Japan's lost decades after the 1990 crash: a lesson for Asian investors

How large was Japan’s 1990 stock market crash?

The Nikkei 225 closed at a record 38,915.87 on 29 December 1989. In 1990 it fell from 37,189 to 23,849, a loss of more than 35%, and it lost more than 43% within a year to a low of 21,902 on 5 December 1990 (Wikipedia). On 28 October 2008 it reached an intraday low of 6,994.90, about 82% below its peak nearly 19 years earlier.

These numbers show two clocks. The first fall took 1990. The long decline took until 2008. A crash has a fast phase, and then a slow phase in which prices wait for earnings and balance sheets to repair. Plan for both phases, since the second decides how long capital stays below its earlier level.

What built the bubble before Japan’s stock market crash?

Land and stock prices roughly tripled during the 1980s (Wikipedia). Commercial land prices in the 6 major cities rose 302.9% from 1985 to 1991, and the Tokyo Imperial Palace grounds were estimated at the peak to be worth more than all real estate in California. Prices that move this far from income need ever more borrowing to hold up.

The Bank of Japan raised its discount rate in five steps, from 2.5% in May 1989 to 6.00% in August 1990 (Wikipedia). The Bank’s own historical paper confirms the first step, from 2.5% to 3.25%. Tightening and a falling market arrived together, and borrowers who had bought assets with loans found their collateral losing value.

How did bad loans prolong Japan’s stock market crash?

Overleveraged banks and insurers carried large bad debts, and government capital, central bank credit, and delayed loss recognition kept many banks alive as zombie banks (Wikipedia). The Bank of Japan’s own review counts cumulative losses of 54.7 trillion yen from disposing of bad loans at major banks between fiscal 1992 and 1999.

In November 1997 Sanyo Securities, Hokkaido Takushoku Bank, and Yamaichi Securities failed, and the Long-Term Credit Bank of Japan and Nippon Credit Bank failed in late 1998. The government injected 9.3 trillion yen of public funds into major banks in March 1998 and March 1999. Slow clean-up kept credit tight for years, so the speed of bank repair is the first indicator to compare across Asian markets today.

How slow was growth after Japan’s 1990 stock market crash?

GDP grew 1.14% a year on average from 1991 to 2003 (Wikipedia). The IMF describes a recession that continued through 1998, with output 5% below its early-1997 peak by the last quarter of 1998, which it calls Japan’s worst postwar recession. Short-term interest rates fell to virtually zero by early 1999, and the general government deficit reached almost 10% of GDP.

The Bank of Japan adopted its zero interest rate policy in February 1999, ahead of other central banks. Deflation persisted, and near-zero rates stayed ineffective because a 0% nominal rate meant a positive real rate when prices fell. GDP in 2017 was only 2.6% above 1997, an annualised growth rate of 0.13% (Wikipedia).

How long did the Nikkei 225 take to recover from Japan’s 1990 crash?

The Nikkei 225 closed at 39,098.68 on 22 February 2024, above the record of 38,915.87 set on 29 December 1989 (Wikipedia). The recovery took 34 years, and it came after a low of 6,994.90 in October 2008.

Property recovered later still. Nationwide average land prices rose year over year for the first time in 2018, by 0.1%. A holder who bought the index at the peak waited 34 years to see the same level again, before any account of inflation or dividends. Equity levels and land prices recovered on different calendars.

What does Japan’s lost decades record teach Asian investors?

Four lessons fit the record. Valuations far above income raise the size of a later fall. Fast bank clean-up shortens the slow phase. Deflation changes the meaning of interest rates. Time below a previous level can last decades. Each lesson maps to an indicator: valuation ratios, non-performing loans, price trends, and the age of the last peak.

Each Asian market has its own debt, demographics, and exchange rate regime, so a reader should treat the Japanese record as a stress case, not a forecast. The pages on Hong Kong stock market crashes and on Singapore in this series show other stress cases from the region.

How can you test an Asian market for a Japan-style crash?

Check five items each quarter as of October 2026: valuation ratios against their own history, the growth of property prices against income, non-performing loans at banks, the direction of consumer prices, and the age of the last index peak. Record each number with its date.

The EQOS Asia method page explains how a structural score combines such inputs. The Λ simulator tests a soft landing, a correction, and a crash for your own exposure, and a 30-minute pre-diagnostic through the contact page turns the result into a ranked set of options.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile