Japan Real Estate Bubble: What It Means for Buyers Now

What actually happened between 1986 and 1991

The bubble is usually summed up in one sentence: prices exploded, then collapsed. The real sequence is far more useful, because it explains the market structure you are buying into today.

Very loose money, then a hard stop

After the 1985 Plaza Accord and the yen appreciation that followed, the Bank of Japan cut its official discount rate from 5.00% (January 1986) to 2.50% (February 1987) and held it there for more than two years. Credit became abundant and real estate absorbed a large share of it.

The reversal was then brutal: five hikes in fifteen months took the discount rate from 2.50% to 6.00% on 30 August 1990.

Date	Official discount rate	Context
February 1987	2.50%	Record low, held for 27 months
May 1989	3.25%	First tightening move
December 1989	4.25%	Nikkei peaks on 29 December
March 1990	5.25%	Property lending curbs imposed on 27 March
August 1990	6.00%	Top of the tightening cycle

The measure that actually broke the land market

On 27 March 1990 the Ministry of Finance issued administrative guidance known as sōryō kisei (総量規制, quantitative restriction on property lending): growth in loans to the real estate sector could no longer exceed growth in a bank's total loan book. It stayed in force until December 1991.

That closed tap, more than the level of interest rates, is what stopped land speculation dead. Keep the mechanism in mind: Japanese land prices have always been highly sensitive to domestic credit conditions. That is still true, and it is one reason a foreign buyer paying cash, with no access to a Japanese mortgage reserved for salaried residents, is not exposed to the same cycle as a leveraged Japanese household.

The peak did not happen everywhere at once

The reference study by the Japan Real Estate Institute, presented to the transport ministry's policy research institute, tracked the single most expensive commercial site in 174 cities. The lag it reveals is striking:

• Kansai (Osaka, Kyoto, Kobe): peak in September 1990, followed by an immediate and violent drop.
• Tokyo's 23 wards: a plateau from September 1990 to September 1991.
• Nagoya, Sendai, Hiroshima, Fukuoka: peak in March or September 1991.
• Regional towns: some did not start falling until 1993, 1994 or even 1996.

Not until September 1996 were all tracked cities falling at the same time. The bubble did not deflate in one event: it took six years to cross the country.

How big the run-up really was, city by city

National averages hide everything. On prime commercial land, with March 1983 as an index of 100, the peak multiples are both spectacular and wildly uneven.

City	Index at peak (March 1983 = 100)	Reading
Sapporo	957	More than 9.5 times higher
Takatsuki (Osaka)	944	A commuter suburb, not a business district
Yokohama	934	More than 9 times higher
Nagoya	861	More than 8.5 times higher
Sendai	789	Regional capital of Tōhoku
Chiba	733	Tokyo commuter belt
Kyoto	701	7 times higher

For comparison, the 1989 government land white paper put cumulative commercial land growth at 312.9 for the Tokyo area, 245.8 for Osaka and 163.8 for Nagoya (1983 = 100). On the most sought-after individual sites, the surge was therefore far sharper than regional averages suggest.

Half the country barely moved

The same work shows that 50 of the 174 cities stayed below an index of 150 at the height of the bubble, meaning less than 50% growth over eight years. Fourteen cities stayed below 110, essentially flat: Tsuruoka, Uwajima, Niihama, Kitami, Sakata, Saijō and Kanoya among them. No city in the three main metropolitan areas, and no prefectural capital, appears on that list.

In other words, the bubble was a big-city and regional-capital phenomenon. Much of rural Japan, where most of the akiya listings in our catalogue are located, never took part in it.

The fact almost nobody quotes: towns with no bubble crashed too

This is the single most important point in this article, and the one that most changes how you should read today's market.

If the 1990s decline had simply been a bubble deflating, towns that never rose should not have fallen. The opposite happened. By March 1999, on prime commercial land, still with March 1983 = 100:

City	Index at bubble peak	Index in March 1999
Kanoya (Kagoshima)	below 110	30
Niihama (Ehime)	below 110	48
Satsumasendai (Kagoshima)	below 110	51
Sakata (Yamagata)	below 110	58
Saijō (Ehime)	below 110	76

These towns did not lose a bubble: they lost two thirds of their pre-bubble land value. More broadly, of the 131 cities tracked continuously from March 1983 to March 1999, 66 ended up below their 1983 price, roughly half the sample. And as late as 2016, some regional cities were still at just 10% to 20% of their March 1983 level on their best commercial site.

What the researchers concluded

The analysis identifies three cumulative causes for that long decline, only one of which is the bubble:

1. The bubble bursting, dominant in large cities and in the early years.
2. A structural change in land pricing, a lasting correction of price expectations that had never been recognised as a bubble in the first place.
3. A structural change in towns themselves: mass car ownership, retail moving to the outskirts, traditional town centres hollowing out.

