preventing financial crises in developing countries...the savings and loan crisis in the united...

FINANCIAL CRISES OCCUR WHEN FINANCIAL SYSTEMS BECOME ILLIQUID

or insolvent. Such crises have recurred throughout the history ofcapitalism. A collapse in investor confidence, usually after aperiod of market euphoria, marks such crises—examples include

the Dutch tulip mania crisis of 1637–38, the Indian cotton futures marketcrash of 1866, and the Great Depression of 1929. When foreign lendersare involved, cross-border payments problems arise as well.

The East Asian crisis belongs to the class of twin financial crises, in-volving both banking and currency problems. According to modern eco-nomic theory, information asymmetries and financial market failures arecentral in explaining macroeconomic fluctuations and financial crises.1

Because lenders know less than borrowers about the use of their funds andcannot compel borrowers to act in the lenders’ best interests, lenders canpanic and withdraw their funds when they perceive increased risks, in theabsence of adequate public regulation and safeguards. That can trigger

121

.

3Preventing

Financial Crises in

Developing Countries

G L O B A L E C O N O M I C P R O S P E C T S

much wider financial crises, with spiralingreal-sector effects. The costs can be severe.Such crises can bring down the financialsystem, cause asset prices to collapse, andbankrupt sound as well as unsound banksand corporations. The East Asian crisis isexpected to cause output in Indonesia, theRepublic of Korea, and Thailand to drop12–24 percent in 1998 (from previoustrend levels), throwing millions into unem-ployment and poverty.

Over the past 100 years industrialcountries have reduced the incidence andseverity of systemic crises through publicpolicy and institutional reforms. They havenot eliminated them entirely, however, asthe savings and loan crisis in the UnitedStates in the 1980s, the banking crises inNordic countries in the early 1990s, and theunfolding financial sector problems in Japanillustrate. In developing countries there isoften a mismatch between public policiesand the institutional structures (which areslow to change) intended to prevent finan-cial crises, and their integration with worldfinancial markets. Thus the number of suchcrises remains large and their costs havebeen growing. Reducing their incidence callsfor policy and institutional reforms in bothnational and international settings.

Until the surge in private capital flowsin the 1990s, most crises in developing

countries (including the sovereign debt cri-sis of the 1980s in Latin America) stemmedfrom macroeconomic mismanagement,including excessive public deficits and over-borrowing abroad. As evident in the lightof recent events in Russia, reforms andpolicies to avoid such sovereign debt crisesare important and still relevant in develop-ing countries. The focus of this chapter,however, is on the type of crisis whichinvolves private-to-private capital flows,and the role of domestic and internationalfinancial systems in intermediating suchflows. The international setting is impor-tant because these crises (East Asia in 1997,Mexico in 1994, and Chile in 1982) areclosely connected to rapidly rising cross-border private capital flows. These flowshave grown massively in the past decade,but without the improvements in institu-tions and public regulation needed for theirsafe management.

The analysis of financial crises and theappropriate policies needed to preventthem highlights the way various factorsinteract and amplify risks. These includeinadequate macroeconomic policies, surgesin capital flows, fragility of domestic finan-cial systems, weak corporate governance,and ill-prepared financial and capitalaccount liberalization. Policymakers needto be concerned about these interactions inthe sequencing and timing of policy andinstitutional reforms.

This chapter’s key messages:• The number and costs of financial

crises have risen in developing coun-tries since the 1980s, partly because rel-atively small economies are moreexposed to the risks of internationalcapital flow reversals. Many recent

122

Over the past 100 years industrialcountries have reduced the incidenceand severity of systemic crises throughpublic policy and institutionalreforms—although they have noteliminated crises entirely.

P R E V E N T I N G F I N A N C I A L C R I S E S I N D E V E L O P I N G C O U N T R I E S

crises are in fact both currency andbanking crises, including the East Asiancrisis (1997) and the Mexican peso cri-sis (1994). Developing countries haverecently been exposed to a wave of cap-ital inflows but have little experiencewith the institutional and prudentialsafeguards needed to minimize associ-ated risks. The easier availability ofcheap international capital in goodtimes encourages excessive private risktaking, which can turn into a majorproblem when favorable financial sen-timent erodes. The institutions neededto minimize the risks of such crises takea long time to develop, while the politi-cal constraints on prompt policyactions to avert them are often severe.However, the building of such requiredinstitutions and safeguards needs toproceed vigorously in all countries, sothat the potential benefits of globaliza-tion can be realized with fewer risks.

• Poor macroeconomic policies leave acountry vulnerable to financial crisis,and prudent policies are the first line ofdefense. In the presence of large capitalinflows and weak financial systems,however, the room for maneuver in set-ting appropriate macroeconomic poli-cies to control excessive private bor-rowing and risk taking is constrainedbecause of the presence of numeroustradeoffs and their distributional conse-quences. Fixed or pegged exchangerates help anchor expectations andreduce uncertainty. But they may alsoprovide unintended incentives to theprivate sector to overborrow (as inThailand), while sterilizing capitalinflows may be costly and ineffective

and shift the composition to short-termand volatile inflows. Flexible exchangerates help regain autonomy for mone-tary policy, improve risk perceptions,and reduce incentives for excessive bor-rowing, but they are not always enoughto avoid crises and may result in volatileand misaligned real exchange rates.Countercyclical fiscal policy is impor-tant, but it too has tradeoffs (fewerschools and roads, for instance, toaccommodate more shopping malls andoffice towers). What is needed is a mul-tidimensional approach, often withmore flexible exchange rates, greaterreliance on fiscal policy, and better andtighter domestic financial regulation(and, where necessary, restrictions oncapital flows) to reduce excessive capi-tal inflows, domestic lending booms,and risks of financial crises.

• Financial sector liberalization, whichcan significantly boost the risk of crisis(particularly in conjunction with opencapital accounts), should proceed care-fully and in step with the capacity todesign and enforce tighter regulationand supervision. At the same time,however, efforts to improve prudentialsafeguards and banking operations willneed to be accelerated. There is strongevidence of a higher probability of cri-sis following liberalization withoutstepped-up prudential safeguards (evenin industrial countries). Regulationsthat increase safety and stability areneeded. Banking and capital marketreforms, oriented toward better riskmanagement, are critical in any strat-egy to prevent financial crises. Publicpolicy and institutional reforms that

123


clamp down on connected bank lend-ing and improve corporate governanceare equally essential to support thesafety of the financial system.

• Capital account liberalization shouldalso proceed cautiously, in an orderlyand progressive manner, given the largerisks of financial crises—heightened byinternational capital market failures—in developing countries. Benefits ofcapital account liberalization andincreased capital flows have to beweighed against the likelihood of crisesand their costs. Clearly the benefitsfrom foreign direct investment (FDI)and longer-term capital inflows out-weigh the costs associated with theincreased likelihood of financial crisis,and developing countries should pur-sue a policy of openness. But for morevolatile debt portfolio and interbankshort-term debt flows and the relatedpolicy of full capital account convert-ibility, there are higher associated risksof financial crisis and greater uncer-tainty about the benefits. Tighter pru-dential regulations on banks, and,where the domestic regulatory and pru-dential safeguards are weak, restric-tions on more volatile short-terminflows that minimize distortions andare as market-oriented as possible(through taxes, for instance), mayreduce the risk of financial crisis. Forcountries that are reintroducing suchrestrictions on capital inflows, theseactions will need to be managed care-fully so as not lead to a loss of confi-dence; their reintroduction for capitaloutflows during a crisis may pose diffi-cult problems (not considered here).

• Changes are needed in the architectureof the international financial system inview of the excessive volatility (euphoriaand panics), strong contagion effects,and increased scope for moral hazard ininternational financial markets. Themost pressing issue is to develop bettermechanisms to facilitate private-to-pri-vate debt workouts—including, undersome conditions, “standstills” on exter-nal debt—and help resume capital flowsand increase international liquidity tocountries in crisis. Although there aresome compelling arguments in favor ofa lender of last resort, appropriate bur-den-sharing, rules for intervention, andmoral hazard remain difficult and unre-solved problems. Improved regulationby creditor-country authorities and bet-ter risk management of bank lending toemerging markets should also helpreduce the probability of crisis. Moretimely and reliable information is desir-able, but complete transparency andbetter information alone will not pre-vent crises. Still, better use of warningindicators may help governments takecorrective actions early enough toreduce the extent and cost of crises. Theissues are undergoing debate and con-sideration in different forums.

Costs and causes offinancial crises

Financial crises have become more fre-quent in developing countries since the

start of the 1980s (figure 3-1). They havetaken three main forms: currency crises,banking crises, or both. Currency crises areusually attacks on the domestic currency that

124


end with a large fall in its value, althoughthey can include speculative attacks that aresuccessfully warded off by the authorities.2

Banking crises refer to bank runs or otherevents that lead to closure, merger, takeover,or large-scale assistance by the government toone or more financial institutions.

Sometimes, both currency and bankingcrises occur around the same time—the so-called twin crises. The 1997 financial crisisin East Asia is the most recent example—with Indonesia, the Republic of Korea,Malaysia, and Thailand all experiencingcurrency turbulence along with seriousbanking sector problems. Earlier examplesinclude the Southern Cone countries—Argentina (1981), Uruguay (1982), andChile (1982). More recently, Mexico(1994), Argentina (1995), and the CzechRepublic (1997), as well as Finland, Nor-way, and Sweden in 1991 and 1992 have

experienced similar problems (Kaminskyand Reinhart 1997). While these crises havebeen associated with large volumes of pri-vate-to-private capital inflows, many othercurrency or twin crises in developing coun-tries, including most recently in Russia, areof the traditional type where excessive pub-lic borrowing plays a central role.

Financial crises can entail large costs(in lost output and welfare) and distribu-tional effects, which are substantially mag-nified in a twin crisis. Banking crises exac-erbate the negative impacts on the economythrough a reduction in the volume of loans,the misallocation of financial resources,and the ensuing contraction in credit andcutbacks in investment (box 3-1).

The greater frequency and cost of cur-rency and twin crises have been associatedwith surges in international capital in-flows—especially private-to-private flows—

125

Increased incidence of financial crises since the 1980sFigure 3-1 Incidence of financial crises, 1970–97

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

0

5

10

1522

Currency crisis Banking crisis

Number of crises

1997

Source: Caprio and Klingebiel 1996a; Frankel and Rose 1996; and Kaminsky and Reinhart 1997.


to developing countries and the growingintegration of these economies with worldfinancial markets (see below and Kaminskyand Reinhart 1997).

Private capital flows have surgedPrivate capital flows to developing coun-tries rose from about $42 billion in 1990 to

roughly $250 billion in 1996. Long-termprivate capital flows went from less than 1percent of developing countries’ GDP in1990 to a peak of 3.7 percent in 1993, andabout 2.8 percent in 1996 (figure 3-2). Thesurge in private capital flows is unprece-dented (at least since the end of the FirstWorld War),3 being twice as high as the pre-

126

Box 3-1 Costs of financial crises

AWorld Bank study found for a sample of 14banking crises a 5.2 percent average decline inoutput growth after crisis (World Bank 1997b).Another study found in emerging markets an

average cost in lost output (over to trend output) of 14.6percent of gross domestic product (GDP) per crisis (IMF1998b). Yet another study found that both output growthand efficiency fall after a banking crisis, with exchangerate volatility and currency crisis common in their after-

math (Lindgren and others 1996). Such crises can alsoresult in significant resolution costs, stretched over manyyears. A study in Latin America found that at least 4 to 5years are required to resolve banking crises (Rojas-Suárezand Weisbrod 1996). The direct fiscal or quasi-fiscal out-lays for bank restructuring vary between industrial coun-tries and emerging markets and between individual coun-tries from 1.5 percent of GDP for U.S. commercial banksin 1989 to 45 percent for Kuwait in 1995 (box figure

Costs of crises can be huge…Cost estimates of bank restructuring

Industrial countries (1980 to 1990s)

Emerging markets (1980s)

Emerging markets (1980s)a

Emerging markets (1990s)b

Argentina (1980–82) Chile (1983–85) Uruguay (1981–84) Kuwait (1992)

0

5

10

15

20

25

30

35

40

45

50

Percent of GDP

Notes: The midpoint cost estimate for each country-episode was selected. The sample includes seven industrial and 34 emerging countries.a. Excludes Argentina, Chile, and Uruguay.b. Excludes Kuwait.Source: World Bank staff estimates based on Caprio and Klingebiel 1996a; Lindgren and others 1996; Rojas-Suárez and Weisbrod 1996; Alexan-der and others 1997.


127

vious peak in private capital flows duringthe oil-boom years (1978–82). Further set-ting them apart, much of recent capitalflows has been private-to-private, ratherthan private-to-public, flows. Some surgesin capital inflows have been particularlymassive: in 1989–96 cumulative privatecapital inflows reached 55 percent of GDP

in Thailand and 50 percent in Malaysia;and in 1993–96, they reached 43 percent inHungary and 35 percent in the CzechRepublic. By contrast, the peak for Organi-sation for Economic Co-operation andDevelopment (OECD) countries in 1989was 2 percent of GDP per year (on aweighted average basis).

…and are greater in developing countriesCumulative loss of output per crisis for industrial and emerging economies

Currency crises Currency crashes Banking crises Currency and banking crises

0

2

4

6

8

10

12

14

16

18

20

Percent of GDP

Industrial Emerging

Note: Only crises with output losses are represented.Source: International Monetary Fund 1998b.

below). In general, restructuring costs are higher in devel-oping countries—where they range between 3 percent and25 percent in Africa, between 1.8 percent and 13.2 per-cent in Asia, between 0.3 percent and 41.2 percent inLatin America, and between 1 percent and 15 percent inEurope and Central Asia—than in industrial countries,where they average less than 6 percent.

Costs of currency crises are higher for emerging mar-kets than for industrial countries, and even higher in casesof currency crashes (see box figure below).

These costs are also much higher for twin crisesreaching 18 percent of GDP in 26 emerging markets, andfor developing countries than for industrial countries.Moreover, the average recovery time back to trend growthrates is longer for such crises (2.6 years, compared with1.5 years for currency crises and 1.9 years for bankingcrises). Calculations for selected individual countries (forthis report) find that the cumulative loss of output (rela-tive to trend) ranges from a low of 0.2 percent of GDP(Mexico 1976) to a high of 30.6 percent (Chile 1971).


High volatility of private capital flowsPrivate capital flows have also been volatileand subject to large reversals. This is seenin the decline during the debt crisis of the1980s, and in the reversal after Mexico’scrisis in 1994, and after East Asia’s crisis in1997. FDI has been more stable and rosesteadily throughout the various crises indeveloping countries (figure 3-2). Thus therecent rapid increases in FDI flows mightbe construed as being of the “jet-airplane”variety, bringing benefits with fewer risks.4

Non-FDI flows show far greater volatility,with sudden reversals.5 Analysis (see below)of non-FDI flows (portfolio equity, bondsand other debt securities, and bank loans)shows that medium-term bank loans havedeclined and have been replaced by portfo-lio flows, which show greater volatilitythan FDI, and sudden reversals, as evident

after the 1994 Mexican crisis. Short-termbank loans are even more volatile, as wit-nessed in the East Asian crisis.

In the context of the increasing integra-tion of developing economies with worldfinancial markets, the fundamental causesof twin crises of the type seen in East Asialie both in domestic policies and institu-tions and in international capital marketfailures. The analysis in the rest of thechapter focuses on these causes of financialcrises and appropriate policies to preventthem.6 The discussion highlights the inter-action of these factors, especially the inter-action of international capital markets withdomestic financial vulnerabilities, whichamplify the risks of a crisis:• Macroeconomic policies may either

exacerbate financial risks or fail to pre-vent boom-bust cycles, often a cause offinancial crises.

128

Private capital flows are volatile, but FDI has been risingsteadilyFigure 3-2 Net private capital flows to developing countries, 1975–96

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Percent of GDP

Total flows

FDI

Non-FDI

1975

1996

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

Note: Weighted average.Source: World Bank 1998c.


• Inadequate prudential regulation andpremature liberalization of domesticfinancial systems (along with poor cor-porate governance) may create condi-tions for excessive risk taking bylenders and borrowers.

• These two factors, coupled with short-term private-to-private capital inflowssurges (as in East Asia) and prematurecapital account liberalization, can cre-ate even greater risks, and increase thelikelihood of financial crises.

• Reliance on capital inflows exposesdeveloping countries to external panicsthat may cause sudden and massivereversals in capital inflows, deep illiq-uidity, and strong contagion effects.Minimizing these risks and dealing moreeffectively with such financial criseswould require a better architecture ofthe international financial system.Moreover, political economy con-

straints may also prevent governments from

acting decisively to prevent a crisis, evenwhen there are warning signals of vulnera-bility (often for many variables at the sametime) and a crisis is known to be brewing—as in Thailand and Mexico (box 3-2).

