76801

        Are All the Sacred Cows Dead?
    Implications of the Financial Crisis for
        Macro- and Financial Policies




                                                                                                                             Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
                                              ¨c
                                   Aslı Demirgu ¸ -Kunt † Luis Serve
                                                                   ´n

The recent global �nancial crisis has shaken the con�dence of industrial and developing
countries alike in the very blueprint of the �nancial and macropolicies that underlie the
Western capitalist systems. In an effort to contain the crisis from spreading, the auth-
orities in the United States and many European governments have taken unprecedented
steps of providing extensive liquidity, giving assurances to bank depositors and creditors
that include blanket guarantees, structuring bail-out programs that include taking large
ownership stakes in �nancial institutions, and establishing programs for direct provision
of credit to non�nancial institutions. Emphasizing the importance of incentives and ten-
sions between short term and longer term policy responses to crisis management, the
authors draw on a large body of research evidence and country experiences to discuss the
implications of the current crisis for �nancial and macroeconomic policies going
forward. JEL codes: G01, G21, G28, G32, E52, E58, F32


The �nancial turmoil that started as a meltdown in structured securitization
instruments in the summer of 2007 in the United States and the United Kingdom
has quickly spread to the rest of the industrial world, and has now become a full-
blown global �nancial crisis. In an effort to contain the crisis from spreading, the
authorities in the United States and many European governments have taken
unprecedented steps of providing extensive liquidity, giving assurances to bank
depositors and creditors that include blanket guarantees, structuring bail-out pro-
grams that include taking large ownership stakes in �nancial institutions, and
establishing programs for direct provision of credit to non�nancial institutions. In
some developing countries, there is talk about reintroducing capital controls as
a policy of last resort in the event of extensive bank runs and capital outflows.

The World Bank Research Observer
# The Author 2010. Published by Oxford University Press on behalf of the International Bank for Reconstruction and
Development / THE WORLD BANK. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org
doi;10.1093/wbro/lkp027           Advance Access publication February 4, 2010                                  25:91–124
Also questioned is the wisdom of a monetary policy narrowly focused on goods
prices without taking into account asset price inflation, and prudential regulation
that does not recognize systemic vulnerabilities.
   Hence the crisis has shaken the con�dence of industrial and developing
countries alike in the very blueprint of the �nancial and macropolicies that
underlie the Western capitalist systems. It is not surprising that many analysts are
already declaring capitalism, and the mainstream policy view associated with it,
to be dead.




                                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
   We argue that the “sacred cows� of �nancial and macropolicies are very much
alive. We seek to make clear that (i) the ongoing crisis does not simply reflect a
failure of free markets, but is a reaction of market participants to distorted incen-
tives; and (ii) managing a systemic crisis requires policy decisions to be made at
different stages of the crisis—the immediate containment stage as well as the
longer-term resolution and structural reforms that follow—which often entail dif-
�cult trade-offs between re-establishing con�dence in the short term and contain-
ing moral hazard in the long term.
   Keeping in mind the importance of incentives and tensions between short-term
and longer-term policy responses to crisis management, we address the following
questions about the implications of this crisis for �nancial and macroeconomic
policies going forward:
     †   Are blanket guarantees inevitable to halt a systemic crisis?
     †   Should governments bail out and take ownership of �nancial institutions?
     †   Should governments regulate �nance much more aggressively given the fail-
         ures in market discipline?
     †   Should monetary policy target asset prices?
     †   Should countries resort to capital controls to contain the crisis?
   Crises recur in part because people forget the lessons from previous crises.
While every crisis is different, past crises also provide important lessons that need
to be learned to prevent policymakers from reinventing the wheel every time a
new crisis erupts. We draw on past research and country experiences to address
the issues that are at the forefront of policy debate today.



Use of Blanket Guarantees in Containing a Systemic Crisis
Crises go through different stages. The �rst is called the “containment� stage in
which crises often emerge unexpectedly and evolve very quickly.1 This is the stage
that attracts the most attention, with bank runs, emergency liquidity loans, and
weekend crisis meetings. Time is of the essence and the need for speed generally
takes over good judgment. How should the authorities judge whether liquidity

92                                    The World Bank Research Observer, vol. 25, no. 1 (February 2010)
support or of�cial guarantees should be employed in the hope of preventing col-
lapse, but potentially at high long-term costs?
   Part of the answer lies in better crisis preparedness. Systemic crises are
infrequent events. Hence, the incumbent policymakers often claim there is “no
playbook� for handling crises. This lack of experience and knowledge leads
to trial-and-error policymaking and copying of policy responses—and often
mistakes—that are being employed elsewhere.
   In the latest crisis the U.S. and European governments provided extensive assur-




                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
ances to bank depositors and creditors that included blanket guarantees in many
cases. Some developing countries copied these arrangements, providing blanket
guarantees in order to prevent capital outflows and assure the public about the
safety of their banking systems.2 But other countries have resisted the need to
do so.3
   It is important to recognize that careful crisis containment strategies are very
dif�cult to devise in the midst of an actual turmoil. Political pressures to rescue
powerful interests are often too dif�cult for the authorities to resist. Because the
crisis seriously threatens the political future of the incumbent governments, the
usual short-termism in policy decisionmaking is even more exaggerated. Avoiding
such mistakes requires that crisis-management decisions are made in an open
debate outside of an actual crisis. Accountability would be improved by requiring
that regulators establish and regularly test a well-publicized benchmark plan for
crisis resolution (Caprio, Demirgu   ¸ -Kunt, and Kane 2008).
                                    ¨c

What Should Be Done in the Containment Stage?
Walter Bagehot’s (1894) classic policy advice for managing liquidity during a sys-
temic crisis is for the central bank to lend freely to solvent banks—but in order to
minimize the subsidizing of risk-taking (moral hazard), the loans are to be made
at penalty interest rates and only on good collateral. Put differently, the advice is
for governments to avoid lending to insolvent banks at all, even on good collat-
eral, and certainly not at below-market interest rates. Unfortunately, as the recent
events illustrate, modern governments pay only lip service to this principle. For
governments to embrace Bagehot’s advice, they need to be able to distinguish
relatively quickly between deeply insolvent banks and those that are solvent
enough to be salvageable. They also need to have the strength to resist the press-
ures that a crisis usually generates to rescue powerful interests.
   In a banking crisis, just like at a battle�eld, regulatory authorities need to run
to the aid of wounded deposit-losing institutions and temporarily stabilize their
condition by providing liquidity. Effective crisis containment requires effective
triage: treatment-worthy institutions need to be identi�ed and provided with
enough liquidity to restore public con�dence in their ability to continue in

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                           93
operation. Unless emergency response teams are assembled and trained in
advance, it is dif�cult to conduct good triage—a point which again stresses the
importance of crisis preparedness.
   Also very important are information problems, as the recent subprime crisis
amply illustrated. Distinguishing viable institutions at short notice becomes more
challenging in environments characterized by low levels of transparency.
Therefore it is the duty of regulators to identify and remedy gaps in information
well in advance, and recognize the gradual reduction in transparency that comes




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
with �nancial engineering and regulatory arbitrage and to nip it in the bud by
demanding improvements. Regulators also need to encourage the use of instru-
ments and the development of markets that would help to yield more accurate
assessment of risks, both in and out of crisis situations.4
   When transparency is allowed to deteriorate, information problems can tie the
hands of authorities and limit their ability to engage in ef�cient containment
strategies. For example, advocates of bailing out insolvent institutions to halt a
systemic crisis argue that only sweeping guarantees and extensive support can
stop the panicky flight of depositors and other creditors. This is of course true if
the crisis entails a series of self-ful�lling runs as envisioned in Diamond and
Dybvig (1983). However, most modern �nancial crises, including the recent one,
are driven instead by fundamental weaknesses in economic balance sheets, which
reveal themselves initially as liquidity problems.
   It must be recognized that the short-term bene�ts of guarantees will vary with
the �scal strength of the guaranteeing government. To hasten the end of an insol-
vency-driven banking crisis and to constrain the spread of insolvencies in the
short term, the government must manifest a substantial capacity for absorbing
losses. This is not a luxury most countries can afford since most governments do
not have the required �scal capacity. Depending on the depth of the systemic
insolvency, such support may not even halt the spread of the crisis, but merely
delay healthy adjustments. This begs the question of whether social costs and
adverse distribution effects could be reduced by following an alternative strategy.
   Even in the midst of a �nancial crisis, it is inef�cient to set aside long-term
goals completely. The manner in which a crisis is resolved affects the frequency
and depth of future crises through the signi�cant impact it has on market disci-
pline. Providing extensive liquidity support and guarantees to insolvent
institutions subsidizes their gambling on their own resurrection and distorts risk-
taking incentives, undermining market discipline and spawning future crises
(Kane and Klingebiel 2004; Calomiris, Klingebiel, and Laeven 2005). If insti-
tutions can count on crisis resolution to be handled in this way then they will be
more willing to risk insolvency, and safety-net subsidies will mainly flow to insti-
tutions that take excessive risks at the expense of taxpayers. The short-term
bene�ts of such bailouts have been oversold. Such policies seldom actually speed

94                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
the recovery of a nation’s real economy from a �nancial crisis or lessen the
decline in aggregate output. Instead, providing liquidity support for insolvent
institutions often prolongs a crisis. It does this by distorting risk-taking incentives
so extensively that sound investments and healthy exits are delayed and additional
output loss is generated. Honohan and Klingebiel (2003) and Claessens,
Klingebiel, and Laeven (2005) measure the impact of different crisis management
strategies on the ultimate cost of resolving �nancial distress in a broad set of
countries. They �nd that providing generous support—in terms of open-ended




                                                                                          Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
liquidity support and blanket deposit guarantees—not only increases the ultimate
�scal cost of resolving crises, but also that it does not speed the recovery, instead
prolonging the duration of the crisis.
   Using a sample of 42 banking crises Laeven and Valencia (2008) also show
that blanket guarantees have little impact on domestic deposits (they pretty much
continue their trend), but more often than not exacerbate the decline in foreign
liabilities. Indeed, announcement of a blanket guarantee is almost taken as a
signal to start a run on a currency. Not surprisingly guarantees are often
accompanied by increases in liquidity support, increasing the �scal cost signi�-
cantly. Indeed it is important to note that empirical evidence suggests that govern-
ments incur most of the �scal costs of resolving the crisis during the containment
phase. Honohan and Klingebiel (2003) show that much of the variation in the
�scal costs of crises is explained by differences in the way a government handled
its liquidity crisis, with the highest costs associated with governments that pro-
vided open-ended liquidity support and blanket deposit guarantees.


