Buying Land to Build Rental Property in Japan: What Has Changed
What the Model Looked Like When It Worked
There was a period — not so long ago — when the arithmetic of residential rental development in Japan was straightforward enough that a smaller operator with a track record could take an 11% yield to the bank and borrow out to 110% of asset value. That sentence is worth sitting with. Positive yield at entry. Demonstrated performance over time. A lending relationship built on both. The result was a financing position that allowed capital to be recycled and portfolios to be built without requiring the operator to find equity for every new project.
That is the model we built our own portfolio on. Land acquired at a significant discount to surrounding values — in our case, stigmatised industrial land in Kawanishi at a fraction of the neighbouring residential price. Construction at costs that left room between the finished project and the rental income it would produce. Yield that the bank could underwrite. Leverage that made the next project possible without waiting for the previous one to fully pay down.
Yield and IRR are not the same measure, and the distinction matters here. Yield is the income produced relative to total cost. It is immediate. It is visible at entry. IRR — internal rate of return — is a time-weighted measure that accounts for financing, holding period, and eventual exit. At the scale we operate, yield is the primary test. If the project does not produce positive yield at entry, it does not proceed. IRR is relevant, but it is a secondary calculation. You cannot finance on projected IRR alone when you are working without institutional capital behind you.
The 11% yield figure was not exceptional at the time. It reflected a market where land prices in certain areas had not yet moved to reflect their underlying value, where construction costs were lower in real terms, and where rental income — while never spectacular in suburban Japan — was sufficient to produce a workable margin above total project cost. When all three of those conditions held simultaneously, the model produced assets that have continued to perform for decades.
The bank relationship was built on that performance. A track record of occupied buildings, consistent rental income, and debt being serviced on time. That relationship is what made 110% LTV possible. The lender was not financing the asset alone. They were financing the operator’s demonstrated ability to produce yield on a new asset in the same way they had on previous ones. That distinction matters enormously when you understand what has happened since.
The Land Price Reality in My Area of Focus
I want to be specific about what has happened in my area of focus, because the numbers tell the story better than any general observation can.
Seven years ago, land in my target acquisition zone was trading at around ¥150,000 per tsubo — approximately 3.3 square metres. Today, the same land trades at around ¥1,200,000 per tsubo. That is an eightfold increase in seven years.
To be clear about what that means in practice: land that would have cost ¥15 million for 100 tsubo now costs ¥120 million for the same plot. The construction cost on top of that has not moved in the opposite direction. It has increased in its own right — substantially, and across every component of a residential build.
This is not a national average. Japan’s national land price statistics show meaningful increases in urban cores and popular regional markets, but nothing approaching this rate across the board. My area of focus experienced this repricing for specific reasons — a municipal redevelopment programme, infrastructure investment, and the gradual dissolution of historical stigma that had kept prices suppressed for years. I understood those reasons when I was acquiring. They are the same framework I described in the Disparity Arbitrage article.
The point is not to celebrate the appreciation — though the portfolio has benefited from it directly. The point is to understand what it means for anyone trying to enter the same market today. The window that existed when land was priced at ¥150,000 per tsubo has closed. An investor arriving now faces a completely different entry point, on both sides of the equation. Land at eight times the price. Construction materials at significantly higher cost. Rental income that has moved, but not at anything approaching the same rate.
At ¥150,000 per tsubo, you could acquire land, build, and still find a yield that a bank would finance. At ¥1,200,000 per tsubo in the same location, that calculation does not close. The rental income that the completed building can produce — given what tenants in the area can and will pay — does not cover the combined cost of land and construction at a yield level that makes financing viable and the investment rational.
I am not speaking theoretically. I have run these numbers on my own area of focus. An investor at my scale, operating without institutional capital, cannot currently make a new land acquisition in the area I built my portfolio in produce a workable yield. That is not a projection. It is the present reality.
Building a Custom Home in Japan — design, coordination, and build execution.
Subscribe to receive new articles from Smith Realty Japan.
Construction Cost: Not One Event, a Compounding Series
The land price movement would be challenging enough on its own. It is not on its own.
Construction costs have moved structurally across the same period — and the pace of that movement has accelerated. This is not a temporary spike driven by a single supply shock. It is a sustained reset driven by materials, logistics, and labour costs that show no sign of reversing.
The April 2025 round of supplier-led increases affected every component of a residential build. Toilets increased by approximately twenty-three percent. Washbasins by approximately forty percent. Bathrooms and kitchens by approximately six percent. Sashes and entrance doors between three and twelve percent. Exterior materials by approximately eight percent. Faucets by approximately eight percent. A further round in April 2026 continued this movement — sashes up a further ten percent, with selected products higher, skylights up approximately twenty percent, sliding doors up approximately ten percent. Suppliers were explicit: raw materials, purchased components, parts, and logistics costs could no longer be absorbed internally. (1)
Interior finishing materials followed. Wallpaper — a cost that appears in every unit of every residential build — increased by approximately thirty percent across multiple suppliers. (2) This is not a premium product category. This is standard finishing material applied to every room of every apartment.
Structural cost moved as well. The Osaka Wide-Area Ready-Mix Cooperative announced a concrete price increase of ¥8,500 per cubic metre from April 2026, taking the standard price from ¥25,500 to ¥34,000 — an increase of approximately thirty-three percent in a single adjustment. (3) Concrete is not optional. It is the base structure of a reinforced concrete residential building. There is no substitution.