Macroeconomic policies tomanage capital flows andreduce financial risksMacroeconomic policies designed to

avoid large external and internalimbalances are a first line of defense in theprevention of financial crises. Crises areoften a result of boom-bust cycles, with sig-nificant interaction between macroeco-nomic factors and weaknesses in financialand corporate sectors. For instance, an eco-nomic boom may result when weak regula-tion and government guarantees of finan-cial liabilities lead financial institutions toengage in excessively risky lending (Krug-man 1998; Corsetti, Pesenti, and Roubini

129

Preventive measures to avert a crisis are obviouslypreferable to waiting for one to occur. Policymak-ers, however, often respond to other pressures.Governments may fail to act because politicians

give greater weight to short-term costs and less to long-term gains. Bad information and analyses may also play arole. Interest groups likely to lose out from policies thatwould minimize the risks of crises lobby to protect theirinterests. For example, measures to correct banking sys-tem fragility hurt bank owners, managers, shareholders,and well-connected firms (as in Indonesia) almost immedi-ately, while benefits are long-term and diffused. Countriesthat have not experienced financial crisis lack a realisticnotion of their costs. The lessons of crises influence thebehavior of policymakers. The hyper-inflation experienceof the 1920s has left German policymakers extremely sen-

sitive to inflationary signs, while policymakers in indus-trial countries have a strong collective memory of theeffects of the Great Depression. Finally, policy conflictsabound, and the process of policymaking within countriesis sometimes flawed (in Korea and Thailand, for example,with limited information sharing between the central bankand the Ministry of Finance).

Some of the most effective banking sector reformshave taken place only in the wake of major crises, as inChile in the 1980s and Argentina after the 1994–95 Mexi-can crisis. Once the domestic financial system is in deeptrouble, with large external borrowing requirements, theconflicts in policy may no longer be manageable. Bailingout domestic banks only results in more pressure on theexternal situation. A “soft landing” scenario may nolonger be practical.

Box 3-2 Political economy and financial crises


1998). Macroeconomic policies can eitherlessen or aggravate these risks. Surges incapital inflows can create the conditions forboom-bust cycles and compound macro-economic and financial management prob-lems—especially in small, open developingeconomies with fragile financial systems(McKinnon and Pill 1997; Corbo and Her-nandez 1996).7 Moreover, real exchangerate movements affect resource allocation,particularly between tradable (export andimport-competing) and nontradable (forinstance, real estate) sectors. Loss of com-petitiveness for tradables and booms innontradables can contribute to strains inthe domestic financial system while aggra-vating external imbalances.

Government’s room for macroeco-nomic policy maneuver is often restrictedby important tradeoffs and ineffectiveinstruments. Fiscal policy may be too bluntto offset the effects of volatile capitalinflows, while reducing public spendingmay conflict with other goals. Tighter mon-etary policies and sterilization may evenincrease capital inflows, particularly vola-tile short-term flows, while an exchange-rate peg eliminates the effectiveness ofmonetary policy and increases incentivesfor private borrowing abroad. A shift toflexible exchange rates increases the lati-tude in monetary policy maneuver but byitself may be insufficient to control over-borrowing and may lead to greaterexchange rate volatility. Flexible exchange

rates can also result in big losses of compet-itiveness and misalignments when capitalflows surge to high levels.

Fixed exchange rates andsterilization of inflowsPolicymakers use fixed or quasi-fixedexchange rates to reduce uncertainty aboutexchange rates, avoid nominal appreciationand maintain external competitiveness, andprovide a nominal anchor to preserve domes-tic stability.

When private capital inflows surge, how-ever, fixed or pegged exchange rates maybecome untenable and costly because of theimplications for domestic macroeconomicgoals (such as reducing unemployment), theinflationary pressures they generate, and theincentives they create for private agents tooverborrow. The pursuit of a pegged nominalexchange rate contributed to the East Asiancrisis, especially in Thailand.

The typical initial response to a surge incapital flows is to increase official reserves tomaintain the exchange rate and guard againstsudden reversals. An analysis of 27 episodesof capital inflows shows that, on averageover the period of surge, one-third of the cap-ital account surplus was absorbed by reserveaccumulation. This ratio rises to 50 percentwhen the capital inflow is at the lower (0–3percent of GDP) or the higher range (morethan 9 percent of GDP; box 3-3).8

Under a currency board arrangement, anextreme form of fixed exchange rate, a coun-try formally gives up its autonomy in mone-tary policy and strongly anchors its currencyto a fixed rate (box 3-4). With a peggedexchange rate, however, the authorities tendto retain some autonomy in monetary policyin the pursuit of domestic objectives. During

130

Sterilization of capital inflows maywork in the short term, but it isincreasingly costly over time.


the early stages of a surge in capital inflows,the authorities buy foreign exchange (whichimmediately expands the supply of domestichigh-powered money) and simultaneouslysell domestic bonds or increase reserverequirements to sterilize the effects of theinflows on domestic money supply. Withoutsuch sterilization, capital inflows wouldexpand the domestic monetary base, creatinga temporary economic and lending boom

and increasing financial system fragility. Eco-nomic agents would lose confidence in theauthorities’ ability to maintain the peg, andexpectations of a devaluation would increase,possibly leading to an attack on the currency.

The fundamental problem with steril-ization is the “inconsistent trio” or “openeconomy trilemma”: any two, but not allthree, features of macroeconomic policy—afixed exchange rate, full capital mobility,

131

Monetary authorities frequently intervene toincrease foreign reserves when capital inflowsbegin to surge, in order to preserve stability ofthe exchange rate (when the domestic currency

is implicitly or explicitly anchored to an exchange ratepeg) and to reduce market uncertainty. When capitalbegan to flow into Morocco in 1990, its foreign reserveswere low, and the authorities absorbed 75 percent of theincoming capital in the first three years. Similarly, in theCzech Republic in 1993, the authorities built up reservesof roughly 70 percent of capital surpluses for the follow-ing three years. In 27 inflow episodes in 21 developingcountries, reserve accumulation absorbed an average of32 percent of the change in the capital account surplus,with the extent of reserve accumulation depending on thesize of flows.

At low levels of capital inflows, reserve accumulationabsorbs close to half of the capital account surplus (boxfigure). This may be due to the buildup of reserves fortrade purposes during the initial phase of capital inflowsand determination of authorities to defend the exchangerate. At intermediate levels of inflow (3–9 percent of GDP)reserve accumulation falls to about 20 percent of the capi-tal account surplus. When inflows are large (exceeding 9percent of GDP), the authorities once again interveneaggressively, possibly because of increased perceived risksof reversals. The rate of reserve accumulation is alsoclearly related to the composition of capital inflows. Thelarger the non-FDI component of the increase in capitalinflows, the higher the reserve accumulation.

Accumulation of reserves has a social cost (differentfrom the cost of sterilization) measured by the differencebetween the cost of servicing the capital inflow equivalent

to the accumulated reserves and the income earned onthese reserves. This estimated cost for some East Asiancountries (Malaysia and the Philippines) reached about0.1 percent of GDP a year for many years. But this costmay be significantly higher: it was 0.16 percent of GDP ayear for the Czech Republic and 0.12 percent for Peru.

A sizeable share of capital inflowsgoes to reserve accumulationReserve accumulation and capital inflows

0–3 3–6 6–9 >9

0

10

20

30

40

50

60

Percent

Capital inflow range (percent of GDP)

Note: Average reserve accumulation as percentage of capitalaccount surplus during inflow surge period.Source: World Bank staff calculations based on data from IMFInternational Financial Statistics.

Box 3-3 Capital inflows and reserve accumulation


132

Under a currency board, a country gives up itsdiscretionary power over monetary policy, com-mitting itself to issue no money that is notbacked by reserves and to tolerate the interest

rates that result. The hope is that the very strong com-mitment to maintaining the value of the currencyreduces its susceptibility to attack, helping to sustain afixed exchange rate and creating greater confidence. Butdoes it work?

A currency board is different from a peggedexchange-rate system primarily because the authoritieshave chosen—through legislation (Argentina and Esto-nia) or other arrangements (an automatic link arrange-ment, as in Hong Kong [China])—to subordinate domes-tic policy and objectives to policies to maintain a fixedexchange rate. Under a pegged exchange rate, the mone-tary authority commits to the currency peg as a mecha-nism to maintain low inflation, but can abandon the pegin the event of a large shock to output (Obstfeld andRogoff 1996). The authority weighs the costs of main-taining the peg (lower output, higher unemployment)against the costs of abandoning it (loss of credibility,higher inflation).

Speculative attacks on the peg can happen undereither arrangement, and the required response to suchattacks is to raise domestic interest rates and squeezedomestic credit high enough to stop the attacks. But if thecosts to domestic output (primarily in nontradable sec-tors) are severe—as they tend to be if interest rates remainhigh for a long period—chances are the peg will be aban-doned. By ruling out this possibility a currency board cre-ates greater credibility for the arrangement. Currencyboards appear to work best for only two groups ofeconomies: small, open ones with large tradable sectors(Hong Kong [China] and Estonia); and economies thathave been extremely unstable, where a currency boardwould restore badly needed credibility to domestic mone-tary policy (Argentina and Bulgaria). Other requirementsfor successful currency board arrangements include tightfiscal policies, with substantial fiscal surpluses and flexi-bility in fiscal policy; labor market flexibility; successfulhigh interest rate defense against previous attacks (with-out large residual costs to the economy); and enough ini-tial reserves to make the system credible.

Source: ADB and World Bank 1998; Obstfeld and Rogoff 1996.

and monetary policy independence—arefeasible (Mundell 1963; Wyplosz 1998;Obstfeld and Taylor 1998).9 Sterilizationpresupposes that independent domesticmonetary policies can be pursued effec-tively (to control domestic money supply)under conditions of international capitalmobility. But when exchange rates are fixedor pegged and there is a large degree of cap-ital mobility (that is, when a country’sfinancial assets issued in its currency arereasonably substitutable (in private port-folios) for other internationally acceptedassets), sterilization policies may be ineffec-tive, because any contraction or expansionof the domestic assets of the central bankwill give rise to an offsetting capital inflowor outflow (Montiel 1993).10

Sterilization may work in the shortterm, but it is increasingly costly over time.If inflows persist, this strategy becomes evenharder to maintain because of rising fiscalcosts, reflecting the fact that interest rateson domestic bonds exceed the interest thatcentral banks earn on foreign depositsabroad.11 Moreover, sterilization leads tohigher domestic interest rates, which attractfurther inflows of capital (figure 3-3). Short-term capital flows—which tend to be themost sensitive to interest rate differentials—increase, raising the vulnerability to liquid-ity crises (Montiel and Reinhart 1997).

Pegged nominal exchange rates can cre-ate unintended incentives to domestic resi-dents to overborrow, thereby fueling surgesin capital inflows. Maintaining the peg (as

Box 3-4 Currency boards—when do they work?


long as it is credible), as in Thailand priorto the recent crisis and in Chile in the late1970s, effectively guarantees against anyexchange rate risk to domestic borrowersacquiring foreign liabilities.12 It lowers thecost of borrowing by socializing theexchange rate risk and allowing privateborrowing without currency hedging. Nor-mally, a prudent borrower facing exchangerate risks (and without a natural hedge,such as exports) would be expected topartly or fully hedge those risks in forwardexchange markets, thereby lowering incen-tives to borrow abroad.

Shifting to flexible exchange ratesPlacing restrictions on capital mobility canreturn autonomy to monetary policyunder a fixed exchange-rate regime.

Switching to a flexible exchange rate alsoreturns autonomy to monetary policy andprovides incentives that, in a world ofgreater capital mobility, may reduce thelikelihood of crises (Goldstein 1995;Corbo and Hernandez 1996). Flexibleexchange rates—whether managed floats,exchange-rate bands (usually with acrawling peg; Williamson 1996), or fullyfloating exchange rates—offer several ben-efits. Through nominal and real apprecia-tion, exchange rates take the brunt of theadjustment to large capital flows andallow greater independence for domesticmonetary policy (that is, more effectiveapplication of sterilization policies) andlower inflation.13 Unintended incentives tooverborrow are avoided because marketparticipants are unsure about the futuredirection of exchange rates. By minimizingthe impact of capital inflows on the exter-nal component of high-powered money,14

flexible exchange rates limit the effects ofcapital flows on the (potentially fragile)domestic banking system (Goldstein1995). Malaysia and Chile, for example,managed surges in capital flows betterthan most other countries because ofwider targets for exchange rates and capi-tal controls (Corbo and Hernandez 1996;Goldstein 1995).

While shifting to a floating exchangerate may limit some of the boom-busteffects of capital flows, it may create otherproblems (Gavin and Hausmann 1996) inthe process of reaching equilibrium.Exchange rates and interest rates maybecome more volatile. A large appreciationwill worsen external competitiveness (espe-cially if trade reforms require a deprecia-tion), with potentially severe consequences

133

Sterilization means higherinterest rates and moreshort-term capital inflowsFigure 3-3 Interest rates and sterilization policies, Indonesia and Chile

Indonesia Chile

0

5

10

15

20

25

30

Pre-inflow (1988/89–1989/90)

Capital inflows and heavy sterilization 1990/91–1992)

Capital inflows and partial sterilization (1993–94)

Percent

Source: Asian Development Bank and World Bank1998.


for the sustainability of capital flows inhighly open and small economies.15 Further,when exchange rates appreciate, they feedexpectations of a lasting boom, reducedomestic interest rates, boost the demandfor credit, lower the costs of—and raisedemand for—foreign borrowing (in domes-tic currency), and raise the returns ondomestic assets (stock markets, for instance)to foreign investors, thereby encouragingmore capital inflows. All these elements cancontinue to support a boom-bust cycle.

Indeed, it is the underlying real appreci-ation (the price mechanism that can operateeither through nominal exchange ratechanges or through domestic nontradableprices) that puts the boom-bust cycle inplace (Corbo and Hernandez 1996). Shift-ing to a flexible exchange rate does not pre-clude a crisis (Khatkhate 1998; IMF1997a). Indeed, crises are as likely to occurunder flexible exchange rates as under fixedexchange rates, especially if other condi-tions, such as adequate prudential and reg-ulatory safeguards on the financial sectorare not in place (figure 3-4).

Under flexible exchange rate regimes,the monetary authorities may attempt tosterilize a surge in capital inflows or theymay opt not to. If they choose to sterilizeflows, domestic inflation is moderated,domestic interest rates do not fall rapidly,and capital flows continue, but the fiscalcosts may be high. If they choose not to ster-ilize, interest rates fall more sharply, reduc-ing incentives for foreign borrowing, butinflation may rise. Often, however, domesticinterest rates will remain persistently high indeveloping countries.16 The incentive forincreased capital inflows thus remains, con-tributing to vulnerability. There are also

other shortcomings with flexible exchangerate regimes, notably the loss of a nominalanchor and lower inflation gains.

Countercyclical fiscal policyGiven large and potentially destabilizingcapital flows, a tightening of fiscal policycan help curb borrowing from abroad andreduce appreciation of the real exchangerate, but only if higher public savings arenot offset by lower private savings.17 Inpractice, few countries take significantcountercyclical fiscal action to temper acapital inflow boom (Schadler et al. 1993).That places too great a burden on mone-tary policy to restrain aggregate demand,which leads to accumulation of short-termliabilities and increases vulnerability.

Three factors make it difficult to takethe required fiscal adjustment measures:

134

Crises are more frequentunder flexible exchangerate regimesFigure 3-4 Frequency of crises underflexible and fixed exchange rate regimes

1975–81 1982–89 1990–960

10

20

30

40

50

60Crises under flexible regimesCrises under fixed regimes

Percent

Note: Ratio of number of crises during period tonumber of countries with flexible and fixed exchangerates, respectively.Source: World Bank staff estimates based on datafrom International Monetary Fund 1998b, and IMFExchange Arrangements and Exchange Restrictions,various issues.


first, the state is to some extent heldhostage to private capital inflows; second,the fiscal process is inflexible relative to thevolatility of capital flows; and third,demand is not met for many essential pub-lic goods and services—often those (such ashuman resources and physical infrastruc-ture) that might be essential to increase thelonger-term efficiency of the economy andthe absorption capacity of resources fromabroad. The result is that fiscal policy isoften procyclical, which makes the situa-tion even worse (as happened in East Asiarecently).

Resorting to other instrumentsIf the surge in capital inflows is large, thestandard tools of macroeconomic policy—shifting to flexible exchange rates, avoid-ing strong sterilization efforts, and imple-menting strong countercyclical fiscaladjustment—may prove ineffective orimpractical. Other instruments may beneeded. Indeed, since fragile financial sec-tors are a prime vulnerability of develop-ing economies, policies should support theconduct of prudent macroeconomic poli-cies by improving and tightening the pru-dential regulatory framework of the finan-cial system (and implementing othermeasures related to external financialliberalization).

Financial liberalization,domestic banking reforms,and corporate governance

In the past two decades developing coun-tries have been encouraged to liberalize

their domestic financial sectors—lift con-trols on domestic interest rates and credit

allocation, privatize financial institutions,and allow entry and competition from newprivate institutions. A growing body of evi-dence shows the importance of strong finan-cial systems and financial deepening forlong-run growth and development (Kingand Levine 1993; Levine 1997; Levine andZervos 1998). Demirgüç-Kunt and Detra-giache (1998) find that moving from finan-cial repression to liberalization of domesticfinancial systems results in faster long-rungrowth of almost 1 percent per year.