Alternatives to Blanket Guarantees
Providing blanket guarantees poses many challenges to government authorities.
The �rst important challenge is to convince creditors and depositors that they
have the political will and �scal capacity to afford the cost of such guarantees.
This may be quite dif�cult for many governments of developing countries. If the
emergency response is seen as inadequate, it may quickly compound the pro-
blems, requiring emergency funding from external sources, such as resource-rich
governments or the International Monetary Fund.
   Even if the guarantees provided are deemed credible and the crisis is contained,
the governments still face a second set of important challenges. These include:
the need to control the additional (guaranteed) debt that insolvent institutions
will continue to attract; to make sure the guaranteed institutions invest these new
resources prudently; to reduce or eliminate the guarantees once the containment
stage of the crisis is over. Regulators are likely to �nd it very dif�cult to address
the moral hazard created by these guarantees. Because fully guaranteed insti-
tutions can attract funding independent of the risks they take, managers of

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                            95
insolvent institutions can be easily tempted to abuse their government assistance
by gambling on their resurrection.
   There is already a large literature which establishes that in normal times overly
generous safety-net policies and deposit insurance will lead to moral hazard and
�nancial instability.5 The longer the guarantee remains in place, and insolvent
banks are allowed to operate, the more dif�cult it will be to curb these excessive
risk-taking incentives. Furthermore, once installed, such guarantees are dif�cult
to claw back and, more importantly, they seriously undermine the credibility with




                                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
which future safety net arrangements can be limited.
   While introduction of blanket guarantees may be tempting for policymakers in
the short term, they are not inevitable. Provision of such extensive guarantees
represents an extreme measure that is best resisted before other alternatives are
exhausted. Ideally, policymakers must be ready to take the time to separate hope-
lessly insolvent institutions from potentially viable ones, and to provide liquidity
support, guarantees, or “haircuts� in a way that would protect taxpayer interests.
Hence an alternative strategy to an indiscriminate blanket guarantee is to take a
“banking holiday� over several days in order to identify insolvent institutions and
to recommend and impose preliminary haircuts on uninsured depositors and non-
deposit creditors before they can liquidate their claims (Kane and Klingebiel
2004). Using the holiday to prepare a program of limited guarantees and to write
down the uninsured debt can restore public con�dence both in the government’s
ability to deal with the crisis and in the banking system itself.
   Baer and Klingebiel (1995) examine the aftermath of pre-1992 systemic crises
in a number of countries and �nd that in cases where the governments assigned
losses to depositors of insolvent banks, the positive effects of reducing depositor
uncertainty quickly overcame the negative effects that surviving banks experi-
enced from depositor write-downs. Fairness requires that small depositors—who
are often more than covered by explicit deposit insurance schemes—have immedi-
ate access to their funds. By the same token, at the end of the holiday, larger
uninsured depositors should also be allowed immediate fractional access to their
transaction balances. Clearly the speed with which the authorities can deal and
resolve these situations depends on the extent to which they have engaged in con-
tingency-planning and crisis-management simulations.
   A holiday that lasts for weeks or months is called a “deposit freeze,� and this
reduces depositors’ liquidity and the nation’s aggregate money supply; it may also
have long-term adverse effects on depositor con�dence.6 Hence to minimize these
adverse effects, insured depositors should be granted access to their funds as soon
as this is feasible. It is also important to note that it is not necessary for banks to
close for triage to begin. In a systemic crisis, bank holidays may be prolonged
because it may be dif�cult to complete the investigation of all banks quickly.
Hence, banks may continue to operate during inspections, though the deposit

96                                    The World Bank Research Observer, vol. 25, no. 1 (February 2010)
insurer may be given a right to void large loans and withdrawals made from a
bank within a certain period of its closure. This would make large transactions
subject to clawback, but it would not interfere with the business of ordinary
households while the inspections are going on.
   Broader time-out strategies for creditors that follow bankruptcy proceedings
can also be employed. Productive assets can be conserved by instituting a grace
period during which major creditors do not receive payments of principal or inter-
est due on existing bond or loan contracts, but use the time to work out a replace-




                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
ment contract structure with the help of courts, mediators, or both. Forcing
private creditors to renegotiate unenforceable contracts is called “bailing-in,� and,
like haircuts imposed on uninsured depositors, is intended to trap creditors that
�nanced weak institutions into participating more fully in loss-sharing.
   Whichever strategy is used, ultimately the damage the crisis causes to the
country’s �nancial sector and its real economy is reduced by separating the insol-
vent from the viable institutions as quickly as possible and by providing support
and allocating losses in ways that protect taxpayers and avoid subsidizing
the go-for-broke strategies of insolvent institutions. Prior crisis planning and com-
mitment to these plans are important steps in being able to retain a long-term
perspective in the containment stage of any crisis.



The Role of the State in the Financial System
After the panic abates, con�dence is restored, and markets start functioning
again, the crisis moves into its resolution stage. With the crisis contained, policies
must now be chosen to deal with the undercapitalization and insolvencies that
are often revealed. Past research and experience provide many valuable lessons on
how best to deal with systemic insolvencies, an issue we discuss below.
   Given the intensity of the recent crisis, direct interventions in the �nancial
system have been so massive that by the end of 2008 governments will be the
largest shareholders in most developed economies’ �nancial industries, reversing
a trend of state retreat over the last 20 years.7 With $500 billion or more
invested, this is equivalent to roughly state ownership of a quarter of the indus-
try’s market value, which strongly contrasts with the ideology of Western capital-
ist systems. Indeed, the extent of these interventions and the increasing
government stake in �nancial institutions have led to questions as to whether
�nancial systems should be privately owned in the �rst place, shaking the com-
mitment of developing country policymakers to ongoing bank privatization pro-
grams. If private banks are prone to excessive risk-taking leading to crises and
costly bailouts, is it not better to have the governments own and operate the
�nancial system in the �rst place?

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                           97
   In answering this question it is important to recognize that bank nationaliza-
tions are very common ways of dealing with systemic �nancial crises. Indeed, gov-
ernments have always taken ownership positions in banking either deliberately or
indirectly as a result of banking crises. In almost every major banking crisis in
recent history—in East Asia, Latin America, and other countries—governments
have become temporary caretakers of �nancial institutions.8 Hence, the recent
crisis is not the exception in this regard. But the way in which this is done has
important implications for maintaining and restoring a functioning �nancial




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
system and minimizing the short term and long term costs of such interventions.9
But the fact that governments �nd themselves involved in resolution of insolven-
cies does not imply this role should be permanent. On the contrary, all evidence
suggests that they would be well-advised to do otherwise.


Bureaucrats as Bankers?
Government ownership of banking has been popular throughout history.10 Early
proponents of state control argued that the government can better allocate capital
to highly productive investments. Gerschenkron (1962) was among the �rst to
argue that private banks would not be able to overcome de�ciencies in infor-
mation and contracting in weak institutional environments. State ownership also
makes appropriation of the surplus from �nance and directing credit much easier,
making it attractive for policymakers. Moreover, there is the concern that private
ownership and concentration in banking may lead to limited access to credit by
different parts of society, so hampering the development process. Indeed, govern-
ment banks are often expected to expand access, making �nancial services more
broadly available. A �nal argument is rooted in the fragility of �nance, and that
private �nancial institutions are too prone to excessive risk-taking, which is dif�-
cult to keep under check; a popular sentiment that resurfaces with every signi�-
cant crisis.
   By now there is a signi�cant amount of empirical research that suggests state
ownership of banks is associated with less �nancial sector development, lower
growth and lower productivity, and that these negative effects are more pro-
nounced at lower levels of income with less �nancial sector development and
with weaker property rights’ protection (Barth, Caprio, and Levine 2001; La
Porta, Lopez-de-Silanes, and Shleifer 2002). Despite explicit mandates for govern-
ment banks to expand outreach, in banking systems dominated by state banks
there are fewer bank branches and automated teller machines. Customers in such
systems face lower fees but they also experience poorer service quality (Beck,
Demirgu ¨c¸ -Kunt, and Martinez Peria 2007, 2008). There are some exceptions in
that certain state-owned banks, designated as development �nance institutions,
have become effective providers of know-how and have had a useful catalytic

98                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
function in “kick-starting� certain �nancial services, for example in Latin
America (De la Torre, Gozzi, and Schmukler 2007). But even then the most suc-
cessful of these initiatives are expected to be privatized.
   Instead, the bulk of the evidence suggests that state-owned banks tend to lend
to cronies, especially around the time of elections, as vividly illustrated by the
recent empirical studies of Cole (2004), Dinc      ¸ (2005), and Khwaja and Mian
(2005). For example, Cole has shown that state banks in Indian states ramped up
agricultural lending in tightly contested districts in election years. Dinc¸ showed




                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
that increased lending by government-owned banks right before elections is not
speci�c to India but can be observed in data from 22 developing countries.
Drilling down to the individual loan level, Khwaja and Mian found evidence from
Pakistan that politically active borrowers were able to secure larger and cheaper
loans from state-owned banks and that they defaulted on these loans much more
than other business borrowers.
   Given the extent to which lending policies are politicized, it is not surprising
that state ownership appears to heighten the risk of crises instead of reducing it.
If anything, research suggests that greater state ownership is associated with
various measures of �nancial instability, including a greater probability of
banking crises (Caprio and Martinez Peria 2000; Barth, Caprio, and Levine 2001;
La Porta, Lopez-de-Silanes, and Shleifer 2002).
   These results and other related evidence explain why many countries embarked
on privatization programs, selling their state banks. Indeed, evidence also suggests
bank privatization, if well-designed, can signi�cantly increase bank performance
(Clarke, Cull, and Shirley 2005; Megginson 2005). To be sure privatization is dif�-
cult and can lead to problems in weaker institutional environments. Hence the
sequencing is important: moving slowly but deliberately with bank privatization,
while preparing state banks for sale and addressing the weaknesses in the overall
incentive environment and regulations, seems to be the preferred strategy.
Ultimately, gains from privatization—if designed properly—can be substantial
since the alternative of maintaining large state ownership can signi�cantly under-
mine real sector reforms and deter economic development.11
   So what explains the poor performance of bureaucrats as bankers? As often the
case in �nance, incentive problems are at the root of this issue since bureaucrats
do not face incentives designed to reward ef�cient resource allocation. Not only
do government of�cials often lack the expertise to be effective managers, they also
face conflicts of interest due to their desire to secure their political base and
reward supporters, which often goes against ef�cient resource allocation. These
problems become worse with fewer checks and balances and in poorer insti-
tutional environments, explaining why state ownership is more damaging at
lower per capita income levels ( Keefer, 2000; La Porta, Lopez-de-Silanes, and
Shleifer 2002).

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          99
   Hence while governments may inevitably �nd themselves as stakeholders in
�nancial institutions as the outcome of systemic crises, they would do well to see
this as a temporary arrangement and plan for their exit as part of the crisis resol-
ution exercise. Despite its weaknesses, a well-functioning private �nancial system
is crucial for promoting development, and substituting government provision of
�nancial services for that of the market is likely to lead to inferior outcomes.
Economic growth does not resume on a sustainable basis until productive assets
and banks are back in the hands of well-capitalized private parties.




                                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
   But how should �nancial crises be resolved and solvency of institutions
restored? In other words, how should bail-out programs be designed in the event
of crises, and what should be the government’s role in this process?


Bureaucrats as Caretakers
If bureaucrats are not good bankers in good times, they are not likely to do better
in bad times, given that the tendency for political forces to dominate economic
judgments will be even stronger. As we have seen in the latest crisis, systemic
crises often involve the injection of substantial sums and the determining of the
�nancial fate of powerful interests. But systemic crises often require comprehen-
sive solutions by the government. So how should governments behave?
   While governments should be prepared to act in a systemic crisis, the approach
and the actions they take still need to be designed to reduce conditions for moral
hazard and the likelihood of a subsequent crisis. This should be done by imposing
real costs on all responsible parties and getting the resources back in productive
use as soon as possible (Demirgu  ¸ -Kunt, 2008).
                                 ¨c
   To see the importance of paying attention to incentives, consider what happens
to �rms outside the �nancial sector that fall into a state of insolvency. Those in
control of the �rm �nd it dif�cult or impossible to raise new funds in any form.
They can no longer act on pro�table investment opportunities and may be forced
to sell important assets. Creditors recognize this situation may distort the incen-
tives of managers, making them more susceptible to fraud and moral hazard,
which at the very least reduces the incentives of the owner and managers to exert
effort. In a market economy, the solution is bankruptcy. As long as the �rm is
economically viable, it makes sense to continue its operations, but only after
restructuring. This restructuring generally wipes out existing equity holders,
while debt holders often have a portion of their claims converted to equity.
Alternatively debt holders can agree to cut down the face value of their debt in
exchange for some warrants. Old management is often replaced, a substantial
portion of the �rm’s assets are sold, and workers are laid off. This restructuring is
not a mere reworking of the �rm’s balance sheet, but represents very real
changes to the way it does business, perhaps even in the business it does.

