The Japan Asset Price Bubble (1986-1991)

Origins: the Plaza Accord and the Yen Problem

The bubble’s origins lie at the Plaza Hotel in New York on 22 September 1985, where the finance ministers of France, West Germany, Japan, the United Kingdom and the United States agreed to coordinate intervention to weaken the U.S. dollar against the yen and the Deutsche Mark. The accord was, by the lights of its drafters, a success rather quickly: the yen rose from roughly 240 per dollar to under 160 within a year, and the dollar’s broad trade-weighted index fell on schedule. Whether it was a success on any longer horizon is a more delicate question, and one to which the answer remains contested forty years on.

The yen’s appreciation was a problem for an export-led economy with thin domestic demand and a manufacturing base whose competitiveness had been calibrated to a weaker currency. Profit margins at the great Japanese exporters—Toyota, Sony, Matsushita, the steel and shipbuilding combines—came under pressure of a sort the country had not seen since the early 1970s. The Bank of Japan, charged with cushioning the domestic economy from what was effectively an externally imposed deflationary shock, did what central banks tend to do in such moments. It eased.

Monetary Easing and the Credit Boom

Between January 1986 and February 1987 the [[Bank of Japan]] cut its official discount rate in five steps, from 5 percent to 2.5 percent. The 2.5 percent figure was the lowest postwar level on record and was held there until May 1989. Combined with a parallel programme of financial deregulation that expanded the lending capacity of Japanese commercial banks, the cuts produced an expansion of [[money supply]] and credit aggregates of a kind one finds, in the historical record, principally on the eve of difficulty.

The mechanics of where the credit went are worth dwelling on, because the standard narrative tends to compress them. Japanese commercial banks had, throughout the postwar period, lent principally to large industrial corporations through the keiretsu system. By the mid-1980s, those corporations were increasingly bypassing the banks and raising capital directly in the bond and equity markets, both domestically and offshore in London and Zurich. The banks, faced with the loss of their traditional clients, sought new ones—real-estate developers, small and medium-sized enterprises, individual mortgage borrowers, and ultimately the speculative property vehicles that would come to define the era. Loan-to-value ratios on property-backed lending rose. Cross-shareholdings among corporations encouraged the use of equity holdings as [[collateral]] for further borrowing, which produced a circular structure: banks lent against shares whose value was supported, in part, by the lending the banks themselves were doing. One has seen this structure since, at lower stakes.

The Equity Surge

The Nikkei 225 stood at approximately 13,000 in early 1986. It closed at 38,915.87 on 29 December 1989, the final trading session of the year, having nearly tripled over four years. At that peak, the aggregate market capitalisation of Japanese equities was approximately 4 trillion U.S. dollars, more than 40 percent of global equity-market value, and Japanese companies—Nippon Telegraph and Telephone, the city banks, the trading houses, the leading industrial conglomerates—dominated rankings of the world’s largest firms by market value.

[[Price-to-earnings ratio]] readings on the broad market exceeded 60 times earnings on certain measures. To put that figure in context: the U.S. market at the height of the dot-com peak in March 2000 traded at roughly 30 times earnings on the same basis, and the secular peak of the S&P 500 multiple in 1929 was around 33. A Japanese P/E of 60-plus was, on any reading the profession found acceptable, statistically anomalous. A small number of analysts in Tokyo argued at the time that traditional multiples did not apply because Japanese accounting understated earnings, because cross-shareholdings produced double-counting, because the Japanese growth model was qualitatively different. Some of these arguments were correct in detail and wrong in aggregate, which is roughly the most dangerous combination of properties an argument can possess.

Land, the Imperial Palace, and Zaitech

Property prices rose alongside equities and, in certain segments, outpaced them. In Tokyo’s six central wards, residential land prices climbed approximately 150 percent between 1985 and 1990, with comparable surges in Osaka and the other major metropolitan areas. The valuations on golf-club memberships and resort properties moved into territory that would, in calmer years, be reserved for fine art and Burgundy grand cru. (One thinks of the great vintages of the late 1980s in the Côte de Nuits, which were genuinely exceptional and have aged well; the Japanese property of the same period was neither.)

The Imperial Palace comparison—that the land beneath it was worth more than the entire state of California, that aggregate Japanese land values exceeded the combined land values of the United States despite Japan having a fraction of the area—became standard reference points in the subsequent literature. They sound like the embellishments one expects in retrospective accounts of manias. They were not embellishments. The Bank for International Settlements ran the numbers more than once.

The phenomenon of zaitech—financial engineering, the literal translation, though the spirit was rather closer to "treasury speculation"—saw nonfinancial Japanese corporations devote substantial resources to securities and property trading in pursuit of investment income. Manufacturing firms reported larger contributions from financial activities than from their underlying businesses in certain quarters. This is, on the historical record, a reasonably reliable late-cycle marker.

The Policy Reversal

By 1989, the social consequences of the asset-price inflation—in particular, the impossibility for ordinary Tokyo households of acquiring residential property within commuting distance of central employment—had become a political problem of a sort the LDP government could no longer ignore. Yasushi Mieno became Governor of the Bank of Japan in December 1989 and accelerated a tightening cycle that had begun under his predecessor in May. The official discount rate was raised in successive steps to 6 percent by August 1990.

In April 1990, the Ministry of Finance introduced quantitative restrictions on real-estate lending: banks were instructed to limit the growth of property loans to no more than the growth rate of their total loan books. The combination of higher [[interest rates]] and direct credit controls removed the principal supports of the bubble. The peak in equities, as it happened, had already passed; the peak in land prices followed roughly eighteen months later. The two markets did not collapse simultaneously, which is one of the things that made the diagnosis difficult in real time.