Domestic financial liberalizationFinancial liberalization also requiresmore—not less—and effective prudentialregulation to ensure the safety and sound-ness of financial systems. It also requiresbetter corporate governance structures andarm’s-length relationships between banksand corporations. These arrangements taketime to build, and without them, financialliberalization associated with surges in cap-ital inflows often leads to financial crises.These crises are not limited to developingcountries—Scandinavian countries thatundertook financial liberalization at theend of the 1980s and early 1990s also expe-rienced these problems.18

A study of the relationship betweenbanking crises and financial liberalization byDemirgüç-Kunt and Detragiache (1998) for53 countries between 1980 and 1995 foundthat crises are more likely in liberalizedfinancial systems19 (figure 3-5). Severalmechanisms link deregulation and liberaliza-tion to crisis (Goldstein and Turner 1996):• Increased competition among financial

institutions (from existing banks, theentry of new banks, development ofnonbanks, and expansion of capital

135


markets) may lower bank margins,profitability, and franchise values oreffective capital base (Asian Develop-ment Bank [ADB] and World Bank1998). Empirical evidence of this effecton franchise values is found byDemirgüç-Kunt and Detragiache(1998; figure 3-6). The decline in bankmargins and profits may be an objec-tive of financial liberalization, but ifexcessive competition leads to sharpdeclines in franchise values, it mayreduce the incentives for prudent bank-ing and lead to excessive risk taking bybank managers. Sheng (1996) findsthese factors to be responsible for bankfailures in Argentina, Chile, Kenya,Spain, and Uruguay.

• Higher real interest rates often emergefollowing liberalization (Galbis 1993). Iffirms are operating with high debt-equity ratios, a hike in interest rates canlead to distress borrowing and an inelas-tic demand for credit, which perpetuatehigh interest rates.20 A bidding up ofdeposit rates may also weaken banks.

• Rapid credit expansion due to reducedreserve requirements and a largermoney multiplier released pent-updemand for credit or easier access toforeign resources, and expanded banklending to boom-bust prone activities(Caprio, Atiyas, and Hanson 1994).Bank credit managers trained in a con-trolled environment may not have theskills needed for a riskier environment.

• Freeing deposit rates with weakbanks, in developing countries andeven more in transition economies,leads to higher deposit and lendingrates to reflect the higher risk. Thistends to attract riskier investors andincreases the overall portfolio risk ofbanks. An increase in systemic riskfurther pushes up interest rates (Mas-sad 1994).

• Many episodes of banking crises areassociated with the entry of bankowners bent on engaging in risky andquestionable activities (as in Chile inthe 1970s and many transit ioneconomies in recent years).

136

Banking crises occur more in liberalized financial systemsFigure 3-5 Interest rate liberalization and probability of crisis

Chile India Malaysia (1985) Paraguay (1995) Portugal (1986) Turkey (1991)0

0.05

0.10

0.15

0.20

0.25With liberalization Without liberalization

Probability of occurrence

Note: Countries are classified as crisis cases if the predicted probability is greater than 0.05, which is actual frequency ofcrisis.Source: Demirgüç-Kunt and Detragiache 1998.


In view of the benefits and risks fromdomestic financial liberalization, the tran-sition needs to be carefully managed.21

Supervisory capacity has to be devel-oped quickly and should precede liberaliza-tion. New bank owners and managers needto meet the criteria for prudent professionalbankers. Similarly, bank managers, loanofficers, and other professional staff needto be properly trained. Entry of foreignbanks may help achieve this objective. Butlifting restrictions on domestic and foreignentry to increase competition and innova-tion needs to be monitored to avoid largedeclines in the franchise values of banksand excessive risk taking.

Authorities should be vigilant in curb-ing lending booms following liberalization,

for example through higher reserve and cap-ital requirements.22 Developing countriesmight temporarily impose limits on creditgrowth to avoid the risks associated withcredit booms, especially during rapid trans-formation of the banking system, when thesupervisory system is insufficiently devel-oped.23 Alternately, countries may wait tolift constraints or decide to impose morestringent and explicit limits and restrictionson risky lending activities (and concentra-tion of risk), such as real estate, securities,and foreign exchange exposure.

Finally, careful sequencing of domesticand external liberalization is called for.Restrictions on the capital account, espe-cially on the more volatile capital flows,should be lifted only after the domesticfinancial sector has been strengthened.

Supporting the financial systemand improving corporategovernanceHow well the financial system functionsalso depends on the legal framework toenforce contracts and protect propertyrights and the state of corporate gover-nance. While these measures are not dis-cussed in detail here, improvements in thisarea are nevertheless of vital importance.When transparency is lacking and corporategovernance is weak, both banking systemsand corporate sectors are more fragile.24

Cozy relations among banks, government,and corporations weaken market discipline,encourage connected lending, increase thescope for moral hazard, and foster ineffi-cient outcomes. Other signs of weaknessesare loose financial accounting and disclo-sure and high leveraging, which facilitatesexcessive risk taking. Concentration of

137

More competition maylower bank margins andfranchise valuesFigure 3-6 Financial liberalization andbank franchise value

Return on equity Capital Liquidity

-0.20

-0.15

-0.10

-0.05

0

0.05

0.10

0.15

0.20

0.25

Correlation coefficient

Note: Pearson correlation coefficient between dummyfor financial liberalization and variable (based onBank-level variables which are averaged by countryfor the period 1988–95).Source: Demirgüç-Kunt and Detragiache 1998.


power in family dominated and politicallyconnected companies, with weak protectionof minority shareholders, is also common indeveloping countries.

These factors have contributed to weakperformance and banking sector distress inEast Asia, their effects intensified byincreased access to foreign resources anddomestic financial liberalization. Debt-to-equity ratios rose significantly in manycountries, and economic efficiency andprofitability declined.

The policy prescriptions to supportdevelopment of the financial system andimprove corporate governance are straight-forward:25

• Developing accounting, auditing, anddisclosure standards to increase theflow of information, and enhance effi-ciency by improving the quality ofinvestment, reducing misallocations,correcting mistakes rapidly, andstrengthening business risk assessmentand the accountability of managers toshareholders.

• Setting up the legal infrastructure—bankruptcy laws, debt workout proce-dures, enforcement of collateral andguarantees—to write and enforce con-tracts confidently and to protect andbalance the interests of creditors, share-holders, and managers, thereby creat-

ing a credit culture in which trust andexpectations of repayment and transac-tion costs are reduced.

• Restricting connected lending practices. Such policies would also support the devel-opment of capital markets and alternatives(equity and long-term debt) to short-termdebt finance and reduce the extent of lever-age and vulnerability of firms to shocks.26

Strengthening domestic banksthrough better regulation andmarket incentivesBanking reform, strongly oriented towardrisk management, is a key ingredient of anylong-term strategy to minimize the risksand costs of financial crises. An efficientbanking sector with effective supervisionand regulation helps reduce the distortionsthat increase vulnerability to potentialcrises.27 The central aim should be toreduce information asymmetries anddevelop a risk management culture in thebanking sector. Internal systems of riskmanagement have to be developed andstrengthened, and best practice techniquesused. Bank supervisors tend to prefer ensur-ing the adequacy of a bank’s internal con-trols to directly assessing financial condi-tions.28 This is important in developingcountries, since the risks facing the bankingsector are especially great because of prob-lems in the state of development and com-petitiveness of domestic financial markets,corporate law and governance, contractenforcement and bankruptcy, sophistica-tion of bankers and their regulators, extentof political connections between institu-tions and governments, and susceptibilityof the economy to domestic and interna-tional economic shocks.29

138

Banking reform, strongly orientedtoward risk management, is a keyingredient of any long-term strategyto minimize the risks and costs offinancial crises.


Developing country regulation shouldtake account of the strengths and weak-nesses of financial systems and of regula-tors. Regulations, controls, and restrictionsthat produce inefficiencies and distortionsshould be abandoned. However, the vul-nerabilities and frequent failures of finan-cial systems indicate that some restraintsare needed (World Bank 1998b, chapter 6).For instance, mild restraints on depositinterest aimed at creating franchise valuefor banks may induce outcomes that aremore efficient than financial repression(where interest rates are kept at low nega-tive real levels, inducing inefficient deepen-ing and misallocation of resources) orimmediate financial liberalization (Hell-man, Murdock, and Stiglitz 1996, 1997;World Bank 1998b). Financial restraintfeatures played a significant role inimproving stability in East Asian countriesin the past; while some of the other fea-tures such as market incentives have beenimplemented with some success in Chile,New Zealand, and the United States(Nicholl 1997; Caprio and Klingebiel1996b; Goldstein and Turner 1996).

Banking regulation and supervision.Weak regulation and supervision are themost widely recognized sources of vulnera-bility in developing countries’ banking sys-tems.30 Most industrial countries subscribeto “Core Principles for Effective Supervi-sion” of the Basle Committee in the designof banking regulation and supervision toreduce vulnerability of the financial sys-tem.31 In addition to macroeconomic sta-bility, the building blocks include:• Higher standards of competence and

integrity of bank management, as wellas effective management controls.

• More transparency and adequate infor-mation on the soundness of banks.

• Public financial safety nets that boostconfidence in the financial system butalso limit induced distortions, such asexplicit or implicit government guaran-tees, that encourage excessive risk taking.

• Effective regulatory and supervisoryoversight for controlling risk and limit-ing the adverse impact of official safetynets.

• Transparent ownership structure thatenhances competitive behavior, andlimits on connected lending.The Basle committee’s core principles

can also be extended to include accountingand information disclosure, loan classifica-tion, and bankruptcy regimes. Internationalaccounting and auditing standards are alsoavailable.32 International financial institu-tions can help countries adopt and imple-ment these regulations.33

Recommending that developing coun-tries build a sound and healthy financialsystem according to these principles is notsufficient, however. Building such systemstakes a long time, and it is hard to deter-mine a minimal requirement for the qualityof the banking sector. Also, adjustments areneeded to take account of specific featuresin developing countries.

Incentives and market discipline. Rely-ing heavily on regulation and supervisionto control excessive risk taking is of ques-tionable efficacy, particularly for develop-ing countries (Caprio and Klingebiel1996b; Caprio 1997; Goldstein and Turner1996). It takes too long to develop supervi-sory capacity and skills. Moreover, super-visors are often unable to detect riskybehavior and take action against troubled

139


banks, because the kinds of behavior tendto change over time and supervisors arenot prepared for them. They may also beprevented by policymakers from takingaction.

These factors argue for relying more onmarket-imposed discipline and improvingincentives for prudent banking. In additionto raising capital adequacy ratios, this strat-egy could include:34 increasing the financialand personal liability of managers anddirectors (going to unlimited liability), orintroducing mutual liability; requiring a tierof uninsured subordinated debt for individ-ual banks (to increase the incentives for pri-vate monitoring of banks); and requiringbanks to regularly publish key information,such as credit ratings. Clear exit policiesand resolution mechanisms should also bespelled out, covering automatic or struc-tured early intervention and graduatedresponses by the authorities as bank capitalreaches some predetermined thresholds(Caprio 1997).

Volatility and the need for more strin-gent regulations. Developing countriesshare structural characteristics that subjectthem to greater volatility. These include anunstable macroeconomic environment, con-centrated economic activity and exports,and susceptibility to greater shocks—termsof trade, weather, interest rates, and policyvolatility (figure 3-7).35

Vulnerability to external shocks and,especially, to changes in international inter-est rates, has been shown to be the mostimportant factor in banking crises.36 Areversal of macroeconomic conditions incapital-exporting countries leads to higherinterest rates, curtailed capital inflows, andslower growth of bank lending.

These structural features have implica-tions for the institutional framework of thebanking sector in developing countries,such as a need for higher capital-adequacyratios than the Basle international standard(see Goldstein and Turner 1996; and Hono-han 1997) and for more stringent limita-tions on the concentration of risks (such asloans for real estate or securities).37

The role of government, guarantees,and moral hazard. Government often playsa pervasive role in the banking sector indeveloping countries. This generates seriousconflicts, especially moral hazard problems,that are a major underlying factor in riskylending. Necessary reforms include:• Above all, severely limiting the govern-

ment’s role in directly running andmanaging banks. This can mean priva-tization of banks, which has to be han-dled carefully, especially with respect to

140

Volatility is linked tobanking crisesFigure 3-7 Bank crises and volatility,1980–94

Grossdomestic product

Consumerprice index

Termsof trade

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Average coefficient of variation

Systemic banking cases

Borderline cases

Noncrisis cases

Source: Caprio and Klingebiel 1996b.


pricing (and its effect on franchise val-ues), treatment of risks, capital ade-quacy, and management.

• Reducing the government’s directinvolvement in allocating credit and inproviding guarantees to commercialenterprises so as to enhance market-ori-ented banking behavior. Because ofmoral hazard, implicit or explicit gov-ernment guarantees can lead to exces-sive risk taking. Banks tend to raisemoney at safe rates and lend at premiumrates to finance speculative investmentsbeyond prudent levels (McKinnon andPill 1997; Krugman 1998).

• Setting up a formal deposit insurancescheme to deal with the negative exter-nalities that individual failures may haveon the rest of the banking system. Insureddepositors, however, have little incentiveto monitor banks, and regulators mayengage in regulatory forbearance anddelay action against troubled banks.Indeed, Demirgüc-Kunt and Detragiache(1997) find that deposit insurance has asignificant positive effect on the likeli-hood of a banking crisis. Thus, supervi-sion, minimum capital requirements, andmandatory issues of subordinated debtwould help reduce moral hazard andinduce banks to reduce their risks. Inaddition, there should be limits on theamounts insured, and co-insuranceshould be required (that is, covering lessthan 100 percent of deposits), as well ascharging risk-weighted deposit insurancepremiums. Policymakers should also con-sider mutual liability for banks, clear pro-cedures for closing insolvent banks, andpossibly private provision and manage-ment of the insurance program.

Benefits and associatedrisks of capital accountliberalization

Capital account (or external) liberaliza-tion and financial integration with

world capital markets can potentially bringlarge benefits, and both have been advo-cated for developing countries for that rea-son.38 Letting domestic agents trade finan-cial assets with foreign economic agentsmay increase access to capital and lower itscost. Productivity improvements, risk diver-sification, and consumption-smoothing areother potential benefits.

Many developing countries have liber-alized capital accounts in the past decade(box 3-5), but recent experience suggeststhat such liberalization and increased finan-cial integration can sharply raise the risksof financial crisis (box 3-6).

In fact, the duality of benefits and risksof international capital mobility isinescapable in a world of asymmetric infor-mation (Obstfeld 1998), where lenders donot know as much as borrowers about theuses of their money and are therefore proneto panic. Thus, the benefits of capital accountliberalization and increased capital flowshave to be weighed against the likelihood ofsuch crises and their costs. Recent discussionsat international forums have heightened therecognition of the issues, especially in rela-tion to volatile short-term flows.

141

Many developing countries haveliberalized capital accounts in the pastdecade, but recent experience suggeststhis can sharply raise the risks offinancial crisis.


Evidence suggests that for FDI and sim-ilar long-term foreign capital flows, thebenefits are significant and the risks low.The benefits of capital account openness toshort-term debt and other volatile non-FDIflows are less certain; the greater volatilityof such flows is strongly associated withfinancial crises. While the clear demarca-

tion between these two categories ofinflows is not watertight (see furtherbelow), in practice, the effects are clearlydifferentiated. There are larger benefits andfewer risks for FDI-type flows, which tendto be more resilient in times of crises and tocarry important benefits beyond finance,than for short-term flows. Thus, developing

142

The OECD’s Code of Liberalization of CapitalMovements of 1961 (extended to include all capi-tal account transactions by 1989) and the Euro-pean Union’s 1988 Second Directive on Liberal-

ization of Capital Movements were milestones in theliberalization of industrial countries’ capital accounts. It isonly since the late 1980s and early 1990s that most indus-trial countries have accelerated the pace of capital-accountliberalization. The number of industrial countries withneither separate exchange rates nor restrictions on pay-ments for capital transactions increased from 3 in 1975 to9 in 1985 and 21 in 1995. The number increased in devel-oping countries as well, from 20 in 1975 to 31 in 1995.

Most industrial and developing countries still hadsome type of capital controls at the end of 1997,mainly on direct investment (143 countries), realestate transactions (128), and capital market securities(127). In addition, most countries implement provi-sions specific to commercial banks and other creditinstitutions (152).

Only a few industrial countries (Luxembourg andthe Netherlands) and developing countries (Armenia,Djibouti, El Salvador, Panama, and Peru) report nocapital controls, and a few report just one type of con-trol (Canada, Denmark, Mauritius, Uganda, andParaguay).

Most countries still have some form of capital controlsControls on capital-account transactions, year-end 1997

Total Developing countries Industrial countries

Number of IMF member countries 184 157 27Controls

Capital-market securities 127 112 15Money-market instruments 111 102 9Collective investment securities 102 97 5Derivatives and other instruments 82 77 5Commercial credits 110 107 3Financial credits 114 112 2Guarantees, sureties, and financial backup facilities 88 86 2Direct investment 143 126 17Liquidation of direct investment 54 54 0Real estate transactions 128 115 13Personal capital movements 64 61 3

Provisions specific to:Commercial banks and other credit institutions 152 137 15Institutional investors 68 54 14

Sources: Quirk and others 1995; Mathieson and Rojas-Suárez 1993; International Monetary Fund 1996, 1997b, and 1998c.

