Shelagh heffernan modern banking phần 6 ppsx

[ 344 ]
M ODERN B ANKING
level.
90
A rescheduling variable (RESC), the dependent variable lagged by one year, was
also included as an explanatory variable, and found to be significant at 1%, indicating the
presence of positive serial correlation, Thus, if a country rescheduled one year, it is likelier
to do so the next.
When the model is estimated for the later period, Rivoli and Brewer find its overall
explanatory power falls. The parameter estimates are smaller, so their usefulness as explana-
tory variables is reduced, and the ratio of reserves to imports is no longer significant. The
correct overall prediction rate was found to be 20% higher in the earlier period.
The next step of the investigation involved introducing the political variables. When
added to the earlier model (1980–85), they are found to have little impact on overall
performance. The only political variables found to be significant are the presence and
length of armed conflict. According to the authors, armed conflict will place heavy demands
on government budgets and often require large-sum hard currency expenditures, hence
it could raise the probability of rescheduling. Adding the short and long-term political
variables improved the prediction rate for rescheduling by 18% and 35%, respectively.
Overall, the explanatory power of the economic and political variables was greater for
the early period, 1980–85, compared to the later period. However, by adding the political
variables (a political economic model), the correct rescheduling prediction rate improved by
9% (short-term measures of political instability) and 12% (long-term measures) in the early
period. For the later period (1986–90) the correct prediction rate rose by 18% (short-term
measures) and 35% (long-term measures).
Recall that armed conflict was found to be the significant variable in the current study,
which differs from the authors’ 1990 results, when government instability was found to
affect bankers’ perceptions of a country’s creditworthiness and armed conflict did not. Part
of the findings may be explained by the differences in dates of estimation: the early and late
1980s. Also, the dependent variable was different. More research is needed on how political
factors affect a country’s probability of default.

Balkan (1992) used a probit
91
model of rescheduling to examine the role of political
(in addition to economic) factors in explaining a developing country’s probability of
rescheduling. Two political variables were included in the model. A ‘‘political instability’’
variable is an index which measures the amount of social unrest that occurred in a given
year. The ‘‘democracy’’ variable, reflecting the level of democracy, is measured by an index
which, in turn, is captured by two components of the political system: participation (the
extent to which the executive and legislative branches of government reflect popular will)
and competitiveness (the degree of exclusion of political parties from the system and the
ability of the largest party to dominate national elections). Balkan also included some
standard economic variables in his model, such as the ratios of debt service to exports,
interest payments to exports, and so on. In common with most studies all the explanatory
variables were lagged by one year to minimise simultaneity problems. The sample period
ran from 1970 to 1984 and used annual data from 33 developing nations. Balkan found
90
The lower the ratio, the smaller the amount of external debt being repaid, which means interest accruals will
be building up.
91
Probit differs from logit in that it assumes the error terms follow a normal distribution, whereas in logit the
cumulative distribution of the error term is logistic.
[ 345 ]
B ANKING IN E MERGING E CONOMIES
the democracy variable was significantly negative: the probability of rescheduling fell as
democracy levels rose. The probability of rescheduling rose with the level of political
instability. The number of type I and type II errors fell when the political variables were
included in the model.
6.5.5. Rescheduling and Debt Conversion Schemes
Once a country has defaulted on sovereign debt, banks can hardly foreclose on the loans, put
the country into receivership or insist on collateral – some of the nation’s assets. Though

‘‘gunboat diplomacy’’ was not unheard of as a means of putting pressure on a sovereign
state in earlier centuries, it has not been considered an acceptable way of resolving such
matters for many years. Since the Mexican announcement in August 1982, many indebted
countries have entered into or completed renegotiations for the repayment of their loans.
The International Monetary Fund (and, to a lesser extent, the World Bank) plays a critical
intermediary role. Rescheduling agreements share a number of features in common:
ž
The agreement is reached between the debtor, the borrowing bank and the IMF. It
typically involves rescheduling the total value of the outstanding external debt, with the
debt repayment postponed.
ž
Bridging loans often feature, as does an IMF guarantee of interbank and trade facilities,
sometimes suspended when a country announced that it was unable to service its
external debt.
ž
The private banks normally agree to provide ‘‘new money’’ to allow the debtor country
to keep up interest payments, raising the total amount of the outstanding debt. The IMF
usually insists on increased exposure by the banks in exchange for IMF loans.
ž
The debtor country is required to implement an IMF macroeconomic adjustment
programme, which will vary according to the economic problems the country faces.
Governments are required to remove subsidies that distort domestic markets, meet strict
inflation and budget deficit targets, and reduce trade barriers. Note the country loses some
‘‘macroeconomic sovereignty’’ because its government is now limited in its choice of
economic policy. Thus, ex post, it can be argued that sovereign borrowing exposed these
countries to high interference costs.
Debt–equity swaps are another means of dealing with a sovereign debt problem, usually
as part of or to complement an IMF rescheduling package. A debt–equity swap involves
the sale of the debt by a bank to a corporation at the debt’s secondary market price. The
corporation exchanges the debt for domestic currency through the central bank of the

emerging market, usually at a preferential exchange rate. It is used to purchase equity in a
domestic firm. It has proved unpopular with some countries because it can be inflationary,
and the country loses some microeconomic sovereignty. Similar debt conversion schemes
in the private sector have allowed firms to reduce their external debt obligations.
Other types of swaps include debt–currency swaps, where foreign currency denominated
debt is exchanged for the local currency debt of the debtor government, thereby increasing
the domestic currency debt. A debt–debt swap consists of the exchange of LDC debt by
[ 346 ]
M ODERN B ANKING
one bank for the debt of another LDC by another bank. Debt–trade swaps grew between
emerging markets as a means of settling debt obligations between them. They are a form
of counter-trade because the borrower gives the lending country (or firm) home-produced
commodities. Alternatively, a country agrees to buy imports in exchange for the seller
agreeing to buy some of the country’s external debt on the secondary markets.
Debt–bond swaps or ‘‘exit’’ bonds allow lenders to swap the original loan for long-term
fixed rate bonds, reducing the debtor’s exposure to interest rate risk. In a period of sustained
rising interest rates, the fixed rate bonds will lower debt servicing costs for the borrowing
country. The Mexican restructuring agreement of March 1990 was an early example of the
new options offered to lenders. In addition to the option of injecting new money, banks
could participate in two debt reduction schemes; either an exchange of discount bonds
against outstanding debt or a par bond, that is, an exchange of bonds against outstanding
debt without any discount, but with a fixed rate of interest (6.25%). The bonds are to
be repaid in full in 2019 and the principal is secured by US Treasury zero-coupon bonds.
Participating banks can also take part in a debt–equity swap programme linked to the
privatisation of state firms–13% opted for the new money, 40% the discount bond (at 65%
of par) and 47% the par bond.
Exit bonds are now known as Brady bonds, because they were an integral part of
the Brady Plan introduced in 1989. This plan superseded the earlier Baker Plan (1985),
which had identified the ‘‘Baker 15’’, the most heavily indebted LDCs, as the key focus
of action.

92
The Baker Plan also called for improved collaboration between the IMF and
the World Bank, stressed the importance of IMF stabilisation policies to promote growth,
and encouraged private commercial lenders to increase their exposure. The Brady Plan
reiterated the Baker Plan but explicitly acknowledged the need for banks to reduce their
sovereign debt exposure. The IMF and World Bank were asked to encourage debt reduction
schemes, either by guaranteeing interest payments on exit bonds or by providing new loans.
The plan called for a change in regulations (e.g. tax rules) to increase the incentive of the
privatebankstowriteoffthedebt.
Brady bonds are now a common part of loan rescheduling, and very simply, are a means by
which banks can exchange dollar loans for dollar bonds. These bonds have a longer maturity
(10 to 30 years) and lower interest (coupon) payment than the loan they replace – the
interest rate can be fixed, floating or step. They can include warrants for raw materials of
the country of issue and other options. The borrowing country normally backs the principal
with US Treasury bonds, which the bond holders get if the country defaults. However, as
the Mexican case in 1995–96 illustrates all too well, outright defaults are rare. As of 2001,
about $300 billion worth of debt had been converted into Brady bonds.
In 1996, the first sovereign bonds were issued by governments of emerging market
countries after their economic conditions improved. Essentially this involves buying back
Brady bonds: they are either repurchased or swapped for sovereign bonds. A secondary
market for trading emerging market debt including Brady and sovereign bonds emerged in
the mid-1980s. Most of it is traded between the well-known commercial and investment
92
Both Richard Baker and Tom Brady were Treasury Secretaries in the 1980s. They played no formal role in
resolving emerging market debt problems but their ideas were influential.
[ 347 ]
B ANKING IN E MERGING E CONOMIES
banks based in London and New York, as well as hedge funds and other institutional
investors. The market allows banks to move the assets off their balance sheets, and for those
with continuing exposure, it is possible to price these assets.