100                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
The happy result for the economy is that resources continue in their best use,
while all responsible parties incur some costs for the �rm’s poor performance.
   To a large extent dealing with �nancial insolvencies should follow the same
general principles. Admittedly the �nancial sector is special due to its particular
fragility, the possibility of contagion, and the major macro-implications when it is
in systemic distress. However, although these differences may justify different
approaches, they do not suggest that incentives matter any less. Any government
involvement should be designed to protect the interests of the taxpayers, impose




                                                                                            Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
losses on the responsible parties, and use the private sector to pick the winners
and losers. For example any plan that purchases bad assets from troubled �nan-
cial institutions, or recapitalizes without extracting some claim from the insti-
tutions, amounts to a transfer from taxpayers to shareholders, which is the group
that keeps the residual value of the entity.
   Recapitalization of the banking system should be designed so that those banks
that need assistance with recapitalization are helped in an incentive-compatible
way. Such a plan would limit taxpayers’ loss exposure, and, in environments with
good information and contract enforcement, leave resolution of bad assets to the
banks themselves. But how should the banks be recapitalized? Just as bureaucrats
are likely to make poor bankers, the selection of individual winners and losers is
also what markets, not governments, do best.
   One way is for the authorities to inject funding only to those institutions that
are able to meet the following criteria (World Bank 2001; Honohan 2005):
   †   Those institutions that are in a strong position should be able to raise capital
       privately, say from a syndicate of private banks. If the institution has dif�culty
       raising capital, it should at least be able to obtain some proportion of it—at
       least half of it or more—from the private sector before applying for govern-
       ment recapitalization assistance. Only those institutions that can secure
       private sector funds would be eligible for the plan. This will ensure that the
       private sector plays an important role in picking the survivors.12
   †   Those institutions that are eligible receive government assistance in the form
       of preferred stock. Preferred stock status would force private parties rather
       than taxpayers to bear the �rst-tier losses.
   †   Any bank participating in the plan will have to suspend all dividend pay-
       ments (and restrict the amount and structure of its compensation plans for
       its senior managers) until the government is fully “bought out.� The bank
       will also agree to comply with strict regulations on its leverage, risk-taking,
       transparency, and disclosure in order to participate in the plan. This will give
       incentives to banks to retire their preferred stock as soon as possible.
 These are tough criteria and only desperate banks will agree to these terms.
Which is exactly the point: government assistance should only be injected into

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                             101
banks in dire straits, yet simultaneously to those with a real chance of survival.
As long as the amount of funding is such that some banks fail, this approach
removes government from decisions as to which banks survive. The availability of
private sector funding serves to identify the candidates, and restrictions on differ-
ent ways to take these funds out of the banks, combined with greater transpar-
ency, makes it more likely that the banks will not be looted or engage in gambling
with taxpayers’ money. By openly stating the terms on which it will assist banks,
and ensuring that those terms provide good incentives for the restructured bank




                                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
going forward, the government will have made the best use of market forces while
minimizing its direct ownership involvement. Many of these features characterized
the U.S. Reconstruction Finance Corporation’s (RFC) program of taking temporary
preferred equity positions in banks after the great depression.13
   Another alternative that has gained popularity after its successful application
in Spain in the early 1980 s is a centralized approach where banks’ nonperform-
ing loans are carved off into an asset management company (AMC) which con-
tinues the restructuring process. The carving-out of an insolvent bank’s bad loan
portfolio and its organizational restructuring under new management and owner-
ship may be appropriate when large parts of the bank’s information capital is dys-
functional. The bad loan portfolio may be sold back into the market or disposed
of by a government-owned AMC.
   A survey by Klingebiel (2000) shows, however, that the AMCs formed in devel-
oping countries have not been as successful in restructuring nonperforming
loans. The effectiveness of AMCs has been quite mixed, better when the assets to
be disposed are primarily real estate, less good when they are loans to large politi-
cally connected �rms. In countries with developed private markets and insti-
tutions, individual banks—if well capitalized—would be in a superior position to
engage in restructuring their own assets. But policymakers in weak institutions
should not expect to achieve the same level of success in restructuring as those in
more developed economies, and they would do well to design simple resolution
mechanisms that offer little discretion to government of�cials (Honohan and
Laeven 2005).
   In summary, governments are not good at providing �nancial services in
normal times or in crises, and those governments that �nd themselves as bankers
as a result of such crises should focus on their exit strategy as quickly as possible,
using the market to identify winners and losers. And although drawing on public
funds to recapitalize some banks may be unavoidable in systemic crises, they must
be used sparingly for leveraging private funds and incentives (World Bank 2001).
Hence, although �nance is prone to excesses and crises, there is a substantial lit-
erature that suggests government ownership is not the answer. Where govern-
ments have a very important role to play in �nance, it is in providing the



102                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
regulatory and supervisory arrangements that help reinforce incentives that limit
this excessive risk-taking and fraudulent behavior.



Design of Prudential Regulations
One important implication of the recent crisis is the widespread calls for reforms
of regulation and supervision. The initial reaction to the emerging crisis was one




                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
of disbelief: how could the crisis emerge in countries whose supervision of credit
risk had been thought to be the best in the world? Indeed, the regulatory stan-
dards and protocols of these countries were in the process of being emulated
worldwide through the international capital accords, known as the Basel
standards.
   Basel II—which is currently in its implementation stage—grew out of concern
that the Basel I accord was unable to address the range of risks in bank activities,
as evidenced by the growth of securitization.14 Basel II is built on three pillars:
(1) minimum regulatory capital requirements for credit risk, operational risk, and
market risk; (2) the supervisory review process; and (3) market discipline and dis-
closure. The minimum capital requirements are determined by either external
ratings from ratings agencies for smaller banks or by outputs from the larger
banks’ own internal ratings models.15
   Many interpreted the crisis as a vivid example of market failure, evidence that
there is no such thing as market discipline, reinforcing calls for stronger regu-
lations through improvements in the Basel II accord. But the crisis also spawned
a growing argument about the role that the Basel I accord may have played in
causing the crisis. Indeed, it is no secret that Basel I contributed to the growth in
securitization by assigning lower capital charges and thus giving incentives to
institutions to move their assets into off-balance-sheet securitization vehicles.
While advocates claimed that Basel II, had it been implemented earlier, could
have lessened or prevented the turmoil, critics of the Basel approach to capital
regulation pointed out that the crisis has simply recon�rmed fundamental flaws
that have been evident in this approach.16
   The �nancial turmoil challenged the Basel II framework in important ways.
First, the events raised serious questions about setting capital requirements based
on external ratings. These ratings proved excessively optimistic and con�rmed
long-standing concerns about the conflicts of interest that arise out of having the
issuers pay the agencies for ratings required by regulators. More importantly,
credit ratings are not appropriate for setting capital requirements. Ratings are
based on expected default rates, yet capital is intended for unexpected losses.
Ratings can be useful for establishing loss reserves for particular assets, but they
do not consider correlations among assets, hence they are not helpful in assessing

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          103
how a bank’s net worth or its portfolio of assets may vary. Therefore the amount
of capital required for an institution’s safety has to be linked explicitly to measures
of volatility of its earnings, which is not information that ratings provide.
   Second, the crisis raised serious concerns about the accuracy of internal risk
models employed by even the largest and most sophisticated market participants.
These models proved inadequate and illustrated how �nancial models and data-
sets can be manipulated to provide desired outcomes.
   Third, the way the crisis emerged in subprime lending but spread to other secu-




                                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
rities revealed problems with inadequate documentation, the disconnect between
the source of risks and the bearers of those risks, and lack of knowledge about
how risks are ultimately distributed.17 All these point to weaknesses in the dis-
closure provisions included in the market discipline pillar of the Basel accord, and
how fundamental issues of transparency were not addressed.
   Given the intensity and persistence of the crisis, many proposals for regulatory
reforms are emerging. Some proposals originate from sources that interpret recent
events as evidence that market discipline is not reliable and want to focus on
designing tougher and more comprehensive regulations. These proposals range
from re-establishing extensive activity restrictions, to speedier implementation of
Basel II, to revising the accord in order to address the weaknesses discussed
above, to enhanced opportunities for further of�cial intervention.18
   Others continue to believe in the reliability of market signals and market disci-
pline, and argue that the Basel approach is fundamentally flawed. On the hypoth-
esis that investors would be quicker to recognize changes in risk and risk
premiums, such sources propose an alternative approach where of�cial supervi-
sion would focus on generating and using better market signals.
   The main features of this approach can be summarized as follows.19 While the
Basel Committee has been responsive to its critics by trying to make the capital
requirements more reflective of the ways in which risks and vulnerabilities are
assessed, this has led to greater and greater complexity. Increased complexity in
bank regulation reduces transparency but increases the scope for regulatory arbit-
rage and forbearance, without necessarily increasing accuracy. Indeed, these
critics argue that while the task of computing correct economic capital for banks
is very dif�cult and complex, bank capital regulation need not be. An alternative
is to rely in large part on the market itself to provide a measurement of risk,
together with the enforcement of simple rules such as the leverage ratio and
prompt corrective action, which are not subject to manipulation. Supervisors
must not only draw on—but also help develop—informative market signals such
as those imbedded in the prices of credit default swaps and subordinated debt. By
requiring large institutions to engage in these markets, regulators would be able
to incorporate market-generated information to their risk assessments. Hence
this alternative combines supervisory and market oversight, helping to generate

104                                   The World Bank Research Observer, vol. 25, no. 1 (February 2010)
and harness the information markets are capable of producing. The job of the
supervisors would be signi�cantly easier if this could be done effectively.
   However, as Caprio, Demirgu      ¸ -Kunt, and Kane (2008) emphasize, prices in
                                   ¨c
credit default swap or subordinated debt markets can be a strong source of disci-
pline only under two conditions: (1) market participants do not expect to be
bailed out when trouble develops, and (2) investors have access to regular flows of
high-quality information. This underlines the importance of establishing incen-
tives that would lead both supervisory authorities and market forces to operate




                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
more effectively.
   Caprio, Demirgu ¨c¸ -Kunt, and Kane argue that this crisis exempli�es not just the
limits of market discipline, but the power of government-induced incentive distor-
tions—and the limits of of�cial supervision as commonly practiced. The failure of
private parties to exercise suf�cient due diligence was rooted in the failure of gov-
ernment supervisors to challenge decisions made by private accountants and
credit-rating organizations. Authorities neglected their duty of examining and
publicizing the implications that these decisions might have for safety-net loss
exposure. By tolerating a decline in transparency, supervisors made it dif�cult to
recognize and price the risk expansion, not only for themselves but also for the
market participants.20
   Hence, Caprio, Demirgu     ¸ -Kunt, and Kane propose reforms that improve the
                             ¨c
chain of incentives under which market discipline and of�cial supervision
operate, which include reforms for lenders, credit rating organizations, securitiza-
tion, accounting, and government of�cials. Perhaps most important among these
are their proposals for enhancing accountability of government of�cials through
better crisis preparedness, greater use of market information to track risks and
subsidies, publicizing estimates of safety-net subsidies, and deferred compensation
schemes.
   First, as already discussed above, crisis preparedness is important to avoid
short-termism in crisis-management. Accountability would be greatly improved
by requiring that regulators establish and regularly test a well-publicized bench-
mark plan for crisis resolution.
   Second, regulators need to draw on market signals to overcome information
problems and improve their ability to track risk in and out of crises. Requiring the
largest banks to issue at regular intervals a series of credit default swaps or unin-
sured subordinated debt would provide such signals, since the holders of these
instruments would apply the market discipline that pillar three of Basel II seeks to
harness.21
   Third, the safety net needs to be strengthened by making authorities more
accountable for its cost. This requires the development of a system of fair-value
accounting for intangible safety-net subsidies to establish political accountability
for controlling them. Important institutions and their supervisors must be

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          105
required to model, estimate, and expose to outside review the value of this intan-
gible source of income—both in individual institutions and in the aggregate.
    Fourth, the decisionmaking horizons of government of�cials can be lengthened
if employment contracts included a fund of deferred compensation that the heads
of supervisory authorities would have to forfeit if a crisis occurred within a couple
of years after leaving their of�ce. Calomiris and Kahn (1996) show that such a
system worked well in the 19th-century Suffolk banking system, where claims to
deferred bonuses were paid only after losses were deducted.