Box 3-5 How far has capital account liberalizationprogressed?


countries should tailor their openness totheir capital inflow needs and their abilityto bear the risks.

In addition to foreign direct invest-ment and trade credits, capital flows canrange from pure debt such as short- andmedium-term bank loans, to long-termbonds, very long-term debt (century andperpetual bonds), quasi-equity (such asconvertible bonds), and portfolio equityflows. The extent of use of these financialinstruments by developing countriesreflects investors’ preferences in terms ofrisk sharing between the parties in thesource and destination countries, currencyexposure and maturity risk to the develop-ing country firm, and extent of diversity ofsources of finance using the instrument(table 3-1).

Derivatives can also reduce the costand risk developing country firms face inaccessing international capital markets.For example, through interest rate swaps,borrowers can assume the kind of liabilitythey prefer (fixed or floating rate) at alower interest rate than through regularborrowing. Currency-swaps enable bor-rowers to match the currency compositionof assets and liabilities.

The analysis below highlights the evi-dence for the benefits and risks of capitalflows, differentiated mainly by FDI andnon-FDI flows. While it is simplified, itillustrates the major issues and their policyimplications, which in practice have to takeaccount of the variety of financial instru-ments, the mechanisms of their intermedia-tion, and the use of the associated resources.

Benefits of capital accountliberalization and capital flowsThe theoretical benefits from capitalaccount liberalization include increasedaccess to capital and faster productivitygrowth, risk diversification, and consump-tion smoothing.39

Capital accumulation and growth. Ben-efits include increased investment and moreefficient allocation of resources, whichresult from taking advantage of differencesamong countries in the productivity of cap-ital and opportunities for risk diversifica-tion. Incomplete risk markets discourageinvestors from undertaking risky projects,many of which have high potential returns.By allowing more risk diversification, moreof these projects will be undertaken, lead-ing to higher expected returns. The

143

In capital markets, the risk depends on the type of borrowingTable 3-1 Financial instruments and their risks

Instrument Risk sharing Currency exposure Maturity risk Diversity of sources

Borrowing facilities Low High High LowSyndicated bank loans Low High Moderate LowStraight bonds Low High Moderate/low HighLeasing Moderate High Moderate LowLimited recourse financing Moderate Moderate/low Moderate LowQuasi-equity instruments Moderate Moderate Moderate HighPortfolio equity investment High Low Low High

Source: World Bank.


expected results are higher capital accumu-lation and productivity growth. Benefitsvary by type of flow.

For FDI and similar long-term, rela-tively stable, flows the benefits are welldocumented. FDI flows accounted for 5–6percent of aggregate investment in develop-ing countries in the 1990s, above the 1–2percent of the previous 15 years (WorldBank 1997a). FDI also tends to “crowd in”more domestic investment: every $1 of FDIin developing countries is associated with$0.50–$1.30 of additional domestic invest-ment. While it is difficult to establishcausality, increased FDI flows are generallyassociated with faster aggregate long-rungrowth (and total factor productivitygrowth), with each percentage pointincrease of FDI in gross domestic product(GDP) associated with a 0.3–0.4 percentagepoint faster growth in per capita GDP(Wacziarg 1998).

For non-FDI—particularly short-termdebt and more volatile flows—the benefitsare less certain.40 Among 18 countries thatreceived significant private capital flows inthe late 1980s and early 1990s, the surge insuch capital flows was associated withincreased investment, as expected; each $1of non-FDI inflows appears to be associ-

ated with just $0.60 of additional invest-ment, however. One reason is that a signifi-cant share of such inflows goes into reserveaccumulation and results in a net socialloss, rather than a gain (see box 3-3).41 Pru-dent behavior also implies that short-termfinancial resources should not be used tofinance long-term investment projects. Tak-ing this into account, and using the sampleaverage capital-output ratio of 2.5 and elas-ticity of output with respect to capital stockof 0.4, a 1 percentage point increase ofnon-FDI capital inflows in GDP would beexpected to generate additional growth ofonly about 0.10 percent of GDP in grossterms and less in net (GNP) terms.42 Somebenefits on productivity could also beexpected, especially in low-savings coun-tries where non-FDI inflows might makepossible highly productive investments.Again, while there are direction of causalityissues and results should be interpretedwith caution, simple correlation on a sam-ple of countries showed little evidence of asignificant positive association betweennon-FDI inflows and productivity growth;in the one case that showed statistically sig-nificant association—the subsample of low-savings countries—the association wasstrongly negative (figure 3-8).

144

Fragile domestic financial systems are often theroot cause of a financial crisis, and while capitalinflows are also blamed, they are but one element.Risks carried by capital inflows and excessive bor-

rowing are important. In the Republic of Korea, exces-sive domestic financial risk taking—including low equityand heavy bank borrowing—was a long-standing prac-tice. What may have tipped the balance in the 1997 cri-sis, however, was capital flows: when in the context of its

entry into the OECD, Korea liberalized the ability of itsbanks to borrow (short-term) abroad (instead of tighten-ing safeguards), there was a massive surge in suchinflows; their reversal subsequently precipitated the crisis.The problems may have been aggravated by Korea’sretaining tight controls on FDI inflows, preventing pre-cisely the type of flows that it should have encouraged—more stable, longer-term flows that would have broughtequity, technology, and better risk management.

Box 3-6 Are capital flows the main culprit?


Benefits depend on the policy environ-ment. The opposing signs and different val-ues of correlation shown in figure 3-8imply that the impact of non-FDI capitalflows and capital account openness maydepend more on the economic and policyenvironment in individual countries thanon the extent of capital account opennessand flows themselves. This is the case forFDI inflows (World Bank 1997b) and offi-cial aid flows (Dollar 1998), for both thesize and direction of impacts on productiv-ity depend on the policy environment.

Some indirect evidence about the bene-fits of non-FDI inflows can also be inferredby looking at the counterfactual case: are

there significant losses in terms of slowergrowth when countries have capital controls,especially on short-term and portfolio flows?Based on the experience of 20 OECD coun-tries in 1950–89, Alesina, Grilli, and Milesi-Ferretti (1994) find no negative impact ofcapital controls on GDP growth. Using asimulation model, Razin and Yuen (1994)show that the long-run effects of liberalizingcapital flows are very modest. Rodrik (1998)uses a GDP per- capita growth equation anda simple index of capital account opennesswith a sample of almost 100 industrial anddeveloping countries for 1975–89 and findsthat capital account convertibility has no sig-nificant effect on growth once other effectsare taken into account (figure 3-9). Car-rasquilla (1998) finds similar results for1985–95 for 19 Latin American countriesusing more direct measures of capital con-trols (figure 3-10).

Risk sharing and consumption smooth-ing. In developing countries characterizedby a concentration of exports and economicactivity, allowing domestic banks to diver-sify their portfolio helps reduce their vulner-ability to external (terms of trade) and inter-nal output shocks. The scope for gains fromopen capital accounts may also arise fromrisk sharing and asset diversification. Theevidence for such gains is based on simula-tion models whose results are mixed. Obst-feld (1995) estimates that the potential gainsmay be very significant, while Tesar (1995)finds them small. Levine and Zervos (1998)find no evidence of significant effects ongrowth of international risk sharing throughincreased integration of stock markets.

Another potential source of welfareimprovement from capital f lows isincreased opportunities for consumption

145

There is little connectionbetween non-FDI inflowsand productivity growthFigure 3-8 Correlation between capitalinflows and total factor productivitygrowth in low- and high-savingscountries

Low-savings High-savings All

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

Bivariate correlation

Non-FDI

FDI

Note: Sample of 18 countries receiving substantialcapital inflows in the late 1980s and early 1990s.Source: World Bank staff estimates.


smoothing in the presence of high incomevolatility.43 A country that is isolated finan-cially would have to accommodate anyexternal shock through changes in con-sumption and investment. In contrast, acountry that is well integrated with worldfinancial markets can lend and borrow andthus maintain consumption and investmentclose to desirable levels—even whennational income is fluctuating. The gainsmay be larger for developing countries withmore income volatility.44 The general obser-vation, however, that capital flows tend tobe procyclical in developing countries indi-cates that consumption smoothing is notsignificant.45 More detailed evidence alsosuggests that while capital inflows may

have reduced the volatility of consumptionrelative to that of income, on average theyare associated with increased volatility.Results for a sample of 17 countries thatgained significantly greater access to pri-vate capital flows show that volatility dur-ing the inflow surge remained higher forconsumption than for income, with the dif-ference increasing in 10 countries. Thus,the gains from consumption smoothingappear uncertain and limited.

Risks associated with capitalaccount opennessThe risks associated with capital account lib-eralization and capital flows for a develop-ing country depend on the ability of policy-making institutions as well as the financialand corporate sectors to adjust to shocks

146

There is no evidence…Figure 3-9 Economic growth and capitalaccount liberalization in 100 countries,1975–89

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆ ◆◆◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆◆

◆◆◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆◆

◆

◆

◆

◆

◆◆

◆

◆

◆

◆

◆◆

◆

◆

◆◆ ◆

◆

◆

◆ ◆

◆

◆

◆

◆

◆

◆◆

◆◆ ◆

◆◆

◆

◆

◆-0.06

-0.04

-0.02

0

0.02

0.04

0.06

-0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1

GDP growth

Capital account liberalization index

Note: The figure is a partial scatter plot (controllingfor per capita income, secondary education, qualityof governmental institutions, and regional dummiesfor East Asia, Latin America and the Caribbean, andSub-Saharan Africa).Source: Rodrik 1998.

…that capital controls slowgrowthFigure 3-10 GDP growth and capitalcontrols in Latin America, 1985–95

◆

◆

◆

◆◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

◆

-2

-1

0

1

2

3

4

5

6

7

0 0.5 1.0 1.5 2.0

GDP growth

Index of capital controls

Source: Carrasquilla 1998.


and absorb risk, as well as on its own volatil-ity. Various financial assets traded interna-tionally differ in their volatility and implica-tions for increased vulnerability to crisis,46

but three arguments link financial integra-tion and increased risks of financial crises.

The first is that openness to capitalflows may increase the risk of currencycrises if surges and reversals of capitalflows (and crises) occur independently of acountry’s policies and actions. 47 Wheninternational interest rates rise, interna-tional investors are likely to cut back theirfinancing to developing countries. At thesame time, the capacity of developing coun-try banks, firms, and governments to ser-vice debt is reduced. There is strong empiri-cal evidence that international interest ratesare a major determinant of non-FDI capitalflows,48 and are a big factor in the proba-bility of crises (Frankel and Rose 1996;Kaminsky and Reinhart 1997). Foreigninterest rates and a volatile external envi-ronment have also been found to be signifi-cant determinants of banking crises and,therefore, indirectly of currency crises (seediscussion of domestic financial sector).

A second argument against financialintegration is that international capitalmarket failures can aggravate domesticfinancial weaknesses and have contagioneffects. A third is that integration, while notthe root cause of financial crises in emerg-ing markets, may contribute to crises whoseorigin is domestic—especially given weakfinancial systems and inappropriate macro-economic policies.49

While there is little direct evidence ofthe role of capital account liberalization orcapital inflows in financial crises, there issome indirect evidence. Since the 1980s

there has been a negative correlationbetween capital flows and the lifting of cap-ital controls (IMF 1997a). At the sametime, currency crises have increased, whichmay indicate causality between capitalaccount liberalization and currency crises(Wyplosz 1998). Most empirical analyses,however, have failed to find statistical evi-dence linking the volume of capital inflowsto crises (Sachs and others 1996). Anexception is Radelet and Sachs (1998), whofind some evidence of a relationshipbetween crises and capital account deficits(but not current account deficits). Frankeland Rose (1996) also find that higher FDIflows (relative to debt) are associated witha lower probability of crises.

There is additional indirect evidencelinking capital flows with crises for a sampleof 27 capital inflow surges in 21 countries(table 3-2). In 1996 these countriesaccounted for 69 percent of private flows todeveloping countries (or 83 percent, when

China is excluded). The mean ratio of totalprivate-to-private capital flows to GDP overthe inflow periods ranges from 2.2 percentto 11.8 percent. The composition of theseinflows varies considerably, with the meanratio of FDI to non-FDI private-to-privateflows ranging from a negative 0.1 to 3.4. Inabout two-thirds of the cases, there was abanking crisis, currency crisis, or twin crisesin the wake of the surge.

147

The risks associated with capitalaccount liberalization depend on a developing country’s ability to adjust to shocks and absorb risk, and on thevolatility of flows.


Volatility of non-FDI and portfolioflows. The risks associated with capitalaccount liberalization hinge on the volatil-ity of capital flows and the risks of reversalduring bad times, when access to additionalfinancing is especially important. For FDI

flows, the risks are small because theseflows respond more to longer-term consid-erations than to short-term internationalinterest rates, and because they interact lesswith domestic financial markets. The risksof large reversals are even lower because

148

Indirect evidence links capital inflow surges with crisesTable 3-2 Surges in private-to-private net capital inflows and financial crises

Mean ratio of Mean ratio ofannual capital FDI to non-FDI

Country Inflow flows to GDP capital inflows Crisis following inflow episode

Argentina 1991–94 2.5 1.0 1994–95 banking crisis followingMexican devaluation

Brazil 1992–96 3.1 0.2 1995 banking crisisChile 1978–81 11.1 0.1 1982–83 currency and banking crisisChile 1989–96 5.1 0.7 No crisisColombia 1992–96 4.4 1.2 No crisisCosta Rica 1986–95 5.5 1.0 No crisisCzech Republic 1993–96 8.3 0.6 1997 currency crisisEstonia 1993–96 5.4 3.4 1997 near-crisisHungary 1993–95 11.8 1.1 1995 crisisIndia 1994–96 2.5 0.3 No crisisIndonesia 1994–96 3.7 1.1 1997 crisisKorea, Rep. of 1991–96 2.5 –0.1 1997 crisisMalaysia 1982–86 3.1 (a) 1985–88 banking crisis Malaysia 1991–96 9.8 2.5 1997 crisisMexico 1979–81 2.5 0.7 1982 crisisMexico 1989–94 4.5 0.6 1994/95 financial crisisMorocco 1990–96 3.2 0.6 No crisisPakistan 1992–96 3.5 0.4 No crisisPeru 1988–96 6.9 0.4 No crisisPhilippines 1989–96 4.5 0.5 1997 crisisPhilippines 1978–82 3.0 0.0 1981 banking crisis

1983–84 currency crisisSri Lanka 1991–95 5.3 0.3 No crisisThailand 1978–84 3.0 0.3 1983 banking crisis

1984 currency crisisThailand 1988–96 9.4 0.2 1997 crisisTunisia 1992–96 3.6 2.5 No crisisVenezuela 1992–93 2.2 0.0 1993–94 banking crisis

1995 currency crisisVenezuela 1976–79 3.9 –0.1 1980 banking crisis

Note: The inflow episodes were selected based on the length (minimum of two years) and the volume of total private-to-privatecapital flows as a percentage of GDP (minimum ratio of 2 percent).Total private-to-private capital flows = total private capital flows – public and publicly guaranteed private creditors.Total private capital flows = total flows – official flows – net use of IMF credit.Total flows = foreign direct investment + portfolio investment + other investment + net errors and omissions.Total non-FDI private-to-private capital flows = total private-to-private capital flows – foreign direct investment.a. The mean ratio is very high and negative, reflecting a very low negative denominator (non-FDI private-to-private capital flows).Source: IMF Balance of Payments 1996 and 1997; World Bank 1998a; Kaminsky and Reinhart 1997.


FDI inflows are usually invested in longer-term assets (plant and machinery, and ser-vices, for instance) that cannot be liqui-dated quickly. For non-FDI flows, there areclearly more risks of volatility and rever-sals, but they may differ according to vari-ous categories.50

Volatility of non-FDI flows and stabil-ity of FDI flows for three countries—Argentina, Mexico, and Hungary—suggestdifferent characteristics and behavior, par-ticularly in times of downturns (figures 3-11 to 3-13). FDI is far less volatile and lesssubject to reversals.51 It even continues toincrease in downturns. Non-FDI private-to-private flows, in contrast, are much morevolatile. Portfolio equity most closelyresembles FDI, but is more volatile. Debtportfolio investment (including private-to-

public) is volatile and intensifies the sever-ity of financial crises. Non-FDI and debtportfolio flows increase in the years justbefore a crisis, then reverse sharply afterthe crisis occurs. These features magnifyboom-bust cycles and, hence, the severity offinancial crises in small, financially opendeveloping countries.

Volatility of short-term interbankflows. Interbank borrowing also tends to behighly volatile. The reversal in flows fromBank for International Settlements (BIS)reporting banks in Korea and Thailand wasdramatic in the second half of 1997(figure 3-14). The liquidity crisis in bothcountries largely reflects this reversal, par-ticularly in short-term interbank creditlines. In contrast, FDI flows held up, atleast in the first half of 1998 (when they

149

Portfolio equity flows…Figure 3-11 Net capital flows toArgentina, 1990-96

◆

◆

◆ ◆ ◆

◆

◆

1990 1991 1992 1993 1994 1995 1996

-10,000

-5,000

0

5,000

10,000

15,000

20,000

25,000

30,000

Millions of U.S. dollars

FDI

Portfolio equity

Portfolio debt

Non-FDI

Source: International Monetary Fund, World Bank.