6.6. Conclusion
This chapter focused on several areas of banking in emerging market countries. The
objective was to provide the reader with an insight into several key issues: attempts to
resolve problems arising from financial repression through reform of banking systems,
sovereign and political risk analysis, and a review of Islamic banking.
Until the last decade of the 20th century, almost half of the world’s population lived
under communism. Communist regimes had extinguished the conventional, private sector,
independent commercial bank. In country after country, throughout the former Soviet
Union and its Warsaw Pact allies in Central and Eastern Europe, the 1990s were to see
private banks return. China, still led by the Communist Party, also underwent profound
financial changes as part of a broader economic reform, starting as early as 1979 and
gathering pace in the new millennium. In India, the world’s largest democracy where
banking has been subject to a high degree of state control, regulation and ownership, some
cautious steps have been taken in the same direction as Russia and China.
If the demise (or reinterpretation) of communism and socialism has been a great victory
for the concept of the conventional western bank, the later 20th century saw two other
developments that posed it challenges. One of these is the growing perception in many
Muslim countries – and beyond – that the whole basis of the conventional western bank’s
operations, lending at interest, is inconsistent with religious principles. The other was the
periodic but serious issue of how western banks should respond to many emerging market
governments that could not or would not service or repay the debt owed to them: the
problem of non-performing sovereign loans.
This chapter has chronicled these massive emerging market changes and their effects on
the global financial landscape, especially banking. It began by analysing the phenomenon of
financial repression, and exploring the question, a pressing policy issue for many countries,
of whether foreign banks should be allowed to operate within their borders.
Next came a survey of financial systems of Russia, China and India. Each are classic, but
different, examples of financial repression in the late 1990s. The key question was whether
the reforms they introduced were enough to alleviate some of the more serious problems
arising from financial repression. All three countries have enjoyed some degree of success.

Though Russia experienced the economic equivalent of a roller coaster ride, it has gone
the furthest in terms of financial liberalisation, followed by China, provided it lives up to
its promises to allow foreign bank entry by 2007 and liberalises interest rates. India is the
laggard here, with no clean plans to reduce state control of the banking sector, though
other parts of its financial sector have been liberalised.
However, these countries are also experiencing a common problem: the difficulty each
government faces in reducing or eliminating state ownership and control of banks. There
is nothing wrong with state ownership per se, provided banks are free from government
interference, have no special privileges which give them an unfair advantage, and have to
[ 348 ]
M ODERN B ANKING
compete with private banks. Though India has allowed some private banks to open, it has
signalled its intention to maintain a strong state presence in banking with very few policies
aimed at encouraging the development of a vibrant competitive private sector able to take
on the state banks.
Russia has made the most advances – the banking sector was part of its privatisation
programme. Unfortunately, the outcome is unappealing: bank oligarchies which confine
their banking activities to the industrial group they serve/own, and state owned banks
that will be difficult to privatise until something is done about the high percentage of
non-performing loans. The state owned Sberbank has special privileges, which give it
advantages over any potential rivals, especially in retail banking. If it is true that about 30
of the thousand plus banks are financially viable, Russia faces a future of unstable banking
periods, which will do nothing to build up trust in the system, a key ingredient for a
successful banking sector. One ray of light is the absence of restrictions on foreign bank
entry, though efficient foreign entrants could imperil short-term financial stability if they
are a contributing factor to the speedy collapse of costly, inefficient local banks.
In China, scratch the surface of a joint stock bank and state control of the bank is quickly
revealed. Most of China’s banks have a high percentage of non-performing loans, and the
overt political interference by local governments in the credit coops is a symptom of a
more serious dilemma. How can the central bank and regulatory body teach Chinese banks

the rudiments of risk management when policy objectives (be they local, provincial or
national) interfere with lending decisions? If unfettered foreign bank entry is allowed from
2007, their presence will contribute to the development of a more efficient banking system,
but the price could be high in the short run as domestic banks are forced out of business. It
is likely a one-party Communist State will intervene if Chinese banks face closure, which
could threaten some of the market oriented policies it has introduced.
Developing economies and emerging markets in Eastern Europe share similar structural
problems, including inadequate monitoring by supervisory authorities and poorly trained
staff, which have compounded general problems with credit analysis, questionable account-
ing procedures and relatively high operating costs. Many developing countries exhibit signs
of financial distress; the problem for banks in the emerging East European markets is debt
overhang from the former state owned enterprises. A stable financial structure is some way
off for Russia, but many of the transition economies that adopted a gradualist approach to
financial reform have created workable and stable structures. These banks were found to be
much closer to an efficiency frontier compared to Russian banks.
Most Islamic banking is located in emerging markets, and this is the reason it has been
discussed in this chapter. Its growth is important because it relies on a financial system
where the payment of interest is largely absent. The development of financial products
that conform to Shariah law has been impressive, and in this respect, Islamic banking
has something to teach the conventional banking system. Nonetheless, there are several
problems that need to be addressed, such as the tendency to concentrate on one or two
products, which discourages diversification. Also, although moderated in some ways, the
potential for moral hazard remains. Finally, the regulatory authorities and banks need to
work together to come up with an acceptable framework to deal with the unique aspects of
regulation associated with Islamic banking.
[ 349 ]
B ANKING IN E MERGING E CONOMIES
The underdeveloped financial systems of emerging market economies make them depen-
dent on external finance. Bankers face unique problems when it comes to the management
of sovereign risk. The key lesson is that bank managers must assess properly the risks asso-

ciated with any given set of assets. The sovereign lending boom of the 1970s demonstrated
how bankers failed to acknowledge that illiquidity can be as serious as insolvency if there is
no collateral attached to the loan, and this in turn led to poor assessment of sovereign risk.
Since then practitioners and academics have worked to improve sovereign risk assessment
by identifying the significant variables contributing to the probability of default and devel-
oping early warning systems. The protracted process of IMF led rescheduling agreements
rather than a swift bailout of the indebted countries (and associated private banks) serves
as a lesson to bankers and developing countries that even though a sovereign default may
undermine the stability of the world financial system, official assistance comes at a steep
price. The more recent episodes of sovereign debt repayment problems underline these
lessons, though political factors appear to have become more important.
Suspension of interest flows on sovereign loans all too often reflects an economic/financial
crisis in the country that does it. It also triggers difficulties for the creditor banks overseas.
Financial crises were, however, omitted from this chapter. They can affect any country,
developed or emerging, and are the subject of Chapter 8, while Chapter 7 explores the
causes of individual bank failure.

B ANK F AILURES
7
7.1. Introduction
Bank managers, investors, policy makers and regulators share a keen interest in know-
ing what causes banks to fail and in being able to predict which banks will get into
difficulty. Managers often lose their jobs if their bank fails. The issue is also impor-
tant for policy because failing banks may prove costly for the taxpayer; depositors and
investors want to be able to identify potentially weak banks. In this chapter, the rea-
sons why banks fail are explored, using both a qualitative approach and quantitative
analysis. Since bank failures often lead to financial crises, the chapter also looks at
their causes, undertaking a detailed examination of the South East Asian and Japanese
financial crises.
After defining bank failure, section 7.3 discusses a range of key bank failures, from the