                                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
    While the discussion of how best to reform regulation and supervision of �nan-
cial institutions is likely to continue, it is important to keep in mind that ulti-
mately the goal of �nancial regulation and supervision is not to reduce �nancial
institution risk-taking, but to manage the safety net so that private risk-taking is
neither taxed nor subsidized. This goal implies that supervisors have a duty to see
that risks can be fully understood and fairly priced by investors. No one should
expect that, in a risky world, risk-neutral regulation and supervision can elimin-
ate the risk of �nancial crises; what it can do is to reduce their frequency and
cost.



Monetary Policy, Asset Prices and Macroprudential Regulation
Most observers agree that lax monetary policy in the United States in the early
2000s helped fuel the housing bubble, at least in its initial stages. In light of the
devastating effects of its bursting, one major policy question looking forward is
the proper role of monetary policy. Should central banks respond systematically
only to inflation in goods prices, as they do at present, or should they also
respond systematically to inflation in assets prices? To put it more bluntly, should
monetary authorities attempt to “prick bubbles� through monetary tightening?
   Not all bubbles are alike. Some create risks to the �nancial system, while others
do not.22 Bubbles that threaten �nancial stability typically involve a feedback loop
between asset prices and credit conditions: a credit expansion raises asset prices,
which in turn encourages further lending and hence further asset price rises, and
so on. When the bubble bursts, asset prices collapse and the loop goes in reverse,
eroding the balance sheets of �nancial institutions, causing a credit crunch and a
fall in economic activity. But not all bubbles have these features—for example the
dot-com bubble of the late 1990s in the United States was not associated with a
credit boom, and its crash did not weaken lenders’ balance sheets to a signi�cant
extent. Hence the pursuit of �nancial stability poses a greater need for the monet-
ary authority to react to some bubbles than to others.
   Attempts by the authorities to deflate an asset price bubble—through monet-
ary policy or other means—in the absence of obvious symptoms of inflationary or

106                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
�nancial distress face serious political-economy obstacles. Lenders and borrowers
riding the bubble euphoria are likely to condemn such seemingly unwarranted
courses of action, and politicians may oppose it under the view that, rather than
a bubble, rising asset prices actually reflect the bene�cial effects of improved pol-
icies.23 Deflating a bubble amounts to suppressing an event, and it is impossible
to prove ex post that the event would have occurred in the absence of policy
action (Wyplosz 2009).
   Political economy concerns aside, the monetary policy response to bubbles has




                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
been debated mainly in industrial countries in the context of inflation targeting
regimes, under which monetary authorities react solely to inflation forecasts and
the output gap (the “Taylor rule�). But the relevance of the question is much
broader. It is of special signi�cance to a growing number of emerging economies
that have in recent years adopted inflation targets to guide monetary policy. More
generally it matters to all countries whose monetary authorities are entrusted
with inflation control and macroeconomic stability.
   The conventional view of inflation targeting assigns no role to asset prices in
the conduct of monetary policy, except to the extent that changes in asset prices
signal changes in expected inflation (Bernanke and Gertler 2000). In this view,
the authorities should deal with asset price bubbles only as their consequences, if
any, arise in terms of the inflation objective—for example by supplying the needed
liquidity in the event of a bubble burst. In other words, the authorities should not
attempt to “prick� bubbles.24
   An alternative view advocates a more proactive response of monetary policy to
asset prices. Of course, asset prices may rise or fall for many reasons, including
changes in fundamentals—that is, reassessments of the anticipated future pro-
ductivity of the assets—so in principle monetary policy should react to deviations
of asset prices from their underlying fundamentals, that is, to asset bubbles,
rather than to deviations from any particular target level (Cecchetti, Genberg,
and Wadhwani 2003). Formally, while the monetary authorities’ objective func-
tion would continue to be de�ned only in terms of goods inflation, their reaction
function should include not only inflation forecasts and the output gap, but also
measures of asset price misalignment. In theory, adding more arguments to the
monetary policy reaction function should allow the authorities to do no worse,
and possibly to do better, than in the standard framework. But the approach
remains untested, and several objections have been raised against its practical
feasibility.
   First, asset price bubbles are not easy to spot in real time—at least at their
early stages when policy action might be most useful to stop them. Identifying a
misalignment of asset prices requires a reliable assessment of their fundamentals.
But available models of asset price fundamentals are inherently imperfect, and in
addition some fundamentals—for example investors’ perceptions of risk and

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                         107
future asset productivity—are not readily observable. A simpler alternative would
be to use arbitrary rules of thumb, such as a threshold above which asset price
increases would be deemed to involve bubbles. However, simple rules of thumb
would likely lead to frequent identi�cation errors, especially in narrow and illiquid
asset markets (such as those of most emerging countries), characterized by high
asset price volatility.25 And reacting to a misidenti�ed bubble may be very
costly—for example tightening in response to an asset price rise that in reality is
driven by improving fundamentals will both hamper growth and interfere with




                                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
the role of asset prices in allocating resources. Of course, one possible solution to
this problem is to react only to bubbles once they become self-evident, but by
then it may be too late for monetary policy to mitigate their effects. Indeed, if the
authorities tighten when the bubble is already nearing collapse on its own, the
contractionary effects of the tightening could compound, rather than mitigate,
those of the bubble collapse. Yet while this view that bubbles are hard to spot ex
ante is certainly correct, one could argue that other standard tasks of monetary
policy, such as forecasting output and inflation over a multiyear horizon, are not
less dif�cult.
   Second, what asset prices should the monetary authority watch? Not all asset
prices are synchronized, and at any one time bubbles may be present in just a few
of the many assets available in the economy. Some observers suggest that the
price of housing may be the best candidate for close monitoring, given that
housing values have major wealth effects on spending—and hence are more rel-
evant than other asset prices for inflation and the output gap—and that housing
cycles of boom and bust are more frequent than equity market cycles.26 A
broader view holds that the monetary authority should focus on the prices of
assets held by highly leveraged �nancial intermediaries, given their key role in the
boom-bust credit cycle and the propagation of �nancial turmoil in the current
crisis (Adrian and Shin 2009).
   Third, little is known about the timing and magnitude of the effects of monetary
policy on asset prices, and hence its ability to prick bubbles. For example, some
research suggests that very large interest rate hikes may be necessary to bring
housing prices down to any noticeable extent—so large in fact as to result in huge
output losses (Assenmacher-Wesche and Gerlach 2008). But other recent research
indicates that monetary policy has a sizable effect on the acquisition of assets by
leveraged intermediaries (Adrian and Shin 2009) and on bank lending standards
(Maddaloni, Peydro   ´ , and Scopei 2008). The implication is that a timely monetary
tightening in the boom might have been effective at containing the cyclical expan-
sion of leverage, credit, and asset prices, and prevented the deterioration of lending
standards, two key ingredients in the gestation of the current crisis.
   Overall, the usefulness of monetary policy to prick bubbles remains controversial.
One additional concern is that gearing monetary policy to deflate hard-to-identify

108                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
bubbles may detract from its transparency and predictability, especially in emerging
markets that are still at the early stage of establishing the credibility of their inflation
targeting regimes or, more generally, their commitment to price stability.
   To approach the problem from a different angle, it is important to recall that
monetary authorities in most countries face two different objectives: price stability
and �nancial stability. With monetary policy as their only instrument, the auth-
orities can �nd themselves in situations where the objective of price stability
requires a policy change in one direction, while that of �nancial stability points in




                                                                                              Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
the opposite direction. Ideally a second instrument should be devoted to �nancial
stability. The best option is a prudential regime capable of dampening �nancial
cycles of boom and bust by preventing feedback loops between asset prices and
credit supply. This is what has been termed “macroprudential� regulation. Its
thrust is to complement regulators’ traditional focus on the risk management of
the individual �nancial institution with a focus on the risk management of the
�nancial system as a whole. Macroprudential regulation aims at dealing with
increases in systemic vulnerabilities due to (i) periodic business downturns that
effect all �nancial institutions; (ii) increases in the number of �nancial institutions
that have become too large, too interlinked, and therefore too dif�cult to fail and
unwind. The �rst is an age-old problem, which is exacerbated by the amplitude of
the business cycle, as discussed above. The second has become increasingly more
important as safety-net subsidies provided to large, interconnected �rms have
increased over time, giving them more incentives to become even larger and more
interlinked.
   While monetary authorities possess a variety of tools to pick up the pieces after
a �nancial crash, they lack the regulatory instruments to contain the lending
boom that usually precedes it. However, provisions, leverage ratios, loan-to-value
ratios, and additional capital buffers can all be designed to be countercyclical,
that is, to move inversely with the business cycle in order to make �nancial inter-
mediaries hold more liquid assets in good times so that they can be run down in
bad times (Goodhart 2008a, 2008b; Goodhart and Persuad 2008). The idea is to
switch the basis of capital adequacy requirements from levels of risk-weighted
assets to their rates of growth—hence requiring additional capital and liquidity
when bank lending and asset prices are rising fast—and relaxing such require-
ments in the downturns. This can be seen as an alternative (or a complement) to
monetary policies to prevent the growth of asset bubbles (and their busts). To
date, however, this kind of instrument has seen little use, except for Spain’s coun-
tercyclical “dynamic provisioning,� as well as the introduction of time-varying,
loan-to-value ratios in a few small economies.
   Aside from practical questions—for example, over what periods applicable
credit growth rates should be calculated—these proposals also raise other poten-
tial dif�culties that require deeper consideration.27 As prudential requirements

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                               109
are tightened in the upswing, they will encourage disintermediation to less-regu-
lated entities (or countries), and this may weaken the effectiveness of the regu-
lations. In addition, the cost of intermediation will rise in the boom, possibly
constraining growth and �nancial innovation. Lastly, macroprudential regulation
adds to the informational burden on the regulators and the complexity of their
task. As with monetary policy, signi�cant political will is required to enable the
authorities to tighten regulation in the upswing. Although reliance on well-
de�ned rules (rather than reliance on discretionary regulatory changes) and inde-




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
pendence from government may be helpful in this regard, it is still questionable
whether without appropriate incentive reforms this will be possible.
   Finally, to deal with too large or too interconnected �rms, reform proposals
aim to increase supervisory scrutiny ( potentially through a college of supervi-
sors) or additional capital charges imposed on these institutions. However,
proper identi�cation and increased regulation of these institutions require cred-
ible estimates of the safety-net subsidies that are provided to them. Estimating
and publicizing safety-net subsidies as advocated by Caprio, Demirgu        ¸ -Kunt,
                                                                           ¨c
and Kane (2008), as discussed in the previous section, would be one step in
this direction.
   These and similar ideas underlie an active debate on macroprudential regu-
lation that has already generated a number of reform proposals (Kashyap, Rajan,
and Stein 2008; Acharya and Richardson 2009; Brunnermeier and others
2009). However, their speci�cs differ, and a consensus on the best way to go has
yet to be achieved.
   In the meantime, what can monetary policy do? An emerging view holds that,
even if bubbles cannot be accurately identi�ed, monetary policy can contribute to
�nancial stability by becoming more “symmetric� over the �nancial cycle—that
is, becoming more restrictive during a �nancial market boom, just like it almost
invariably becomes accommodative at times of asset price crashes (Papademos
2009). This view is supported by theoretical models in which monetary policy
reacts to credit growth or other indicators of the �nancial cycle to dampen
boom–bust episodes (Bordo and Jeanne 2002a, 2002b; Christiano and others
2008). In light also of the recent empirical �ndings cited earlier, that monetary
policy affects signi�cantly the volume and the quality of the assets held by lever-
aged �nancial institutions, the implication is that there might be a role for
balance sheet aggregates in the determination of monetary policy. But that role
would be due to reasons of �nancial stability, rather than the classic reasons of
price stability behind the once-popular targeting of monetary aggregates. (Adrian
and Shin 2009) However, at this stage little is known about the complexities of
implementing such policy or about its likely effects, for the obvious reason that it
has never been applied. These questions are currently the subject of active
research.

