Financial Risks, Stability, and Globalization

13 Changes in Bank Behavior During Japan’s Financial Crisis

Omotunde Johnson
Published Date:
April 2002
  • ShareShare
Show Summary Details

In the 1990s, Japan faced a serious financial crisis due to the formation and bursting of a bubble. In terms of magnitude, it was a crisis no other country had ever experienced. Though essentially a serious domestic problem, because Japan was playing an increasingly important role in the international financial arena and also because of the size of her economy, it became a global concern out of fear it could trigger a cross-border financial crisis.

As the central bank, the Bank of Japan (BOJ) took every possible measure to contain the financial crisis and, indeed, prevented it from adversely affecting the international financial scene and global economy. However, crisis prevention costs were considerable.

The BOJ is now examining its crisis management and trying to glean lessons from the experience for other countries or future generations that may face the same problem. This paper deals with the behavior of banks during the financial crisis, especially focusing on how to detect the spread of a financial crisis at an early stage.

Outline of the Financial Crisis in Japan

Asset prices (i.e., land and stocks) rose steeply during the late 1980s (the so-called “bubble era”), fell sharply in the early 1990s, and have since continued falling. For a chronology of the financial crisis, see Table 13.1 As a result, Japanese deposit-taking institutions incurred a huge amount of nonperforming assets, and many saw their capital impaired along with the disposal of such nonperforming assets.

Table 13.1.Japan’s Financial System: Chronology of Events
YearMonthEventsLaws, Measures, and Policies
  • Ministry of Finance (MOF) regulates the total amount of loans and loans to three sectors (real estate, construction, finance companies)

1991JulyResolution of Toho Sogo Bank (first financial assistance extended by DIC) announced
  • MOF releases “Policies for Financial Administration” to encourage liquidation of assets/ collateral, etc.

  • Cooperative Credit Purchasing Company established

  • Bankers’ Federation announces revision of its disclosure standards (disclosure of the non-accrual amounts and past due loans)

  • MOF announces that the “capital ratio of financial institutions should be 8 percent or higher”

MayResolution of Kamaishi Shinkin announced
  • Jusen companies implement restructuring plans (February to June 1993)

1994DecemberResolution of Tokyo Kyowa and Anzen credit cooperatives announced
1995JanuaryTokyo Kyodo Bank established
  • Financial System Research Council, Disclosure WG publishes report on expanding range of information to be disclosed about the soundness of assets of financial institutions (disclosure extended to loans on which interest payment has been waived, and loans to borrowers whose business restoration is supported by banks)

  • MOF releases “Revitalization of Financial Functions of the Financial System”

JulyCosmo Credit Cooperative ordered to suspend operations
AugustKizu Credit Cooperative ordered to suspend operations Resolution of Hyogo Bank announced
SeptemberDaiwa Bank announces loss at its New York branch
  • Financial System Research Council releases final report on measures for the maintenance of financial system stability;

  • Cabinet decides on concrete measures to address the jusen problem

1996MarchResolution of Taiheiyou Bank announced
  • Three bills (Law to Implement Measures for Ensuring the Sound Management of Financial Institutions, Law to Provide Special Procedures for Reorganizing Financial Institutions, and amended Deposit Insurance Law) and Jusen Law passed

  • Housing Loan Administration Corporation established

  • Tokyo Kyodou Bank reorganized into Resolution Collection Bank (as an assuming institution)

NovemberHanwa Bank ordered to
  • Financial system reform (Japanese suspend operations version of the Big Bang) announced

1997AprilNCB announces plans to restructure
Hokkaido Takushoku Bank announces plans to merge with Hokkaido Bank
OctoberMerger of Fukutoku Bank and Naniwa Bank announced; resolution of Kyoto Kyoei Bank announced
NovemberSanyo Securities files application for rehabilitation

Resolution of Hokkaido Takushoku Bank announced

Yamaichi Securities announces suspension of operations

Resolution of Tokuyo City Bank announced
  • A bill to amend the Deposit Insurance Law passed (special merger)

  • Two bills concerning financial bank holding companies passed

  • Two bills concerning financial system stabilization (Amended Deposit Insurance Law and Emergency Measures for Financial Functions Stabilization Law) passed

  • Issue of preferred stock and subordinated bonds based on Emergency Measures for Financial Stabilization Law approved (21 banks: ¥1.8 trillion)

  • Prompt Corrective Action (PCA) introduced

MayMerger of Hanshin Bank and Midori Bank announced
  • Bankers’ Federations revise unified disclosure guidelines for the preparation of disclosure reports

Osaka announces restructuring plans for credit cooperatives within the prefecture(disclosure of loans past due 180 days or more and restructured assets added)
JuneLTCB and Sumitomo Trust Bank announce plans to merge
  • Financial Supervisory Agency (FSA) inaugurated;

  • Four bills (Financial Reform Law, SPC Law, Revision of 13 laws to comply with SPC Law, Obligation Netting Law) passed

JulyFinancial Supervisory Agency and BOJ conduct intensive on-site inspections and examinations
  • Government and ruling parties release second report on “Comprehensive Plan for Financial Revitalization (Total Plan)”

  • Government and ruling party submit a bill concerning the “Total Plan” to the Diet

OctoberLTCB placed under Special Public Administration
  • “Financial Reconstruction Law” and “Financial Functions Early Strengthening Law” became effective