collapse of Overend Gurney in 1866 to the well-publicised failures such as the Bank of
Credit and Commerce International and Barings bank over a century later. Based on these
case studies, some qualitative lessons on the causes of bank failure are drawn in section 7.4.
A review of econometric studies on bank failure is found in section 7.5, most of which use
a logit model to identify the significant variables that increase or reduce the probability
of a bank failing. The quantitative results are compared to the qualitative contributors.
Section 7.6 concludes.
7.2. Bank Failure – Definitions
Normally, the failure of a profit-maximising firm is defined as the point of insolvency,
where the company’s liabilities exceed its assets, and its net worth turns negative. Unlike
certain countries that default of their debt, some banks do fail and are liquidated. Recall
from the discussion in Chapter 5 that the USA, with its prompt corrective action and
least-cost approach, has a well-prescribed procedure in law for closing and liquidating
failed banks. In other countries, notably Japan (though it is attempting to move towards
a US-type approach) and some European states, relatively few insolvent banks have
been closed in the post-war period, because of real or imagined concerns about the
systemic aspects of bank failure. Thus, for reasons which will become apparent, most
practitioners and policy makers adopt a broader definition of bank failure:abankis
deemed to have ‘‘failed’’ if it is liquidated, merged with a healthy bank (or purchased and
[ 352 ]
M ODERN B ANKING
acquired
1
) under central government supervision/pressure, or rescued with state financial
support.
There is a widerange of opinions about this definition. Some think a failing bankshould be
treated the same way as a failing firm in any other industry. Others claim that failure justifies
government protection of the banking system, perhaps in the form of a 100% safety net,
because of its potential for devastating systemic effects on an economy. In between is support
for varying degrees of intervention, including deposit insurance, a policy of ambiguity as to

which bank should be rescued, merging failing and healthy banks, and so on.
The debate among academics is reflected in the different government policies around the
world. The authorities in Japan (until very recently) and some European states subscribe to
the view that virtually every problem bank should be bailed out, or merged with a healthy
bank. In Britain, the tradition has been a policy of ambiguity but most observers agree the
top four or five commercial banks
2
and all but the smallest banks would be bailed out.
The United States has, in the past, tended to confine rescues to the largest commercial
banks. However, since 1991, legislation
3
has required the authorities to adopt a ‘‘least cost’’
approach (from the standpoint of the taxpayer) to resolve bank failures, which should mean
most troubled banks will be closed, unless a healthy bank is willing to engage in a takeover,
including taking on the bad loan portfolio or any other problem that got the bank into
trouble in the first place.
There are three ways regulators can deal with the problem of failing banks.
1. Put the bank in receivership and liquidate it. Insured depositors are paid off, and assets
sold. This approach is most frequent in the USA, but even there, as will be observed,
some banks have been bailed out.
2. Merge a failing bank with a healthy bank. The healthy bank is often given incentives,
the most common being allowing it to purchase the bank without the bad assets. Often
this involves the creation of an agency which acquires the bad assets, then attempts
to sell them off. See the ‘‘good bank/bad bank’’ discussion in Box 8.1 of Chapter 8. A
similar type of takeover has emerged in recent years, known as purchase and acquisition
(P&A). Under P&A, assets are purchased and liabilities are assumed by the acquirer.
Often a state or state-run resolution pays the difference between assets or liabilities. If
the P&A is partial, uninsured creditors will lose out.
3. Government intervention, ranging from emergence of lending assistance, guarantees for
claims on bad assets or even nationalisation of the bank.

These different forms of intervention are discussed in more detail below.
The question of what causes failure will always be of interest to investors, unprotected
depositors and the bank employees who lose their jobs. However, if the state intervention
school of thought prevails, then identifying the determinants of bank failure is of added
1
See Table 7.1.
2
The big four: HSBC, Barclays, Royal Bank of Scotland and HBOS. Lloyds-TSB has been in fifth place (measured
by asset size and tier 1 capital) since the merger of the Halifax and Bank of Scotland – HBOS.
3
This rule is part of the Federal Deposit Insurance Corporation Improvement Act, 1991 – see Chapter 4 for
more detail.
[ 353 ]
B ANK F AILURES
importance because public funds are being used to single out banks for special regulation, bail
out/merge banks and protect depositors. For example, the rescue of failing American thrifts
in the 1980s is estimated to have cost the US taxpayer between $250 and $300 billion.
In Japan, one reason the authorities shied away from early intervention was a hostile
taxpaying public. However, the use of public funds to inject capital into weak banks and
nationalise others, together with the need to extend ‘‘temporary’’ 100% deposit insurance
well past its ‘‘end by’’ date, had raised the cost of the bailout to an estimated 70 trillion yen
($560 billion).
4
Posen (2002) estimates the direct cost of the Japanese bailout to the year
2001 to be 15% of a year’s GDP, compared to just 3% for the US saving and loan debacle.
7.2.1. How to Deal with Failed Banks: The Controversies
Most academics, politicians (representing the taxpayer), depositors, and investors accept the
idea that the banking sector is different. Banks play such a critical role in the economy that
they need to be singled out for more intense regulation than other sectors. The presence
of asymmetric information is at the heart of the problem. A bank’s managers, owners,

customers, regulators and investors have different sets of information about its financial
health. Small depositors are the least likely to have information and for this reason, they
are usually covered by a deposit insurance scheme, creating a moral hazard problem (see pp.
6–7 for more detail). Regulators have another information set, based on their examinations,
and investors will scrutinise external audits.
Managers of a bank have more information about its financial health than depositors,
regulators, shareholders or auditors. The well-known principal agent problem arises because
of the information wedge between managers and shareholders. Once shareholders delegate
the running of a firm to managers, they have some discretion to act in their own interest
rather than the owners’. Bank profits depend partly on what managers do, but also on other
factors unseen by the owners. Under these conditions, the best managerial contracts owners
can devise will lead to various types of inefficiency, and could even tempt managers into
taking on too much risky business, either on- or off-balance sheet.
However, these types of agency problems can arise in any industry. The difference in the
banking sector, it is argued, is that asymmetric information, agency problems and moral
hazard, taken together, can be responsible for the collapse of the financial system, a massive
negative externality. Though covered in depth in Chapter 4, it is worth summarising how a
bank run might commence – recall a core banking function, intermediation. Put simply,
banks pay interest on deposits and lend the funds to borrowers, charging a higher rate of
interest to include administration costs, a risk premium and a profit margin for the bank.
All banks maintain a liquidity ratio, the ratio of liquid assets to total assets, meaning only a
fraction of deposits is available to be paid out to customers at any point in time. However,
there is a gap between socially optimal liquidity from a safety standpoint, and the ratio a
profit-maximising bank will choose.
5
4
Sources: ‘‘Notes’’, The Financial Regulator, 4, 2000, p. 8 and The Banker, January 1999, p. 10.
5
For example, in the UK, mutually owned building societies maintain quite high liquidity ratios, in the order of
about 15–20%, compared to around 10% for profit oriented banks.

[ 354 ]
M ODERN B ANKING
Given that banks, even healthy ones, only have a fraction of their deposits available at
any one time, an unexpected sudden surge in the withdrawal of deposits will mean they soon
run out of money in the branches. Asymmetric information means rumours (ill-founded or
not) of financial difficulties at a bank will result in uninsured depositors withdrawing their
deposits, and investors selling their stock. Contagion arises when healthy banks become the
target of runs, because depositors and investors, in the absence of information to distinguish
between healthy and weak banks, rush to liquidity.
Governments may also impose a reserve ratio, requiring banks to place a fraction of
non-interest earning deposits at the central bank. This is, effectively, a tax on banking
activity. The amount paid is the interest foregone multiplied by the volume of funds held as
reserves. Since the nominal rate of interest incorporates inflation expectations, the reserve
ratio is often loosely thought of as a source of inflation tax revenue. In recent decades many
western governments have reduced or eliminated this reserve ratio. For example, in the
UK, the reserve ratio has been reduced from a cash ratio of 8% before 1971 to one of just
0.4% today. In developing and emerging markets, the reserve ratio imposed on banks can
be as high as 20%, as an inexpensive form of government revenue.
Asking all banks to set aside capital as a percentage of their assets (capital assets ratio) or
risk weighted assets (the ratio of capital to weighted risk assets) is now the preferred method
for ensuring banks have a cushion against shocks to credit, market and operational risks
which could threaten the viability of the bank. The Basel 1 and 2 agreements are examples
of an application of this approach. They were discussed at length in Chapter 4 and are
noted here to illustrate the role they play in averting bank failures and crises.
In the absence of intervention by the central bank to provide the liquidity necessary to
meet the depositors’ demands, the bank’s liquidity problem (unable to meet its liabilities
as they fall due) can turn into one of insolvency, or negative net worth. Normally, if the
central bank and/or other regulators believe that but for the liquidity problem, the bank is
sound, it will intervene, providing the necessary liquidity (at a penalty rate) to keep the
bank afloat. Once depositors are satisfied they can get their money, the panic subsides and