110                                 The World Bank Research Observer, vol. 25, no. 1 (February 2010)
Capital Controls
The �nancial turmoil of the current crisis propagated rapidly to emerging markets
around the world, which suffered, to varying degrees, sharp increases in external
borrowing premiums and abrupt declines of capital inflows, as well as large falls
in their currencies in the face of a flight to safety by international investors. Thus,
even though the crisis originated in the North, the �nancial symptoms in emer-
ging markets have been similar—albeit in general less acute—to those of the




                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
homegrown crises of the 1990s. Propagation of this �nancial shock—which adds
to the real shock accruing through the global slowdown—has been facilitated by
the deepening of direct and indirect �nancial links across economies brought
about by international �nancial integration.28
   As in previous episodes of global turmoil, the sudden stop in capital flows has
put emerging markets under stress and has hit especially hard countries that
were running large current account de�cits, had developed large currency and
maturity mismatches, or both—as was the case in several Eastern European econ-
omies. In this context, some observers have advocated controls on capital outflows
to stem the creditor run and relieve the pressure on foreign reserves and exchange
rates.29 Indeed, rigorous exchange controls were a key feature of Iceland’s 2008
emergency package (supported by the IMF), amid the collapse of its currency and
banking system. Among emerging markets, only a few have resorted to controls
on outflows so far. Ukraine, Russia, and Belarus introduced some restrictions on
outflows, while Ecuador imposed an exit tax.30
   Controls on capital outflows, which date back to Germany in the 1930s, have
a long tradition as a crisis containment tool. They seek to prevent the disorderly
retreat of investors at times of turmoil31 and thereby protect the stock of reserves
of the Central Bank, relieve pressure on the exchange rate, and provide some
room for monetary policy—which is severely constrained when capital mobility is
high and policymakers also try to pursue exchange rate targets.
   There have been numerous episodes of outflow controls in crisis times—for
example Spain, in the context of the 1992 ERM crisis; Venezuela, at the time of
the banking crisis of 1994; Malaysia and Thailand, on occasion of the 1997 –98
East Asian and Russian crises; and Argentina, at the time of collapse of the
Convertibility Plan in 2001.
   These controls took a wide variety of forms. They frequently targeted “speculat-
ive� transactions and exempted current account transactions, flows related to
foreign direct investment, or both. They ranged from unremunerated deposit
requirements on banks’ foreign asset holdings (Spain) to comprehensive prohibi-
tion of outflows (Venezuela and Argentina). But how effective were they? The
question has been the focus of a multitude of studies: see for example the over-
views by Ariyoshi and others (2000) and Magud, Reinhart, and Rogoff (2007).

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          111
The case of Malaysia is perhaps the one that has attracted the most attention.
Observers agree that the controls succeeded in limiting outflows and segmenting
onshore and offshore markets, providing some breathing space for monetary and
�nancial policy. There is much less agreement regarding the effective contribution
of the controls for easing the cost of the crisis, as some observers contend that the
worst of the global turmoil was already over at the time the controls were insti-
tuted and that Malaysia’s subsequent recovery path was no better than that of
countries that did not introduce controls, while others argue that the controls did




                                                                                                        Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
allow a speedier recovery than would have otherwise occurred given the higher
vulnerability of the Malaysian economy relative to that of other East Asian
countries.32
   Analyses of other episodes of controls on outflows yield mixed conclusions. In
Spain and Thailand, the controls helped contain outflows and relieve pressure on
the exchange rate only temporarily, and eventually both countries had to
abandon their pegs—Spain through an ERM realignment, and Thailand by float-
ing the exchange rate. In Venezuela, the controls did not fully succeed in contain-
ing outflows either, but there is evidence that they helped to ease pressure on the
exchange rate and gain some degree of monetary policy autonomy which,
through lower interest rates, reduced the immediate �scal cost of the country’s
banking crisis.
   A concrete way to assess the action of controls is by examining the differential
between the prices of identical assets traded in onshore and offshore markets.
Among such assets are the American Depositary Receipts (ADRs) issued abroad
by an increasing number of large emerging market �rms. The onshore– offshore
price differentials on ADRs have been recently examined by Levy-Yeyati,
Schmukler, and van Horen (2009) for several episodes of inflow and outflow con-
trols. The �nding is that controls on outflows generate a positive differential
between onshore and offshore prices as investors buy equity at home and sell it
abroad in order to transfer their wealth out of the country (inflow controls have
the opposite effect).33 This suggests that even if capital controls fail to bring out-
flows to a halt, they usually do succeed in temporarily segmenting asset markets,
by creating “no-arbitrage� bands that effectively decouple rates of return at home
from those prevailing abroad.
   However, there is also evidence that this market segmentation weakens with
the passage of time (Kaminsky and Schmukler 2001). The reason is that capital
controls develop leakages as investors �nd arbitrage strategies and loopholes to
circumvent them.34 The search for loopholes at times of crisis is driven by inves-
tors’ anticipations of big return differentials between domestic and foreign assets,
especially in the event of a large exchange rate depreciation. These anticipated
differentials often dwarf the increase in the cost of taking capital out imposed by
the controls. Evasion mechanisms for controls on outflows range from the

112                                  The World Bank Research Observer, vol. 25, no. 1 (February 2010)
traditional overinvoicing of imports and underinvoicing of exports to misreporting
of capital account transactions in order to make them �t under those categories
of flows left unrestricted by the authorities. Deeper international �nancial inte-
gration, as well as a higher degree of development of domestic �nancial markets,
make the enforcement of controls harder, as they broaden the menu of evasion
strategies available to sophisticated (and typically large) investors.35 As a result,
effective enforcement of controls typically requires continuous monitoring and
adaptation efforts by the authorities. In this regard, the international experience




                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
shows that comprehensive controls are easier to enforce because they are not vul-
nerable to evasion through those outflows left unrestricted. However, comprehen-
sive controls also deter “nonspeculative� transactions, such as those related to
foreign direct investment, as was seen in the Malaysia episode.
   Controls on outflows also entail costs in terms of investors’ con�dence. Because
they are usually imposed ex post, as an emergency measure, they typically lead to
violation of explicit or implicit contracts. As a result they may raise the perceived
riskiness of inward investment and reduce access to foreign capital, and this
adverse effect may last well after the removal of the controls on outflows, as inves-
tors considering future capital inflows factor into their risk-return assessment an
increased perception of the likelihood that they may be unable to take their
capital out when needed, at least without cost. While this credibility effect is hard
to quantify, there are strong indications that it was at work in the experience
of Malaysia, in which the evidence suggests that the controls on outflows
subsequently had a deterrent effect on capital inflows, including long-term ones
(Goh 2005).
   Lastly, capital controls create rents whose allocation is at the discretion of
public of�cials and which therefore opens the door to cronyism and corruption
around exceptions and loopholes. In the Malaysia episode, for example, there is
evidence that politically connected �rms did much better under the controls than
unconnected �rms.36
   To summarize, the evidence indicates that controls on capital outflows are
effective only temporarily and are more vulnerable to evasion when they are selec-
tive rather than comprehensive. They also involve costs in terms of investor con�-
dence and augmented potential for rent-seeking. All this, however, does not mean
that controls on outflows should never be used under any circumstances. In fact,
when a run on domestic assets gets under way, the authorities may have no
option but to impose controls to prevent the collapse of the �nancial system, a
free fall of the currency, or both. The lesson instead is that controls should only
be used as a last resort, and only for brief periods to give the authorities a respite
from corrective measures. But the more frequent the resort to controls, and the
longer they are held in place, the more likely they are to become a substitute for
the actual measures and lead eventually to an even bigger crash.

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          113
   From a broader perspective, controls on capital outflows are just a rudimentary
tool of last resort to cope with the vulnerabilities associated with international
�nancial integration. While integration offers countries new opportunities in
terms of access to foreign �nancing and enhanced risk diversi�cation, it also
increases countries’ vulnerability to external shocks—as the current crisis has
clearly shown. How to manage this vulnerability remains a hotly debated issue.37
   Conceptually, the fragilities associated with capital account openness arise from
two main sources: currency mismatches and maturity mismatches. For develop-




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
ing countries, currency mismatches are due to their inability to borrow externally
in domestic currency (the “original sin� of Eichengreen and Hausmann 1999),
which forces them to bear the exchange risk associated with foreign-currency
borrowing.
   In turn, maturity mismatches can leave borrowing countries vulnerable to
creditor runs in a way that is analogous to bank runs. These runs were at the
core of the East Asian crisis and have been a recurrent factor in other emerging
market crises (Broner, Lorenzoni and Schmukler 2007). They played also a key
role in the propagation and ampli�cation of the subprime crisis in the United
States (Brunnermeier 2009). Maturity mismatches may reflect the reluctance of
lenders to enter into long-term contracts when moral hazard and asymmetric
information are pervasive, or when the macro�nancial framework is deemed
fragile. The appeal to public and private borrowers is that in such conditions
short-term borrowing is cheaper than long-term borrowing.38 However, private
external borrowing decisions fail to take into account their social costs—that is
their contribution to raising the aggregate risk of a creditor run or a currency
collapse (Caballero and Krishnamurty 2003; Korinek 2009).
   In fact, major currency and maturity mismatches did arise in a number of
countries in the run-up to the global crisis. For example, in some Eastern
European countries households actively engaged in the “reverse carry trade� by
borrowing in low-interest rate currencies, oblivious to the exchange risk involved;
hence foreign-currency �nancing of residential mortgages became ubiquitous.
Worse yet, corporations often borrowed short term in foreign currency to �nance
long-term domestic-currency assets. Banks lent in foreign currency to borrowers
lacking foreign currency assets or income, effectively replacing currency risk with
credit risk in their balance sheets (Buiter 2009). These mismatches have contribu-
ted to the magni�cation of the adverse impact of the �nancial turbulence trig-
gered off by the crisis.
   The lesson, however, is not that external borrowing should be restricted. In the
case of currency risk, the key is to achieve an ef�cient distribution of currency
risk within the economy by ensuring, through regulatory means, that it is appro-
priately priced and therefore borne by those best able to do so (for example agents
holding foreign currency assets, including exporters). For instance, regulations