DecemberNCB placed under Special Public Administration
  • Financial Reconstruction Commission (FRC) inaugurated

  • Capital injection pursuant to the Early Strengthening Law (15 banks: ¥7.45 trillion)

AprilKokumin Bank placed under management of the Financial Reorganization Administrator (FRA)
  • “Final Report” of the WG on Financial Inspection Manuals released

MayKoufuku Bank placed under management of the FRA
JuneTokyo Sowa Bank placed under management of the FRA
AugustNamihaya Bank placed under management of the FRA
OctoberNiigata Chuo Bank placed under management of the FRA
  • Financial System Council released “Report on the Framework of the Deposit Insurance System and Resolution of Failed Financial Institutions after the Termination of Special Measures”

  • Three ruling parties agreed to postpone lifting of special measures concerning the Deposit Insurance System (end-March 2001 to end-March 2002)

However, the measures taken, at least in the early 1990s, to deal with nonperforming assets were not sufficient or timely. Examining why the authorities were so slow to respond, we find four reasons.

  • No one had expected that land and stock prices would fall so sharply and continuously. There was a vague expectation that the drop in asset prices would soon stop once the economy picked up.

  • There was a fear of social unrest. The public was not used to financial disruption as very few deposit-taking institutions had failed in Japan since World War II. In addition, because the disclosure system was inadequate, the public was not fully aware of the situation at deposit-taking institutions, and if they had known of the problems, panic might have ensued.

  • If a deposit-taking institution failed against such a background, it could have led to contagion. However, the resolution framework at that time was insufficient, and crisis management was constrained by very limited financial and human resources.

  • Due to insufficient infrastructure with respect to accounting, public disclosure, and corporate governance, market discipline was not effective. While some measures were taken to promote provisioning and the write-offs of nonperforming assets—such as the relaxation of write-off rules using a special account and the establishment of the Cooperative Credit Purchasing Company to purchase nonperforming loans (the difference between book value and purchase price was tax free)—they were not enough.

The Japanese financial authorities came to recognize that the situation was much more serious than first thought and began dealing with the nonperforming asset problem in a more comprehensive way after the failure of two credit cooperatives in December 1994. Their first priority was to prevent social unrest caused by contagion and then to fully protect depositors. It was also thought necessary that shareholders and managers be held responsible for failed deposit-taking institutions in order to contain moral hazard.

In the process of resolving failed deposit-taking institutions, the authorities often faced cases where it was difficult to take adequate action within the existing framework. The Deposit Insurance Corporation (DIC) at that time could only effect payoff (a deposit of no less than ¥10 million would be subject to haircutting) or provide financial assistance to facilitate transfer of business. Because it was deemed inappropriate for depositors to share losses, the only method available was financial assistance, which was restricted to within the payoff cost (the net cost required of the DIC to protect deposits up to ¥10 million). As the financial contribution provided by the DIC was restricted, external support from the financial sectors, called houga-cho system, was needed to fully protect depositors. But because this external support was voluntary, it became unsustainable when the failure of deposit-taking institutions increased during 1995.

In June 1996, to eliminate the restriction mentioned above, the Deposit Insurance Law was amended as follows.

  • “Special financial assistance,” which enabled the DIC to fully assume losses incurred by failed deposit-taking institutions regardless of the payoff cost, was introduced. The framework for the full protection of all deposits and liabilities was thus put in place.

  • The insurance premium was raised from 0.012 percent to 0.084 percent (for insured deposits) to strengthen DIC funding. As a result, annual insurance revenues increased to ¥460 billion.

  • A special account for credit cooperatives was set up as a temporary measure until end-March 2001, and a framework was established enabling the injection of public funds to deal with failed credit cooperatives.

The amendment of the Deposit Insurance Law in 1996 contributed to ensuring flexibility of the deposit insurance system. However, the system did not provide a perfect solution to the nonperforming asset problem, which only worsened. Reasons for this are as below.

  • Loan-loss provisioning and write-off amounts tended to be smaller compared with the actual amount of nonperforming loans since no comprehensive standards had been established.

  • The financial authorities believed overemphasizing the problem would destabilize the financial system. Thus, public funds were to be injected only into credit cooperatives.

  • Before the nonperforming asset problem surfaced, public funds had been injected into jusen (housing loan companies). The injection of public funds (¥680 billion) into jusen was approved in the same session of the Diet as the amended Deposit Insurance Law. During discussions taxpayers expressed strong opposition because jusen were not depository institutions and thus were remote from taxpayers. Hence, the politicians became extremely sensitive to injecting further public funds.

In autumn 1997, internationally active financial institutions such as Hokkaido Takushoku Bank and Yamaichi Securities failed. Market participants became extremely risk averse, and the market began losing liquidity very rapidly.2 As a result, many financial institutions faced funding difficulties.

The Bank of Japan (BOJ) began extending loans on special terms (hereafter, “special loans”) to Hokkaido Takushoku Bank, Yamaichi Securities, and others based on four criteria (see below). From the viewpoint of extending such special loans, the BOJ believed there was a difference between “a failure of depository institution” and “maintenance of the stability of financial functions.” In other words, the BOJ extended special loans because it thought it necessary to maintain financial functions such as the payment of deposits, extension of loans, and payment services in situations where a deposit-taking institution failed and managers and shareholders were held responsible. However, liquidity provision by central banks, if provided without restriction, might erode self-discipline on the part of managers of deposit-taking institutions, depositors, and creditors, which in turn would lead to a weakening of the financial system. To prevent a moral hazard problem from arising, the following four conditions were set for the provision of special loans (which the BOJ has applied in all recent cases):

  • there must be a strong likelihood that systemic risk may materialize;

  • there must be no alternative to the provision of the central bank funds;

  • all responsible parties are required to take clear responsibility to avoid moral hazard; and

  • the financial soundness of the Bank of Japan should not be impaired.

