Before you start
What you'll be able to answer
- How did the push to cut the major banks' bad loans from 2002 work?
- How did these years of weak growth and bank trouble end?
- How far had Japan's government debt risen by 2005, and why?
Where this sits
Japan's Lost Decade — 1989 to 2006 · this module is lit
- Dec 1989Share prices in Japan peak
- Aug 1990The Bank of Japan's last rise in its discount rate, the interest rate at which it lent to banks
- Aug 1992The government announces the first of six spending packages up to 1995
- 1995-96The first direct use of taxpayers' money against financial instability covers losses at failed housing-loan companies, the jusen
- 1 Apr 1997The consumption tax rate, a tax on what people buy, is raised
- Nov 1997Sanyo Securities, Hokkaido Takushoku Bank and Yamaichi Securities fail
- Oct-Dec 1998Japan's parliament, the Diet, makes more public money available to deal with banks' bad loans, and the state takes over the Long-Term Credit Bank and Nippon Credit Bank
- Feb 1999The Bank of Japan begins its zero interest rate policy, pushing its overnight rate as low as possible
- Mar 1999Public capital, new capital paid for by the state, is put into the major banks
- Mar 2001The Bank of Japan begins quantitative easing, a target for the money banks keep in their accounts with it
- Oct 2002The Financial Services Agency, the banking regulator, launches its programme to cut bad loans
- Mar 2006The Bank of Japan ends quantitative easing
In 2002 Japan's big banks still held large bad loans
By 2002 Japan's government had spent years propping up its banks. It had made public money available to them. It had taken two failing banks into public ownership. It had also put state money into the major banks as new capital. Yet the banks still held large bad loans, loans that were not being repaid. The economy barely grew that year. In October 2002 the Financial Services Agency (FSA), the government body that supervises banks, set the major banks a deadline to cut their bad loans.
A bank has set no money aside against a large loan. It decides the loan will never be repaid and writes it off, taking it off its books as lost. What does that do to the bank?
The loss comes out of its profit, then its capital. A bank's capital is its owners' money. It absorbs losses before depositors lose anything. Writing off a bad loan uses up some of that capital. So a bank that cuts its bad loans also needs more capital. If a bank had put money aside for the loan earlier, it took the loss then. If it put nothing aside, the loss shows in its accounts only when it writes the loan off.
The FSA set the major banks a deadline to halve their bad-loan ratio
The target was in the FSA's Program for Financial Revival of 30 October 2002. The FSA measured each major bank's bad loans as a share of all its lending: the bad-loan ratio. The major banks' ratio was to fall to about half its March 2002 level by March 2005. March 2005 was the end of the 2004 financial year. The stated aim was to end the bad-loan problem by then.
The FSA used three kinds of pressure. Banks had to value their loans more strictly. They had to strengthen their capital. Their management was tightened, and outside auditors checked their accounts more strictly. The FSA also wanted a banking system that people could rely on.
The FSA made the banks value their loans more strictly
Each bank graded its borrowers, from normal, through "in danger of bankruptcy", down to the worst. It set money aside for the loans it expected to lose. The FSA tightened that grading. For large borrowers that were weak but not yet in danger of bankruptcy, a major bank had to estimate the cash each borrower could repay, loan by loan, and set money aside on that basis. Special inspections in 2003 and 2004 checked the banks' grading. The FSA published how far the banks' grades differed from its inspectors' grades.
Part of a bank's capital was deferred tax assets: tax savings it expected on profits it hoped to make later. If those profits never came, the savings were worth nothing. The FSA called them a less solid kind of capital and said it would value them strictly.
Check yourself
The FSA published the gap between how each major bank graded its borrowers and how the FSA's inspectors graded them. What was that meant to push the banks to do?
The bad-loan ratio fell, and the banks wrote bad loans off
The major banks' bad-loan ratio was 8.4 per cent in March 2002, 5.2 per cent in March 2004 and 4.7 per cent by September 2004, on the FSA's figures. In the year to March 2003 alone their bad loans fell by ¥6.5 trillion, to ¥20.2 trillion. The Cabinet Office, part of Japan's government, reported that fall.
In that year the banks wrote off loans to borrowers in danger of bankruptcy or worse. Fewer new bad loans appeared, the Cabinet Office said. Clearing the old loans was costly. The cost, it said, was more than the banks' core operating profit, the profit from their ordinary banking business.
Check yourself
Work from this section: the FSA's target was a ratio of about half the March 2002 ratio. Roughly what ratio was it aiming for, in per cent?
4.2 per cent
About 4.2 per cent: half of the March 2002 ratio. The FSA said "about half", so 4.2 per cent is approximate.
Check yourself
Three of these were part of the FSA's 2002 programme for the major banks. Which one was not?