Translated for a 2026 buyer: the low price of a rural Japanese house is not a 1991 discount waiting to correct. It is the price of a place losing residents and shops. Our article on Japanese demographics and the property market maps that mechanism municipality by municipality.

Why Japanese houses are still cheap in 2026

Three forces stack up, and only one is a direct legacy of the bubble.

1. Buildings depreciate, land holds the value

A Japanese timber house is depreciated over 22 years for tax purposes and its market value tends towards zero between 20 and 30 years, regardless of its actual condition. A well-kept 1985 house can therefore sell for little more than the land beneath it. This is not a bubble effect but a permanent feature of the market, covered in our article on building depreciation and land value in Japan.

2. Nine million vacant homes

The 2023 Housing and Land Survey counts roughly 9 million vacant dwellings in a stock of 65.02 million, a record vacancy rate of 13.8%. Of those, 3.85 million are neither for sale, nor for rent, nor second homes: these are genuinely abandoned houses, up 370,000 since 2018. Supply structurally exceeds demand across much of the country.

3. New build keeps coming

Japan keeps building new housing, including in municipalities with shrinking populations, which accelerates the obsolescence of older stock. The new versus second-hand question is covered in our shinchiku versus chūko comparison.

Together these explain why complete houses trade below 5,000,000 JPY (30,864 EUR), and why we can run a whole selection of Japanese houses under 50,000 euros. They are not ruins by definition: they are properties sitting where local demand disappeared.

Where the market stands in 2026: a two-speed country

Since 2021 the national trend has reversed. The kōji chika (公示地価, official land prices published every year by the land ministry) released on 17 March 2026 records a fifth consecutive annual rise, and the strongest increase since 1992.

Segment	2026 change	2025 change
National average, all uses	+2.8%	+2.7%
Commercial land, national	+4.3%	+3.9%
Residential land, national	+2.1%	+2.1%
Tokyo, commercial	+12.2%	+10.4%
Tokyo, residential	+6.5%	+5.7%
Tokyo 23 wards, commercial	+13.8%	+11.8%

A factor of roughly six separates the national average from central Tokyo. That is the definition of a two-speed market, and it is exactly the lesson of the 1990s: in Japan the national average describes almost no actual property.

Where the growth is concentrated

Rising areas share three traits: employment density, dense rail service, and tourist footfall. That covers metropolitan cores, international ski resorts, and well-connected hot spring towns. Our review of Japanese property prices in 2026 maps those pockets, while our Tokyo property market page and Kyoto property market page show what that means in live listings.

Elsewhere, prices keep falling. It is the same geography the bubble study described in 1999, with thirty more years of evidence behind it.

Is there a new property bubble in Japan in 2026

The question comes up constantly, especially in the face of that +13.8% in central Tokyo. Compare the two periods on the criteria that actually created the bubble.

Criterion	1989 to 1990	2026
Bank of Japan policy rate	Discount rate raised to 6.00%	Policy rate at 1.00% since June 2026
Property lending policy	Quantitative curbs imposed in March 1990	No quantitative curbs in force
Annual land price growth, national	Double digits for several years	+2.8% across all uses
Breadth of the rise	Nationwide, down to mid-sized towns	Concentrated in a few metros and tourist areas
Demographics	Population still growing	Population falling since 2010

The two situations have little in common. Today's rise is narrow, moderate at national level, and happening in a country that is losing people, which mechanically limits how far speculation can generalise the way it did in 1988.

The genuine 2026 risk is not a national bubble but local overheating in segments driven by tourism and foreign buyers, in a weak-yen context we cover in our analysis of the weak yen and Japanese real estate. Paying today for record tourist flows in a small town with no local economy is repeating, on a small scale, the mistake made by Takatsuki buyers in 1990.

What the bubble means for your purchase in practice

Five practical rules follow directly from the data above. They form the core of how we screen properties.

1. Buy a location, never a discount

Towns that fell to 30% of their 1983 level were not cheap by accident: they were losing their economic function. The operating test has not changed in thirty years: a served railway station within 20 to 30 minutes, open shops, a hospital, a school. A low price without those four things is not a bargain.

2. Think in land value

Since the building trends towards zero, your exit value will essentially be land. Check plot size, buildability and access before you check floor area.

3. Expect no automatic capital gain

Japanese history offers no mean reversion: in half the tracked cities, the 1999 price was below the 1983 price. Rental yield is the engine, resale is a bonus. Our rental yield simulator computes that yield with running costs included.

4. Keep transaction costs in view

Total purchase costs in Japan remain 6% of the price or less. That is low, and it holds on small tickets too, but fixed fees weigh proportionally more on a 3,000,000 JPY property than on a 30,000,000 JPY one. The breakdown is in our article on property purchase costs in Japan.

