listed it as built in May 1991, roughly 35 years old.

That is not a typo you can wave away. A 2021 build and a 1991 build are two entirely different assets wearing the same price tag. And the exterior photography did not look like a five-year-old building. This matters because the two versions imply two opposite conclusions. If the building is genuinely from 2021, then ¥53M for a steel-frame structure in the 23 wards at a 10% yield is a near-anomaly — the kind of number that should make you ask what else is wrong, because well-built new stock in Tokyo does not casually yield double digits. If it is from 1991, the 10% explains itself instantly: a 35-year-old steel-frame building, fully vacant, with a retail unit, should price at a high yield, because the market is pricing the age and the risk.

The same building can appear with different data across portals for mundane reasons — a data-entry error, a confused re-list, a mix-up between the original permit date and a later renovation — but the reason does not matter until the fact is settled. Three documents settle it, and none of them is the listing: the property registry (登記簿, tōki-bo), which records the building's legal particulars; the fixed-asset tax valuation certificate (固定資産税評価証明, kotei-shisan-zei hyōka shōmei), which the tax office issues against the assessed structure; and the confirmation certificate (確認済証, kakunin-zumi-shō), the building-permit document issued at construction. Until those agree with each other, we treat the build year as unknown — and an unknown build year makes every downstream number, from depreciation to financing to exit, provisional. Our working assumption, given the exterior and the yield, is that 1991 is the more likely truth. But an assumption is not a document.

Red Flag #2 — A "Full-Occupancy Yield" on a Building That Is 0% Occupied

The ¥5,328,000 annual rent is not income. It is a full-occupancy assumption (満室想定, manshitsu sōtei) — the rent the building would produce if every unit were let at the seller's projected figure. Today the building produces zero, because all three units are vacant. The 10.05% is arithmetic performed on a projection, not a track record.

This distinction is the one foreign buyers relying on broker-translated summaries most often miss, because the English gloss usually reads "yield: 10.05%" with no asterisk. A satisfied-occupancy yield attached to an empty building is a hypothesis about the future dressed as a statement about the present. The seller's rent number is the number that makes the deal look best; it is precisely the number that requires the most evidence before you trust it.

And "why is it empty?" is not a single question — it has at least five answers, and each one prices the deal differently. It could be a coincidental simultaneous move-out, in which case the vacancy is noise. It could be that the units have been marketed for a long time without takers, which says the assumed rent is above market. It could be that major repair work is pending, which is a capex liability the yield ignores. It could be that the previous owner simply never marketed the units, which is neutral-to-good. Or it could be a physical problem with the building or its systems, which is a red flag of a different order. A fully vacant building is not automatically a bad building — but you cannot underwrite one until you know which of those five stories is true, because they range from "harmless" to "deal-breaker." The document that begins to answer this is the rent roll paired with each unit's last rent, last tenancy date, and move-out reason.

Red Flag #3 — The Retail Unit Is Carrying the Yield

Break the ¥5,328,000 down and the arithmetic exposes where the risk concentrates. The assumed rent works out to roughly ¥444,000 per month. Two residential 3DK units in this area — based on comparable listings we could see for similar-age, similar-size units near the same station — realistically let for something like ¥110,000–130,000 per month each, so call it ¥220,000–260,000 combined. (Those are estimates from visible comparables, not confirmed lease figures.)

Do the subtraction and the ground-floor retail/office unit has to produce roughly ¥180,000–220,000 per month on its own for the ¥444,000 total to hold. In other words, close to half the entire yield rests on one commercial unit. That is the fragile part of the structure, and here is why: residential rent in a Tokyo ward is one of the most predictable numbers in real estate — there is deep, continuous tenant demand, and comparables are dense. Retail rent in a residential neighborhood is not. This location reads as residential rather than a station-front commercial strip, which means demand for a shop or office space is thinner, more idiosyncratic, and far harder to underwrite. A retail unit can sit empty for six months in a way a cheap 3DK rarely does.

So the honest read is that the 10.05% is not a residential yield with a bonus shop attached. It is a residential yield of perhaps 5–6% plus a commercial bet, and the commercial bet is the part we have the least ability to verify from a broker summary. If the retail unit under-lets or sits vacant, the headline number does not shade down — it collapses.

Red Flag #4 — No Building Confirmation Number on the Listing

The listing showed no building confirmation number — the field was simply "–". For an older building this can be innocent: portals frequently omit the number on aged stock. But the absence is a prompt to confirm that the underlying paperwork exists, because two documents here reach well beyond the question of legality.

The completion certificate (検査済証, kensa-zumi-shō) attests that the finished building passed its final inspection, and the building plan summary (建築計画概要書, kenchiku keikaku gaiyō-sho) records the approved plans on file with the ward. When these are missing — or when the physical building no longer matches the confirmation drawings — the problem is not primarily that you might be fined. The problem is financing and exit. Lenders lean on a clean paper trail; a building without a completion certificate, or with a footprint that diverges from its approved plans, is harder to finance, which shrinks the pool of buyers who can purchase it from you later. We have seen deals where a mismatch between the recorded and actual building specs turned an easy loan into a hard decline. On a mixed-use building where the retail unit's configuration may have changed over three decades, this is exactly the kind of thing we want confirmed before discussions go further, not discovered during the buyer's loan review.

Red Flag #5 — Flood Risk

Katsushika Ward sits in Tokyo's low-lying east, threaded between major rivers, and it is an area where flood risk belongs in the underwriting rather than the footnotes. The ward publishes flood hazard maps (ハザードマップ, hazādo mappu) for the surrounding river systems, and because this listing included precise coordinates, a document request lets us check the projected inundation depth for the exact point rather than the neighborhood average.