Under the then existing framework to fully protect all depositors, special loans to deposit-taking institutions have been collected in full3 when financial assistance was provided by the DIC.

In addition to providing special loans to individual deposit-taking institutions, the BOJ also provided ample liquidity to the market, in an amount equivalent to more than ¥1 trillion in excess of required reserves. As a result, a breakdown of market functions was successfully avoided.

Though the financial crisis of autumn 1997 was overcome as a consequence of a set of policies including BOJ measures, it nevertheless shook the whole financial system. That the DIC funds would soon dry up in dealing with the successive failure of deposit-taking institutions was recognized, as was the fact that an increase in loan-loss provisioning and write-offs would undermine the capital base of solvent institutions. In addition, the continuing economic downturn worsened the profitability of Japanese deposit-taking institutions and public distrust prevented them from increasing capital.

During this period, as large deposit-taking institutions failed in succession, many came to realize how serious the nonperforming asset problem had become, which enabled a consensus to be reached with respect to injection of public funds. Hence, with revision of the Deposit Insurance Law and enactment of the Financial Functions Stabilization Law, a framework for capital injection using public funds was established in February 1998, enabling ¥30 trillion in public funds (of which ¥17 trillion was for the resolution of failed deposit-taking institutions and ¥13 trillion for capital injection) to become available.

However, at that time, deposit-taking institutions did not want the public to think that a large capital injection using public funds meant there was a large amount of potential nonperforming assets. Moreover, the Financial Crisis Management Commission, an entity authorized to approve the injection of public funds, was not given power to examine and control deposit-taking institutions, and thus did not possess a sufficient legal basis. As a result, in March 1998, capital of about the same amount was injected into 21 banks, but was not sufficient. The total amount of funds injected was no more than ¥1.8 trillion.

In March 1998, as a result of the injection of public funds in small amounts, the financial system temporarily restored stability; but in June, sentiment worsened following the market’s hardening views of the problems at the Long-Term Credit Bank of Japan (LTCB).4 LTCB was an internationally active bank with assets of ¥26 trillion and notional principal of ¥50 trillion from off-balance trading. If it failed in a disorderly manner, it was considered unavoidable that the international financial market would be affected. Thus, to maintain its financial functions, the financial authorities sought the possible merger with another deposit-taking institution. However, this merger plan did not work out, and the LTCB problem was left to Diet discussions in the summer session of 1998. Meanwhile, the BOJ strengthened monitoring of its funding operations in order to maintain LTCB’s financial functions. The BOJ also insisted that measures to maintain financial functions be explicit in the new framework of financial failure resolution.

In October 1998, the Financial Reconstruction Law and the Early Strengthening Law were enacted (see below). Public funds were doubled (from 30 trillion yen to 60 trillion yen). The deposit insurance system became, at this point, a very comprehensive and empowered one from both financial and functional perspectives.5

The Financial Reconstruction Law provides for

  • management by a Financial Reorganization Administrator;

  • a bridge bank—to assume the business of failed banks;

  • special public administration (temporary nationalization); and

  • emergency measures concerning the purchase of assets.

The Early Strengthening Law provides for

  • capital injection into deposit-taking institutions using public funds.

The Financial Reconstruction Law provided methods to maintain financial functions. Based on the law, LTCB (in October 1998) and the Nippon Credit Bank (in December 1999) were placed under special public administration. Moreover, five regional banks were placed under the management of Financial Reorganization Administrators in 1999. Between 1992 and end-March 2000, a total of 110 deposit-taking institutions (including planned ones) were resolved under the Deposit Insurance Law (and revisions).

Pursuant to the Early Strengthening Law, ¥7.5 trillion in public funds was injected into 15 large banks in March 1999. In addition, in September of the same year, ¥300 billion was injected into four regional banks. The amount of funds injected was determined based on new loan-loss provisioning and write-off standards,6 which aim at an adequate amount of capital being maintained even after possible additional provisioning or write-off and the disposal of unrealized losses on securities. As a result of these measures, confidence in Japan’s financial system seemed to have, though not fully, recovered.

Problems Behind the Financial Crisis in Japan

Japan’s financial system has been seriously shaken. Many consider that various factors—for example, the macroeconomic environment—are involved. This paper focuses on the behavior of deposit-taking institutions. The changes in behavior were both causes and results of the financial crisis. This paper intends to identify in what way the changes in the banking behavior were connected to the financial crisis, with particular attention on some of the aspects that predicted, or at least reflected, the deepening of the crisis.