5. Check the build year before anything else

The seismic code revised in 1981, the shin taishin (shin-taishin, new earthquake resistance standard), splits the housing stock in two. Many very cheap properties predate it. That is not disqualifying, it is a retrofit budget to price in, as explained in our article on the 1981 seismic standard.

A worked example

Two houses of the same vintage, both listed at 6,000,000 JPY (37,037 EUR).

Item	Property A: served town, station 15 min away	Property B: no station, car essential
Asking price	6,000,000 JPY (37,037 EUR)	6,000,000 JPY (37,037 EUR)
Purchase costs (6% or less)	360,000 JPY (2,222 EUR)	360,000 JPY (2,222 EUR)
Estimated refurbishment	4,000,000 JPY (24,691 EUR)	4,000,000 JPY (24,691 EUR)
Total invested	10,360,000 JPY (63,950 EUR)	10,360,000 JPY (63,950 EUR)
Rental demand	Residents and short stay	Short stay only, seasonal
Exit value at 10 years	Land-driven, active market	Land with no identified buyer

For identical money, the building is not what separates the two: the depth of the local resale market is. That is precisely the variable 1990 buyers underestimated. Our guide to buying property in Japan walks through those checks step by step, and our machiya for sale selection shows how the same screen applies to traditional townhouses.

Common mistakes to avoid

• Assuming the Japanese market is falling. Nationally it has not been since 2021. It is simply very uneven.
• Assuming the opposite and buying anywhere. The +13.8% in central Tokyo says nothing about a house two hours from the nearest station.
• Waiting for a catch-up. No Japanese dataset supports a mechanical return to 1991 levels, or even to 1983 levels in half of the regional cities.
• Confusing a low price with a bargain. See rule one: the discount usually reflects a very accurate local economic reality.
• Believing that buying grants residency. Owning property in Japan grants no visa of any kind, whatever the amount invested.
• Counting on a Japanese mortgage. Japanese home loans require you to be resident and salaried in Japan. Otherwise the purchase is cash.
• Trusting a national average. It has never described a real property in Japan, not in 1990 and not in 2026.

If you need a specific property assessed, our buying support in Japan starts with a first written exchange, with no commitment, and our costed case studies document two real deals from search to keys.

In short: the bubble explains the past, not today's price

Japan's real estate bubble was a metropolitan phenomenon, peaking between September 1990 and September 1991, followed by a decline that took six years to reach every city in the country. Its collapse explains what happened to large cities in the 1990s. It does not explain rural prices today: those come from a shrinking population, hollowed-out town centres, a building stock that depreciates by accounting convention, and 9 million vacant homes.

For a buyer in 2026 the consequence is simple and reasonably encouraging. The market is neither wreckage to scavenge nor a bubble to flee: it is a segmented market where location decides everything, where entry costs stay at 6% or less, and where the mistakes of 1990 remain perfectly repeatable on a small scale if you buy a price instead of a place.

The right instinct is the same as it was thirty-five years ago: look at the station, the shops and the population before you look at the price tag.

Frequently asked questions

When did the Japanese real estate bubble burst?

There is no single date. Land prices peaked in September 1990 in Kansai, between September 1990 and September 1991 in Tokyo, and in March or September 1991 in Nagoya, Sendai and Fukuoka. Some regional towns did not begin falling until 1993, 1994 or 1996.

How much did Japanese property prices fall after the bubble?

It depends entirely on location. On prime commercial land, with March 1983 set at 100, half of the 131 tracked cities were below 100 by March 1999. Some towns in Kyūshū and Shikoku fell to 30 or 50, and a few were still at 10% to 20% of their 1983 level in 2016.

Why are houses so cheap in Japan?

Three reasons stack up, only one of which relates to the bubble: timber buildings depreciate to almost nothing over 20 to 30 years, rural populations are shrinking, and the country counted about 9 million vacant homes in 2023, a vacancy rate of 13.8%.

Are Japanese property prices recovering?

Yes, but very unevenly. Official land prices published in March 2026 show a fifth consecutive annual rise, at +2.8% nationally, driven by Tokyo where commercial land gained 12.2%. Many rural municipalities are still falling.

Is there a new property bubble in Japan?

Nothing suggests one at national level. The policy rate has been at 1.00% since June 2026, no quantitative curbs on property lending are in force, national growth is +2.8%, and the population is shrinking. The real risk is local overheating in tourist-driven segments, not a general bubble.

Should I wait for prices to drop before buying in Japan?

Waiting assumes a future correction that Japanese data does not show anywhere. In rising areas, waiting is expensive; in falling areas, waiting does not fix the underlying issue, which is the absence of local demand.

Does buying property in Japan give you a visa?

No. Property ownership grants no residence status in Japan, whatever the amount invested. A non-resident foreigner can buy freely but must pay cash: Japanese mortgages are reserved for salaried residents in Japan.