Expected inundation depth is not an abstraction — it feeds three concrete things. It affects insurance, where mapped high-risk locations can face higher premiums or narrower coverage. It affects lending, because some institutions weigh hazard exposure in their collateral assessment. And it affects resale, since the next buyer will run the same check we are running. The fact that the ground floor here is a retail/office unit rather than living space is marginally better than a ground-floor apartment — a flooded shop is a worse day than a flooded home is a disaster — but "marginally better" is not "irrelevant." Flood exposure is a permanent attribute of the site, and it should be priced in, not assumed away.

From Gross to Net

Suppose, generously, that the ¥444,000/month full-occupancy assumption turns out to be genuinely achievable. Even then, the 10.05% is a gross figure, and the gap between gross and net is where the real return lives. Here is the walk-down we would model, with the operating lines as reasoned estimates rather than confirmed figures.

Against a ¥53M price, that NOI implies a net yield of roughly 6.6–7.4%. For a building inside the 23 wards, that is genuinely not a bad number — which is the point worth sitting with. The 10.05% was never the return; the mid-6s to low-7s is the plausible return if everything holds.

And now the downside case, which is not a stretch given Red Flag #3. If the units actually let for closer to ¥400,000/month in total — entirely possible if the retail unit under-performs — gross rent falls to about ¥4,800,000, and the same cost stack drags net yield down into the low-6% range. Let the retail space then sit empty for half a year, and the number keeps sliding. The distance between the flyer's 10.05% and a realistic bad-but-not-catastrophic outcome in the low 6s is the entire risk premium of this deal, and it is invisible if you read only the headline.

What Price Would Make This Work

We would not buy at ¥53M before the documents are in. But the deal is not uninteresting — it is mispriced for us at asking, which is a different verdict from "avoid."

The logic runs through the risk stack. If the seller can confirm the 1991 build year, no legal issues, no imminent major repairs, that the two residential units can each let at roughly ¥110,000–130,000, that the retail unit has a real leasing history in the ¥150,000–180,000 range, and that the building has actually collected something near ¥5.3M in rent in the past — then this becomes a legitimately interesting cash-flow asset. Absent that confirmation, we are being asked to pay a full price for an unproven income stream on an aging, fully vacant, mixed-use building.

Our entry range would start around ¥45,000,000–48,000,000. At ¥47,000,000, the same ¥5,328,000 assumed rent implies a gross yield of 11.34% (¥5,328,000 ÷ ¥47,000,000). That extra margin is not greed; it is the compensation we would want for carrying the specific risks — 35-year-old structure, total vacancy, and meaningful retail exposure — that the ¥53M price asks us to absorb for free. The discount is the price of the risk.

The mirror image is worth stating plainly: if the building genuinely is from 2021, the entire analysis inverts. A 2021 steel-frame building in the 23 wards at ¥53M and 10% would be anomalous enough that our first instinct would be suspicion — what undisclosed problem produces that number? — rather than enthusiasm. Either way, the build year is the hinge, which is why it sits at the top of the document request.

The Document Request

Before any further discussion of price, this is the packet we would ask the broker to assemble. Nothing here is exotic; all of it exists for a legitimately marketed building.

• The rent roll (レントロール) — current and historical.
• A 3-year move-in / move-out history across all units.
• The last rent achieved per unit, with tenancy dates.
• Property tax statements (固定資産税) for the recent years.
• The repair history — what has been done, when, and what is pending.
• The confirmation and completion certificates (確認済証 / 検査済証).
• The property registry (登記簿) for building and land.
• Written confirmation of the build year, reconciled against the above.

If that packet comes back clean, we can model the financed cash flow — actual cash left after debt service — and produce a defensible maximum bid. If it does not come back, that is itself an answer.

Verdict

Reusing the scoring we apply across The Underwriting Files:

• Interest as an asset: ★★★★☆
• Buy at the ¥53M asking price today: ★★☆☆☆
• Worth requesting the full document set: ★★★★★

Unlike a passive central-ward condo, this is a property you buy to work — a cash-flow play where the upside is real if the risks resolve favorably. But the 10.05% is the least important number on the page. The deal turns entirely on three things: settling the build-year contradiction, explaining the total vacancy, and verifying the retail rent. Clear those, and it gets genuinely interesting. Leave them open, and the yield is just a headline.

Closing

This is what due diligence looks like before you ever book a flight or walk a property. Five red flags, an NOI walk-down, and a document request — all from a one-page broker summary and a second listing that happened to disagree with it. None of that work is visible in the "10.05%" that made us stop scrolling, and none of it is something an AI summary of the listing would surface on its own. The number tells you where to look. The reasoning tells you whether to buy.

If you are looking at a listing right now and want this same read run against the exact address — the build-year reconciliation, the gross-to-net rebuild, the flood-map check on the specific coordinates — that is what our Property Price Check is for. Get the read before you reply to the broker: start a Price Check.

Related Reading

• A 16% Yield Apartment Near Japan's Semiconductor Hub — A Due-Diligence Case Study — the same discipline applied to a regional listing where the land registry broke the deal
• Japan Property Due Diligence Checklist for Foreign Buyers — the full pre-purchase verification framework
• Japan Property Valuation: How to Assess Fair Value as a Foreigner — building a defensible number from comparables
• Cap Rate & Yield: Japan Real Estate Investment Metrics — why gross yield and net yield are different objects
• Japan Property Running Costs for Foreign Owners — the cost stack behind the gross-to-net walk

Property details have been anonymized and some figures (residential and retail comparable rents, the operating-cost lines in the NOI table) are estimates stated as such, not confirmed figures. Build-year, registry, and boundary questions in Japan should be verified with the relevant certificates and, where needed, a licensed judicial scrivener (司法書士). This article is general information, not investment or legal advice.