Sangetsu, Japan’s dominant interior materials supplier, announced increases of eighteen to thirty percent across wallcoverings, flooring, fabrics, and adhesives — with a warning that further increases may follow. (2)
Taken individually, each of these increases can be managed. Taken together, across a single residential construction project, they alter the structure of the cost base in a way that is not recoverable through value engineering or specification adjustment. You cannot remove a bathroom. You cannot avoid concrete. You cannot skip the wallpaper. These are embedded costs, and they have all moved in the same direction at the same time.
The result is that a project which would have produced an 11% yield under earlier cost conditions now produces something materially lower — in many cases, below the threshold at which financing is available and the investment makes sense. The 11% that built the relationship with the bank does not appear at the entry point any more. Not in my area of focus. Not at current land prices. Not at current build costs.
The Yield Problem at Ground Level
A recurring question at present is whether buying land and building residential rental property in Japan remains a workable model.
At the level where smaller operators are active — working directly across land acquisition, design, and construction without institutional backing — the outcome is clear. If rental income does not support the combined cost of land and construction at entry, the project does not proceed. There is no portfolio to absorb mistakes. There is no scale to hide misjudgment. The relationship between land cost, build cost, and rental income must hold at the point of commitment. If it does not, the project is declined.
Financing sits alongside this. Operators with a track record can obtain leverage and structure projects accordingly. New entrants generally cannot. At current land prices, that difference is often decisive. A project that works with leverage may not work without it. And the leverage that was available when yield was 11% is not available when yield has compressed to a level that no longer satisfies the lender’s underwriting.
This shift is not theoretical. It is visible in daily work. We are asked less often about residential rental builds. Over the past three years, we have accepted multiple build-to-sell projects. This is not a shift in preference. It is a shift in what works.
At REIT scale and institutional level, IRR-based underwriting over long holding periods with access to cheap debt changes the arithmetic. These are not the conditions most owner-operators work under. But they explain why activity at that scale continues in certain markets, while the investor profile we represent finds the entry point increasingly difficult to justify.
What Is Replacing the Old Model
Two models remain active at the owner-operator level. Both are visible in current work.
The first is build-to-sell. Developers acquiring land now are often doing so with intent to complete and sell rather than hold. Return is realised at completion rather than accumulated over time through rental income. This is a direct response to conditions where the yield required to justify a long-term hold does not appear at current entry pricing. What is no longer broadly visible is investor-led buy-and-build-for-rent at current entry pricing. The model has not disappeared. But it has narrowed significantly.
The second model is more specific to Japan and deserves to be understood on its own terms. An aging cohort of landowners — people who acquired land years or decades ago at historical cost — continues to build residential rental property. Their position is fundamentally different from an investor acquiring land today. Land cost is historical. Construction converts an existing asset into an income-producing one. The yield calculation looks different when land is already held.
But the motivation in many of these cases goes beyond yield. Residential rental construction in Japan is used as an inheritance planning tool. A landowner borrows to build. The debt reduces the taxable value of the estate. The building produces income. The asset passes to the next generation with a reduced tax liability and a cash-flowing property attached. This is a well-established structure in Japanese estate planning, and it is driving a meaningful share of residential construction activity in the current environment — not because the yield economics are exceptional, but because the inheritance logic is compelling.
For existing portfolios built under earlier conditions, the position is different again. Buildings delivered when land and construction costs were lower continue to perform. Rental income has reduced debt over time. Exposure has decreased. Income remains. These assets demonstrate that the model works — not as a historical curiosity, but as a live example of what the arithmetic produces when the entry point is correct. The issue is not the model. The issue is the entry point.
The Decision That Now Matters Most
The question is not whether rental property works in Japan. The question is whether buying land and building rental property works at current entry pricing, for an operator at a specific scale, with a specific financing position.
At current land prices in my area of focus — ¥1,200,000 per tsubo against ¥150,000 seven years ago — the answer for a new entrant without historical land holdings is: not easily. Not at my scale. Not without a financing position that most new entrants cannot access.
Projects that move forward today share common characteristics: access to finance, control of cost, the right site at the right entry price, clear tenant demand, and realistic assumptions. Without these, the model does not hold.
The ability to decline has become as important as the ability to build. Not every site works. Not every project should proceed. This is where experience becomes visible — not in design, not in construction, but in the decision at the front end.
A precise way to state it: at current land prices, investor-led buy-and-build-for-rent is no longer a broad entry model. It is limited. Selective. Conditional. The model still works. But the entry point has changed — and in my area of focus, it has changed by a factor of eight.
Our Insights reflect how we think about investing in Japanese real estate — the questions we ask, the trends we watch, and the reasoning behind the decisions we make for our own portfolio. We share them in the hope they’re useful food for thought, but they are not advice — just one active investor’s view of the market.
Sources
(1) LIXIL Corporation — Price Revision Announcement (2025–2026)
https://www.biz-lixil.com/pdf/lixil_kakakukaitei.pdf
https://newsroom.lixil.com/ja/2025110701
(2) Sangetsu Corporation — Interior Materials Price Increase (2026)
https://this-c.com/news/lixilpriceup202304/
(3) Osaka Wide-Area Ready-Mix Cooperative — Concrete Price Increase (2026)
https://www.beton.co.jp/archives/12598