the bank run is stopped. However, if the regulators decide the bank is insolvent and should
not be rescued, the run on deposits continues, and it is forced to close its doors.
If the authorities do intervene, but fail to convince depositors the problem is confined to
the one bank, contagion results in systemic problems affecting other banks, and perhaps all
banks, putting the sector in danger of collapsing. In the extreme, the corresponding loss of
intermediation and the payments system could reduce the country to a barter economy. A
‘‘bank holiday’’ may be declared in an effort to stop the run on banks, using the time to meet
with the stricken banks and decide how to curb the withdrawals, usually by an agreement
to supply unlimited liquidity to solvent banks when their doors open after the holiday.
If the bank holiday agreement fails to reassure depositors, or no agreement is reached,
the outcome is a classic negative externality because what began as a run on one bank (or a
few small banks) can lead to the collapse of the country’s financial system. The economic
well-being of all the agents in an economy has been adversely affected by the actions of
less than perfectly informed depositors and investors, who, with or without good reason,
decided their bank was in trouble and sought to get their funds. The negative externality is
[ 355 ]
B ANK F AILURES
a type of market failure, as is the presence of asymmetric information. Market failure is a
classic argument for a sector to be singled out for government intervention and regulation.
The evidence on whether bank contagion is a serious matter is mixed and controversial.
In a comprehensive paper, Kaufman (1994) reviews a number of contagion theories, and
related evidence. His main findings are as follows.
ž
Bank contagion spreads faster in the banking sector compared to other sectors. Based on
US studies, compared to non-banks, there is evidence that contagion appears faster and
spreads to a larger proportion of the sector.
ž
Bank contagion is both industry and/or firm-specific because compared to other industries,
depositors tendto be less well informed about the performance oftheir bankor the banking
sector. Kaufman’s survey of the evidence suggests that bank contagion and bank runs

are largely firm-specific and rational, that is, depositors and investors can differentiate
between healthy and unhealthy banks. The costs of failure also appear to be lower.
ž
Bank contagion results in a larger number of failures: compared to other sectors, contagion
does cause a larger percentage of failures.
ž
Contagion results in larger losses for depositors, but the evidence suggests the losses are
smaller than losses to creditors in other sectors. For example, during the 1980s when
there were a large number of thrift and bank failures in the USA, the solution of the
problem was more efficient when compared to the closure of insolvent non-bank firms
through bankruptcy procedures.
ž
Kaufman’s review finds little evidence to support the view that runs on banks cause
insolvency among solvent banks; nor does it spread to other parts of the financial sector
or the rest of the macroeconomy.
These findings might lead the reader to conclude that an inordinate amount of resources
may be directed to the protection of a sector that does not need it. However, it is worth
stressing that most of Kaufman’s evidence comes from the United States, which has one of
the most generous deposit insurance schemes in the world,
6
about which customers are well
informed. But even with deposit insurance, some failing US banks are rescued, and ‘‘too
big to fail’’ policies often apply. Such an environment naturally instils confidence among
depositors and creditors, which reduces the likelihood of contagion and makes it difficult to
quantify its effects.
Close supervision by regulators, and perhaps intervention too, contributes to managerial
incentives to gamble. Senior management is normally the first to recognise their bank is,
or will be, in serious trouble. They have the option of taking no action, letting it fail, and
losing their jobs. However, if they are the only ones with information on the true state
of the bank, downside risk is truncated. If a gamble fails, the bank fails with a larger net

loss, but bad as this event is for managers, its marginal effect on them is zero. If a gamble
succeeds, the bank, and their jobs, are saved. Returns are convexified, encouraging gambling
to increase their survival probability and resurrect the bank. Thus, they will undertake
highly risky investments, even with negative expected returns. Likewise, ‘‘looting’’ (defined
6
The USA was the first country to introduce deposit insurance, in 1933.
[ 356 ]
M ODERN B ANKING
in Chapter 4) may be seen as a way of saving the bank, and if not, providing a comfortable
payoff for the unemployed managers. Given the presence of contagion in the sector, such
behaviour should be guarded against through effective monitoring.
Lack of competition will arise in a highly concentrated banking sector and can also be a
source of market failure. Depositors are paid less interest and borrowers are charged more
than marginal operating costs could justify. However, anti-competitive behaviour occurs
in other sectors, and is usually monitored by official bodies with the power to act, should
the behaviour of a firm (or firms) be deemed insufficiently competitive. Chapter 9 e xplores
competitive issues in banking.
To summarise, the banking system needs to be more closely regulated than other markets
in the economy because of market failure, which can be caused by asymmetric information
and negative externalities. The special regulation can take a number of forms, including
deposit insurance (funded by bank premiums being set aside in an insurance fund), capital
requirements (e.g. Basel 1 and Basel 2), the licensing and regular examination of banks,
intervention by the authorities at an early stage of a problem bank, and lifeboat rescues.
In fact, the transition from the failure of an individual bank to the complete collapse of a
country’s banking/financial system is rare. The US (1930–33) and British (1866) cases have
already been discussed in some detail. Proponents of special regulation of banks, and timely
intervention if a bank or banks encounter difficulties, would argue that the presence of
strict regulation of the banking sector has prevented any serious threats to financial systems
of the developed economies (with the arguable exception of Scandinavia and Japan – see
Chapter 8). However, there have been frequent systemic crises in developing and emerging

market economies. As will be seen in Chapter 8, contagion was responsible for the spread of
the threat of financial crises from Thailand to Korea, Indonesia, Malaysia, the Philippines
and beyond, to Russia and Brazil.
Unfortunately, there is a downside to the regulation of banks. The key problem is one
of moral hazard, defined and discussed in Chapters 1 and 4. In banking, moral hazard
arises in the presence of deposit insurance and/or if a central bank provides liquidity to a
bank in difficulties. If a deposit is backed by insurance, then the depositor is unlikely to
withdraw the deposit if there is some question raised about the health of the bank. Hence,
bank runs are less likely, effectively putting an end to the possibility of systemic failure of
the banking system. Blanket (100%) coverage of depositors (in some cases, creditors too)
is often deemed necessary to stop bank runs. However, deposit insurance is costly, and
normally governments limit its coverage to the retail depositor, on the grounds that this
group lack the resources to be fully informed about the health of a bank.
Restricted forms of deposit insurance do not eliminate the possibility of bank runs
because wholesale depositors and others (e.g. non-residents, or those holding funds in
foreign currencies) are usually excluded. For example, in 1998 the Japanese authorities had
to extend the insurance to 100% coverage of most deposits because of its persistent banking
problems, which reduced depositor confidence and caused runs (see Chapter 8 for more
detail). The next section gives the background to the rescue of Continental Illinois Bank
in 1984. This bank was heavily dependent on the interbank markets, and suffered from a
withdrawal of funding by uninsured wholesale depositors.
[ 357 ]
B ANK F AILURES
On the other hand, the more extensive the deposit insurance coverage and/or central
bank intervention, the more pronounced the moral hazard problem because if agents know
a bank will be supported in the event of problems, they have little incentive to monitor the
banks, making it easier for senior management to undertake risks greater than they might
have in the absence of closer scrutiny by their customers. The looting hypothesis, discussed
in Chapter 4, is more likely to be a problem, especially if bank managers who know their
bank is in trouble undertake highly risky investments in an attempt to rescue themselves

from the problem.
Counter-arguments made by proponents of deposit insurance are that even in its presence,
the incentives of senior management are not altered to such a degree that they undertake
riskier investments. Management is still answerable to shareholders, who do have an
incentive to monitor their investments, which are unprotected in the event of failure. Also,
if the bank does fail, it is the managers and employees who lose their livelihoods.
A final argument relates to the fund itself. Normally it is the banks which pay the
insurance premia to fund the deposit insurance. In most countries, all banks pay the same
premium but in the United States, regulators rank banks according to their risk profile,
which is kept confidential. The riskier the bank, the higher the premium paid (see Chapter 5
for more detail). Being answerable to shareholders, linking the deposit insurance premium
to banks’ risk profiles and loss of employment will help to reduce the problem of moral
hazard on the part of bank management.
Another way of dealing with bank runs is for the central bank to supply liquidity to
illiquid banks caught up in bank runs, provided most of the deposits are denominated in
the home currency. This is discussed in more detail in the section on lender of last resort
in Chapter 8. It is raised here for completeness, and because some interesting work has
been done on the fiscal costs of resolving bank crises. Hoggarth et al. (2003) look at the
impact of liquidity support and government guarantees on output losses, controlling for the
degree of bank intermediation in a given country. They find open-ended liquidity support
has a significantly negative effect on output, that is, during a banking crisis, the greater the
liquidity support, the bigger the fall in output. By contrast, the deposit guarantees appear to
have no effect on output. This result is similar to the findings reported by Bordo et al. (2001),
who looked at crises in 29 countries from 1973 to 1997. They suggest open-ended liquidity
support could mean more insolvent banks survive, increasing moral hazard, encouraging
banks to increase risk-taking activities in the hope the gamble is successful, and allow
loss-making agents to continue to borrow.
This section reviewed the controversies related to the methods used to rescue banks in the
event of a banking crisis. Proponents of government intervention are by far the majority,
though some researchers argue that intervention is only justified if the benefits should