Concentration of Loans to Particular Industries Without Proper Risk Control

In the early 1980s, large corporations began decreasing their funding dependency on deposit-taking institutions due to financial deregulation and liberalization. Japanese deposit-taking institutions increased their exposure to real estate-related loans via finance companies without implementing proper credit risk controls (namely, risk diversification and risk-adjusted pricing), which contributed to the formation of the bubble. Such concentration of risks had not been resolved when the bubble burst, and deposit-taking institutions were left with a huge and unprecedented total amount of nonperforming loans.7

As evidenced by loans extended by various categories of deposit-taking institutions for the past 20 years (Figure 13.1), city banks, long-term credit banks, and trust banks saw their loan growth rate rise in the early 1980s. While the rate peaked in the late 1980s at city banks, it peaked a little later at long-term credit banks and trust banks. The contribution of loans by long-term credit banks and trust banks to three industries (real estate, construction, and finance companies, hereafter referred to as “three industries”) to the change in total loans from a year earlier was greater than those by city banks. In the late 1980s and the early 1990s, regional banks and regional banks II expanded loans mainly to the three industries, and thus the loan growth rate rose. (Most deposit-taking institutions, except city banks, suffered heavier losses as they expanded loans right up until the bubble burst.)

Figure 13.1.Loans by Type of Industry

(Percent change from previous year; contribution of loans to three industry sectors—construction, real estate, finance and insurance—to overall increase/decrease)

Many deposit-taking institutions overweighted the real estate collateral; as shown in Figure 13.2, they expanded credit by taking real estate as collateral and, moreover, sharply increased lending without collecting proper risk premiums (Figure 13.3).

Possible explanations for such behavior include the following.

  • Based on excessively bullish expectations of economic growth and past experience, there was a generally believed myth that land prices would never fall.

  • As for the three industries, Japanese deposit-taking institutions did not fully recognize that a fall in land prices would not only lower the value of collateral, but also deteriorate financial conditions of the borrowers.

  • Japanese deposit-taking institutions took unrealized gains on securities into account when evaluating capital, although they were exposed to significant risk with respect to their stockholdings compared to core capital (Tier I).

  • Loans via finance companies resulted in deposit-taking institutions being less concerned about risk concentration.

Figure 13.2.Loans by Type of Collateral Contribution to Overall Increase/Decrease by Type of Collateral


Figure 13.3.Loans by Type of Industry

(All Japanese Banks)

Even when the nonperforming asset problem drew public attention, Japanese deposit-taking institutions failed to change their views on land prices because of expectations of continuing economic growth. Though the financial authorities recognized the potential risk caused by such lax behavior on the part of banks, the probability of losses on real estate—related loans seemed low judging from historical records, and the profitability of deposit-taking institutions appeared stable because they were realizing part of the gains on stocks. Thus the authorities were not successful in giving effective warnings.

Later, deposit-taking institutions came to realize that they had to dispose of nonperforming loans as soon as possible. Due to intensive inspections and examinations conducted by the authorities, the publication of “Financial Inspection Handbook (Financial Supervisory Agency)” and “Viewpoint on the Write-offs and Allowances in Association with the Capital Injection (Financial Reconstruction Commission),” and greater disclosure of nonperforming assets in 1999, the disposal of nonperforming assets was promoted, at least on an accounting basis. From end-March 1992 to end-September 1999, the cumulative losses stemming from the disposal of nonperforming assets (= loan-loss provisioning + direct write-offs) at all banks was ¥61.0 trillion. From end-March 1992 to end-March 1998 (total of ¥58.8 trillion), such loan-loss provisioning and write-offs were funded by net operating profit (¥35.1 trillion), capital (¥7.2 trillion), and realized gains on stocks (¥12.8 trillion). See Figure 13.4. Unrealized gains on stocks declined significantly as stock prices declined.

Figure 13.4.Losses on the Disposal of Non-Performing Loans and Capital

(All Japanese Banks)

However, removing nonperforming loans from the balance sheet, which would improve cash flow (i.e., securitization), is yet to be seen in full scale because8

  • many deposit-taking institutions are unable to liquidate nonperforming assets because securitization might incur additional losses as the securitization market is not sufficiently liquid;

  • effective direct write-off methods (i.e., consigning a servicer for asset collection and improving the auction process) had been established only recently; and

  • the cost of holding nonperforming assets is now relatively small due to monetary easing and the resulting low market rates.

Meanwhile, large Japanese deposit-taking institutions are beefing up credit risk management by implementing credit risk models and risk-adjusted pricing. Thus, the concentration of credit risk seems less likely in the future. In fact, after the collapse of the bubble, loans to the three industries declined considerably. The loan spread widened from 1991, and deposit-taking institutions adjusted risk premiums appropriately.


Japanese deposit-taking institutions traditionally held shares of corporations to ensure a long-term business relationship and, during the bubble era, increased their shareholdings for two reasons (Figure 13.5):

Figure 13.5.Number of Shares Held by Shareholders

  • to maintain business relationships with large corporations, which had heavily relied on banks for funding but had begun shifting to equity financing; and

  • to facilitate their own equity financing, which they did to cope with the Bank for International Settlements (BIS) capital adequacy standards (1998), by selling their stocks to corporations; in return, deposit-taking institutions bought the stocks of corporations.

The original purposes of cross-shareholdings were to prevent buyouts, maintain keiretsu relationships, and ensure a stable supply of funds (often referred to as the basis of the so-called “main bank system”). As equity financing by both corporations and deposit-taking institutions expanded during the bubble era, cross-shareholdings quickly expanded.

The percentage of stockholdings by Japanese deposit-taking institutions of their total assets was 6 percent (at city banks, 7 percent), higher than that of European banks9 allowed to hold stocks (3 percent). (U.S. banks were basically prohibited from holding stocks—See Table 13.2).

