Financial Liberalization: What Went Right,What Went Wrong?

THE FINANCIAL LIBERALIZATION

that took place in the developingcountries in the 1980s and 1990s

was part of the general move toward giving marketsa greater role in development. It was also a reactionto several factors specific to finance: the costs, cor-ruption, and inefficiencies associated with usingfinance as an instrument of populist, state-led devel-opment; a desire for more financial resources; citi-zens’ demands for better finance and lower implicittaxes and subsidies; and the pressures exerted onrepressed financial systems by greater internationaltrade, travel, migration, and better communications.

The financial reforms went beyond the interestrate liberalization that had been recommended bythe so-called Washington Consensus. To varyingdegrees, governments also allowed the use of foreigncurrency instruments and opened up capitalaccounts. Domestic markets developed in centralbank and government debt, and international mar-kets expanded in government and private bonds.Capital markets developed, but less rapidly, and weremost successful in the larger,already rapidly growing,East and South Asian countries. State banks contin-ued to have a major role for much of the 1990s; theirprivatization was gradual and often proved costly.Central banks moved away from trying to financedevelopment; they became more independent andsuccessfully focused on keeping inflation low, buttheir debt increasingly absorbed bank deposits.

Certainly the reforms produced some gains. Butthe growth benefits of the financial and nonfinan-

cial reforms in the 1990s were less than expected.Financial crises raised questions of whether finan-cial liberalization was the wrong model, what hadgone wrong, and the appropriate direction of futurefinancial sector policy. Overall, the 1990s is proba-bly best considered a precursor of better things thatwill take some time to achieve.

Section 1 of this chapter describes why and howfinancial liberalization occurred. Section 2 discussesthe outcomes of financial liberalization during the1990s, including the crises that occurred and theirrelation to macroeconomic policies, financial liber-alization, and the overhangs of old economic andpolitical systems. Section 3 summarizes the lessonsfrom the experience of the 1990s, and section 4draws suggestions for future policy. Section 5 con-cludes the chapter.

1. From Financial Repression toFinancial Liberalization

The financial repression that prevailed in develop-ing and transition countries in the 1970s and 1980sreflected a mix of state-led development, national-ism, populism, politics, and corruption.The finan-cial system was treated as an instrument of thetreasury: governments allocated credit at below-market interest rates, used monetary policy instru-ments and state-guaranteed external borrowing toensure supplies of credit for themselves and publicsector firms, and directed part of the resources that

Financial Liberalization:What Went Right,What Went Wrong?

207

Chapter 7

low returns, crowded out more efficient potentialusers, and encouraged wasteful use of capital.

Financial repression also worsened income distri-bution.Subsidies on directed credits were often large,particularly in periods of high inflation, and actualallocations often went to large borrowers.4 The lowinterest rates led to corruption and to the diversion ofcredits to powerful parties.Diversions tended to growover time, particularly when inflation reduced realinterest rates on credits, and rising fiscal deficits anddirected credits absorbed more of the limited deposits.

High CostsThe repressed systems were costly.Banks,particularlystate banks and development banks, periodically

E C O N O M I C G ROW T H I N T H E 1 9 9 0 s208

were left to sectors they favored. State banks wereconsidered necessary to carry out the directedcredit allocations,1 as well as to reduce dependenceon foreigners. Bank supervisors focused on com-plying with the often intricate requirements ofdirected credit rather than with prudential regula-tions. Interest rates to depositors were kept low tokeep the costs of loans low. In some cases, lowdeposit and loan rates were also populist measuresintended to improve income distribution.2

Repressed finance was thus an implicit tax andsubsidy system through which governments trans-ferred resources from depositors receiving lowinterest rates (and from those borrowers not receiv-ing directed credits) to borrowers paying low ratesin the public sector and to favored parts of the pri-vate sector. Governments had to allocate creditbecause they set interest rates that generated excessdemand for credits. Capital controls were needednot (as often argued) to protect national saving, butto limit capital outflows fleeing low interest ratesand macroeconomic instability, and to increase thereturns from the inflation tax.3 In effect, capitalcontrols were a tax on those unwilling or unable toavoid them and they encouraged corruption (Han-son 1994).

Factors behind Financial LiberalizationThree general factors provided an impetus for themove to financial liberalization: poor results, highcosts, and pressures from globalization.This sectiondiscusses each in turn.