exceed any costs. The work by Hoggarth et al. (2003) and Bordo et al. (2001) suggests the
type of intervention is important. Their results indicate deposit guarantees have no impact
on output, whereas liquidity support reduces output. These findings indicate the type of
bank rescue needs to be carefully considered, keeping in mind that while intervention
may have fiscal costs, the absence of any support also has consequences – conceivably,
systemic meltdown.
TEAM FLY
[ 358 ]
M ODERN B ANKING
Table 7.1 Failed Bank Resolution Strategies and Who Loses
Resolution
options
Shareholders –
Lose Money
Creditors –
Lose Money
5
Taxpayers –
Government
Injection
Managers –
Lose Jobs
Employees –
Lose Jobs
Capital injection by
shareholders
Yes – in the
short term,
but could be
made up if

the bank
recovers
No No Yes – likely Possibly if
sharehold-
ers demand
cost cuts
Government injection
1
Likely Likely Yes Likely Likely
M&A – state funded
2
Partly Possibly Yes Yes Likely
M&A – private Likely Likely No Yes Yes
P&A
3
Yes Yes – if
uninsured
Possibly Yes Yes
Bridge
bank/nationalisation
4
Yes Possibly Yes Yes Possibly
Liquidation Yes Yes – if
uninsured
No Yes Yes
Notes:
1
Government injection usually comes with conditions for bank restructuring which are likely to cause managers
and some employees to lose their jobs. Some creditors could lose out in any financial restructuring.
2

Some merger or acquisitions involve the state agreeing to take on the dud assets and/or inject funds.
3
P&A: purchase and acquisition. Assets are purchased and liabilities are assumed by the acquirer. Often the
state/state run resolution pays the difference between assets or liabilities. If the P&A is partial, uninsured creditors
will lose out.
4
The state will take over the bank temporarily (the bridge bank) until a strategy for resolving the bank’s problems
is agreed. The bank is later sold, though it may be several years later. Uninsured creditors may lose out, depending
on the option, unless a government issues a blanket guarantee for depositors and creditors.
Source: Hoggarth et al. (2003), table 1.
Hoggarth et al. (2003) provide a useful table summarising the different options available
for troubled banks, and the trade-offs involved. The table is reproduced here (Table 7.1),
with some adaptations.
At the other extreme are the free bankers, discussed in Chapter 4. Those who support
special regulation of the banking sector expect governments/regulators to use the determi-
nants of bank failure to achieve an optimum where the marginal benefit of regulation/rescue
is equal to its marginal cost. Even free bankers have an interest in what causes a bank to
fail, if only because investors and depositors lose out when a bank goes under. Either way,
it is an important question, and the next few sections attempt to answer it.
7.3. Case Studies on Bank Failure
Bank failures, broadly defined, have occurred in virtually every country throughout history.
In the 14th century the Bardi family of Florentine bankers was ruined by the failure of
[ 359 ]
B ANK F AILURES
Edward III to meet outstanding loan obligations – the only time in history, to date, that
an English government failed to honour its debts. Some failures seriously undermine the
stability of the financial system (as happened, for example, in the UK in 1866 and the USA
in 1933). Others do not. In some cases, state support of problem banks proves costly. For
example, the taxpayers’ bill for the US thrift bailout is put at around $250–$300 billion,
while recent problems with the Japanese banking system has cost the taxpayer about

$560 billion to date. In this section, bank failures are examined on a case by case basis,
the objective being to identify the qualitative causes of bank failure. After a brief historical
review, the main focus is on modern bank failures, commencing with the failure of Bankhaus
Herstatt in 1974.
7.3.1. Historical Overview
This subsection is a selective, brief review of well-known bank failures in Victorian England
and between 1930 and 1933 in the USA. In England, there were two major bank failures
in the 19th century: Overend Gurney and Company Ltd
7
in 1866, and Baring Brothers in
1890. Overend Gurney originated as a discount house but by the 1850s was a prosperous
financial firm, involved in banking and bill broking. After changes in management in 1856
and 1857 it began to take on bills of dubious quality, and lending with poor collateral to
back the loans. By 1865 the firm was reporting losses of £3–£4 million. In 1866, a number
of speculative firms and associated contracting firms, linked to Overend Gurney through
finance bills, failed. London-based depositors began to suspect Overend was bankrupt; the
consequence was a drawing down of deposits and a fall in the firm’s stock market price. On
10 May 1866 the firm sought assistance from the Bank of England, which was refused. The
bank was declared insolvent the same afternoon.
Overend Gurney was a large bank: by balance sheet it was about ten times the size of
the next largest bank in the country – the Midland.
8
Its failure precipitated the collapse of
a number of country banks and firms associated with it. Contagion spread: country banks
withdrew deposits from other London banks and finance houses, which in turn caused a
run on the Bank of England. Several banks and finance houses, both unsound and healthy,
failed. The 1844 Bank Charter Act was suspended to enable the Bank of England to
augment a note supply, which was enough to allow the panic to subside. Overend Gurney
was liquidated, and though the Bank Act was not amended, the episode made it clear that
henceforth the Bank of England was to intervene as lender of last resort in situations of

severe panic.
Baring Brothers was a large international merchant bank which failed in1890. Baringshad
been founded in 1762, largely to finance the textile trade in Europe. After the Napoleonic
wars, Barings began to finance for public projects in foreign countries; initially the long-term
lending to foreign governments was concentrated in Europe and North America, but in
1821–22 the loan portfolio was expanded to include Mexico and Latin America, notably
Chile, Colombia and Brazil. Even though these loans were non-performing, Barings granted
7
Legislation passed in 1858 allowed limited liability and Overend Gurney became a limited liability company
in 1862.
8
Wood (2003), p. 69.
[ 360 ]
M ODERN B ANKING
additional, large loans to the governments of Argentina and Uruguay between 1888 and
1890. By the end of 1890, these loans made up three-quarters of Barings’ total loan portfolio.
Problems with key banks in Argentina and Uruguay led to suspended payments and bank
runs. Barings’ Argentine securities dropped in value by one-third; the firm also faced a
drop in income from loan repayments and liabilities arising from a failed utility. Barings
borrowed heavily from London banks in an effort to contain the problem, but in November
1890 was forced to report the crisis to the Bank of England. The Governor of the Bank of
England organised subscriptions to a fund – London’s key merchant banks contributed, and
the fund guaranteed Barings’ liabilities for three years. Eight days after Barings reported its
problems to the Bank, its illiquidity had become public knowledge. But there was no run
of any significance, and no other banks failed. Though put into liquidation, Barings was
refloated as a limited liability company, with capital from the Baring family and friends.
Both banks underwent notable changes in bank management in the years leading up
to the failures. The collapse of the banks was due largely to mismanagement of assets,
leading to a weak loan portfolio in the case of Barings, and for Gurneys, the issue of poor
quality finance bills. Batchelor (1986, pp. 68–69) argued that, unlike Barings, the Gurney

failure caused a serious bank run because the public lacked crucial information about the
state of the bank’s financial affairs. The Latin American exposure of Barings was well
known, but there was no run because of its historical reputation for financial health in the
banking world.
One of the most important series of bank failures occurred in the USA between 1930
and 1933.
9
The stock market crash of October 1929 precipitated a serious depression and
created a general climate of uncertainty. The first US banking crisis began in November
1930, when 256 banks failed; contagion spread throughout the USA, with 352 more bank
failures in December. The Bank of the USA was the most notable bank failure. It was the
largest commercial bank, measured by deposits. It was a member of the Federal Reserve
System, but an attempt by the Federal Reserve Bank of New York to organise a ‘‘lifeboat’’
rescue with the support of clearing house banks failed. It was followed by a second round of
failures in March 1931.
Other countries also suffered bank failures, largely because the depression in the USA
had wide-reaching global effects. The largest private bank in Austria, Kreditanstalt, failed
in May 1931, and in other European states, particularly Germany, banks were closed.
Meanwhile, in the USA, another relapse followed a temporary recovery, and in the last
quarter of 1932 there were widespread bank failures in the Midwest and Far West of the
USA. By January 1933 bank failures had spread to other areas; by 3 March, half the states
were required to declare bank holidays to halt the withdrawals of deposits. On 6 March
1933, President Roosevelt declared a nation-wide bank holiday, which closed all banks
until some time between the 13th and 15th of March, depending on location. There were
17 800 commercial banks prior to the bank holiday period, but fewer than 12 000 were
allowed to open, under new federal/state authority licensing requirements. About 3000 of
the unlicensed banks were eventually allowed to remain open, but another 2000 were either
9
The details of US bank failures in the early 1930s are taken from Friedman and Schwartz (1963), pp. 332–349,
351–353.