114                                 The World Bank Research Observer, vol. 25, no. 1 (February 2010)
should discourage banks from lending foreign currency to borrowers without
foreign currency assets or income. Likewise, adequate prudential regulation of
large �nancial and non�nancial borrowers can help limit maturity mismatches.
Alternatively, the government could contain aggregate exposure to currency and
maturity mismatches by accumulating short-term foreign assets to offset, at least
in part, private sector actions—a strategy adopted by a number of emerging
markets after the East Asia crisis.
    In the longer term, institutional changes supporting credible nominal stability




                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
should help open the door to domestic-currency borrowing, even if only margin-
ally at �rst—as seen in countries like Chile, Mexico, Poland, and South Africa—
thereby helping to mitigate currency mismatches. Credible macro�nancial stab-
ility should also allow enhanced access to borrowing at longer maturities.
    Controls on capital inflows are often advocated as a potential alternative (or a
complement) to prudential regulatory measures for preventing the build-up of
currency and time mismatches. Depending on the speci�cs, controls on inflows
could help deter overborrowing—limiting exchange rate appreciation, asset price
bubbles, and excessive risk-taking at times of large inflows—and/or alter the com-
position of flows against short-term inflows (“hot money�) susceptible to rapid
reversal.
    Since the early 1990s, recourse to controls on capital inflows has become fre-
quent as a number of emerging countries have tried to navigate global inflow
booms. A partial list includes Indonesia (where controls were introduced in
1991), Malaysia (1994), Thailand (1995, and again in 2006), Brazil (1994),
Chile (1991), Colombia (1993, and again since 2004), Mexico (1990), and the
Czech Republic (1992). As in the case of outflows, controls on inflows included a
variety of instruments—from explicit taxes on foreign borrowing by local �rms
and �xed-income investments by foreigners (Brazil) to unremunerated reserve and
minimum-stay requirements for foreign borrowing by domestic �rms (Chile and
Colombia, as well as Thailand over 2006–08)—and limits on banks’ external
borrowing (Indonesia, Malaysia, Thailand, the Czech Republic, and Mexico).
Several of these countries matched the introduction of controls on inflows with a
removal of barriers to outflows and with various reforms in the prudential frame-
work of the domestic �nancial system.
    The case of Chile in the 1990s has been closely scrutinized by a host of empiri-
cal studies. The controls were based on a mandatory unremunerated one-year
deposit (or encaje) at the Central Bank equal to a given fraction (initially set at 20
percent) of eligible inflows. The two distinguishing features of the scheme were (i)
its implicit cost was higher for short-term inflows than for long-term ones,
and (ii) the cost was determined by the prevailing level of world interest rates.
Initially the regime was applied only to foreign loans (except trade credit). As
investors succeeded in �nding a variety of loopholes, it was gradually extended to

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          115
encompass nondebt creating flows as well—even including foreign direct invest-
ments deemed “speculative.� The scheme was eventually eliminated as inflows
evaporated at the time of the 1998 Russian crisis.39
    In spite of the often-heard view (especially outside Chile) that the inflow controls
were highly successful, their effectiveness remains disputed. On the whole, the large
literature concerned with the Chilean experience suggests that (i) the controls were
successful at giving the Central Bank some degree of monetary autonomy, allowing
it to keep domestic interest rates above international levels—although by most esti-




                                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
mates this effect was quantitatively small and temporary; (ii) the composition of
inflows was somewhat altered in favor of longer-term flows—although here the evi-
dence is weak; (iii) no clear result emerges regarding the impact on total capital
inflows; (iv) there was no discernible effect on exchange rates.
    Evidence from other episodes is mixed. Colombia’s unremunerated reserve
requirement, which was in effect over 1993–2000, and then again in 2007, was
designed along the same lines as Chile’s. Moreover, during 2004– 06, as well as
in 2007, Colombia also imposed restrictions on short-term portfolio investment
(initially including its outright prohibition).40 But there is little evidence that
these measures made any difference for the volume of inflows, the share of short-
term flows in the total, or the level of the exchange rate. In contrast, there is
some indication that short-term inflows were reduced in the episodes of Malaysia
and Thailand in the 1990s. In Brazil the controls were seemingly ineffective in all
dimensions. Thailand’s recent imposition of unremunerated reserve requirements
in 2006 was followed by a shift of portfolio inflows away from debt instruments
and toward equity, and a substantial differential between offshore and onshore
interest rates.41
    As with controls on outflows, these episodes show that controls on inflows
quickly develop leakages. The most common evasion mechanism was the mis-
statement of the purpose of the inflow; see Carvalho and Garcia (2006) for the
case of Brazil and Nadal-de-Simone and Sorsa (1999) for Chile. The result was a
weakening of the controls over time—especially rapid in the case of Brazil—in
spite of the authorities’ constant efforts to close loopholes and deter unwanted
flows.
    Controls on capital inflows also entail other costs. Persistent barriers to capital
inflows may deter the development of local �nancial markets, which in turn may
hamper ef�ciency and growth.42 And they also raise domestic �nancing costs by
allowing local interest rates to remain above international levels. Microeconomic
studies have found large adverse effects of Chile’s higher domestic interest rates
on the availability and cost of �rms’ �nancing.43 Moreover, because larger �rms
could better afford the cost of evasion strategies, or had direct access to foreign
�nancing, the cost of the controls—in the form of a higher cost of borrowing—
was disproportionately borne by small and medium-sized �rms.

116                                   The World Bank Research Observer, vol. 25, no. 1 (February 2010)
   Finally, what about the ability of controls on capital inflows to prevent volatility
and crises, which was the main rationale for their use? Overall, there is not much
evidence that they were of great help in this regard, although identifying the
counterfactual is admittedly hard. Many observers have noted that the encaje did
not prevent Chile from suffering a major sudden stop on occasion of the Russia
crisis in 1998, although the scheme may have helped Chile mitigate the
“normal� ups and downs of capital inflows in the 1990s (Edwards 1999). In ret-
rospect perhaps this should not be surprising. At times of acute �nancial turmoil,




                                                                                         Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
the run for the exit is not limited to the unwinding of short-term domestic asset
positions—whose build-up the encaje and similar mechanisms try to deter—but
includes also that of longer-term ones, and involves domestic and international
investors alike. The only deterrent to investor exit then is the possibility that the
entry cost would have to be incurred again in the future, should investors decide
to rebuild their positions. But in past crisis episodes most countries have in fact
dismantled their inflow controls, so not even this deterrent remains. This puts
into perspective the capacity of inflow controls to deliver stability in times of
extreme �nancial turmoil such as is the case at present.



Conclusions
The intensity and the breadth of the on-going �nancial crisis have surprised
nearly everyone. Perhaps, more importantly, the policy responses to the crisis
have led to considerable confusion and shaken the con�dence of the development
community in the wisdom of �nancial and macroeconomic policies that underpin
Western capitalist systems.
   We have drawn on a large body of analytical research, econometric evidence,
and country experience to argue that the “sacred cows� of �nancial and macropo-
licies are still very much alive. For the most part, the confusion arises from not
being able to recognize incentive conflicts and trade-offs inherent in short-term
and long-term responses to a systemic crisis. Policies employed to contain a
crisis—often in haste to re-establish con�dence and with disregard of long-term
costs—should not be interpreted as permanent deviations from well-established
policy positions. The fact that governments may end up providing blanket guaran-
tees or owning large stakes in the �nancial sector in an effort to contain and deal
with the crisis does not negate the fact that generous guarantees over the long
term are likely to back�re or that government of�cials make poor bankers.
   Financial crises often do expose weaknesses in the underlying incentive frame-
works and the regulation and supervision systems that are supposed to reinforce
them. But �nance is risky business and it is naive to think that regulation and
supervision can, or should, completely eliminate the risk of crises, although they

Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                          117
can make crises less frequent and less costly. Neither monetary policy nor capital
controls can substitute for well-designed prudential regulation.
  Despite their inherent fragility, �nancial systems underpin economic develop-
ment. The challenge of �nancial sector policies is to align private incentives with
public interest without taxing or subsidizing private risk-taking. Public ownership
or too aggressive regulation would simply hamper �nancial development and
growth. But striking this balance is becoming increasingly complex in an ever
more integrated and globalized �nancial system.




                                                                                                                Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
Notes
Aslı Demirgu    ¸ -Kunt is Chief Economist, Financial and Private Sector Development Network and
               ¨c
Senior Research Manager, Development Research Group, the World Bank, 1818 H Street NW                  ,
Washington DC 20433, USA; email address: Ademirguckunt@worldbank.org. Luis Serve           ´ n is Senior
Adviser for Macroeconomics and Growth, the World Bank; email address: Lserven@worldbank.org.
Useful comments from Alan Gelb, Edward Kane, Michael Klein, Justin Lin, Martin Ravallion, and
Sergio Schmukler are gratefully acknowledged. Annamaria Kokenyne kindly contributed infor-
mation on recent capital control measures. The authors’ �ndings, interpretations, and conclusions
do not represent the views of the World Bank, their Executive Directors, or the countries they
represent.
    1. Honohan and Laeven (2005) bring together research and �rst-hand crisis experience to
catalog lessons on a variety of issues that regularly arise in crises.
    2. Hungary and Jordan are such examples.
    3. See for example Switzerland, the Czech Republic, Poland, and Slovakia.
    4. See ‘Design of Prudential Regulations’ below for further discussion of these instruments and
markets.
    5. See Demirgu    ¸ -Kunt and Detragiache (1998, 2002), Demirgu
                     ¨c                                                  ¸ -Kunt and Huizinga (2004),
                                                                        ¨c
Demirgu ¨c¸ -Kunt and Kane (2002), and Demirgu     ¸ -Kunt, Kane, and Laeven (2008a, 2008b).
                                                  ¨c
    6. Argentina’s reduced deposit mobilization after its 2001 deposit freeze is an example.
    7. “Bank Bail-outs: Quids pro-quo,� The Economist, November 22, 2008, pp.84–6.
    8. Indonesia, the Philippines, South Korea, Mexico, and Chile are examples. See Caprio and
Klingebiel (1996) for a complete list.
    9. Whether or not the control of the bank completely passes into public hands during the resol-
ution stage is secondary to the fact that after the resolution the bank should be well-capitalized.
Problems often emerge because governments are unable to do the �nancial restructuring properly;
and banks that continue to operate with insuf�cient capital have every incentive to resume reckless
risk-taking.
    10. See World Bank (2001) for a discussion.
    11. Similar trade-offs are considered in liberalizing �nancial systems, since premature liberaliza-
tion without adequate attention to developing institutions increases �nancial fragility. See
Demirgu ¨c¸ -Kunt and Detragiache (1998).
    12. Notice there was also signi�cant private equity injected into �nancial institutions in the
latest crisis. The �gures reached US$500 billion by September, 2008, which is larger than in any
other crisis in history.
    13. It is dif�cult to evaluate the success of programs such as the RFC, but in this case it is cred-
ited with having contributed to a recovery of con�dence and output (until monetary tightening
reversed both in 1937). The government recovered its initial capital and did not bail out nonviable
banks. See Mason (2000) for an analysis.