The Collapse

The Nikkei 225 began to decline in the first trading session of 1990 and fell sharply throughout the year. By August 1992 the index had lost approximately 63 percent of its peak value. The descent was not uninterrupted—there were several rallies of 20 percent or more, each of which was followed by a lower low, which is the texture of every serious bear market and one of the reasons such markets are so effective at destroying capital. The Nikkei reached an interim low of 7,054.98 in March 2009, more than 81 percent below its 1989 peak.

Real-estate values fell more slowly than equities and ultimately more deeply in many segments. Commercial property in central Tokyo declined for more than a decade, with some districts losing more than 80 percent of peak value before stabilising. Residential land followed a similar trajectory on a longer timeline. Golf-club memberships, which had traded at prices comparable to luxury houses in Surrey or Connecticut, lost the majority of their market value and most of their secondary-market liquidity along with it.

The Banking Crisis

The collapse in collateral values left Japanese banks holding hundreds of trillions of yen in non-performing loans. The most exposed borrowers were real-estate developers and the housing-loan companies known as jusen, which had been established in the 1970s to extend mortgage credit and which had drifted, by the late 1980s, into speculative property finance. Many of these borrowers could not service debts that had been extended on the assumption that asset prices would continue to rise.

The Japanese authorities initially responded with forbearance—permitting banks to recognise losses gradually rather than forcing immediate write-downs. The intention was to preserve financial stability, and at the margin it probably did. The cost was an extended period of impaired bank balance sheets, a credit channel that did not transmit monetary easing efficiently, and the emergence of [[zombie companies]]: insolvent borrowers kept alive by continued bank credit, occupying capital and labour that would, in a less forgiving system, have been redeployed.

The full scale of the problem was not addressed until the late 1990s and early 2000s, when a sequence of bank failures, mergers, capital injections, and the establishment of the Resolution and Collection Corporation accelerated the disposal of bad loans. The Long-Term Credit Bank of Japan and the Nippon Credit Bank were nationalised in 1998. The non-performing loan problem was not considered substantially resolved until the early 2000s, more than a decade after the bubble’s peak. There is a school of thought, with which one has some sympathy, that the forbearance approach traded a sharp recession for a long stagnation and got rather the worse of the bargain. There is another school which argues that the political economy of a sharp recession in Japan in 1992 would have been ungovernable. Both schools have a point and the question is not, in the strict sense, decidable.

The Lost Decade and Its Sequels

The period from 1991 to 2001 became known as the Lost Decade. Average real GDP growth ran at roughly 1 percent. [[Deflation]] became a persistent feature of the price level, wages stagnated, and unemployment rose to levels that were, by Japanese historical standards, alarming, even if by Western European standards they would have been read as benign. The term Lost Decades, in the plural, came into use as the stagnation extended through the 2000s and the 2010s.

Successive Japanese governments deployed large fiscal stimulus packages financed by sovereign borrowing, raising the ratio of government debt to GDP to among the highest in the developed world. The Bank of Japan cut its policy rate to zero in 1999 and pioneered [[quantitative easing]] in 2001, becoming the first major central bank to adopt the policy formally. Neither set of measures restored the inflation rates or the asset valuations of the pre-bubble era within the following two decades.

Demographic factors compounded the economic effects. Japan’s working-age population peaked in the mid-1990s and began a sustained decline, which reduced potential growth rates and complicated every effort to escape deflation. Corporate behaviour shifted toward debt reduction and cash accumulation—a pattern the economist Richard Koo characterised as a balance-sheet recession, in which firms that have been burned by collapsing collateral values prioritise the repair of their balance sheets over investment, even when borrowing costs approach zero. Koo’s framework is contested in academic monetary economics; it is rather more respected by practitioners who lived through the period, which is suggestive without being decisive.

The Round Trip

The Nikkei 225 did not return to its December 1989 closing peak until February 2024, a span of more than 34 years. The eventual recovery occurred against a backdrop of sustained corporate-governance reforms (the Abe-era stewardship and governance codes mattered more than is sometimes acknowledged), large-scale equity purchases by the Bank of Japan and the Government Pension Investment Fund, and a global rotation of capital that benefited Japanese exporters once the yen weakened materially in the 2022-2024 cycle. The duration of the round trip—a generation and a third—has become a benchmark frequently invoked in discussions of long-horizon equity risk, particularly by those of us who find the standard "stocks always recover within a decade" framing to be culturally specific to U.S. markets in a way the assertion does not advertise.

What the Episode Teaches

The Japanese experience illustrates the dynamics of a credit-driven asset-price bubble in which lending, collateral values, and market prices reinforce one another in a procyclical manner—a structure that Hyman Minsky described in the abstract and that Japan provided in the empirical. It is a case study in debt deflation as Irving Fisher described it: falling asset prices and rising real debt burdens compounding one another through a financial system whose lending capacity is itself impaired by the process. And it remains the cleanest available demonstration of the limits of conventional [[monetary policy]] when policy rates approach the zero lower bound and the financial system is carrying legacy bad debt the authorities are reluctant to force into the open.

Subsequent episodes—the United States housing bubble of the mid-2000s, the European sovereign debt crisis of 2010-2012—have been analysed with reference to features of the Japanese precedent, in some cases by policymakers who served their apprenticeships studying it. Ben Bernanke’s Federal Reserve, in particular, made a deliberate effort to avoid what it understood as the Japanese errors: forbearance, slow recapitalisation, insufficient monetary aggression at the lower bound. Whether that effort succeeded on its own terms is a question to which the next decade or two will provide the better part of the answer; I would not prejudge it. The episode and its aftermath remain among the most extensively studied in modern financial history, and one imagines they will continue to be, since the alternative is that the next bubble proceeds without the benefit of the precedent, which seems unlikely though not impossible.