[ 361 ]
B ANK F AILURES
liquidated or merged with other banks. The suspended operations and failures caused losses
of $2.5 billion for stockholders, depositors and other creditors. Friedman and Schwartz
(1963) argued a poor quality loan book and other bad investments was the principal cause
of some bank failures in 1930,
10
but later on, the failures were due largely to bank runs,
which forced banks to divest their assets at a large discount.
7.3.2. Bankhaus Herstatt
This West German bank collapsed in June 1974 because of losses from foreign exchange
trading, which were originally estimated at £83 million but rose to £200 million. At the
time it was unclear how the bank had managed to run up such losses. The bank’s failure
is famous because it exposed a weakness in the system related to liquidity risk. Bankhaus
Herstatt was due to settle the purchase of Deutsche marks (DMs, in exchange for dollars) on
26 June. On that day, the German correspondent banks, on instruction from the American
banks, debited their German accounts and deposited the DMs in the Landes Central bank
(which was acting as a clearing house). The American banks expected to be repaid in
dollars, but Bankhaus Herstatt was closed at 4 p.m., German time. It was only 10 a.m. on
the US east coast, causing these banks to lose out because they were caught in the middle
of a transaction. The US payments system was put under severe strain. The risk associated
with the failure to meet interbank payment obligations has since become known as Herstatt
risk. In February 1984, the chairman of the bank was convicted of fraudulently concealing
foreign exchange losses of DM 100 million in the bank’s 1973 accounts.
7.3.3. Franklin National Bank
In May 1974 Franklin National Bank (FNB), the 20th largest bank in the USA (deposits
close to $3 billion), faced a crisis. The authorities had been aware of the problem since the
beginning of May, when the Federal Reserve refused FNB’s request to take over another
financial institution and instructed the bank to retrench its operations because it had
expanded too quickly. A few days later, FNB announced it had suffered very large foreign

exchange losses and could not pay its quarterly dividend. It transpired that in addition
to these losses, the bank had made a large volume of unsound loans, as part of a rapid
growth strategy.
These revelations caused large depositors to withdraw their deposits and other banks
refused to lend to the bank. FNB offset the deposit outflows by borrowing $1.75 billion
from the Federal Reserve. Small depositors, protected by the FDIC, did not withdraw their
deposits, otherwise the run would have been more serious. In October 1974, its remains
were taken over by a consortium of seven European banks, European American.
10
Friedman and Schwartz (1963), pp. 354–355 distinguish between the ex ante and ex post quality of bank assets.
Ex ante, banks’ loan and other investment decisions were similar in the early 1920s and the late 1920s. The key
difference was that the loans/investments of the late 1920s had to be repaid/matured in the Great Depression.
Thus, they argue, with the exception of foreign lending, the number of bank failures caused by poor investment
decisions is debatable.
[ 362 ]
M ODERN B ANKING
FNB had been used by its biggest shareholder, Michele Sindona, to channel funds illegally
around the world. In March 1985 he died from poisoning, a few days after being sentenced to
life imprisonment in Italy for arranging the murder of an investigator of his banking empire.
7.3.4. Banco Ambrosiano
Banco Ambrosiano (BA) was a commercial bank based in Milan and quoted on the Milan
Stock Exchange. It had a number of foreign subsidiaries and companies located overseas,
in Luxembourg, Nassau, Nicaragua and Peru. The Luxembourg subsidiary was called Banco
Ambrosiano Holdings (BAH). The parent, Banco Ambrosiano, owned 69% of BAH. BAH
was active on the interbank market, taking eurocurrency deposits from international banks
which were on-lent to other non-Italian companies in the BA group.
The parent bank, BA, collapsed in June 1982, following a crisis of confidence among
depositors after its Chairman, Roberto Calvi, was found hanging from Blackfriars Bridge in
London, 10 days after he had disappeared from Milan. Losses amounted to £800 million,
some of them linked to offshore investments involving the Vatican’s bank, the Institute for

the Works of Religion. The Bank of Italy launched a lifeboat rescue operation; seven Italian
banks provided around $325 million in funds to fill the gap left by the flight of deposits,
and BA was declared bankrupt by a Milan court in late August 1982. A new bank, Nuovo
Banco Ambrosiano (NBA), was created to take over the bank’s Italian operations. The
Luxembourg subsidiary, BAH, also suffered from a loss of deposits, but the Bank of Italy
refused to launch a similar lifeboat rescue operation, causing BAH to default on its loans
and deposits.
The main cause of the insolvency appears to have been fraud on a massive scale, though
there were other factors whose contribution is unclear. The BA affair revealed a number
of gaps in the supervision of international banks. The Bank of Italy authorities lacked the
statutory power to supervise Italian banks. Nor was there a close relationship between senior
management and the central bank, as in the UK at the time. It appears that Sig. Calvi’s
abrupt departure may have been precipitated by a letter sent to him by the surveillance
department of the Bank of Italy seeking explanations for the extensive overseas exposure,
asking for it to be reduced and requesting that the contents of the letter be shown to other
directors of the bank. This activity suggests the regulatory authorities were aware of the
problem. The Bank of Italy refused to protect depositors of the subsidiary in Luxembourg
because BA was not held responsible for BAH debts; it owned 69% of the subsidiary. The
Bank of Italy also pointed out that neither it nor the Luxembourg authorities could be
responsible for loans made from one offshore centre (Luxembourg) to another (Panama)
via a third, again in Latin America.
In 1981, the Luxembourg Banking Commission revised some of its rules to relax bank
secrecy and allow the items on the asset side of a bank’s balance sheet to be freely passed
through the parent bank to the parent authority, though bank secrecy is still upheld for
non-bank customers holding deposits at Luxembourg banks. The authorities in Luxembourg
also obtained guarantees from the six Italian banks with branches in Luxembourg that
they would be responsible for the debts of their branches. The 1975 Basel Concordat was
revised in 1983 (see Chapter 5) to cover gaps in the supervision of foreign branches and
[ 363 ]
B ANK F AILURES

subsidiaries. In July 1994 the former Prime Minister of Italy, Bettino Craxi, was convicted
of fraud in relation to the collapse of Banco Ambrosiano.
Over 20 years later, questions relating to the death of Roberto Calvi continue. The first
Coroner’s Inquest judged the death to be suicide, but the family has always protested this
verdict, pointing to evidence such as bricks stuffed in the pockets of the deceased. A second
inquest recorded an open verdict. The City of London police decided to investigate and in
2003, one woman was arrested on suspicion of perjury and conspiring to pervert the course
of justice.
In March 2004,
11
four people went on trial in Rome for the murder of Roberto Calvi. One
of them, a former Mafia boss, is already in prison. A member of the Mafia turned informer
named the four accused. The prosecution claims the Mafia ordered his murder because
Mr Calvi bungled attempts to launder bonds stolen by the Mafia and was blackmailing
associates with links to Vatican and Italian society. A masonic lodge (P2) where Mr Calvi
was a member also appears to be involved.
7.3.5. Penn Square and Continental Illinois
As will become apparent, the collapse of these two banks was connected. Penn Square Bank,
located in Oklahoma City, had opened in 1960, as a one-office retail bank.
12
On 5 July
1982, the bank collapsed, with $470.4 million in deposits and $526.8 million in assets. It
embarked on an aggressive lending policy to the oil and gas sector – its assets grew more
than eightfold between 1977 and 1982. It sold the majority interest in these loans to other
banks, but remained responsible for their servicing. From the outset, loan documentation
was poor and loan decisions were based solely on the value of the collateral (oil and gas)
rather than assessing the borrower’s ability to repay. From May 1977 onward, the Office
of the Comptroller of Currency (OCC), the main regulatory authority, expressed concern
about a host of problems: poorly trained staff, low capital, lack of liquidity, weak loans and
increasing problems with the loan portfolio. The external auditors signed qualified opinion