Table 13.2.Laws Relating to Bank Investments in Nonfinancial Corporations
JapanCorporations with capital exceeding ¥35 billion or net worth exceeding ¥140 billion are not allowed to hold shares in domestic corporations in excess of the higher of their capital or net worth in total (Antitrust Law).
Banking groups are not allowed to hold shares in non-financial corporations in excess of 5 percent (15 percent in the case of bank holding companies) of the outstanding shares of such corporations (Banking Law).
Finance companies are not allowed to hold more than 5 percent of the outstanding shares of a domestic corporation. (Antitrust Law)
United StatesBanks are basically not allowed to hold shares in corporations (National Bank Act).
Bank holding companies are not allowed to hold more than 5 percent of the voting shares of a nonfinancial corporation (Bank Holding Company Act.
European UnionBanks are allowed to hold shares in nonfinancial corporations up to 60 percent of their own funds (Second Banking Coordination Directive).
United KingdomThere is no legal provision prohibiting banks from holding shares in nonfinancial corporations. However, banks generally do not hold such shares for lack of economic rationale.
GermanyBanks are allowed to hold shares in nonfinancial corporations up to 60 percent of their own funds (Gesetz für das Kreditwesen).

Several problems for depository institutions stemming from cross-shareholdings could be identified.

  • Insufficient awareness of high credit and market risks inherent in cross-shareholdings. Japanese banks extended loans to the issuers of the stocks they held; thus, both credit and market risks simultaneously materialized during the economic downturn. Japanese banks may not have paid full attention to the procyclicality aspect of their behavior.

  • Excessive risk-taking due to the huge amount of unrealized gains on stocks. As mentioned before, Japanese deposit-taking institutions lacked efficient risk control systems and awareness of the risks involved in cross-shareholdings. Nevertheless, they took on an excessive amount of risk depending on the amount of unrealized gains on stocks at the beginning of the bubble era (Figure 13.6), which left a huge amount of nonperforming assets. Although identifying unrealized gains on stocks with capital was consistent with the idea of fair value accounting, the fact that the unrealized gain could not necessarily be realized at any time was overlooked by the Japanese deposit-taking institutions. At the same time, they did not fully take credit and market risks into account, as mentioned above. This encouraged them to engage in risk-taking activities.

  • Capital structure easily affected by stock price. As stock prices declined as the bubble collapsed, the capital base of banks became extremely vulnerable through two channels: (1) a decrease in Tier I capital caused by an increase in write-offs of stockholding when “the lower of cost or market” method was applied (Figure 13.7); and (2) a decrease in Tier II capital (unrealized gain in stocks). In addition, to ensure funds for writing off nonperforming assets, deposit-taking institutions realized some gains by selling stockholdings and repurchasing them at the market value (Figure 13.8). As a result, the book value of stockholdings rose; thus, the capital base became even more vulnerable.

  • Low profitability. Earnings per share on cross-shareholdings stayed low compared with funding costs, putting downward pressure on profit. From the formation to the bursting of the bubble (FY 1989 to FY 1999), cross-shareholdings-related costs are estimated at ¥10 trillion10 (all banks).

  • Conflict of interest arising from a main bank being both a shareholder and creditor at the same time. Main banks chose to provide steady funds to corporations, rather than to act as shareholders. Thus, they might have failed to set proper pricing for the lending. Based on this experience, each deposit-taking institution changed capital structure after the bursting of the bubble. More specifically, instead of being highly dependent on unrealized gains on stocks, institutions shifted to emphasize core capital. In addition, deposit-taking institutions have begun unwinding cross-shareholdings, taking the following changes in the financial environment into account.

Figure 13.6.Actual Amount of Capital

(All Japanese banks)

Figure 13.7.Write-Offs of Stockholdings

(All Japanese banks)

Figure 13.8.Unrealized Gains on Stocks

(All Japanese banks)
  • Due to the introduction of mark-to-market accounting and the emergence of mega banks, deposit-taking institutions became aware of the importance of the comprehensive control of their overall portfolios from the viewpoint of risk management and improving profitability.

  • Sharing the same view, corporations ceased to view the liquidation of cross-shareholdings as taboo.

Insufficient Disclosure of Actual Financial Conditions

From 1991, the actual amount of nonperforming assets probably grew dramatically, accompanied by a sharp decline in asset prices. Nevertheless, the amount that was covered by disclosure standards, and, thus, publicly announced, was very limited. Therefore, there might have been a big gap between the actual amount and the amount disclosed. As the definition of nonperforming assets was gradually enlarged, the amount of nonperforming loans became larger. As each revision of definition resulted in larger numbers, “the announced amount of nonperforming loans” became less reliable for depositors and market participants.

Having dealt with financial system problems more recently, non-performing assets are currently defined and classified more clearly (see Figure 13.9), and disclosure is on a par with other developed countries.

Figure 13.9.Stricter Classification Rules Applied to Disclosure of Nonperforming Assets

Notes. For end-March 1993, disclosure of only loans on a nonconsolidated basis. Since end-March 1999, disclosure of loans on both a consolidated and nonconsolidated basis.

Self-assessments are carried out with respect to all credits (loans, securities loaned, guarantees, etc.) on a nonconsolidated basis.

Since end-March 1998, the Financial Supervisory Agency aggregates and discloses the results of self-assessments by financial institutions. Only a number of institutions disclose results of self-assessments individually.