…closely resemble FDI…Figure 3-12 Net capital flows to Mexico,1990–96

◆

◆◆

◆

◆

◆

◆

1990 1991 1992 1993 1994 1995 1996

-25,000

-20,000

-15,000

-10,000

-5,000

0

5,000

10,000

15,000

20,000

25,000

FDI

Portfolio equity

Portfolio debt

Non-FDI




were higher than in 1997), even though theprospects for the following period will alsodepend on the global situation.

What are the implications?The larger risks and uncertain benefits ofportfolio and short-term flows for coun-tries with weak institutional capability andfinancial systems suggest proceeding care-fully with capital account convertibility.Because the risks stem largely from the dis-tortions and externalities associated withinternational borrowing and from thewedge between social and private rates ofreturn, and social and private risk, policyshould attack distortions at or close to theirsource. Since the capacity to implementsuch policies and their effectiveness maynot be perfect, this approach must be prag-matic and take account of developing coun-tries’ specific conditions.

The first step is to eliminate tax incen-tives and other distortions that encourageshort-term capital inflows. Another is to useprudential regulations on currency andmaturity positions by banks. The Basle CorePrinciples for Effective Banking Supervisionrecommend only that banking supervisors

150

…but debt investment isvolatileFigure 3-13 Net capital flows to Hungary,1990-96

◆◆

◆

◆

◆

◆

◆

1990 1991 1992 1993 1994 1995 1996

-4,000

-3,000

-2,000

-1,000

0

1,000

2,000

3,000

4,000

5,000FDI

Portfolio equity

Portfolio debt

Non-FDI



The reversal in short-term credit in the second half of 1997was dramaticFigure 3-14 Rate of change of total debt outstanding by BIS-reporting banks to banks andnonbanks

Banks Nonbanks Banks Nonbanks-35

-25

-15

-5

5

15

25

35

Percent

Dec. 1995 – June 1996 June 1996 – Dec. 1996 Dec. 1996 – June 1997 June 1997 – Dec. 1997

Rep. of Korea Thailand

Source: Bank for International Settlements.


ensure that bank managers set appropriatelimits and implement adequate internal con-trols on foreign currency exposure. Butdeveloping countries should also introducespecific limits on currency and maturitymismatches (for example, requiring mini-mum liquid foreign currency assets to covershort-term foreign currency liabilities), andprudential regulations limiting the aggregateopen currency positions of banks, includingderivatives. But because even well-managedfirms and financial institutions have runinto severe losses through the use of suchinstruments, there is also a need for bettersupervision of these regulations and of riskmanagement procedures. Countries mayalso introduce more stringent liquidityrequirements in terms of foreign assets rela-tive to foreign liabilities than for domesticcurrency liabilities.

Prudential regulations of banks andfinancial institutions does not resolve therisk of excessive exposure by the corporatesector. Banks may satisfy foreign currencyexposure requirements by borrowing in for-eign currency and lending in foreign cur-rency to domestic firms. If domestic corpo-rations do not have foreign exchange cover,the currency risk for banks is transformedinto a credit risk. Thus, additional measuresare needed for domestic corporations. Thesemay include requiring disclosure of short-term and unhedged borrowing, reducing thetax deductibility of such borrowing, and therating of firms raising funds abroad and list-ing on the domestic stock exchange.52

When the domestic regulatory andsupervisory system for banks is weak, con-trols over corporations are ineffective, andaccess lender of last resort is uncertain,restrictions on capital flows may be useful.

This often implies maintaining or reinforc-ing capital account restrictions. For coun-tries that are reintroducing such restric-tions, this may mean loss of credibility, sosuch actions have to be managed in a way

that does not lead to even greater loss ofconfidence. The imposition of capitalaccount restrictions, as part of a preventivepackage to minimize the risks of financialcrisis, is concerned mainly with capitalinflows. Their reintroduction for capitaloutflows during a crisis poses many diffi-cult problems, not considered here.

Restrictions on capital flows shouldminimize distortions and be as market-oriented as possible. One way is explicittaxes or reserve requirements on foreignexchange liabilities according to holdingperiod. In Chile and Colombia implicit taxeshave substantially shifted the composition ofsuch flows and discouraged short-term flowswithout having much impact on the volumeof flows (box 3-7; World Bank 1997b; Mon-tiel and Reinhart 1997). Restrictions on cap-ital flows have to reflect specific factors,such as administrative capability, and haveto balance the need to be comprehensive inorder to minimize distortions and evasionswith the need to discriminate between capi-tal inflow categories, according to the bene-fits and risks associated with such flows.

151

The larger risks and uncertain benefitsof portfolio and short-term flows forcountries with weak institutionalcapability and financial systems suggestproceeding carefully with capitalaccount convertibility.


The international financialsystem

The international environment plays animportant part in financial crises in

emerging markets. Volatility in interna-tional interest rates and economic growthin industrial countries affect the allocationof assets to emerging markets and createrisks of booms and reversals in capitalflows. Other characteristics, such as volatil-ity and sudden shifts in market sentimentassociated with euphoria, panics, herdbehavior, and contagion are also influen-tial. These failures in international financialmarkets have implications for internationalfinancial institutions.

Proposals for reforming the interna-tional financial system architecture havebeen under discussion since the Mexicancrisis. A working group, under the auspicesof the Group of 10 (G-10) industrial coun-tries, drafted the Resolution of SovereignLiquidity Crises, which focuses on sovereignbonds.53 Discussions have gained urgency

with the outbreak of the East Asian crisisand its global spread. The Group of 22countries (G-22) established three workinggroups on enhancing transparency andaccountability; strengthening financial sys-tems; and managing international financialcrises. These working groups have nowfinalized and submitted their reports,54 anddiscussions of these proposals in officialforums began in early October 1998. TheG-7 countries have since agreed on a num-ber of specific initiatives to strengthen theinternational financial system (Group ofSeven 1998). These include, in the immedi-ate context, an enhanced IMF facility toprovide a precautionary line of credit and aWorld Bank emergency facility to providesupport to countries at times of crisis for theprotection of vulnerable groups and finan-cial sector restructuring; and, in the longer-term, agreement on other principles tostrengthen the global financial system,including greater transparency, enhancedsurveillance, orderly and progressive capital

152

Chile introduced restrictions on capital inflows in1991 through unremunerated reserve require-ments (World Bank 1997b). These reserves,which have to be maintained for one year regard-

less of loan maturity, constitute an implicit tax on foreignborrowing that varies inversely with the holding period. In1995, reserve requirements were extended to all types offoreign financial investments, including American deposi-tory receipts. Colombia introduced capital controls in1993 through unremunerated reserve requirements ondirect external borrowing with a maturity of less than 18months. These were subsequently tightened, requiringreserves for all loans with maturities of less than five years.

Chile has since lowered the reserve requirement to zero.It is difficult to gauge the effects these restrictions

have on the volume of flows, as a change in flows couldalso be caused by other macroeconomic and financialdevelopments. The restrictions in Chile and Colombia canbe thought of as an implicit tax that significantly increasedthe interest differential between domestic and foreignshort-term interest rates. Econometric studies that use thisapproach to estimate their effects suggest that they sub-stantially changed in the term structure of external bor-rowing—discouraging short-term inflows—and encour-aged equity investment in Chile and Colombia (Cardenasand Barrera 1997; Quirk and others 1995).

Box 3-7 Restrictions on capital flows in Chile and Colombia


account liberalization, orderly resolution offuture crises, and the need for good prac-tices in social policy to protect the most vul-nerable. Other announcements include theneed to pursue further proposals forstrengthening prudential regulations inindustrial countries to promote safe andsustainable capital flows, strengtheningfinancial systems in emerging markets, andimprovements in other related areas. Thissection considers five main issues that arestill evolving and remain subject to somedebate.55

An international lender of lastresort?Arguments have been advanced for an inter-national of lender of last resort, but sucharguments also raise unresolved issues.

The traditional argument concerns thepossibility of systemic risk. If a countryfails to serve bank debt—whether sovereignor private—it may undermine the liquidityand even the solvency of banking systemsin creditor countries. This risk was clearlypresent during the debt crisis of the 1980s,when BIS reporting banks’ direct exposureto major emerging markets exceeded theircapital, but it was much weaker in the 1997East Asian crisis (table 3-3). This argumenthas lost some of its force with the greaterrisk diversification by banks and use ofnon–bank-based financial instruments.56

A second argument is based on theabsence of an effective national lender oflast resort (Mishkin 1998), whetherbecause the country has chosen a currencyboard or because of the intrinsic difficultyin a small, highly open economy of aninternal resolution of the liquidity problemsof the domestic financial system.57

A third argument is based on the riskscaused by contagion and the potential spreadof panic among international investors (theAsian crisis provided yet another strikingexample of this). When a vulnerable currencyis attacked, the attack may spread to othercountries’ currencies, even when their funda-mentals are sound. A lender of last resortwould provide reserves to emerging marketsthreatened by speculative attacks and thusprevent a currency collapse.

A final justification for a lender of lastresort is on social welfare grounds. Whilemarket participants should bear the conse-quences of their actions and incur the costsof a crisis, some costs are borne by groupsnot responsible for the crisis, particularlythe more vulnerable.

The G-7 adopted the principle of estab-lishing a precautionary bilateral and multi-lateral line of credit to countries that are atrisk and pursuing strong IMF-approvedpolicies—to be drawn only in the event of a

153

BIS banks’ exposure to emergingmarkets is much less than in the1980sTable 3-3 Commercial (BIS-reporting) banks’exposure to emerging markets(debt as percentage of banks’ capital)

MajorEast Latin emerging

Asia-5a America-5b marketsc

All BIS-reporting banksEnd 1982 19.1 58.1 101.1June 1997 18.8 14.2 50.0

German banks 17.0 13.7Japanese banks 39.2 5.2U.S. banks 6.8 14.5

a. Indonesia, Republic of Korea, Malaysia, Philippines, andThailand.b. Argentina, Brazil, Chile, Mexico, and Venezuela.c. Major emerging markets: East Asia-5, Latin America-5,China, Colombia, Czech Republic, Hungary.Source: Bank for International Settlements and Organisationfor Economic Co-Operation and Development


liquidity need. This would have potentiallyimportant benefits in helping to avert crisesby reducing perceptions of uncertaintyabout international support and securingcountry policy improvements. There arealso potential caveats to the effectiveness ofthis proposal, the problems being the sameas those that apply to a lender of last resort.

Bailouts, moral hazard, and riskand burden sharingWhether through a formal lender of lastresort or ad hoc rescue packages, bailoutscreate moral hazard. Three types are possi-ble: the first type relates to expectations by

developing country governments of abailout, which can reduce incentives toimplement better policies. In most circum-stances, however, the economic, social, andpolitical costs of a financial crisis are toohigh for such moral hazard to operate. Infact, governments may delay calling oninternational financial institutions—despite the fact that a prompt responsewould reduce the costs of a crisis. A secondtype of moral hazard can arise becauseinternational creditors expect to be pro-tected if a crisis occurs. A third type canarise because banks and private corpora-tions undertaking risky activities expect tobe bailed out under workouts of foreigndebts, leading to the domestic socializationof these debts.

Hard evidence about the extent ofmoral hazard in international lending is

elusive. It has been argued that the Mexicobailout may have contributed to excessiverisk taking in Asia, but the very large gener-alized decline in spreads on lending duringthe period preceding the East Asian crisisacross all emerging markets may have alsoowed significantly to a generalized climateof euphoria. Still, it would be hard con-clude that moral hazard has not been play-ing a significant role in influencing investorand borrower behavior in recent times,especially in the case of Russia before theimmediate runup to the crisis (whenspreads were still moderate). The abruptcut-off in capital flows and sharply higherspreads to all emerging markets as a risk-class following Russia’s collapse may alsobe partly ascribed to the realization thatbailouts were no longer certain.

A supervisory role would be requiredfor an international lender of last resort tominimize moral hazard. This implies usingconditionalities for prudent macroeco-nomic management, implementing institu-tional reforms to reduce risks of crisis, andsupporting measures that reduce incentivesfor (and introduce restrictions on) excessiverisk taking.58 Imposing this supervisory roleon sovereign governments poses many chal-lenges, however (Obstfeld 1998).

Bailouts also require dealing with riskand burden sharing issues, which meansadopting clear rules to make sure that pri-vate operators bear some of the costs of theirrisky behavior. For domestic debtors, guar-antees may be justified only for commercialbanks in order to protect the payments sys-tem. These guarantees have to be pairedwith significant debt-reduction concessionsby private creditors (Goldstein 1998), whichmust bear some of the costs of a crisis and

154

Whether through a formal lender oflast resort or ad hoc rescue packages,bailouts create moral hazard.


not be the only ones bailed out through theintervention of the lender of last resort.59

The size of rescue packages hasincreased dramatically in recent years.Large amounts are thought to be necessaryto quiet down markets as they panic. Butsuch “rescue creep” has risks. No reason-able amount of public money can stop ajustified speculative attack. By themselves,larger packages worsen moral hazard prob-lems and may lead to excessively toughconditions, defeating the end objectives.

In the final analysis, recourse to alender of last resort depends on resolving aseries of issues: the political concerns asso-ciated with the need to supervise sovereigngovernments, the tradeoff between theshort-run benefits of avoidance or reduc-tion in the severity of crisis and the long-run risks from moral hazard, and the avail-ability of alternatives to official newlending. In the present international archi-tecture, the mandate and correspondingresources to play this role are lacking.Given such limits, better national risk man-agement in private and public spheres willremain a key.

Complements to new officiallending, and involvement of theprivate sector in crisisprevention and resolution A first cushion against a reversal in capitalflows is adequate international reserves, acommon but costly policy. Another possi-bility is to enter into private marketarrangements that guarantee liquidity up toa predetermined limit. Argentina has such acontingent repo facility with internationalbanks.60 Indonesia had standby creditoptions, but the amounts were far too small

to cope with the country’s financial crisis in1997.

Another alternative is to promotedebtor-creditor negotiations to reach restruc-turing agreements allowing rollovers, exten-sion of maturities, and reduction of debt. Ifclear and predictable, such workouts canhelp reduce lending distortions and inducebetter pricing of risk.61 The main implemen-tation issue is collective action by creditors.Every creditor has an incentive to try to getout first or to “free-ride” on others’ accep-

tance of workout arrangements. Negotia-tions are difficult to initiate, protracted, andhard to enforce because of informationasymmetries and transactions costs (Eichen-green and Portes 1995). The collectiveaction problem is much more challenging ina crisis that involves mostly private-to-pri-vate debt (as in East Asia) than in oneinvolving public debt (as in the crisis of the1980s).62 Further complicating the processis the much greater number of creditors anddebtors than in the past and the centrality ofexchange risk, as recent debt workouts inKorea and Indonesia show.

Three types of contract clauses in debtinstruments can be used to improve credi-tor coordination: collective representation,such as in bondholder councils; qualifiedmajority voting; and sharing clauses thatdiscourage dissident creditors from engag-ing in disruptive legal proceedings (Eichen-green and Portes 1995, Goldstein 1998).63

155

By themselves, larger packages worsenmoral hazard problems and may lead toexcessively tough conditions, defeatingthe end objectives.


Debtor and creditor country govern-ments are the central players in orderlyworkouts of private debts, as well as sover-eign debt (Aggarwal 1998). The debtorcountry usually has primary responsibilityfor setting negotiations, especially given theadjustment policies they will have to imple-ment. This should not lead to provision ofgovernment guarantees and the socializa-tion of private debts, however, which hap-pens frequently (including in the recentKorean and Indonesian agreements) andexacerbates moral hazard. Creditor govern-ments play a crucial role by forcing theirfinancial institutions to the negotiatingtable, to see that the private sector bearssome of the costs of risk taking. Interna-tional financial institutions, which haverestrictions on lending into arrears, need toformalize the conditions for exceptions(which are frequent in practice) if they areto negotiate with private creditors in pro-viding additional liquidity.

The external debt workout must alsobe properly sequenced with domestic debtrestructuring. External creditors should notreceive undue precedence or seniority.

The most critical aspect of a debtworkout, however, is the temporary sus-pension of debt payments, which helps stopthe decline in the currency and buys time toput in place a credible adjustment programand to organize debtor-creditor negotia-tions. By allowing an orderly debt restruc-turing, it could result in better outcomes forboth the debtor country and creditors.

To be effective, however, the standstillhas to come at the right time. That timinghas to take into account three factors: one isthat governments may delay declaring a debtstandstill, fearing a loss of confidence and

credibility and thereby greatly reduce thebenefits. This seems to have been the case inEast Asia in 1997. A second consideration isthe need to prevent debtor governments withweak reputations from making excessive useof standstills; debt standstills should be pos-sible only under exceptional circumstancesand in extreme distress. The third is the needto get a standstill in place at the earliest pos-sible date, so that all (or at least most) credi-tors share in the costs of restructuring.