in 1977 and 1981.
The way Penn Square’s failure was dealt with marked an apparent change in FDIC policy.
Of the 38 banksthat failed since 1980, only eight were actually closed with insured depositors
paid off. The other 30 had been the subject of purchase and assumption transactions,
whereby the deposits, insured and uninsured, were passed to the acquiring institution. Of
the $470.4 million in deposits at Penn Square, only 44% were insured. The uninsured
deposits were mainly funds from other banks. The FDIC paid off the insured depositors, and
in August 1983 the Charter National Bank purchased the remaining deposits.
At the time of its collapse, Continental Illinois National Bank (CI) was the seventh
largest US bank and the largest correspondent bank, involving about 2300 banks.
13
Though
its problems were well known by regulators, they were caught out by the speed of the bank’s
collapse. In the summer of 1984, a number of CI customers were having trouble repaying
11
‘‘Four go on Trial for the Murder of God’s Banker’’, The Guardian, 17 March 2004.
12
At the time, branching was prohibited in the State of Oklahoma. The details on Penn Square come from FDIC
(2001). Penn Square was one of the early failures – one of many between 1980 and 1994. See below.
13
Kaufman (1994, 2002).
[ 364 ]
M ODERN B ANKING
their loans because of the drop in oil prices. The decline in oil prices also undermined
the value of the collateral securing these loans, much of it in real estate in centres of
oil production.
Penn Square had a close connection to Continental Illinois. The bank was one of fivelarge
banks around the country that purchased participations in oil and gas loans. Shortly after the
collapse of Penn Square, CI announced a second-quarter loss of $63.1 million and revealed
its non-performing loans had more than doubled to $1.3 billion. In subsequent quarters,

the bank was slow to recover and its non-performing loans held steady at approximately
$2 billion, even though the non-performing loans related to the Penn Square connection
had declined.
The first-quarter results of 1984 (17th April) revealed the bank’s non-performing loans
had risen to $2.3 billion, representing 7.7% of its loans. Increasingly, CI had been relying
on the overseas markets to fund its domestic loan portfolio. On the eve of the crisis, 60% of
its funds were being raised in the form of short-term deposits from overseas. This reliance
on uninsured short-term deposits, along with its financial troubles, made it especially
vulnerable to a run.
Rumours about the solvency of the bank were rife in the early days of May 1984,
thereby undermining the ability of the bank to fund itself. On 10 May the rumours were
so serious that the US OCC took the unusual step of rebutting the rumours, though
the normal procedure was a terse ‘‘no comment’’. The statement merely served to fuel
more anxiety and the next day, CI was forced to approach the Chicago Reserve Bank
for emergency support, borrowing approximately $44.5 billion. Over the weekend, the
Chairman of Morgan Guaranty organised US bank support for CI: by Monday 14 May,
16 banks made $4.5 billion available under which CI could purchase federal funds on an
overnight basis. However, the private lifeboat facility was not enough. The run on the
bank continued and the bank saw $6 billion disappear, equivalent to 75% of its overnight
funding needs.
On 17 May the Comptroller, the Federal Deposit Insurance Corporation (FDIC) and
the Federal Reserve Bank announced a financial assistance programme. The package had
four features. First, there was a $2 billion injection of capital by the FDIC and seven US
banks, with $1.5 billion of this coming from the FDIC. The capital injection took the form
of a subordinated demand loan and was made available to CI for the period necessary to
enhance the bank’s permanent capital, by merger or otherwise. The rate of interest was
100 basis points above the one-year Treasury bill rate. Second, 28 US banks provided a
$5.5 billion federal funds back-up line to meet CI’s immediate liquidity requirements, to be
in place until a permanent solution was found. It had a spread of 0.25% above the Federal
funds rate. Third, the Federal Reserve gave an assurance that it was prepared to meet any

extraordinary liquidity requirements of CI. Finally, the FDIC guaranteed all depositors and
other general creditors of the bank full protection, with no interruption in the service to
the bank’s customers.
In return for the package, all directors of CI were asked to resign and the FDIC took direct
management control of the bank. The FDIC bought, at book value, $3.5 billion of CI’s debt.
The Federal Reserve injected about $1 billion in new capital. The bank’s holding company,
CI Corporation, issued 32 million preference shares to the FDIC, that on sale converted
[ 365 ]
B ANK F AILURES
into 160 million common shares in CI and $320 million in interest-bearing preferred stock.
It also had an option on another 40.3 million shares in 1989, if losses on doubtful loans
exceeded $800 million. It was estimated they exceeded $1 billion. Effectively, the bank was
nationalised, at a cost of $1.1 billion. A new team of senior managers was appointed by the
FDIC, which also, from time to time, sold some shares to the public. By 1991 it was back
in private hands, and in 1994 it was taken over by Bank America Corp. (Kaufman, 2002,
p. 425).
Continental Illinois got into problems for a number of reasons. First, it lacked a rigorous
procedure for vetting new loans, resulting in poor-quality loans to the US corporate sector,
the energy sector and the real estate sector. This included participation in low-quality loans
to the energy sector, bought from Penn Square. Second, CI failed to classify bad loans as
non-performing quickly, and the delay made depositors suspicious of what the bank was
hiding. Third, the restricted deposit base of a single branch system forced the bank to rely
on wholesale funds as it fought to expand. Fourth, supervisors should have been paying
closer attention to liability management, in addition to internal credit control procedures.
Regulators were concerned about CI’s dependence on global funding. This made it
imperative for the FED and FDIC to act as lender of last resort, to head off any risk of a run
by foreign depositors on other US banks. Continental Illinois was also the first American
example of regulators using a ‘‘too big to fail’’ policy. The three key US regulatory bodies
were all of the view that allowing CI to go under would risk a national or even global
financial crisis, because CI’s correspondent bank relationships left it (and the correspondent

banks) highly exposed on the interbank and Federal funds markets. The regulators claimed
the exposure of 65 banks was equivalent to 100% of their capital; another 101 had between
50% and 100% of their capital exposed. However, Kaufman (1985, 1994) reports on a
Congressional investigation of the collapse, which showed that only 1% of Continental’s
correspondent banks would have become legally insolvent if losses at CI had been 60 cents
per dollar. In fact, actual losses turned out to be less than 5 cents on the dollar, and no
bank suffered losses high enough to threaten its solvency. The Economist (1995) argues that
regulators got their sums wrong, and reports that some privately believed the bank did not
need to be rescued. However, it is worth noting that the correspondent banks were not
privy to this information at the time of the crisis, and would have been concerned about
any losses they incurred, even if their solvency was not under threat. Given the rumours,
it was quite rational for them to withdraw all uninsured deposits, thereby worsening the
position of CI.
Furthermore, the episode did initiate a too big to fail policy, which was used sporadically
throughout the 1980s. Some applications were highly questionable. For example, in 1990
the FDIC protected both national and off-shore (Bahamas) depositors at the National Bank
of Washington, D.C., ranked 250th in terms of asset size. The policy came to an end with
the 1991 FDIC Improvement Act (FDICIA), which required all regulators to use prompt
corrective action and the least cost approach when dealing with problem banks. However,
the ‘‘systemic risk’’ exception in FDICIA has given the FDIC a loophole to apply too big to
fail.
14
14
See Chapter 5 for more detail on FDICIA.
[ 366 ]
M ODERN B ANKING
7.3.6. Johnson Matthey Bankers
Johnson Matthey Bankers (JMB) is the banking arm of Johnson Matthey, dealers in gold
bullion and precious metals. JMB was rescued in October 1984, following an approach to
the Bank of England by the directors of JM, who believed the problems with JMB might