Yet, disclosure of quantitative credit risk information as a whole and qualitative information, including the management of risks other than credit risk, is still insufficient, and there is room for improvement.

Unstable Funding Structure

The major players—for example, city banks—in the financial system faced the following structural problems, which meant that they were susceptible to funding difficulties once reputational risk or market confusion materialized (Figure 13.10).

Figure 13.10.Balance Sheet of All City Banks

(September 1999)

Notes: All branches except large-denonimation time deposits of foreign branches. Borrowed money includes direct dealing.

Borrowed money includes subordinated loans and thus does not necessarily include short-term money market funds.

  • The amount of funds invested (core investment) substantially exceeded the amount of funds raised (core deposits).12 This resulted in huge dependence on funds raised in the wholesale market, rather than from other sources. The funding of deposit-taking institutions will become more stable when they depend on “sticky” small deposits that are less risk-sensitive (more deposit surplus), rather than on negotiable funds that are short-term and risk-sensitive.

  • The term structure of funds raised in the market was predominantly overnight.12

  • Banks were passive in liquidating assets.

Those structural problems mentioned above exacerbated funding difficulties experienced by Japanese deposit-taking institutions during the financial crisis of autumn 1997 when confidence in Japan’s financial system was undermined. It was observed that deposit-taking institutions that experienced deterioration in financial condition went through different stages13 in terms of funding difficulties, and some institutions eventually failed.

Meanwhile, the BOJ, engaged in the daily monitoring of the funding of deposit-taking institutions, grasped changes at an early stage, and asked the administrative authorities to draft resolution plans and monitor more intensively activities concerning funding operations.

Stage 1. Risk-sensitive market participants and large depositors became more selective and reluctant to do business with troubled deposit-taking institutions. As a result, higher risk premiums were attached to such troubled depository institutions, increasing fund-raising costs (Figure 13.11).

Figure 13.11.Average Interest Rates on Time Deposits1

(Difference between failed and sound banks)

1 No less than ¥10 million new receipts

Stage 2. As information about troubled deposit-taking institutions spread, providers of funds in the market became more risk averse and avoided making long-term deposits. Thus, troubled deposit-taking institutions were faced with higher rates and shortened maturities (Figure 13.12).

Figure 13.12.Average Term on Time Deposits1

(Difference between failed and sound banks)

1 No less than ¥10 million new receipts.

Stage 3. As the reputation of troubled deposit-taking institutions deteriorated further, providers of funds began to refuse further credit. Moreover, small depositors began to lose confidence and core deposits started to be depleted (Figure 13.13), leading to funding difficulties (Figure 13.14—increased volatility of deposit surplus ratio). They thus began recovering or selling loans and securities (Figure 13.15). In addition, more collateral was required against funding from the market.

Figure 13.13.Ratio of Deposit Surplus

Figure 13.14.Volatility of Deposit Surplus Ratio

Figure 13.15.Changes in Japanese Banks Deposits and Loans

Stage 4. With the prospect of failure, troubled deposit-taking institutions became unable to obtain negotiable large-lot time deposits, and the run on retail deposits accelerated. At this stage, the institutions were unable to deal with the situation by selling assets and so relied almost exclusively on funds from the market (which shifted from long-term to short-term overnight). When a deposit-taking institution was no longer able to raise overnight funds in the market, it faced failure.

Based on Japan’s experiences, in order to reduce the vulnerability of individual banks to external shocks, it is important to enhance the “sticky” core deposit base and stabilize funding as a strategy. Such strategy is helpful for Japanese financial institutions because from April 2002, depositors will have to share the cost should a deposit-taking institution fail; hence, large depositors and creditors are likely to become more sensitive to reputation.

Is Early Warning of a Financial Crisis Possible?

Basic Philosophy of Early Warning

As is clearly demonstrated by Japan’s experience described thus far, a financial crisis, once triggered, incurs a tremendous cost to society. The overall cost suffered by Japan until 1999 in regaining confidence in its financial system amounts to ¥77.4 trillion. Included in this figure are ¥61.0 trillion for the disposal of nonperforming loans held by viable institutions, ¥6.8 trillion for failure resolutions, and ¥9.5 trillion for capital injection into banks.

Therefore, it is most important to know the kinds and magnitudes of risks being accumulated within the financial system in order to detect the symptoms of the formation and bursting of a bubble. Moreover, should the materialization of risks become inevitable, it is essential to be well aware of the development of the crisis so that effective countermeasures can be taken and the social cost minimized. Indicators that reflect risk accumulation and crisis development, if available, would be helpful for making effective judgments. In the light of our experience, the behavior of banks clearly changes during the process of both risk accumulation and crisis development. Therefore, changes in their behavior would be a useful source of information in making judgments on the condition of the financial system. Monitoring such behavior, it would be possible to detect some early warning signs. It should be noted, however, that there cannot be any one unique early warning indicator since changes in banks’ behavior surface in various ways. It is necessary to carefully follow different indicators, those related to bank behavior among others, taking into account the economic and financial conditions of the time.