Improved regulation andstepped-up supervision on banklending in creditor countriesAsymmetric informational problems aremore acute in cross-border lending and canlead to less discriminating and more riskylending. This was the case in the runup tothe debt crisis in the 1980s, and seems tohave occurred in the East Asia crisis. Wit-ness the dramatic drop in spreads forKorean and Thai private borrowing, withspreads between bank and corporate bor-rowers nearly equalized by 1996–97 (fig-ures 3-15 and 3-16).

Improved prudential regulation andstepped-up supervision of internationallyactive banks in creditor countries can helpreduce these risks.64 One proposal is torequire a higher risk weight (than the 20 per-cent under the Basle rule) for lending toemerging markets,65 based on the assessmentof the country’s financial system. This wouldraise the cost of borrowing to many develop-ing countries,66 but by improving the pricingof risk, it should reduce the incidence of crisisand the volatility of lending and interest ratespreads, and increase incentives for reform.

Information about and assessment ofnational financial systems, including the

156


quality and effectiveness of supervision andimplementation of domestic regulations orglobal standards, may be valuable. Marketparticipants may make use of it, and regu-lators in creditor countries may require itsuse in risk management by lending banksunder their supervision.

Information and monitoring ofvulnerabilitiesMore good information is always betterthan less. At the same time, complete trans-parency does not exist, and better informa-tion (recognizing the limits of costs in com-p i l ing such in format ion) wi l l notnecessarily prevent crises. Even with elabo-rate disclosure rules, information asymme-try remains, as recent crises in industrialcountries (for example, Republic Bank and

Orange County bankruptcies in the UnitedStates, and financial crises in a number ofScandinavian countries) demonstrate. Still,there are potential benefits to better infor-mation and disclosure and there are twodifferent sets of issues under discussion:improvements in information and disclo-sure standards, and better use of informa-tion to assess national vulnerabilities andundertake measures to forestall crises.

Transparency and accountability. As inthe Mexican crisis in 1994, the East Asiacrisis highlighted weaknesses in the cover-age, frequency, and timeliness of informa-tion available to assess vulnerabilities.Before the onset of the Asian crisis, and forseveral weeks (if not months) after, theamounts of foreign liabilities to which thecountries were exposed were not precisely

157

Spreads dip in Korea…Figure 3-15 Spreads between corporateand bank borrowers in the Republic ofKorea, 1991–97

1991 1992 1993 1994 1995 1996 1997

30

40

50

60

70

80

90

100

110

Private corporations

Commercial banks and financial institutions

Basis points

Source: Euromoney (loanware).

…and in ThailandFigure 3-16 Spreads between corporateand bank borrowers in Thailand, 1990–97

1990 1991 1992 1993 1994 1995 1996 1997

40

50

60

70

80

90

100

110

120

130

Private corporations

Commercial banks and financial institutions

Basis points

Source: Euromoney (loanware).


known. Uncertainty about short-term debtand foreign exchange reserves exacerbatedthe financial panic and crisis. The EastAsian crisis has illustrated the role of pri-vate position-taking by nonfinancial firms,which are difficult to monitor in a liberal-ized environment. Questions have also beenraised about the disclosure of informationand transparency of international agenciesthemselves. Accordingly, improvements are

needed on information and disclosure at alllevels (private sector, national authorities,and international financial institutions), butinternational standards need to be appliedcarefully and progressively over time, rec-ognizing constraints and costs.

Following the Group of Seven (G-7)Halifax proposals, international financialinstitutions, in conjunction with nationalauthorities, are working to improve thequality and timeliness of information oncentral bank reserves, short-term foreigncurrency debt (including central bankderivatives transactions), and domesticfinancial sector indicators (such as nonper-forming loans and short-term debt).67 Fornational authorities, better disclosure andaccounting standards, especially about for-eign exchange liquidity positions, withrespect to their financial institutions andprivate corporations, are also important.For international institutions, the presump-

tion is toward greater release of informa-tion to the public, except in clearly definedcases where confidentiality requirementsoverride the gains from making informa-tion public.

More and better information can alsobe made available from creditor countrysources and from the BIS. This informationcan be used to improve risk assessment.There is also a case for better privateefforts at collecting information, despitethe failure of rating agencies to adequatelyassess risks in Asian countries during therunup to the crisis.68 The G-22 workinggroup on transparency and accountabilityhas also recommended that modalities forcompiling and publishing data on interna-tional exposure of investment banks, hedgefunds, and other institutional investors beexamined.

Warning indicators and manifestationsof vulnerability. Warning indicators areunlikely to predict crises, particularly theirtiming, but they can provide timely andbetter information about impending prob-lems so that policymakers can take preven-tive actions. The literature has used twoapproaches. The first, the signals approach,aims at determining characteristic andabnormal behavior of a set of variables—leading indicators—preceding crises, rela-tive to tranquil periods. The indicators thatpredict the most actual crises and producethe least false alarms are used as leadingindicators. The second is the regressionapproach, which looks at the statistical sig-nificance of various indicators in models ofcrisis determination.69 Warning indicatorsof vulnerability usually flash positive sig-nals for many variables at the same time.There is, however, no uniform, well-defined

158

Warning indicators are unlikely topredict crises, particularly their timing,but they can provide timely and betterinformation about impending problemsso that policymakers can take preventiveactions.


set of indicators. Country conditions tendto be crucial in determining the significanceof specific indicators (IMF 1998b; Gold-stein and Reinhart 1998). Almost all stud-ies have found that traditional indicators ofvulnerability—notably those relating toindebtedness, fiscal policy, sovereign riskratings, and interest spreads—have failed tosend useful warning indicators of currencycrises.

Vulnerability indicators of currencycrises. The most important signals of a cur-rency crisis are real exchange rate apprecia-tion; international illiquidity, as measuredby the ratio of short-term liabilities toreserves, money stock to reserves, or for-eign assets to liabilities (of banks); andlending booms financed by foreign borrow-ing. Other significant indicators are slowerGDP and export growth, higher foreigninterest rates, deteriorating terms of trade,a decline in equity prices, and a bankingcrisis (Kaminsky and Reinhart 1997; Gold-stein and Reinhart 1998).

The indicator most commonly associ-ated with currency crises is the size of thecurrent account deficit.70 Empirical workhas generally failed to find current accountdeficits helpful by themselves in predictingcrisis, however (Frankel and Rose 1996;Sachs and others 1996; Milesi-Ferretti andRazin 1996; Radelet and Sachs 1998). Anexception is Goldstein and Reinhart(1998), who find that ratios of currentaccount deficit to GDP and to investmenttop the list of leading indicators of currencycrises. In any case, current account deficitsremain important in assessing vulnerabilityif complemented by analysis of the causalfactors. Large or fast-increasing deficitsshould always be monitored, since they

usually reflect rising capital inflows.Deficits should also be carefully monitoredif spending is going to consumption ratherthan investment—particularly in the trad-ables sector—since there is presumption oflower risks (because of faster growth ofGDP and exports). The East Asian crisishas shown that the allocation and efficiencyof the increased investment is also relevant.

Warning indicators of banking crises.Most indicators of banking crises aremacroeconomic, and closely related tothose for currency crises (Goldstein andTurner 1996; Demirgüç-Kunt and Detra-giache 1997; Kaminsky and Reinhart1997). Work is being done on developingstructural or microeconomic warning indi-cators. Relevant variables include spreadsbetween deposit and lending rates, access tointerbank loans, changes in the ratio ofcapital to risk-weighted assets, the loans-to-deposits ratio, foreign currency exposure,government ownership, and the proportionof lending at the discretion of banks anddirected by government (Honohan 1997;Rojas-Suárez 1998).

Notes1. Greenwald, Stiglitz, and Weiss (1984); Green-

wald and Stiglitz (1993); Mishkin (1991, 1997);Stiglitz (1998b, 1998c).

2. Many empirical studies consider currencycrises to be episodes of large devaluations (Edwards1989; Edwards and Montiel 1989; Frankel and Rose1996). In contrast, Eichengreen, Rose, and Wyploz(1995) and Kaminsky and Reinhart (1997) favor abroader approach, focusing on devaluations as well asepisodes of unsuccessful speculative attacks. Otkerand Pazarbasioglu (1996) regard crises as includingcases of devaluation, increases in the rate of crawl, andshifts to a more flexible exchange rate system.

3. During 1870–1913 a number of then-emergingeconomies also received large amounts of capital

159


inflows. For 1870–89 and 1890-1913 the largest vol-umes (mean absolute value of current account as per-centage of GDP) were respectively: 18.7 percent and6.2 percent for Argentina, 8.2 percent and 4.1 percentfor Australia, and 7.0 percent and 7.0 percent forCanada (Obstfeld 1998).

4. Stiglitz (1998a); Summers (1998).5. Kumar, Moorthy, and Perraudin (1998) find

that a decline in portfolio flows has a stronger impacton the probability of crisis than a decline in FDI.

6. Other factors include the greater degree ofinherent risks present in developing countries, due totheir narrower economic bases (smaller economiesspecialized in fewer economic activities).

7. Surges in capital inflows can occur eitherexogenously, because of events in the world economyoutside the control of policymakers of the economy inquestion, or endogenously, because of changes incountry policies and circumstances (Hernandez andRudolf 1995; Gavin et al. 1995; Montiel and Rein-hart 1997). They also respond to the macroeconomicpolicy mix of the capital importing country, as well asthe capital market structure (Montiel and Reinhart1997).

8. Fernandez-Arias and Montiel (1995) also findthat in half of a sample of 12 countries experiencingthe largest inflows relative to the size of theireconomies, reserve accumulation accounted for about40 percent of the inflows.

9. In Corsetti, Pesenti, and Roubini (1998) thisinconsistency is between fixed exchange rate, govern-ment bailout guarantees (and their implication formonetary policy), and foreign debt accumulation andcurrent account deficits (or capital mobility).

10. There is evidence showing a significant degreeof both openness and capital mobility in developingcountries. It is based on interest parity tests: Edwardsand Khan (1985); Khor and Rojas-Suárez (1991);Haque and Montiel (1991); Reisen and Yèches (1993);Robinson (1991); and Dasgupta and Dasgupta (1995);and correlation between savings and investment: Doo-ley, Frenkel, and Mathieson (1987); Wong (1988).

11. The cost of sterilization may be significant:from 0.5 to 2 percent of GDP per year in Chile andColombia in the 1990s (Williamson 1996), and 0.3 to0.75 percent of GDP per year for Malaysia, Thailand,and Indonesia in 1990–96 (ADB and World Bank1998).

12. Private borrowers in Latin America in the1990s generally displayed a far greater willingness tohedge their foreign exchange liabilities, while borrow-ers in East Asian countries generally avoided them—partly because historical nominal exchange ratevolatility (and the volatility of financial prices) wasmuch higher in Latin America than in East Asia.

13. The changes in the real exchange rate are par-ticularly welcome if they reflect price adjustments inresponse to fundamental factors such as a permanenttransfer of resources from increased capital inflows,shifts and gains in productivity following reforms,improved terms of trade, correction of earlier excessivedepreciation, or increased levels of consumption toequilibrium levels consistent with higher permanentincome (and the need to incur current account deficits).

14. The impact is minimized even in the absenceof sterilization simply because, as the currency appre-ciates, the extent of the impact on domestic money isreduced by that exact amount of appreciation.

15. Assuming that the prevailing conditions donot justify a rise in long-term equilibrium exchangerates. Measuring whether prevailing exchange ratesare misaligned with fundamentals is, however, notori-ously difficult. In particular, relative purchasing powerparity movements may not always provide the correctpicture of misalignments from equilibrium exchangerates, and there may be other, better measures (Broner,Loayza, and Lopez 1998).

16. Schadler et al. (1993) and Dasgupta and Das-gupta (1995) find such evidence. This may be due tolack of credibility of low-inflation programs (Kaminskyand Leiderman 1998), to a rise in credit demand, or toincreased riskiness of the financial sector.

17. The forms of such tighter fiscal policy mayhave differential effects. An adjustment that curbsspending or raises taxes on nontradables wouldreduce domestic inflation and interest rates. Alterna-tively, one that curbs spending or raises taxes on trad-ables would improve the current account deficit andreduce borrowing from abroad. Cutting spendingwould have a more direct and immediate effect onaggregate demand than raising taxes, because of lags.

18. Kaminsky and Reinhart (1997) find thatmore than half of the 26 banking crises they studiedwere followed by a balance of payments crisis withinthree years. Conversely, only about 1 in 10 of thebalance of payments crises were followed by banking

160


crises within three years. Also, regression of the mea-sure of banking crises against the balance of pay-ments measure indicates that balance of paymentscrises do not help predict banking crises. Sachs, Tor-nell, and Velasco (1996) find that banking sectorfragility is a major determinant of currency crisis.Milesi-Ferreti and Razin (1996) show that the bank-ing sector plays an important role in determining cur-rent account sustainability.

19. However, Eichengreen and Rose (1998) findno evidence for a role of domestic financial fragility inpredicting banking crises.

20. Sundarajan and Balino (1991) provide evi-dence of this effect in the case of the crises in theSouthern Cone countries during the 1980s: Chile(1981–83), Argentina (1980–82), and Uruguay(1982–85).

21. In addition to any measures and regulationson foreign currency exposure and access to foreignborrowing by banks, discussed below.

22. There is some disagreement as to the effec-tiveness of higher reserve requirements as an instru-ment for restraining lending booms.

23. Honohan (1997); Caprio, Atiyas, and Han-son (1994). Such limits may be set at high levels thatwould not normally be reached, but restrain occa-sional bursts of overexuberant and risky expansion(World Bank 1998b).

24. The contributions of weak corporate gover-nance and transparency to the East Asian crisis areanalyzed in ADB and World Bank (1998).

25. World Bank (1998b). Also, work on develop-ing standards for corporate governance is being under-taken within the OECD.

26. For extensive discussion see World Bank(1997b) and ADB and World Bank (1998).

27. See also the G-22 working group report onstrengthening financial systems, October 1998.

28. An example is the more recently developedvalue-at-risk techniques for risk management.

29. These factors determine the balance of bene-fits (in terms of efficiency and stability) according tothe type of banking system structure, ranging from“narrow banking” to “universal banking” (Kaufmanand Kroszner 1997).

30. Caprio and Klingebiel (1996b) find faultysupervision and regulation to be significant in 26 of 29

bank insolvency cases. Poor bank management is afactor in 20 cases.

31. This consensus also constitutes the core ofthe IMF (1998a) guidelines.

32. Published by the International Accounting Stan-dards Committee (IASC) and the International Federa-tion of Accountants (IFAC). The International Organiza-tion of Securities Commissions (IOSCO) is also workingon establishing universal principles for securities marketregulations, improving disclosure requirements, anddeveloping standards for cross-border offerings.

33. Some may also argue the usefulness of com-petition in setting standards as against harmonization.

34. More radical options would entail the aboli-tion of deposit insurance, the adoption of narrowbanking, or the adoption of free banking; see Caprioand Klingebiel (1996b).

35. Goldstein and Turner (1996); Sundarajanand Balino (1991); Kaminsky and Reinhart (1997);Caprio and Klingebiel (1996b); Demirgüç-Kunt andDetragiache (1997); Gavin and Hausmann (1996).

36. Eichengreen and Rose (1998) find a highlysignificant correlation between changes in industrialcountry interest rates and banking crises in emergingmarkets. Also, Kaminsky and Reinhart (1997) findthat foreign-domestic interest rates signaled crises inall 20 cases for which data are available.

37. Banking consolidation, which at first sightshould allow pooling and diversification of risks, doesnot necessarily do that: larger banks may still take onexcessive risks without an adequate managementstructure in place. Bank supervision in the past con-ventionally focused on balance sheets, but much moreattention is now devoted to the soundness of banks’management processes in assessing and managing risks(Mishkin 1996).

38. The ability of foreign financial institutions toenter domestic markets, which may be part of externalfinancial liberalization but not formally part of capitalaccount liberalization, also contributes to financialintegration and provides benefits similar to those oftrade liberalization, in terms of competitive effects andimproved quality of services and reduced prices.Claessens, Demirgüç-Kunt, and Huizinga (1997) pro-vide empirical evidence that broader foreign owner-ship of banks renders domestic banking markets morecompetitive and reduces domestic bank costs. Also,

161


foreign banks that are internationally and in terms oftheir activities more diversified help strengthen thedomestic financial system. The reverse implication isthat the franchise value of domestic banks may fall,inducing more risk taking. While this negative impactis real and the ensuing risk should be managed, on bal-ance the benefits and risks of foreign entry are thesame as those associated with FDI (in the financial sec-tor) and warrant similar treatment.

39. World Bank (1997b) discusses these benefitsat length. Another benefit sometimes cited is thatfinancial openness submits governments to the hardscrutiny of international markets and would restrainany tendencies for mismanagement. Also, financialdeepening through increased capital flows helpsdevelop capital markets and allows more banking sys-tem intermediation, which are shown to affect growthpositively (Levine and Zervos 1998). Increased inter-national competition also enhances the quality of thefinancial system.

40. Of course the associated benefits of someflows such as trade credit which are closely relatedto trade should not be assessed only within thisframework.