118                                          The World Bank Research Observer, vol. 25, no. 1 (February 2010)
    14. The Basel committee itself notes that Basel II arose in response to the growth of securitiza-
tion, giving the impression that this growth was an exogenous development rather than at least in
part a response to Basel I and its “loan-by-loan� approach to assessing a �rm’s overall risk exposure.
    15. See Caprio, Demirgu   ¸ -Kunt, and Kane (2008) for an in-depth discussion of Basel I and Basel
                             ¨c
II accords and recommendations for reforms.
    16. See Caprio, Demirgu   ¨c¸ -Kunt, and Kane (2008) and statements of the Shadow Financial
Regulatory Committee, which can be found on the American Enterprise Institute (AEI) website.
    17. For a discussion of tensions between the goals of expanding access and maintaining stability,
see World Bank (2007).
    18. These include proposals by Krugman (2008), the Institute for International Finance, the




                                                                                                          Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
Counterparty Risk Management Policy Group (2008), and the Financial Stability Forum. Goodhart
(2008a) and Goodhart and Persuad (2008) recommend authorities set countercyclical capital
requirements, introducing a leverage ratio that would move inversely to the business cycle.
    19. See American Enterprise Institute statements and Caprio, Demirgu    ¸ -Kunt, and Kane (2008).
                                                                           ¨c
    20. Investigating the impact of compliance with Basel Core Principals on bank soundness,
Demirgu ¨c¸ -Kunt, Detragiache, and Tressel (2008) show that compliance with information provision
and transparency is the most robustly associated with the �nancial strength of institutions.
    21. Subordinated debt has advantages in terms of improving market discipline since subordi-
nated debt holders have incentives to negotiate covenants that would protect their interests.
However, credit default swaps (CDS) have advantages in tracking risk since, unlike subordinated
debt (which are illiquid instruments), CDS are continuously traded, providing useful signals on a
timely basis.
    22. Indeed, bubbles can be welfare-enhancing to the extent that they provide stores of value that
would otherwise not be available to savers. The best example is that of �at money in Samuelson’s
overlapping-generations model; see also Ventura (2004) for a recent open-economy example.
    23. Goodhart (2008b).
    24. This view is also stated by Mishkin (2008).
    25. How fast should asset prices rise for us to conclude safely that a bubble is at work? Too low
a threshold would lead to the (incorrect) detection of too many bubbles. Even modest changes in
key fundamentals—such as investors discount rates or their anticipations of the rate of growth of
future dividends—can lead to large changes in the prices of long-lived assets over short periods. In
turn, too high a threshold (say 20 or 40 percent per year) could fail to detect many bubbles
altogether. For example during the recent U.S. housing bubble real-estate price increases remained
consistently below 15 percent per annum, according to the Case-Shiller U.S. national home price
index (see http://www.standardandpoors.com/home/en/us).
    26. The role of housing prices is stressed by Bordo and Jeanne (2002a, 2002b) and Cecchetti,
Genberg, and Wadhwani (2003).
    27. Goodhart (2008b).
    28. These links have been documented by extensive theoretical and empirical research. See for
example Calvo (1999), Broner, Gelos, and Reinhart (2006), and Didier, Schmukler, and Mauro
(2008).
    29. Calls for a reconsideration of controls on capital outflows in the context of the current crisis
span a broad range of observers, from leading academics like Calvo (2008) and Subramaniam and
Rodrik (2008) to the World Council of Churches (2008).
    30. Also, Indonesia and China tightened reporting requirements. Argentina has also recently
sought to prevent sales of American Depositary Receipts abroad by domestic residents.
    31. In this sense, controls on outflows are similar in spirit to a bank holiday or a suspension of
stock market trading.
    32. See Ariyoshi and others (2000), Dornbusch (2001), Kaplan and Rodrik (2001) and Magud,
Reinhart, and Rogoff (2007).
    33. Notice that these transactions do not entail a net capital outflow or a reserve loss, as the
gross outflow implied by the nonrepatriation of the proceeds of the asset sale abroad is matched by


Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                                           119
the gross inflow implied by its purchase with foreign funds, without any net effect on the foreign
reserve stock; see Levy-Yeyati, Schmukler, and van Horen (2004). However, this kind of inter-
national asset migration may have consequences for the tax collection capacity of the local auth-
orities, as domestic residents effectively shift assets out of the domestic jurisdiction.
    34. This applies also to the case of inflow controls; see Carvalho and Garcia (2006) for a detailed
account of evasion strategies employed by investors in Brazil.
    35. As evasion strategies often involve setup costs (as well as the possibility of penalties), they
are more easily available to large, wealthy and/or well-connected investors. In fact, evidence from
several crisis episodes in Latin America shows that large investors typically liquidate domestic assets
ahead of the full-blown crisis and the imposition of controls, so that the burden of controls falls




                                                                                                                Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
mainly on small investors—while flight from domestic assets by large investors helps trigger the
eventual crash. This is one of the mechanisms identi�ed by Halac and Schmukler (2004) through
which the cost of �nancial crises is disproportionately borne by small investors.
    36. Johnson and Mitton (2003).
    37. The perils of �nancial integration in a world of imperfect markets have been stressed by
many distinguished observers, going as far back as Tobin (1978), and including Harberger (1980),
Diaz-Alejandro (1985), Bhagwati (1998), Cooper (1998), and Stiglitz (2002).
    38. Broner, Lorenzoni and Schmukler (2007) show that this maturity premium is quite substan-
tial in the case of emerging markets.
    39. Nadal-de-Simone and Sorsa (1999) provide a detailed description of Chile’s controls on
inflows.
    40. Concha, Galindo, and Quevedo (2008) provide a detailed analysis of Colombia’s experience
with controls on capital inflows.
    41. McCauley (2008) documents Thailand’s recent experience, as well as those of other Asian
countries.
    42. See for example Bekaert, Harvey and Lundblad (2003).
    43. Forbes (2007).



References
Acharya, A., and M. Richardson. 2009. Restoring Financial Stability: How to Repair a Failed System.
  New York: Wiley.
Adrian, Tobias, and Hyun Shin. 2009. “Money, Liquidity and Monetary Policy.� American Economic
  Review 99(2): 600 –5.
Ariyoshi, A., K. Habermeier, B. Laurens, and I. Otker-Robe. 2000. “Capital Controls: Country
   Experiences with their Use and Liberalization.� IMF Occasional Paper 190.
Assenmacher-Wesche, K., and S. Gerlach. 2008. “Ensuring Financial Stability: Financial Structure
   and the Impact of Monetary Policy on Asset Prices.� University of Zurich Institute for Empirical
   Research in Economics Working Paper 361.
Baer, Herb, and Daniela Klingebiel. 1995. “Systemic Risk when Depositors Bear Losses: Five Case
   Studies.� In G. Kaufman, ed, Research in Financial Services: Private and Public Policy, vol. 7.
   Stanford, CT: JAI Press: 195–302.
Bagehot, Walter. 1894. Lombard Street. London: Kegan, Paul, Trench, Trubner & Co.
Barth, James R., Gerard Caprio, Jr., and Ross Levine. 2001. “Banking Systems around the Globe: Do
   Regulations and Ownership Affect Performance and Stability?� In Frederic Mishkin, ed., Prudential
   Supervision: What Works and What Doesn’t. Chicago: University of Chicago Press: 31–96.
                   ¨c
Beck, T., A. Demirgu ¸ -Kunt, and M. Martinez Peria. 2007. “Reaching Out: Access to and Use of
   Banking Services across Countries.� Journal of Financial Economics 85:234 –66.


120                                          The World Bank Research Observer, vol. 25, no. 1 (February 2010)
      . 2008. “Banking Services for Everyone? Barriers to Bank Access and Use around the
   World.� World Bank Economic Review, 22(3): 397– 430.
Bekaert, G., C. Harvey, and C. Lundblad. 2003. “Equity Market Liberalization in Emerging Markets.�
   Journal of Financial Research 26:275 –99.
Bernanke, B., and M. Gertler. 2000. “Monetary Policy and Asset Price Volatility.� NBER Working
   Paper 7559.
Bhagwati, J. 1998. “The Capital Myth: The Difference between Trade in Widgets and Trade in
  Dollars.� Foreign Affairs 77:7–12.
Bordo, M., and O. Jeanne. 2002a. “Monetary Policy and Asset Prices: Does ‘Benign Neglect’ Make




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
   Sense?� International Finance 5:139–64.
       . 2002b. “Boom-busts in Asset Prices, Economic Instability and Monetary Policy.� CEPR
   Discussion Paper 3398.
Broner, F., G. Gelos, and C. Reinhart. 2006. “When in Peril, Retrench: Testing the Portfolio Channel
   of Contagion.� Journal of International Economics 69:203– 30.
Broner, F., G. Lorenzoni, and S. Schmukler. 2007. “Why Do Emerging Markets Borrow Short-term?�
   NBER Working Paper 13076.
Brunnermeier, M. 2009. “Deciphering the Liquidity and Credit Crunch 2007– 08.� Journal of
   Economic Perspectives 23(1): 77 –100.
Brunnermeier, M., A. Crockett, C. Goodhart, A. Persaud, and H. Shin. 2009. “The Fundamental
   Principles of Financial Regulation.� Geneva Reports on the World Economy 11.
Buiter, W. 2009. “The Return of Capital Controls. (http://www.voxeu.org/index.php?q=node/3104).
Caballero, R., and A. Krishnamurty. 2003. “Excessive Dollar Debt: Financial Development and
   Underinsurance.� Journal of Finance 58(2): 867 –94.
Calomiris, Charles, and Charles Kahn. 1996. “The Ef�ciency of a Self-Regulated Payments System:
   Learning from the Suffolk Banking System.� Journal of Money, Credit and Banking 28(4): 766 –97.
Calomiris, Charles, Daniela Klingebiel, and Luc Laeven. 2005. “Financial Crisis Policies and
   Resolution Mechanisms: A Taxonomy from Cross-Country Experience.� In Patrick Honohan,
   and Luc Laeven, eds., Systemic Financial Crises: Containment and Resolution. New York: Cambridge
   University Press: 25 –75.
Calvo, G.. 1999. “Contagion in Emerging Markets: When Wall Street Is a Carrier.� (http://www.bsos.
   umd.edu/econ/ciecrp8.pdf ).
       . 2008. “A New Bretton Woods System Should Curb Boom and Bust.� In R. Baldwin, and B.
   Eichengreen, eds., What G20 Leaders Must Do to Stabilize our Economy and Fix the Financial
   System. (http://www.voxeu.org/index.php?q=node/2543).
Caprio, G., and D. Klingebiel. 1996. “Dealing with Bank Insolvencies: Cross Country Experience.�
  Policy Research Working Paper 1620, The World Bank, Washington, D.C. (updated database pub-
  lished in Honohan and Laeven 2005).
Caprio, G., and Maria Soledad Martinez Peria. 2002. “Avoiding Disaster: Policies to Reduce the Risk
  of Banking Crises.� In E. Cardoso, and A. Galal, eds., Monetary Policy and Exchange Regimes:
  Options for the Middle East. The Egyptian Center for Economic Studies, Cairo, Egypt.
Caprio, G., Aslı Demirgu¨c¸ -Kunt, and Edward Kane. 2008, “The 2007 Meltdown in Structured
  Securitization: Searching for Lessons not Scapegoats.� World Bank Working Paper 4756.
Carvalho, B., and M. Garcia. 2006. “Ineffective Controls on Capital Inflows under Sophisticated
   Financial Markets: Brazil in the Nineties.� NBER Working Paper 12283.




Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                                        121
Cecchetti, S., H. Genberg, and S. Wadhwani. 2003. “Asset Prices in a Flexible Inflation Targeting
   Framework.� In W    . Hunter, G. Kaufman, and M. Pomerleano, eds., Asset Price Bubbles.
   Cambridge, MA: MIT Press.
Christiano, L., C. Ilut, R. Motto, and M. Rostagno. 2008. “Monetary Policy and Stock Market
  Bubbles.� ECB Working Paper 955.
Claessens, Stijn, Daniela Klingebiel, and Luc Laeven. 2005. “Crisis Resolution, Policies and
   Institutions: Empirical Evidence.� In Patrick Honohan, and Luc Laeven, eds., Systemic Financial
   Crises: Containment and Resolution. New York: Cambridge University Press: 169– 96.
Clarke, George, Robert Cull, and Mary Shirley. 2005. “Bank Privatization in Developing Countries: A




                                                                                                               Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
   Summary of Lessons and Findings.� Journal of Banking and Finance 29:1905–30.
Cole, S. 2004. “Fixing Market Failures or Fixing Elections? Elections, Banks, and Agricultural Credit
   in India.� Harvard Business School, Cambridge, MA.
Concha, A., A. Galindo, and C. Quevedo. 2008. “An Assessment of Another Decade of Capital
  Controls in Colombia: 1998 –2008.� Unpublished manuscript.
Cooper, R. 1998. “Should Capital Account Convertibility Be a World Objective?� Princeton Essays in
   International Finance 207.
Counterparty Risk Management Policy Group. 2008. “Containing Systemic Risk: The Road to
  Reform.� The Report of the CRMPG III, August. (http://www.crmpolicygroup.org/index.html).
De la Torre, Augusto, Juan Carlos Gozzi, and Sergio Schmukler. 2007. “Innovative Experiences in
   Access to Finance: Market Friendly Roles for the Visible Hand?� Policy Research Working Paper
   4326, World Bank, Washington, DC.
      ¨c
Demirgu ¸ -Kunt, Aslı. 2008. “Resolution of Systemic Financial Problems.� Viewpoint Note, World
  Bank, September.
      ¨c
Demirgu ¸ -Kunt, Aslı, and Enrica Detragiache. 1998. “The Determinants of Banking Crises: Evidence
  from Developing and Developed Countries.� IMF Staff Papers 45:81 –109.
       . 2002. “Does Deposit Insurance Increase Banking System Stability? An Empirical
   Investigation.� Journal of Monetary Economics 49:1373 –406.
      ¨c
Demirgu ¸ -Kunt, Aslı, and Harry Huizinga. 2004. “Market Discipline and Deposit Insurance.�
  Journal of Monetary Economics 51(2): 375 –99.
      ¨c
Demirgu ¸ -Kunt, Aslı, and Edward Kane. 2002. “Deposit Insurance around the Globe? Where Does It
  Work?� Journal of Economic Perspectives 16(2): 175– 95.
      ¨c
Demirgu ¸ -Kunt, Aslı, Enrica Detragiache, and T. Tressel. 2008. “Banking on the Principles:
  Compliance with Basel Core Principles and Bank Soundness.� Journal of Financial Intermediation
  17(4): 511 –42.
Demirgu ¸ -Kunt, Aslı, Edward Kane, and Luc Laeven. 2008a. Deposit Insurance around the World:
       ¨c
  Issues of Design and Implementation. Cambridge, MA: MIT Press.
       . 2008b. “Determinants of Deposit Insurance: Adoption and Design.� Journal of Financial
   Intermediation 17(3): 407– 38.
Diamond, Douglas W      ., and Philip H. Dybvig. 1983. “Bank Runs, Deposit Insurance, and Liquidity.�
   Journal of Political Economy 91:401 –19.
Diaz-Alejandro, C. 1985. “Good-bye Financial Repression, Hello Financial Crash.�., Journal of
   Development Economics 19, 1– 24.
Didier, T., S. Schmukler, and P. Mauro. 2008. “Vanishing Contagion?� Journal of Policy Modeling
   30(5): 775 –91.
   ¸ , Serdar. 2005. “Politicians and Banks: Political Influences on Government-Owned Banks in
Dinc
   Emerging Countries.� Journal of Financial Economics 77(2): 453–79.


122                                         The World Bank Research Observer, vol. 25, no. 1 (February 2010)
Dornbusch, R. 2001. “Malaysia: Was it different?� NBER Working Paper 8325.
Edwards, S. 1999. “How Effective Are Capital Controls?� Journal of Economic Perspectives 13:65 –84.
Eichengreen, B., and R. Hausmann. 1999. “Exchange Rates and Financial Fragility.� Proceedings,
   Federal Reserve Bank of Kansas City: 329– 68.
Forbes, K. 2007. “One Cost of the Chilean Capital Controls: Increased Financial Constraints for
   Smaller Traded Firms.� Journal of International Economics 71:294 –323.
Gerschenkron, Alexander. 1962. Economic Backwardness in Historical Perspective: A Book of Essays.
   Cambridge, MA: Belknap Press of Harvard University Press.
Goh, S. 2005. “New Empirical Evidence on the Effects of Capital Controls on the Composition of




                                                                                                       Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
  Capital Flows in Malaysia.� Applied Economics 37:1491–503.
Goodhart, C. 2008a. “The Regulatory Response to the Financial Crisis.� CESIfo Working Paper
  2257.
       . 2008b. “Central Banks’ Function to Maintain Financial Stability: An Uncompleted Task.�
   In R. Baldwin, and B. Eichengreen, eds., What G20 Leaders Must Do to Stabilize our Economy and
   Fix the Financial System. (http://www.voxeu.org/index.php?q=node/2543).
Goodhart, C., and Avinash Persaud. 2008. “ A Party Pooper’s Guide to Financial Stability.� Letter to
  the Editor, The Financial Times, June 4.
Greenspan, A. 2004. “Risk and Uncertainty in Monetary Policy.� American Economic Review 94:
   33 –40.
Halac, Marina, and Segio Schmukler. 2004. “Distributional Effects of Crises: The Financial
  Channel.� Economia Journal of the Latin American and Caribbean Economic Association 5:1 –67.
Harberger, Arnold C.. 1980. “Vignettes on the World Capital Market.� American Economic Review
  70:331 –7.
Honohan, Patrick. 2005. “Fiscal, Monetary and Incentive Implications of Bank Recapitalization.� In
  Patrick Honohan, and Luc Laeven, eds., Systemic Financial Crises: Containment and Resolution.
  New York: Cambridge University Press: 25 –75.
Honohan, Patrick, and Daniela Klingebiel. 2003. “The �scal Cost Implications of an
  Accommodating Approach to Banking Crises.� Journal of Banking and Finance 27:1539– 60.
Honohan, Patrick, and Luc Laeven. 2005. Systemic Financial Crises: Containment and Resolution.
  New York: Cambridge University Press.
Johnson, S., and T. Mitton. 2003. “Cronyism and Capital Controls: Evidence from Malaysia.� Journal
   of Financial Economics 67:351 –82.
Kaminsky, G., and S. Schmukler. 2001. “Short- and Long-run Integration: Do Capital Controls
  Matter?� In Susan Collins, and Dani Rodrik, eds., Brookings Trade Forum 2000. Washington, DC:
  Brookings Institution Press: 125–78.
Kane, Edward, and Daniela Klingebiel. 2004. “Alternatives to Blanket Guarantees for Containing a
  Systemic Crisis.� Journal of Financial Stability 1:31 –63.
Kaplan, E., and D. Rodrik. 2001. “Did the Malaysian Capital Controls Work?� NBER Working Paper
  8142.
Kashyap, A., R. Rajan, and J. Stein. 2008. “Rethinking Capital Regulation.� Unpublished
  Manuscript.
Keefer, Philip. 2000. “When Do Special Interests Run Rampant? Disentangling the Role of Elections,
   Incomplete Information and checks and Balances in Banking Crises.� Policy Research Working
   Paper 2543, World Bank, Washington, DC.
Khwaja, Asim Ijaz, and Atif Mian. 2005. “Do Lenders Favor Politically Connected Firms? Rent
  Provision in an Emerging Financial Market.� Quarterly Journal of Economics 120(4): 1371–411.


Demirgu
      ¨c               ´n
        ¸-Kunt and Serve                                                                        123
Klingebiel, Daniela. 2000. “The Use of Asset Management Companies in the Resolution of Banking
   Crises: Cross-Country Experience.� World Bank Policy Research Working Paper 2284,
   Washington, DC.
Korinek, A. 2009. “Excessive Dollar Borrowing in Emerging Markets.� Unpublished Manuscript.
Krugman, Paul. 2008. “Partying Like it’s 1929.� New York Times, March 21.
Laeven, Luc, and Fabian Valencia. 2008. “The Use of Blanket Guarantees in Banking Crises.� IMF
   Working Paper WP/08/250.
La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer. 2002. “Government Ownership of
   Commercial Banks.� The Journal of Finance 57(1): 265–301.




                                                                                                                Downloaded from http://wbro.oxfordjournals.org at International Monetary Fund on March 10, 2010
Levy-Yeyati, E., S. Schmukler, and N. van Horen. 2004. “The Price of Inconvertible Deposits: The
   Stock Market Boom During the Argentine Crisis.� Economics Letters 83:7 –14.
      . 2009. “International Integration through the Law of One Price: The Role of Liquidity and
   Capital Controls.� Journal of Financial Intermediation 18:423 –63.
                       ´ , and S. Scopei. 2008. “Does Monetary Policy Affect Bank Credit
Maddaloni, A., J. Peydro
  Standards?� Unpublished Manuscript.
Magud, N., C. Reinhart, and K. Rogoff. 2007. “Capital Controls: Myth and Reality.� Unpublished
  Manuscript.
Mason, Joseph R. 2000. “Reconstruction Finance Corporation Assistance to Financial
  Intermediaries and Commercial and Industrial Enterprise in the U.S., 1932– 1937.� World Bank,
  Washington DC.
McCauley, R.N. 2008. “Managing Recent Hot Money Inflows in Asia.� ADB Institute Discussion
  Paper 99.
Megginson, William. 2005. “The Economics of Bank Privatization.� Journal of Banking and Finance
  29:1931– 80.
Mishkin, R. 2008. “How Should we Respond to Asset Price Bubbles?� Speech delivered at the
   Wharton Financial Institutions Center, May 15.
Nadal-de Simone, F., and P. Sorsa. 1999. “A Review of Capital Account Restrictions in Chile in the
  1990s.� IMF Working paper 99/52.
Papademos, L. 2009. “Monetary Policy and the ‘Great Crisis’: Lessons and Challenges.� (http://
   www.ecb.int/press/key/date/2009/html/sp090514.en.html.).
Stiglitz, J. 2002. Globalization and its Discontents. New York, NY: W.W. Norton.
Subramaniam, A., and D. Rodrik. 2008. “We Must Curb International Flows of Capital.� Financial
  Times, February 25.
Tobin, J. 1978. “A Proposal for International Monetary Reform.� Eastern Economic Journal 4:154–9.
Ventura, J. 2004. “Bubbles and Capital Flows.� Unpublished Manuscript.
World Bank. 2001. Finance for Growth: Policy Choices in a Volatile World. Policy Research Report,
  Washington DC.
      . 2007. Finance for All? Policies and Pitfalls in Expanding Access. Policy Research Report,
   Washington DC.
World Council of Churches. 2008. “14-11-08 Statement on the Global Financial Crisis.� (http://
  www.oikoumene.org/index.php?id=6419).
Wyplosz, C. 2009. “What Do we Know about Monetary Policy that Friedman Did Not Know?�
  Unpublished Manuscript.




124                                          The World Bank Research Observer, vol. 25, no. 1 (February 2010)