In May 2005 the FSA said its target had been met
The major banks reported their results for the year to March 2005 in May. On 25 May the Minister for Financial Services said their combined bad-loan ratio had fallen to 2.9 per cent, and that the goal of halving the March 2002 ratio "has thus been achieved". The judgement is the FSA's own, on a target it had set itself.
The Minister said the bad loans had held back Japan's economy ever since its bubble in share and land prices burst. The FSA said it would now stop relying on emergency measures. It would work on making the banking system sounder, and it would keep watching each bank's loans.
Growth picked up from 2003 and share prices rose from their April 2003 low
At the same time, the economy picked up. GDP is the value of everything the economy produces, with price changes taken out. It barely grew in 2002. It then grew by more than 1 per cent in every year from 2003 to 2006, on figures from the World Bank, an international development lender. Share prices fell to their lowest point of the period in April 2003. By December 2005 they were roughly twice as high. These share prices come from an index of the International Monetary Fund (IMF), which advises and lends to its member countries. Unemployment, at its highest of the period in 2002, fell in each year after.
The Cabinet Office dated the lowest point of the downturn to January 2002, in a report of October 2003. Output began to recover after that. The recovery began with exports, and with firms whose profits were rising, it said.
In March 2006 the Bank of Japan ended quantitative easing, a target for the money banks kept in their accounts with it. Which of these changes had the Bank said it was waiting for?
Consumer prices. In 2001 the Bank had promised to keep quantitative easing until consumer prices, not counting fresh food, were steadily at least as high as a year before. Share prices and jobs had improved too, but the promise named prices.
In March 2006 the Bank of Japan ended quantitative easing
On 9 March 2006 the Bank of Japan, Japan's central bank, ended quantitative easing. It went back to targeting its overnight rate, the rate banks charge each other for one-day loans, and kept it at effectively zero. Consumer prices were above their level a year earlier and were expected to stay above it, the Bank said. So the condition of its 2001 promise was met: consumer prices, not counting fresh food, steadily at least as high as a year before. Over a few months it would cut the money banks kept with it towards the amount they had to hold.
The recovery was continuing steadily, the Bank said. Firms had high profits and invested more, and households spent steadily. Years later the Bank went back to easing, supplying extra money to the economy. It ran a policy it calls Quantitative and Qualitative Monetary Easing from April 2013 to March 2024.
Japan's general government gross debt is the total owed by central government, local government and the social security funds. In 1990 it was a little over half a year's GDP. How big do you think it was by 2005?
About one and a half times a year's GDP, nearly three times its share in 1990. Each year the government spent more than it collected and borrowed the difference. GDP counted at current prices, without taking out price changes, stopped growing after 1997.
Government debt nearly tripled as a share of GDP
The chart shows Japan's general government gross debt as a share of GDP from 1992, on IMF World Economic Outlook data, April 2026. It was 55 per cent of GDP in 1990 and 153 per cent in 2005, its highest point on the chart.
A deficit is the gap between what a government spends and what it collects. The government borrows to cover it, so deficits add to its debt. Continued deficits had fed rapidly growing debt, the IMF said in 2003. The debt ratio is debt divided by GDP, here at current prices. If GDP stops growing, the ratio rises even when debt does not. Prices were falling after 1997, so GDP at current prices shrank in several years, on World Bank figures.
Why did Japan's slump last so long, and why do economists still disagree about it? The last module of the series sets out the arguments.
Source: IMF, World Economic Outlook database, April 2026 (GGXWDG_NGDP), general government gross debt, Japan. IMF data, used with credit.
Check yourself
Look at the chart. In which year did government debt first rise above a year's GDP, 100 per cent?
Check yourself
The key questions
How did the push to cut the major banks' bad loans from 2002 work?
The FSA set the major banks a deadline to halve their bad-loan ratio. It made them value loans and count capital more strictly. The banks wrote bad loans off and the ratio fell year by year. In May 2005 the FSA said its target had been met.
How did these years of weak growth and bank trouble end?
Growth picked up from 2003, share prices rose from their April 2003 low, and unemployment fell from its 2002 peak. In March 2006 the Bank of Japan said consumer prices were above their level a year earlier. It then ended quantitative easing.
How far had Japan's government debt risen by 2005, and why?
It nearly tripled as a share of GDP between 1990 and 2005. The government kept borrowing to cover its deficits. GDP counted at current prices stopped growing after 1997, and that also pushed the debt ratio up.
The numbers
Check yourself
How did the FSA push the major banks to cut their bad loans from 2002?
Check yourself
When the Bank of Japan ended quantitative easing in March 2006, what did it go back to targeting?
Check yourself
What happened to Japan's government debt, as a share of GDP, between 1990 and 2005?