41. In capital inflow surges the increase in totalinflows is due mainly to non-FDI flows, and the con-tribution is larger when the size of capital inflows islarge (greater than 9 percent of GDP). As seen in box3.3, reserve accumulation is also larger.

42. Because these flows have to be serviced, thenet gains are a fraction of the gross. FDI flows toohave to be serviced through profit repatriation, but asignificant part of such profits are reinvested, consis-tent with the long-run nature of such inflows.

43. Low-income developing countries may alsobenefit from long-term consumption smoothing. Theymay borrow and increase their consumption now inview of increased income in the future.

44. The precise welfare improvement associatedwith increased consumption smoothing depends on anumber of factors, such as the time-preference and theshape of the utility function, as well as assumptions aboutmarket structure, country size, and technology. The esti-mates of utility benefits from consumption smoothingvary widely, from nearly 0 percent of lifetime consump-tion (Backus, Kehoe, and Kydland 1992; Cole and Obst-feld 1991; Tesar 1995) to a very significant (15 percent)

fraction of lifetime consumption (Obstfeld 1995; vanWincoop 1994). Typically, models that allow incomegrowth to endogenously depend on diversification appearto arrive at higher estimates of gains than models whereincome is fixed. The alternative sets of assumptions alsohave differing degrees of ability to account for the stylizedfacts about consumption volatility.

45. This ignores longer-term consumption-smoothing effects, which are important.

46. Sachs and others (1996) find that countrieswith large short-term, variable, interest and foreigncurrency–denominated debt are more prone to crisis.Radelet and Sachs (1998) find that the ratio of short-term debt to reserves is strongly associated with theonset of crisis, whereas the ratio of long-term debt toreserves is not. Frankel and Rose (1996) find that thelower the reliance on FDI flows (compared to totaldebt), or the greater the reliance on more volatile capi-tal flows, the higher the probability of crisis.

47. Models of self-fulfilling expectations of cur-rency crisis imply that the intrinsic instability of theinternational financial system is a major contributor tocurrency crisis and, therefore, complete openness ofthe capital account implies greater risks for developingcountries. This issue is discussed below.

48. The evidence is discussed in World Bank(1997b), chapter 2. See also Montiel and Reinhart (1997).

49. For instance, McKinnon and Pill (1997)model how excessive foreign borrowing can take placein a recently liberalized domestic financial system withinadequate supervision, and the presence of moralhazard, possibly due to government guaranties, in thecontext of unrestricted access to external finance.

50. Some argue, however, that such distinctionsare not operational, that is, that volatility of flowscannot be distinguished among capital account cate-gories, due to a high degree of substitution amongthese categories. Claessens and others (1995)researched capital inflows to five developing and fiveindustrial countries over a 15-year period (or longer)and found no evidence of patterns in the volatilityamong components of the capital account. Specifically,long-term flows were as likely to be volatile as short-term flows. Similar research was later conducted byChuhan and others (1996), who also found that vari-ous types of capital flows behave similarly. However,they rejected the notion that flows are essentially the

162


same. They focused on interrelationships of the behav-ior of flows, that is, on the relative responsiveness ofone flow to changes in another. They determined thatthe composition of capital flows does matter, specifi-cally that short-term inflows are more responsive to achange in FDI than the reverse and, therefore, suffermuch more from contagion effects. Some have alsoargued that multinational corporations, for instance,hedge long-term FDI by rolling over opposite short-term currency positions, but there is little empiricalevidence to support that view.

51. This may be partly due to lags in the mea-surement of FDI, with disbursement for new invest-ments spread over many years.

52. A more difficult and controversial measure isto set prudential ratios for firms borrowing abroad—such as a minimum equity to liability ratio, maximumforeign to domestic liability ratio, and maximum openforeign exchange position.

53. The so-called Rey Report (May 1996), whichrecommends that financial systems in emerging mar-kets be strengthened, that collective action clauses beadded to bond contracts to facilitate orderly work-outs, and that international financial institutions con-sider “lending into arrears” on sovereign debt owed toprivate creditors.

54. The three working groups, established by theFinance Ministers and Central Bank Governors of 22systemically significant economies included senior offi-cials from these countries and international financialinstitutions, focused on: increasing transparency anddisclosure; strengthening financial systems; andimproving the management of international financialcrises. Three reports of the working groups were pub-lished in October 1998. Also the G-7 summit (Birm-ingham, May 1998) has considered ways to strengthenthe global financial architecture.

55. Another issue concerns regional arrangements.Contagion tends to have, at least initially, a stronglyregional character as demonstrated in both the Mexicanand East Asian crises. These “neighborhood” spillovereffects may be due to underlying linkages or regionalsimilarities as perceived by investors. This provides anargument for institutional arrangements of a regionalcharacter to improve monitoring and surveillance andhelp initiate and implement policies to prevent financialcrises. Regionally coordinated (and pooled) interventionmay also be useful in responding to crisis.

56. There may be increased bank exposure, how-ever, to the extent counterparty risks have increased inrecent years with the proliferation of hedge funds andinvesting on margins.

57. An expansionary monetary policy or lender oflast resort activity to contain financial crisis and provideliquidity is often counterproductive. Such a policywould cause expected inflation to rise and the domesticcurrency to depreciate. The depreciation of the currencywould aggravate the domestic financial crisis, since itleads to a deterioration in the balance sheets of domesticbanks and firms that have debt denominated in foreigncurrency. It may also lead to a jump in expected infla-tion, which would cause interest rates to rise, worseningthe balance sheets of firms and households and poten-tially causing greater losses to banking institutions. Thetotal net result is a worsening of the situation. An inter-national lender of last resort would help overcome theseproblems and contain the domestic crisis.

58. These include the various rules discussedabove: adequate disclosure requirements for banks,adequate capital standards, penalties and sharing inthe costs by managers and shareholders, careful moni-toring of banks’ risk management procedures, promptcorrective action, and so on.

59. The U.S. Shadow Financial Regulation Com-mittee (1998) has proposed a mandatory loss-sharingsystem imposing “haircuts” on foreign lenders whowithdraw or fail to roll over their claims before IMFloans are paid back.

60. The facility allows the Central Bank ofArgentina the option of issuing short-term dollar-denominated government bonds and provincial loanstotaling $6.7 billion to international banks subject to abuyback clause. It pays a fee as long as the facility isavailable, and a spread is determined if it is used. Themechanism allows the Central Bank to act as a lenderof last resort without resorting to domestic money cre-ation, which is not possible under the currency boardarrangement.

61. International debt workouts are usefullycomplemented by strong domestic bankruptcy lawsand systems of debtor-creditor workouts.

62. The case of sovereign debt is discussed in theRey Report of the G-10.

63. In order to avoid the adverse selection effectsof such contract clauses, industrial countries shouldinclude them in their own government bond contracts.

163


64. There are also questions about the effects ofinternational hedge funds on the volatility of exchangerates and stock markets in small countries.

65. This proposal was made recently by AlanGreenspan, Chairman of the U.S. Federal ReserveBoard, in a speech before the 34th Annual Conferenceon Bank Structure and Competition of the FederalReserve Bank of Chicago.

66. The implicit tax is paid by the borrowingcountry. This is unlike the imposition of taxes on capi-tal flows by the borrowing country, where the tax rev-enues accrue to the government of the borrower.

67. To this end, the IMF has established the Spe-cial Data Dissemination Standards (SDDS) and theGeneral Data Dissemination System (GDDS). See alsothe recommendations of the G-22 report of the work-ing group on transparency and accountability, Octo-ber 1998.

68. It has been argued that rating agencies sufferfrom a conflict of interest, having to respect local sen-sitivities to gain and maintain a foothold in emergingmarkets. It is very unlikely that a rating agency willsurvive if it brazenly misleads its customers.

69. A survey is in IMF (1998b).70. A rule of thumb is that a current account

deficit greater than 5 percent is often an indicator ofvulnerability—for sustainability, the growth of thatdebt which should not exceed the average rate of eco-nomic growth.

ReferencesAlexander, William E., Jeffrey M. Davis, Liam P.

Ebrill, and Carl-Johan Lindgren, eds. 1997. Sys-temic Bank Restructuring and MacroeconomicPolicy. Washington, DC: International MonetaryFund.

ADB (Asian Development Bank) and World Bank.1998. “Managing Global Financial Integration:Emerging Lessons and Prospective Challenges.”Manila (mimeo).

Alesina, Alberto, Vittorio Grilli, and Gian MariaMilesi-Ferretti. 1994. “The Political Economy ofCapital Controls.” In Leonardo Leiderman andAssaf Razin, eds., Capital mobility: the Impact onConsumption and Growth. Cambridge, UK/NewYork: Cambridge University Press.

Aggarwal, Vinod. 1998. “Exorcising Asian Debt:Lessons from Latin American Rollovers, Workouts,and Writedowns.” Paper prepared for a World Bankconference on The Aftermath of the Asian Crisis.Washington, DC: World Bank. (May 13–14).

Backus, David K., Patrick J. Kehoe, and Finn E. Kyd-land. 1992. “International Real Business Cycles.”Journal of Political Economy 100(4):745–75.(August).

BIS (Bank for International Settlements). The Matu-rity, Sectoral, and Nationality Distribution ofInternational Bank Lending. Basle: Bank for Inter-national Settlements, Monetary and EconomicDepartment (various issues).

Bordo, Michael D., Bruce Mizrach, and Anna J.Schwartz. 1995. “Real Versus Pseudo-internationalSystemic Risk: Some Lessons from History.” NBERWorking Paper 5371. Cambridge, MA: NationalBureau of Economic Research. (December).

Broner, Fernando, Norman Loayza, and J. HumbertoLopez. 1998. “Real Exchange Rate Misalignmentin Latin America.” Washington, DC: World Bank(mimeo).

Burki, Shahid, and Sebastian Edwards. 1995. “LatinAmerica After Mexico: Quickening the Pace.”Washington, DC: World Bank (mimeo).

Calvo, Guillermo A. 1996. “Capital Flows andMacroeconomic Management: Tequila Lessons.”International Journal of Finance and Economics.1:207–23.

Calvo, Guillermo A., Morris Goldstein, and EduardHochreiter, eds. 1996. Private Capital Flows toEmerging Markets After the Mexican Crisis. Wash-ington, DC: Institute for International Economicsand Austrian National Bank.

Calvo, Sara, and Carmen Reinhart. 1996. “CapitalInflows to Latin America: Is There Evidence ofContagion Effects?” In Guillermo A. Calvo, Mor-ris Goldstein, and Eduard Hochreiter, eds., PrivateCapital Flows to Emerging Markets After the Mex-ican Crisis. Washington, DC: Institute for Interna-tional Economics.

Caprio, Gerard, Jr. 1997. “Safe and Sound Banking inDeveloping Countries—We’re Not in Kansas Any-more.” Policy Research Working Paper 1739.Washington, DC: World Bank, Policy ResearchDepartment.

164


Caprio, Gerard, Jr., and Daniela Klingebiel. 1996a.“Bank Insolvencies: Cross Country Experience.”Policy Research Working Papers 1620. Washing-ton, DC: World Bank, Policy Research Depart-ment, Finance and Private Sector DevelopmentDivision.

____. 1996b. “Bank Insolvency: Bad Luck, Bad Policyor Bad Banking?” Annual World Bank Conferenceon Development 1996. Washington, DC: WorldBank.

Caprio, Gerard, Jr., Izak Atiyas, and James A. Hanson,eds. 1994. Financial Reform: Theory and Experi-ence. New York, NY: Cambridge University Press.

Carrasquilla, Alberto. 1998. “On Regulating CapitalAccount Transactions.” Inter-American Develop-ment Bank. Paper presented at a World Bank semi-nar. Washington, DC: World Bank, (June 23).

Cardenas, Mauricio, and Felipe Barrera. 1997. “Onthe Effectiveness of Capital Controls in Colombiaduring the 1990s.” Journal of Development Eco-nomics 54:27–57.

Cashin, Paul, Manmohan S. Kumar, and C. JohnMcDermott. 1995. “International Integration ofEquity Markets and Contagion Effects.” WorkingPaper WP/95/110. Washington, DC: InternationalMonetary Fund.

Chuhan, Punam, Gabriel Perz-Quiros, and Helen Pop-per. 1996. “International Capital Flows: Do Short-Term Investment and Direct Investment Differ?” Pol-icy Research Paper 1669. Washington, DC: WorldBank, Policy Research Department. (October).

Claessens, Stijn, Asli Demirgüç-Kunt, and HarryHuizinga. 1997. “How Does Foreign Entry Affectthe Domestic Banking Market?” Washington, DC:World Bank, (June).

Claessens, Stijn, Michael Dooley, and Andrew Warner.1995. “Portfolio Capital Flows: Hot or Cold?”World Bank Economic Review 9(1):153–74.

Cole, Harold L., and Maurice Obstfeld. 1991. “Com-modity Trade and International Risk-Sharing:How Much Do Financial Markets Matter?” Jour-nal of Monetary Economics 28:3–24. (August).

Corbo, Vittorio, and Leonardo Hernandez. 1996.“Macroeconomic Adjustment to Capital Inflows:Lessons From Recent Latin American and EastAsian Experience.” The World Bank ResearchObserver 11 (1):61–85.

Corsetti, Giancarlo, Paolo Pesenti, and NourielRoubini. 1998. “Paper Tigers? A PreliminaryAssessment of the Asian Crisis.” Paper presented atthe NBER International Seminar on Macroeco-nomics, Lisbon. (June 14–15).

Dasgupta, Dipak, and Bejoy Dasgupta. 1995. “Inter-est Rates in Open Economies: Real Interest Parity,Exchange Rates and Country Risk in Industrialand Developing Countries.” Policy ResearchWorking Paper 1283. Washington, DC: WorldBank.

Demirgüç-Kunt, Asli, and Enrica Detragiache. 1997.“The Determinants of Banking Crises: Evidencefrom Industrial and Developing Countries.” WorldBank Policy Research Working Paper 1828. Wash-ington, DC: World Bank Development ResearchGroup and International Monetary Fund ResearchDepartment, World Bank and International Mone-tary Fund. (September).

____. 1998. “Financial Liberalization and FinancialFragility.” Paper presented at the Annual BankConference on Development Economics. Washing-ton, DC: World Bank. (April 20–21).

Dollar, David. 1998. “Dividends and Disappoint-ments: Learning from the Success and Failure ofForeign Aid.” Policy Research Working Paper.Washington, DC: World Bank.

Dooley, Michael, Jeffrey Frankel, and Donald Math-ieson. 1987. “International Capital Mobility:What Do Saving-Investment Correlations Tell Us?”IMF Staff Papers 34:503–30. Washington, DC:International Monetary Fund. (September).

Edwards, Sebastian. 1989. Real Exchange Rates,Devaluation, and Adjustment. Cambridge, MA:MIT Press.

Edwards, Sebastian, and Mohsin S. Khan. 1985.“Interest Rate Determination in Developing Coun-tries: a Conceptual Framework.” IMF Staff Paper32:377–403. Washington, DC: International Mon-etary Fund.

Edwards, Sebastian, and Peter J. Montiel. 1989.“Devaluation Crises and the Macroeconomic Con-sequences of Postponed Adjustment in DevelopingCountries.” IMF Staff Paper 36:875–903. Wash-ington, DC: International Monetary Fund.

Eichengreen, Barry, and Richard Portes. 1995. “Cri-sis, What Crisis? Orderly Workouts for Sovereign

165


Debtors.” London: Centre for Economic PolicyResearch.

Eichengreen, Barry, and Andrew K. Rose. 1998. “Con-tagious Currency Crises: Channels of Con-veyance.” In Takatoshi Ito and Anne O. Krueger,eds., Changes in Exchange Rates in Rapidly Devel-oping Countries (forthcoming).

Eichengreen, Barry, and Charles Wyplosz. 1993. “TheUnstable EMS.” Brookings Papers on EconomicActivity 1(1993):51–124. Washington, DC: Brook-ings Institution.

Eichengreen, Barry, Andrew K. Rose, and CharlesWyplosz. 1995. “Exchange Market Mayhem: theAntecedents and Aftermath of Speculative Attacks.”Economic Policy: A European Forum 21:249–312.London: Centre for Economic Policy Research.(October).

____. 1996. “Contagious Currency Crises.” Scandi-navian Economic Review 98 (4):463–84. Alsoreprinted in Torben M. Andersen and Karl O.Moene, eds., Financial Liberalization and Macro-economic Stability. Malden, MA: Blackwell.

Fernandez-Arias, Eduardo, and Peter J. Montiel.1995. “The Surge in Capital Inflows to DevelopingCountries: Prospects and Policy Response.” PolicyResearch Working Paper 1473. Washington, DC:World Bank, International Economics Department,International Finance Division.

Flood, Robert, and Peter Garber. 1984. “CollapsingExchange-rate Regimes: Some Linear Examples.”Journal of International Economics 17:1–13.

Frankel, Jeffrey, and Andrew K. Rose. 1996. “Cur-rency Crashes in Emerging Markets: An EmpiricalTreatment.” Journal of International Economics41:351–66.

