Why "Buy Japan" Really Means "Buy Tokyo": The Capital Where People, Money, and Jobs Refuse to Leave

TL;DR: “Japan is dying” is true at the national level and false where it matters: people, companies, and money are concentrating into Tokyo, not draining out of it. In 2025, 40 of Japan’s 47 prefectures lost residents to migration while Tokyo posted the country’s single largest net inflow — and that gap between the scary headline and the capital’s reality is exactly the opening for a foreign buyer. It also explains why a 4% net yield in Minato can beat an 8% net yield in rural Gunma on a 10-year total-return basis: the worked example is below.

---

The headline that scares buyers is the wrong map

Open any article on Japanese demographics and you get the same story: an aging, shrinking nation, empty houses, ghost towns. All true — for “Japan” as a single number. The mistake is treating Japan as one market. It isn’t. It’s two: a capital that keeps absorbing people and capital, and roughly 40 prefectures slowly emptying into it.

The numbers force the split. In 2025, only 7 prefectures gained population from internal migration — Tokyo, Saitama, Chiba, Kanagawa, Osaka, Shiga, and Fukuoka. The other ~40 lost people on net (directional, as of writing). This is not a rising tide. It is a drain with one main plughole, and the plughole is Tokyo. When you buy “Japanese property” as an index, you average a winner against four dozen losers. When you buy Tokyo, you buy the winner directly.

Caveat worth stating up front: “Tokyo wins” does not mean every Tokyo address wins. Concentration happens at the metro level; ward and station still decide the outcome. More on that below.

From the desk — In a decade of walking foreign buyers through this city, the pattern I keep seeing is that the ones who fixate on the national “Japan is shrinking” headline talk themselves out of the deal, while the ones who close are the ones who stop arguing about the country and start asking me which ward and which train line. The headline scares the casual buyer off the table, and that is precisely the quiet I watch serious buyers exploit.

Tokyo is where Japan’s people actually go

Tokyo prefecture recorded the largest net in-migration of all 47 prefectures in 2025, roughly +65,000 people, even as the national inflow cooled from prior years (directional, as of writing). Read that twice: in a country losing population overall, the capital still pulled in the most newcomers of anywhere. Every one of those arrivals needs a roof — and Tokyo’s housing supply, hemmed in by land and regulation, does not expand to greet them on cue.

Widen the lens to Greater Tokyo (Tokyo plus Saitama, Chiba, and Kanagawa) and the pull is even starker: a net ~+123,000 internal migrants in 2025, versus only ~+8,700 for the entire Osaka–Kyoto–Hyogo–Nara cluster. That is Tokyo’s metro region drawing roughly 14 times the net inflow of Japan’s #2 metro (directional). The “second city” is not a close second. For a buyer, migration is the most honest demand signal there is — it is people voting with their suitcases, and they keep voting for the same place.

Money and jobs concentrate even harder than people

People follow jobs, and jobs follow money — so look at where the money sits.

• Tokyo prefecture generated roughly ¥120 trillion in FY2022, about 21% of national GDP, on under 1% of Japan’s land area (directional, as of writing). One-fifth of a G7 economy is produced on a sliver of ground. That density is what underwrites rents and land values — there is simply more economic activity per square meter than anywhere else in the country.
• Tokyo hosts about 2,960 large companies (capital of ¥1bn or more) — roughly half of all such firms in Japan — plus around 76% of foreign-affiliated companies, some 2,300-plus head offices.
• Greater Tokyo’s gross metropolitan product was about US$2 trillion in 2022, ranking among the largest city economies on earth, second only to New York.

String those together and you get the thesis in one line: a deep, durable, white-collar tenant base that regional cities structurally cannot replicate. A company headquarters is not a tourist — it signs a long lease, pays salaries every month, and those salaried workers rent and buy the apartments you would own. Osaka and Fukuoka are real economies, but they do not host half the nation’s big firms or three-quarters of its foreign corporates. That concentration is the moat.

The jobs engine is tightening, not loosening

A tenant base only matters if it is growing more crowded, not less. Right now it is tightening.

Tokyo’s all-grade office vacancy fell to roughly 1.5–2.1% in 2025, with Grade-A rents rising at the fastest pace since 2007 (CBRE/Colliers, as of writing). Translate office data into housing terms: vacancy that low means firms are competing for space and expanding headcount, not shedding it. More desks filled means more paychecks, means more demand for the apartments those workers live in. The engine that fills residential units is running hot.

Honest caveat: office and residential are different asset classes, and a hot office market does not mechanically lift every condo. But office tightness is a leading indicator of white-collar employment, and white-collar employment is the bedrock of Tokyo residential demand. When the people who lease the offices are hiring, the residential floor under your investment gets firmer.

What this means for a foreign buyer: the asymmetry

Here is the trade in plain terms. The “Japan is shrinking” narrative is loud, global, and largely accurate — so it depresses sentiment, scares off casual buyers, and helps keep pricing comparatively sane. Meanwhile the asset you would actually buy sits inside the one market that is demographically and economically gaining. You are buying a top-tier global city — second only to New York by metro output — at prices set by the anxieties of a country that is, on aggregate, getting smaller.

That is the asymmetry: national pessimism sets the price; metro-level strength sets the fundamentals. The gap between the two is your margin. Add the structural backdrop foreign buyers already know — a relatively weak yen, no nationality restriction on freehold ownership, and rental yields that often clear borrowing costs — and Tokyo is one of the rare places where the scary story and the good investment are the same story.

The risk is not that Tokyo reverses. It is that “Tokyo” is not granular enough. Buy the wrong ward, the wrong distance from a major line, or an aging building with a thin reserve fund, and you can underperform inside a winning city. Concentration is the macro tailwind; selection is still your job.

The yield trap: why a 4% net yield in Minato can beat 8% in rural Gunma

The macro picture above is exactly why the most common counter-argument fails. Rural Japan offers gross yields of 10–15%. Investors see those numbers and assume they’ve found value. What they’ve usually found is a shrinking population, a building that can’t be sold, and an 8% net yield that turns into a 1% IRR after you account for capital decline, sustained vacancy, and zero exit liquidity. Minato-ku’s 4% net yield often generates a superior 10-year total return. The math is not complicated. The psychology is.

I get emails every month from investors who’ve concluded “rural Japan is undervalued.” They’ve found buildings in Gunma or Akita with 12% gross yields. They calculate the net. They get excited. They’re comparing the wrong number.

Here’s both cases built properly.

Figures below are illustrative — representative deal types at current market conditions, not specific audited transactions.

Property A: Minato-ku Studio, Mita Area

• Purchase price: ¥22,000,000
• Current rent: ¥88,000/month
• Gross yield: 4.80%
• Annual NOI (after all costs): ¥792,000
• Net yield: 3.60% (NOI ÷ total capital including acquisition costs of ¥1.5M ≈ ¥23.5M total)

Property B: Rural Gunma 2LDK, Regional City

• Purchase price: ¥5,800,000
• Current rent: ¥40,000/month
• Gross yield: 8.28%
• Annual NOI: ¥380,000 (after costs)
• Net yield: 6.21% (NOI ÷ ~¥6.1M total cost)

On income yield alone, Gunma wins. 6.21% vs 3.60% — 72% more income yield for 26% of the price. The spreadsheet looks obvious.

Model the full 10 years and it isn’t.

From the desk — The inquiries that cross my desk most often are foreign buyers convinced rural Japan is undervalued because the gross yield reads double Tokyo’s, and the conversation I dread is the one years later when they want to sell. I’ve watched owners list regional units for two years with no offers at any price, while a comparable central studio I’d have moved in months. The yield they fell in love with never had an exit attached to it, and that’s the line item no spreadsheet shows them until it’s too late.