threaten the whole group. The original lifeboat rescue package consisted of the purchase of
JMB and its subsidiaries by the Bank of England for a nominal sum (£1.00), and wrote off a
large proportion of their assets. The bullion dealer, Johnson Matthey, was required to put up
£50 million to allow JMB to continue trading. Charter Consolidated, a substantial investor
in JM, contributed £25 million. Other contributors were the clearing banks (£35 million),
the other four members of the gold ring (£30 million), the accepting houses which were not
members of the gold ring (£10 million) and the Bank of England (£75 million).
On 7 November 1984 an agreed package of indemnities was announced to cover the
possibility that JMB’s loan losses might eventually exceed its capital base of £170 million.
In May 1985 the Bank of England declared that provisions of £245 million were necessary
to cover the loan losses. With this increase in loan provisions, all lifeboat contributions
were raised to make up the shortfall; the Bank of England and other members of the lifeboat
contributing half the amount of the shortfall each. On 22 November 1984, the Bank of
England made a deposit of £100 million to provide additional working funds.
JMB got into trouble because it managed to acquire loan losses of £245 million on a loan
portfolio of only £450 million, so it had to write off over half of its original loan portfolio.
Compare this to the case of Continental Illinois, where non-performing loans were only
7.4% of its total loans. Press reports noted that most of these bad loans were made to traders
involved with Third World countries, especially Nigeria, suggesting a high concentration
of risks. The Bank of England’s guideline on loan concentration (banks should limit loans
to a single borrower or connected group of borrowers to 10% of the capital base) appears to
have been ignored. The Bank of England was aware of some problems in 1983 but did not
act until the full extent of the problems emerged after a special audit in 1984.
The auditors also appeared to be at fault. Under the UK Companies Act, their ultimate
responsibility lies with the shareholders and they are required to report whether the accounts
prepared by the bank’s directors represent a ‘‘true and fair view’’. In assessing the bank,
the auditor reviews the internal audit and inspections systems, and on a random basis
examines the record of transactions to verify that they are authentic, and discusses with the
directors decisions made in highly sensitive areas such as provisions against bad and doubtful
debts. Auditors are not permitted to discuss the audit with bank supervisors, without the

permission of the clients. The auditors can either agree with the directors that the accounts
represent a true and fair view, or they can disagree with the directors, in which case they
must either resign or qualify the accounts. The auditors at JMB signed unqualified reports,
implying all was well. On the other hand, if the auditors had signalled problems by signing
a qualified report or resigning, it might have precipitated a bank run, and the authorities
may not have had enough time to put together a lifeboat operation.
As was noted in Chapter 5, the Bank of England’s system of supervision was flexible.
However, the JMB affair revealed two gaps in the reporting system. First, auditors had no
formal contact with the Bank of England and were unable to register their concerns, unless
they either resigned or qualified their reports. Second, the statistical returns prepared for
[ 367 ]
B ANK F AILURES
the Bank of England, based on management interviews, were not subject to an independent
audit. The 1987 amendment to the Banking Act addressed these problems partially; auditors
were encouraged to warn supervisors of suspected fraud, and were given greater access to
Bank of England information.
The JMB affair prompted the establishment of a committee involving the Treasury, Bank
of England officials and an external expert, to review the bank supervisory procedures,
especially the relationship between the auditor and supervisor. The result of the review of
the affair was an amendment of the Banking Act (1987). However, the effectiveness of
private auditing was again questioned after the BCCI closure (see below).
The JMB case illustrated the use of a lifeboat rescue by the Bank of England, and is a rare
example of where the too big to fail doctrine was extended to protect non-banking arms
of a financial firm. The main point of a rescue is to prevent the spread of the contagion
effect arising from a collapsed bank. Johnson Matthey was one of the five London gold price
fixers. Obviously, the Bank of England was concerned that the failure of the banking arm
would spread to JM, thereby damaging London’s reputation as a major international gold
bullion dealer. The episode suggests the Bank is prepared to engage in a lifeboat rescue
effort to protect an entire conglomerate, provided it is an important enough operator on
global financial markets.

7.3.7. The US Bank and Thrift Crises, 1980–94
Between 1980 and 1994 there were 1295 thrift failures in the USA, with $621 billion in
assets. Over the same period, 1617 banks, with $302.6 billion in assets, ‘‘failed’’ in the sense
that they were either closed, or received FDIC assistance. These institutions accounted for
a fifth of the assets in the banking system. The failures peaked between 1988 and 1992,
when a bank or thrift was, on average, failing once a day.
15
This section will begin with a
review of the thrift failures, followed by the commercial bank failures.
16
Failing thrifts
Thrifts are savings and loan (S&L) banks, either mutuals or shareholder owned, though
by the end of the crisis, the majority were stock owned. Until 1989, they were backed
by deposit insurance provided by the Federal Savings and Loan Insurance Corporation
(FSLIC). The FSLIC was in turn regulated by the Federal Home Loan Bank Board. Both
institutions were dissolved by statute in 1989.
In 1932, Congress passed the Federal Home Loan Bank Act. The Act created 12 Federal
Home Loan Banks, with the Federal Home Loan Bank Board (FHLBB) as their supervisory
agent. The aim was to provide thrifts with an alternative source of funding for home
mortgage lending. In 1933, the government became involved in the chartered savings
15
These figures are from FDIC (2001). ‘‘Failure’’ includes thrifts that were either closed by the Federal Savings
Loan Insurance Corporation (FSLIC) or the Resolution Trust Corporation (RTC), or received financial assistance
from the FSLIC. For a detailed account of the crisis, see White (1991).
16
The author’s account used two excellent publications by the Federal Deposit Insurance Corporation, FDIC
(1997) and FDIC (1998).
[ 368 ]
M ODERN B ANKING
and loans firms. The Home Owner’s Loan Act was passed, authorising the Federal Home

Loan Bank Board (FHLBB) to charter and regulate the savings and loan associations.
The National Housing Act, 1934, created a deposit insurance fund for savings and loan
associations, the Federal Savings and Loan Insurance Corporation (FSLIC). Unlike the
FDIC, which was established as a separate organisation from the Federal Reserve System,
the FSLIC was placed under the auspices of the FHLBB. S&L depositors are insured for up
to $100 000.
The first signs of trouble came in the mid-1960s, when inflation and high interest
rates created funding problems. Regulations prohibited the federally insured savings and
loans from diversifying their portfolios, which were concentrated in long-term fixed rate
mortgages. Deposit rates began to rise above the rates of return on their home loans. In
1966, Congress tried to address the problem by imposing a maximum ceiling on deposit
rates, and thrifts were authorised to pay 0.25% more on deposits than commercial banks
(regulation Q), thereby giving them a distorted comparative advantage. Unfortunately, the
difference was not enough because market interest rates rose well above the deposit rate
ceilings. The system of interest rate controls became unworkable and aggravated the thrifts’
maturity mismatch problems.
The 1980 Depository Institutions Deregulation and Monetary Control Act (DIDMCA)
took the first significant step towards reforming this sector. The DIDMCA allowed interest
rate regulations to be phased out, and permitted thrifts to diversify their asset portfolios
to include consumer loans other than mortgage loans, loans for commercial real estate,
commercial paper and corporate debt securities. But lack of experience meant diversification
contributed to a widespread loan quality crisis by the end of the 1980s.
DIDMCA came too late for thrifts facing the steep rise in interest rates that began
in 1981 and continued in 1982. Federally chartered S&Ls had not been given the legal
authority to make variable rate mortgage loans until 1979, and then only under severe
restrictions. Variable rate mortgages could not be freely negotiated with borrowers until
1981. By that time, deposit rates had risen well above the rates most thrifts were earning
on their outstanding fixed rate mortgage loans. Accounting practices disguised the problem
because thrifts could report their net worth based on historic asset value, rather than the
true market value of their assets.

Policies of regulatory forbearance aggravated the difficulties. Kane and Yu (1994) defined
forbearance as:
‘‘a policy of leniency or indulgence in enforcing a collectable claim against another party’’
(p. 241)
To repeat the definition used in other chapters, regulatory forbearance occurs when the
supervisory/regulatory authorities put the interests of the firms they regulate ahead of the
taxpayer. In the case of the thrifts, supervisory authorities adopted lenient policies in the
enforcement of claims against thrifts, because they had a vested interest in prolonging the
S&Ls’ survival: fewer thrift failures reduced the demands on the FSLIC’s fund. In 1981 and
1982, the FHLBB authorised adjustments in the Regulatory Accounting Principles, thereby
allowing thrift net worth to be reported more leniently than would have been the case had
Generally Accepted Accounting Principles (GAAP) been applied. In 1980 and 1982, the