Frankel, Jeffrey, and Sergio Schmukler. 1996. “Crisis,Contagion, and Country Funds: Effects on EastAsia and Latin America.” Working Paper NumberPB96-04. San Francisco: Federal Reserve Bank ofSan Francisco.

Galbis, Vicente. 1993. “High Real Interest RatesUnder Financial Liberalization: Is There a Prob-lem?” IMF Working Paper 93/7. Washington, DC:International Monetary Fund. (January).

Gavin, Michael. 1998. “Fiscal Policy in EmergingMarket Economies.” Paper presented at the Capi-tal Flows Seminar at College Park, MD, May

1998. Economic Development Institute (WorldBank) and University of Maryland.

Gavin, Michael, and Ricardo Hausmann. 1996. “TheRoots of Banking Crises: The MacroeconomicContext.” Working Paper 318. Washington, DC:Inter-American Development Bank. Also appearedin Ricardo Haussman and Liliana Rojas-Suárez,eds., Banking Crises in Latin America. Washing-ton, DC: Inter-American Development Bank

Gavin, Michael, Ricardo Hausmann, and LeonardoLeiderman. 1995. “The Macroeconomics of Capi-tal Flows to Latin America: Experience and PolicyIssues.” Working Paper 310. Washington, DC:Inter-American Development Bank. (October).

Glick, Reuven, and Andrew Rose. 1998. “Contagionand Trade: Why are Currency Crises Regional?”Prepared for a Centre for Economic PolicyResearch/World Bank Conference Financial Crises:Contagion and Market Volatility, London. (May).

Goldstein, Morris. 1995. “Coping with Too Much ofA Good Thing: Policy Responses for Large CapitalInflows in Developing Countries.” Policy ResearchWorking Paper 1507. Washington, DC: WorldBank, Policy Research Department.

____. 1998. “Limiting the Moral Hazard in Interna-tional Financial Rescues.” Prepared for a Brook-ings symposium. Washington, DC: Brookings Insti-tution. (June).

Goldstein, Morris, and Carmen Reinhart. 1998. “Fore-casting Financial Crises: Early Warning Signals forEmerging Markets.” Washington, DC: Institute forInternational Economics (forthcoming).

Goldstein, Morris, and Philip Turner. 1996. “BankingCrises in Emerging Economies: Origins and PolicyOptions.” BIS Economic Papers No. 46. Basle:Bank for International Settlements. (October).

Greenwald, Bruce, and Joseph E. Stiglitz. 1993.“Financial Market Imperfections and BusinessCycles.” Quarterly Journal of Economics 108 (1):77–114.

Greenwald, Bruce, Joseph E. Stiglitz, and AndrewWeiss. 1984. “Informational Imperfections in theCapital Markets and Macroeconomic Fluctua-tions.” American Economic Review 74 (22):194–99.

Group of Seven. 1998. G-7 Leaders’ Statement on theWorld Economy and Declaration of G-7 Finance

166


Ministers and Central Bank Governors, October30.

Haque, Nadeem U., and Peter J. Montiel. 1991. “Cap-ital Mobility in Developing Countries: SomeEmpirical Tests.” World Development 19:1391–98. (October).

Hellman, Thomas, Kevin Murdock, and Joseph E.Stiglitz. 1996. “Deposit Mobilization throughFinancial Restraint.” In N. Hermes and R.Lensink, eds., Financial Development and Eco-nomic Growth: Theory and Experiences fromDeveloping Economies. London: Routledge.

____. 1997. “Liberalization, Moral Hazard in Bank-ing, and Prudential Regulation: Are CapitalRequirements Enough?” Research Paper No.1466.

Hernandez, Leonardo, and Heinz Rudolf. 1995. “Sus-tainability of Private Capital Flows to DevelopingCountries: Is a General Reversal Likely?” PolicyResearch Working Paper 1518. Washington, DC:World Bank, Policy Research Department.

Honohan, Patrick. 1997. “ Banking System Failures inDeveloping and Transition Countries: Diagnosisand Prediction.” BIS Working Papers No. 39.Basle: Bank for International Settlements, Mone-tary and Economic Department. (January).

International Monetary Fund. 1996. ExchangeArrangements and Exchange Restrictions, AnnualReport. Washington, DC.

____. 1997a. International Capital Markets. Washing-ton, DC. (November).

____. 1997b. Exchange Arrangements and ExchangeRestrictions, Annual Report. Washington, DC.

____. 1998a. Toward a Framework for Financial Sta-bility. Prepared by a Staff Team led by David Folk-erts-Landau and Carl-Johan Lindgren. World Eco-nomic and Financial Surveys series. Washington,DC: International Monetary Fund. (January).

____. 1998b. World Economic Outlook. Washington,DC: International Monetary Fund. (May).

____. 1998c. Exchange Arrangements and ExchangeRestrictions, Annual Report (forthcoming).

Jeanne, Olivier. 1997. “Are Currency Crises Self-Ful-filling: A Test.” Journal of International Econom-ics 43(3/4):163–86.

Kaminsky, Graciela L., and Leonardo Leiderman.1998. “High Real Interest Rates in the Aftermath

of Disinflation: Is it a Lack of Credibility?” Journalof Development Economics 55:191–214.

Kaminsky, Graciela L., and Carmen Reinhart. 1997.“The Twin Crises: The Causes of Banking and Bal-ance-of-Payments Problems.” Working Papers inInternational Economics 37. College Park, MD:University of Maryland, Center for InternationalEconomics, Department of Economics. (December).

Kaufman, George G., and Randall S. Kroszner. 1997.“How Should Financial Institutions and Marketsbe Structured? Analysis and Options for FinancialSystem Design.” Inter-American DevelopmentBank, Working Paper Series 338. Washington, DC:Inter-American Development Bank. (February).

Khatkhate, Deena. 1998. “Chasing Shadows: EastAsian Financial Crisis and the IMF.” Lecture toBank Negara Malaysia.

Khor, Hoe Ee, and Liliana Rojas-Suárez. 1991. “Inter-est Rates in Mexico: the Role of Exchange RateExpectations and International Creditworthiness.”IMF Working Paper WP/91/12. Washington, DC:International Monetary Fund, Research and West-ern Hemisphere Departments.

King, Robert G., and Ross Levine. 1993. “Finance andGrowth: Schumpeter Might be Right.” QuarterlyJournal of Economics 108(3):717–37.

King, Mervyn, and Sushil Whadwhani. 1990. “Trans-mission of Volatility between Stock Markets.”Review of Financial Studies 3 (1):5–33.

Krugman, Paul. 1996. “Are Currency Crises Self-Ful-filling?” NBER Macroeconomics Annual 1996:345–506. Cambridge, MA/London: MIT Press.

____. 1998. “What happened to Asia?” in Paul Krug-man’s web page (http://web.mit.edu/people/krugman/index.html).

Kumar, Manmohan S., Uma Moorthy, and William Per-raudin. 1998. “Determinants of Emerging MarketCurrency Crises and Contagion Effects.” Paper pre-sented at the Centre for Economic Policy Research.Financial Crises: Contagion and Market Volatility: aCEPR/World Bank Conference. London. (May).

Levine, Ross. 1997. “Financial Development and Eco-nomic Growth: Views and Agenda.” Journal ofEconomic Literature: 688–726. (June).

Levine, Ross, and Sara Zervos. 1998. “Stock Markets,Banks, and Economic Growth.” The AmericanEconomic Review. (June).

167


Lindgren, Carl-Johan, Gillian Garcia, and Matthew I.Saal. 1996. Bank Soundness and MacroeconomicPolicy. Washington, DC: International MonetaryFund.

Massad, Carlos. 1994. “Financial Reforms, OpeningUp and Stability: Policy Issues and SequencingConsiderations.” In Shakil Faruqi, Financial SectorReforms, Economic Growth, and Stability. Wash-ington, DC: World Bank.

Masson, Paul. 1998. “Contagion: Monsoonal Effects,Spillovers, and Jumps Between Multiple Equilib-ria.” Prepared for a CEPR/World Bank Conferenceon Financial Crises: Contagion and MarketVolatility. London. (May).

Mathieson, Donald J., and Liliana Rojas-Suárez.1993. “Liberalization of the Capital Account:Experiences and Issues.” IMF Occasional Paper103. Washington, DC: International MonetaryFund. (March).

McKinnon, Ronald I., and Huw Pill. 1997. “CredibleEconomic Liberalizations and Overborrowing.”American Economic Association Papers and Pro-ceedings of 109th Annual Meeting. (January).

Milesi-Ferretti, Gian Maria, and Assaf Razin. 1996.“Current-Account Sustainability.” Princeton Stud-ies in International Finance 81. Princeton, NJ:Princeton University, Department of Economics.(October).

Mishkin, Frederic S. 1991. “Asymmetric Informationand Financial Crises: A Historical Perspective.” InR.Glenn Hubbard, ed., Financial Markets andFinancial Crisis. Chicago: University of ChicagoPress.

____. 1996. “Bank Consolidation: A Central Banker’sPerspective.” NBER Working Paper 5849. Cam-bridge, MA: National Bureau of Economic Research.

____. 1997. “Understanding Financial Crises: ADeveloping Country Perspective.” Annual WorldBank Conference on Development Economics1996. Washington, DC: World Bank.

____. 1998. “Systemic Risk, Moral Hazard and theLender of Last Resort.” Prepared for a WorldBank/Reinventing Bretton Woods Committee Con-ference on the Aftermath of the Asian FinancialCrisis, Washington, DC. (March 13–14)

Montiel, Peter J. 1993. “Capital Mobility in Develop-ing Countries: Some Measurement Issues and

Empirical Estimates.” Policy Research WorkingPaper 1103. Washington, DC: World Bank, PolicyResearch Department.

Montiel, Peter J., and Carmen M. Reinhart. 1997.“The Dynamics of Capital Movements to Emerg-ing Economies during the 1990s.” Paper preparedfor the UNU/WIDER project on Short-term Capi-tal Movements and Balance of Payments Crises.

Mundell, Robert A. 1963. “Capital Mobility and Sta-bilization Policy Under Fixed and FlexibleExchange Rates.” Canadian Journal of Economicsand Political Science: 475–85. (November).

Nicholl, Peter. 1997. “Market-Based Banking Regula-tion.” In Suman K. Bery and Valeriano F. Garcia,eds., Preventing Banking Sector Distress and Crisesin Latin America. Proceedings of a Conferenceheld in Washington, DC. April 15–16, 1996.

Obstfeld, Maurice. 1986. “Rational and Self-fulfillingBalance of Payments Crises.” American EconomicReview 76: 676–81.

____. 1995. “Risk-taking, Global Diversification andGrowth.” American Economic Review 84(5):1310–29.

____. 1996. “Models of Currency Crises with Self-ful-filling Features.” European Economic Review:1037–47. (April).

____. 1998. “The Global Capital Market: Benefactoror Menace?” NBER Working Paper 6559. Cam-bridge, MA: National Bureau of EconomicResearch.

Obstfeld, Maurice, and Kenneth Rogoff. 1996. Foun-dations of International Macroeconomics. Cam-bridge, MA: MIT Press

Obstfeld, Maurice, and Alan Taylor. 1998. “The GreatDepression as a Watershed: International CapitalMobility over the Long Run.” In Bordo, MichaelD., Claudia D. Goldin, and Eugene N. White, eds.,The Defining Moment: The Great Depression andthe American Economy in the Twentieth Century.Chicago: University of Chicago Press.

OECD. Financial Markets Trends. Paris: Organisationfor Economic Co-operation and Development (var-ious issues).

Otker, Inci, and Ceyla Pazarbasioglu. 1996. “Specula-tive Attacks and Currency Crises: The MexicanExperience.” Open Economies Review 7 (Sup. 1):535–52.

168


Quirk, Peter J., Owen Evans, and others. 1995. Capi-tal Account Convertibility: Review of Experienceand Implications for IMF Policies. IMF OccasionalPaper 131. Washington, DC: International Mone-tary Fund. (October).

Radelet, Steven, and Jeffrey D. Sachs. 1998. “The EastAsian Financial Crisis: Diagnosis, Remedies andProspects.” Cambridge, MA: Harvard Institute forInternational Development. (April 20).

Razin, Assaf, and Chi-Wa Yuen. 1994. “Convergencein Growth Rates: a Quantitative Assessment of theRole of Capital Mobility and International Taxa-tion.” In Leonardo Leiderman and Assaf Razin,eds., Capital Mobility: the Impact on Consump-tion, Investment and Growth. Cambridge, UK:Cambridge University Press.

Reinhart, Carmen. 1998. “Capital Flows and Conta-gion: Causes , Consequences , and Pol icyResponses.” Paper presented at the Capital FlowsSeminar at College Park, MD, May 1998. Eco-nomic Development Institute (World Bank) andUniversity of Maryland.

Reisen, Helmut, and Hélène Yèches. 1993. “Time-Varying Estimates on the Openness of the CapitalAccount in Korea and Taiwan,” Journal of Devel-opment Economics 41:285–305.

Robinson, David. 1991. “Thailand: Adjusting to Suc-cess-Current Policy Issues.” IMF Occasional Paper85. Washington, DC: International MonetaryFund.

Rodrik, Dani. 1998. “Who Needs Capital-AccountConvertibility?” Symposium paper to appear inPrinceton Essays in International Finance.

Rojas-Suárez, Liliana. 1998. “Early Warning Indica-tors of Banking Crises: What Works for EmergingMarkets? With Applications to Latin America.”Washington, DC: Inter-American DevelopmentBank.

Rojas-Suárez, Liliana, and Steven R. Weisbrod. 1996.“Banking Crises in Latin America: Experiences andIssues.” In Ricardo Haussmann and Liliana Rojas-Suárez, eds., Banking Crises in Latin America.Washington, DC: Inter-American DevelopmentBank.

Sachs, Jeffrey D., Aaron Tornell, and Adrés Velasco.1996. “Financial Crises in Emerging Markets: TheLessons from 1995.” Brookings Papers on Eco-

nomic Activity 1(1996):147–215. Washington,DC: Brookings Institution.

Schadler, Susan, Maria Carkovic, Adam Bennett, andRobert Kahn. 1993. “Recent Experiences withSurges in Capital Inflows.” IMF Occasional Paper103. Washington, DC: International Monetary Fund.

Sheng, Andrew. 1996. Bank Restructuring: Lessonsfrom the 1980s. Washington, DC: World Bank.

Stiglitz, Joseph E. 1998a. “Boats, Planes, and CapitalFlows.” Financial Times. March 21.

____. 1998b. “Knowledge for Development: Eco-nomic Science, Economic Policy, and EconomicAdvice.” Annual Bank Conference on Develop-ment Economics. (April 20–21).

____. 1998c. “The Role of the Financial System inDevelopment.” Presentation at the Fourth AnnualConference on Development Economics in LatinAmerican and the Caribbean. El Salvador. (June).

Summers, Lawrence. 1998. “Deputy Secretary Sum-mers’ Remarks before the International MonetaryFund.” (March 19). Available at http://www.treas.gov/press/releases/pr2286.htm

Sundarajan, Vasudevan, and Tomas J.T. Balino. 1991.Banking Crises: Cases and Issues. Washington,DC: International Monetary Fund.

Tesar, Linda L. 1995. “Evaluating the Gains fromInternational Risksharing.” Carnegie-RochesterConference Series on Public Policy 42(1995):95–143. Amsterdam: North-Holland.

Valdès, Rodrigo O. 1996. “Emerging Market Conta-gion: Evidence and Theory” (mimeo). Massachu-setts Institute of Technology.

Wacziarg, Romain. 1998. “Measuring the DynamicGains from Trade.” Policy Research WorkingPaper 2001. Washington, DC. (November).

Williamson, John. 1996. The Crawling Band as anExchange Rate Regime: Lessons from Chile,Colombia, and Israel. Washington, DC: Institutefor International Economics. (October).

van Wincoop, Eric. 1994. “Welfare Gains from Inter-national Risk-Sharing.” Journal of Monetary Eco-nomics 34:175–200. (October).

Wolf, Holger C. 1997. “Regional Contagion Effects InEmerging Stock Markets.” Working Papers inInternational Economics G-97-03:1-18. Princeton,NJ: Princeton University, Department of Econom-ics, International Finance Section. (October).

169


Wong, David Y. 1988. “What Do Savings-Invest-ment Correlations Tell Us About CapitalMobility?” Working Paper 88-10. Philadelphia,PA: Federal Reserve Bank of Philadelphia.(May).

World Bank. 1997a. Global Economic Prospectsand the Developing Countries. Washington,DC.

____. 1997b. Private Capital Flows to DevelopingCountries: The Road to Financial Integration.A World Bank Policy Research Report. NewYork: Oxford University Press.

____. 1998a. Global Development Finance. Washing-ton, DC.

____. 1998b. World Development Report 1998/99:Knowledge for Development. New York: OxfordUniversity Press.

____. 1998c. World Development Indicators. Wash-ington, DC.

Wyplosz, Charles. 1998. “International Capital Mar-ket Failures: Sources, Costs, and Solutions.” Back-ground Paper for Global Economic Prospects andthe Developing Countries 1998. Washington, DC:World Bank.

170

preventing financial crises in developing countries...the savings and loan crisis in the united...

Documents