Factor 1: Vacancy — Gunma’s hidden tax

Minato-ku vacancy for a well-located studio: 5–7% across a typical 10-year hold. There is always a pool of young professionals, expats, and medical staff who want to be central. Turnover happens; the unit fills.

Gunma regional city: the population of most non-capital regional cities in Gunma prefecture has declined 8–12% over the past decade. Vacancy in the rental market runs 15–20% for older stock.

Adjusted effective gross income for Gunma at 18% vacancy: ¥40,000 × 12 × (1 − 0.18) = ¥393,600/year gross effective

Previous NOI was calculated on 8% vacancy. Rerun at 18%: NOI drops by approximately ¥46,000/year → revised NOI: ¥334,000 Revised net yield: 5.48%

Still higher than Minato. But the gap has shrunk. And vacancy is not a fixed number — it tends to worsen in declining markets.

Factor 2: Capital value — the 10-year trajectory

This is where the analysis diverges completely.

Minato-ku: Inner Tokyo property has historically tracked or slightly beaten CPI over long periods, with periodic strong appreciation. Conservative modeling assumes flat-to-modest real appreciation. Call it 1.5% nominal annually for illustration.

¥22,000,000 × (1.015)^10 = ¥25,567,000

Rural Gunma: Population decline drives land value decline. Japan’s National Institute of Population and Social Security Research projects ongoing population reduction in most Gunma municipalities outside Maebashi. Building value depreciates on a fixed schedule. Land value tracks population.

Conservative scenario: 1.5% nominal annual decline. ¥5,800,000 × (0.985)^10 = ¥4,980,000

Optimistic scenario for Gunma: flat. ¥5,800,000.

Even being generous to Gunma, you’re looking at zero capital upside and potentially a ¥820,000 loss in asset value over a decade.

The 10-year IRR comparison

Let’s run both scenarios to IRR. All-cash, no leverage, pre-tax for simplicity.

Minato-ku (Property A):

Year	Cash Flow
0	−¥23,500,000
1–9	+¥792,000/year (slight increase modeled)
10	¥792,000 + ¥25,567,000 − ¥900,000 selling costs = ¥25,459,000

Approximate IRR: 5.7%

Rural Gunma (Property B):

Year	Cash Flow
0	−¥6,100,000
1–9	¥334,000–¥310,000/year (worsening vacancy trend)
10	¥310,000 + ¥4,980,000 − ¥250,000 selling costs = ¥5,040,000

Approximate IRR: 3.9%

Minato 5.7%, Gunma 3.9%. The lower-yield property delivered the higher total return.

Now run the pessimistic Gunma scenario: vacancy climbs to 25% in years 7–10, exit price drops to ¥4,200,000, the building needs a ¥400,000 repair in year 6.

IRR falls below 2%. You’ve tied up capital for a decade and earned less than a Japanese government bond — with far more work and risk.

Comparing total return without tax is incomplete. Capital gains taxes in Japan cut into exit proceeds. If Minato appreciation generates a ¥3.5M gain and Gunma generates ¥0, the tax on Minato’s gain costs roughly ¥536,000 (15.315% for 5+ year holds). After-tax IRR gap narrows, though Minato still wins.

Factor 3: Liquidity — can you actually exit?

Minato-ku: deep buyer pool. Domestic investors, family offices, foreign capital looking for Tokyo exposure, J-REIT adjacent buyers. A quality Mita studio will transact. You can exit within 3–6 months at a price close to your ask.

Rural Gunma 2LDK: you may not be able to exit at all. Japan’s akiya (vacant house) problem is concentrated in exactly these markets. There are towns in Gunma where properties are listed at ¥1,000,000 and don’t sell. Your ¥5.8M purchase could be worth whatever anyone will pay — and if nobody is buying, that’s zero liquidity at any price.

I know investors sitting on properties in Niigata and Tochigi listed for two years. The yield looked great at purchase. Exit doesn’t exist.