Such indicators can perhaps be classified into two categories: market information (including processed data), and information obtained in fulfilling the functions of the central bank. The latter includes managerial information pertaining to individual deposit-taking institutions that the BOJ obtains through on-site examination and off-site monitoring. Since systemic risk is typical where the insolvency of one particular deposit-taking institution might spread to other institutions by way of deposit runs and deteriorating market functions, a rational preventive measure would be to maintain close contact with specific institutions that are the potential source of financial unrest. Moreover, aggregating quantitative information (e.g., results of asset evaluation) regarding these institutions might help to reveal the amount of overall risks accumulated within the financial system.

Early Warning Indicators

Japan’s experience suggests the points that should be verified to detect burgeoning problems, and also the indicators that could be used to make related judgments.

Indicators Suggesting the Degree of Risk Accumulation

  • Aren’t there any signs regarding the formation/bursting of a bubble? In a financial structure where indirect financing plays a dominant role, abrupt changes in asset value in general (formation or bursting of a bubble) inevitably hit deposit-taking institutions through increased nonperforming loans. Thus, it is essential to detect the formation and bursting of a bubble at the earliest possible stage. This requires close monitoring of risk concentration indicators (outstanding loans by industry, collateral, and region; interactions among industries, collaterals, and regions when credit risk materializes), leverage indicators, and market information (asset prices, long-term interest rates, return on investment, etc.).

    In any case, a basic requirement is to stabilize the growth expectations of market participants through adequate monetary and other macroeconomic policies, thus forestalling the formation of a bubble. In this respect, fine-tuning of market participants’ growth expectations is of great importance. For this, an effective means available to the central bank is the frequent and informative announcement of the bank’s view on actual monetary and economic conditions.

  • Aren’t there any deposit-taking institutions facing serious managerial problems as a result of the bursting of a bubble? In order to determine the consequences of the bursting of the bubble, it is necessary to quantify the amounts attaching to various risks (credit, market, interest rate, and other risks) as well as potential changes in the solvency of deposit-taking institutions if those risks materialize.

    This, in turn, argues for the central bank to have access to internal risk management information (risk amounts derived from models, risk management techniques, etc.) of individual institutions. In practice, however, timely access to the risk management information of individual institutions is not always possible. Therefore, it is also necessary to assess risk amounts based on such crude information as classified assets or unrealized gains/losses on securities holdings and macro stress testing. In addition to direct information regarding risk amounts, market indicators such as individual stock prices, corporate bond spreads, and credit ratings are also useful for suggesting future managerial problems.

Indicators Suggesting How Far a Financial Crisis Has Progressed

  • How far have the managerial problems of individual deposit-taking institutions developed? According to our experience, banks change their funding behavior as a managerial crisis develops. It would therefore be meaningful to follow changes in the gap between core deposits and core investments (loans/securities), the volatility of the gap, maturity and spreads of borrowed funds, deposits by size, etc.

  • Will problems caused by the bursting of a bubble be confined to the level of individual institutions, or spread to the financial system as a whole? The condition of individual institutions is not the only concern of the central bank. Being responsible for the overall stability of the financial system, the bank cannot be indifferent to the possibility that problems at individual institutions may give rise to systemic risk.

    In order to assess the probability of a systemic problem arising, crude information (classified asset amounts, unrealized gains/losses on securities holdings, etc.) needs to be aggregated, macro stress testing needs to be performed based on such aggregated information, and the solvency distribution of individual institutions needs to be verified. Other useful indicators would include changes in bank stock prices in general, the rating distribution of individual banks.

  • If managerial problems caused by the bursting of a bubble spread to the financial system as a whole, would the adverse impact be so extensive as to affect the real economy through a diminishing financial intermediary function? When assessing repercussions on the real economy, a useful indicator is the diffusion index relating to the lending attitude of deposit-taking institutions as judged by business enterprises

However, the DI (diffusion index) may also reflect conditions of business enterprises, as well as conditions of deposit-taking institutions. To differentiate these two factors, it is necessary to carefully observe the business enterprise side factors and the real economy.

Although the indicators mentioned earlier are useful in making accurate judgments on the status quo, a more important task is to take timely and effective policy measures based on such judgments, to prevent risks within the financial system from accumulating to an unacceptable level, and to minimize disruption to the system. For this to be possible, our experience argues that the roles of the central bank and other authorities responsible for financial system stability (in Japan, these include the Deposit Insurance Corporation, the Financial Service Agency, and the Ministry of Finance) should be clearly defined beforehand. Moreover, a framework should be established whereby necessary information flows to the relevant authorities without delay, so that joint operation by the authorities can be smoothly effected once a financial crisis is triggered.

Japan’s experience suggests that one possible policy measure to minimize disruption to the financial system when a systemic crisis is a threat is the extension of lender-of-last-resort loans either to viable institutions or to wider markets. Institutions whose viability is of serious concern could be subject to prompt corrective action, or early resolution. When a crisis has a potentially significant adverse impact on the financial system as a whole or on the real economy, comprehensive measures might be taken with a view to prevent systemic risk from materializing, such as capital injection, or coverage of all losses incurred by failed institutions. In fact, these measures are incorporated in the new law governing Japan’s deposit insurance system.


In order to ensure financial system stability, it is not enough to examine the probability of individual failures and carry out prompt corrective action or early resolution. More important, an estimation should be made of the probability of systemic risk materializing, the potential adverse impact on the financial intermediary function, etc., and preemptive measures taken. From this viewpoint, it is crucial to make a comprehensive judgment of the overall condition of the financial system by adequately detecting and analyzing behavioral changes at individual institutions, and constantly and carefully verifying a variety of market information, as well as monetary and economic developments influencing those markets.