Liquidity risk never appears in yield calculations. It needs to appear in your decision.

Factor 4: Management costs at distance

Running a property in Mita from overseas: manageable. Tokyo property management companies are professional, English-compatible firms exist, and the ecosystem for foreign landlords is developed.

Running a property in a regional Gunma city from overseas: harder. Fewer management companies. Less competition keeps fees high. When the roof needs inspection or a tenant dispute requires a local visit, you’re either flying in or paying a premium for someone else to handle it. Management costs for regional properties can run 10–15% of rent versus 5–7% in Tokyo. Add that to your cost structure and the yield gap narrows further.

When rural Japan does make sense

Not arguing rural Japan is always wrong. It makes sense when:

You’re already there. If you live in Gunma or have a business reason to be there regularly, the management friction disappears.

You’re buying at extreme discount for specific cash flow. A ¥1.5M property renting for ¥30,000/month has a different risk profile than ¥5.8M. At low enough prices, even a poor exit is survivable.

You’re targeting specific micro-markets. Karuizawa (resort), Nikko (tourism), Kusatsu (hot spring town) — some regional markets have demand drivers that override the demographic trend. Proximity to specific industries or military bases creates pockets of stable demand.

You have a specific operational play. Minpaku in a high-tourism rural area can outperform standard rental yields. But that’s a hospitality business, not passive property investment.

Your next step: stop buying “Japan,” start buying a ward

If you take one thing from this: never make a Japan decision off a Japan number. The national figure averages a structural winner against forty structural losers. Your job is to isolate the winner and then go one level deeper.

Three concrete moves:

1. Reframe the question. Not “should I buy in Japan?” but “which Tokyo ward, and on which line?” That single shift filters out most of the regional decline the headlines are warning you about.
2. Compare wards on the fundamentals that compound — population trend, employment density, transport access, and yield — rather than on photos. Our /wards breakdowns and the side-by-side /compare tool exist for exactly this; the /tools yield and cost calculators let you pressure-test a specific building before you ever contact an agent.
3. Learn the local terms before you negotiate. Costs like reikin (a non-refundable “key money” gift to the landlord) can quietly reshape your returns; the plain-English /glossary covers the ones that matter.

The window here is not “Tokyo is cheap.” It is that the world is pricing Tokyo like it is part of a dying country, while the city itself keeps winning the war for Japan’s people, companies, and money. That gap does not stay open forever. Pick a ward, run the numbers, and act while the headline is still doing your discounting for you.

FAQ

Q: What yield threshold makes rural Japan worth considering? As a rough filter: if net yield is below 7% in a shrinking population area, the risk-reward doesn’t work on a total return basis for most foreign investors. If net yield is above 8% and the property is in a market with a specific demand driver (tourism, industry, university), it warrants deeper analysis.

Q: Is Minato-ku overpriced right now? Probably stretched in parts — Azabu and Hiroo have run hard. “Overpriced” is relative to alternatives. Compared to prime urban property in other major Asian cities, Tokyo central is still competitive on yield and price-per-sqm. Whether it appreciates further from here is a macro call, not a yield calculation.

Q: Can you get financing for rural Japanese property as a foreigner? Rarely. Most Japanese banks won’t finance regional property for non-residents. This means rural investments are typically all-cash, which concentrates your capital risk and eliminates the leverage effect on equity IRR.

Q: What’s the single best argument for rural yield over Tokyo yield? Cash flow today versus capital gain later. If you need income now and have no ability to wait for appreciation, a 7% yielding rural property that cash flows from day one might suit your situation better than a 3.5% net yield Tokyo property where the return is back-loaded into the exit. Know what you need the money to do.

Q: Should I diversify between Tokyo and regional? Not unless you have a strong operational reason to be in both markets. Managing two different property ecosystems doubles your administrative complexity. If you want yield diversification within Japan, J-REITs give you regional exposure without the illiquidity and management burden.