The central bank, which has dealings with individual deposit-taking institutions and faces the markets on a daily basis, is in the best position to monitor early warning indicators detectable in the behavior of institutions and to carry out timely and adequate policy measures in the interest of financial system stability.

As described in this paper, it has been seen that deposit-taking institutions’ funding behavior changes as a crisis deepens. Against this background, the BOJ carefully monitors the liquidity position of individual deposit-taking institutions on an ongoing basis, and this was an important prerequisite for the BOJ to adequately carry out its policy and provide liquidity in a critical situation. Moreover, the central bank is in a position to make an adequate judgment on the interaction between financial system stability and the real economy, since it is endowed with abundant tools for economic analysis. When analyzing and judging how the real economy will be affected by the possible impairment of financial system soundness, for example, it is indispensable to take into account interaction between financial and economic activities. In this regard, the central bank will continue to play an important role in the area of prudential policies.

To date, various studies have been made with respect to early warning indicators that can help determine the probability of individual failures. However, there are few studies regarding indicators that might suggest the probability of systemic risk materializing. With respect to the latter, the BOJ is of the view, at least at this moment, that finding one unique indicator that can warn of a financial crisis at an early stage is difficult. After all, it is necessary to make a judgment in a comprehensive manner, carefully analyzing different indicators suggesting behavioral changes at deposit-taking institutions, and taking into consideration the general climate of the real economy at the same time. The BOJ, which has gone through a financial crisis of an internationally exceptional magnitude, will thoroughly digest its experience and continue its efforts to identify methods and indicators that enable potential financial crises to be detected at an early stage.

Appendix: Quantitative Analysis of “Market View”—EDP Estimated from Stock Prices

One of the tools used to make a quantitative analysis of the “market view” of the credibility of an individual financial institution is EDP (expected default probability) estimated from stock prices.14

EDP is the probability of a firm defaulting, assuming that the preconditions below are met. The default of a firm is defined as a condition when the market value of assets (gross) is less than the market value of liabilities (gross) (= insolvent).


  • The stock market reflects franchise value correctly, and the market value of stocks (gross) can be treated as the market value of assets (net).

  • As for the volatility of the future market value of assets, the market value of assets follows a random walk (more precisely, the Brownian motion process) based on a particular growth rate (= growth trend of assets; see Figure 13.A1).

Figure 13.A1.Expected Default Probability

Given that EDP of individual financial institutions is estimated daily, observing changes in EDP—for example, when a large financial institution fails or before and after a particular measure is taken (i.e., capital injection with using public funds)—would help analyze how these events have affected the credibility of an individual financial institution.


The views expressed in this paper are those of the authors and do not necessarily reflect the views of the bank of Japan. The authors are particularly grateful to Ms. Keiko Sumida for finalizing the paper.

As a result, term interest rates and corporate bonds in the short-term money market jumped, and the Japan premium went up to 100 basis points.

Yamaichi Securities is not eligible under the Deposit Insurance System. Hence there is a possibility that the special loan extended to Yamaichi Securities after 1997 would not be recollected in full as Yamaichi has failed with a tremendous number of debts. However, the Minister of Finance said in a statement of November 24, 1997: “We intend to take appropriate measures to strengthen the framework (of the Compensation Fund for Deposited Securities) to cope with the situation arising in the course of the failure of securities companies, including the ultimate resolution of this particular case. These measures will include the provision of a legal basis for increasing the financial resources of the Fund and strengthening its functions.”.

The credit crunch was not as bad as the one that occurred in 1997, but the Japan premium went up to 70 basis points.

In March 1996 the DIC had only 16 staff and ¥380 billion in funds, but after two years of financial crisis management, it had 2,300 employees and ¥60 trillion in funds.

In January 1999, the Financial Reconstruction Commission released “Viewpoint on the Write-offs and Allowances in Association with the Capital Injection.” According to this, the percentages of loss-provisioning should be: (1) around 70 percent of “in danger of bankruptcy” loans that are neither guaranteed nor collateralized; (2) around 15 percent of “special attention” loans that are neither guaranteed nor collateralized; and (3) for other “need attention” loans, an appropriate provisioning ratio based on historical loss calculated according to the average remaining maturity.

As a result, the ratio of nonperforming loans to nominal GDP in Japan was 6.0 percent (September 1999) at its peak, compared to 2.9 percent in the U.S. at its peak (June 1991).

Other than the above, there are some banks that overestimated future tax deductible amounts concerning taxable loan-loss provisioning (= overevaluation of future revenue).

The total disposal cost of nonperforming loans (NPLs) is ¥67.9 trillion (from March 1992 to September 1999), which includes (1) the amount of cumulative write-offs and loan-loss provisioning for the NPLs of viable deposit-taking institutions and (2) costs for the disposal of failed banks.

Euro area banks and mutual funds.

Cost of cross-shareholdings = (cost of funds - rate of earnings on stocks) × average outstanding shares.

During 1998, the ratio of overnight call money to all call money (city banks) was up to 60 percent.

These stages are not necessarily sequential. There are cases where some stages proceeded simultaneously.

Besides stock prices, LIBOR, and bond spreads are also useful figures when estimating EDP. In this case, stock prices are used for three reasons: (1) stock prices reflect various market participants’ views; (2) more firms are involved; and (3) data are easy to obtain.

    Other Resources Citing This Publication