Poor ResultsTogether, the limited mobilization and inefficientallocation of financial resources slowed economicgrowth (McKinnon 1973; Shaw 1973). Low inter-est rates discouraged the mobilization of finance,and bank deposit growth slowed in the 1980s in themajor countries (figure 7.1). Capital flight occurreddespite capital controls (Dooley et al. 1986).Alloca-tion of scarce domestic credits and external loans togovernment deficits, public sector “white ele-phants,” and unproductive private activities yielded

30%

25%

20%

15%

10%

5%

0%

–5%1960s 1970s 1980s 1990s

Perc

enta

ge o

f GD

P In

crea

seSouth Asia (3 countries)

Latin America (7)

Africa (10)

East Asia (4)

FIGURE 7.1

Increase in Average Deposits/GDP in MajorCountries, by Regions, 1960s–-90s(difference between 3-year average of bank deposits/GDP atthe end of each decade)

Source: IMF, International Financial Statistics.

Note: Countries and regions covered are: East Asia: Indonesia, Republic of Korea, Malaysia, the Philippines,Thailand. Latin America: Argentina, Brazil, Chile, Colombia, Mexico, Peru,República Bolivariana de Venezuela. Africa: Ghana, Kenya, Nigeria, Tanzania, Uganda, Burkina Faso,Cameroon, Côte d’Ivoire, Mali, Senegal. South Asia: Bangladesh, India, Pakistan.

required recapitalization and the takeover of theirexternal debts by governments. Political pressuresand corruption were widespread. Loan repaymentswere weak because loans financed inefficient activi-ties, because loan collection efforts were insufficient,and because borrowers tended to treat loans fromthe state banks simply as transfers.Typically, banksand other intermediaries rolled over their nonper-forming loans until a period of inflation wiped outdepositors’ claims and permitted a general default.Since intermediaries were not forced to follow rea-sonable prudential norms or mark their portfolios tomarket, the losses were nontransparent, even to thegovernments that often owned them. Inflation alsohelped to conceal the problems of commercial banksthrough their earnings on low interest deposits.Thehidden costs of the repressed systems became moreapparent once financial liberalization began.

Pressures from GlobalizationPerhaps most important, financial repression cameunder increasing pressure from the growth of trade,travel, and migration as well as the improvement ofcommunications.5 The increased access to interna-tional financial markets broke down the controls oncapital outflows on which the supply of low-costdeposits had depended.6 Capital controls may beeffective temporarily, but over time mechanisms(such as overinvoicing imports and underinvoicingexports) develop to subvert them (Arioshi et al.2000; Dooley 1996). These mechanisms becamemore accessible as goods and people became moreinternationally mobile.

The Evolution of Financial Liberalization The shift in policies differed in timing, content, andspeed from country to country and included manyreversals. Broadly:

• African countries turned to financial liberaliza-tion in the 1990s, often in the context of stabi-lization and reform programs supported by theInternational Monetary Fund and World Bank,as the costs of financial repression became clear.

F I N A N C I A L L I B E R A L I Z AT I O N : W H AT W E N T R I G H T, W H AT W E N T W RO N G ? 209

• In East Asia, the major countries liberalized in the1980s, though at different times and to differentdegrees. For example, Indonesia, which had lib-eralized capital flows in 1970, liberalized interestrates in 1984, but the Republic of Korea did notliberalize interest rates formally until 1992. Lowinflation generally kept East Asian interest ratesreasonable in real terms, however. In most coun-tries, connected lending within industrial-finan-cial conglomerates and government pressures oncredit allocation remained important.

• In South Asia, financial repression began in the1970s with the nationalization of banks in India(1969) and Pakistan (1974). Interest rates anddirected credit controls were subsequentlyimposed and tightened,but for much of the 1970sand 1980s real interest rates remained reasonable.Liberalization started in the early 1990s with agradual freeing of interest rates; a reduction inreserve,liquidity,and directed credit requirements;and liberalization of equity markets.

• In Latin America, episodes of financial liberaliza-tion occurred in the 1970s but financial repres-sion returned, continued,or even increased in the1980s, with debt crises, high inflation, govern-ment deficits, and the growth of populism (Dorn-busch and Edwards 1991).In the 1990s,substantialfinancial liberalization occurred, although thedegree and timing varied across countries.

• In the transition economies, financial liberaliza-tion took place fairly rapidly in the 1990s in thecontext of the reaction against communism(Bokros, Fleming, and Votava 2001; Sherif, Bor-ish, and Gross 2003).

The earliest policy changes generally focused oninterest rates. In many instances governments raisedinterest rates with a “stroke of the pen” to mobilizemore of the resources needed to finance budgetdeficits and to enable the private sector to play agreater role in development. (Some interest rateincreases, designed to curb capital flight, wereintended more for stabilization than for liberaliza-tion.) New financial instruments were introduced

new domestic and foreign banks and, in somecases, nonbank intermediaries.

In general, however, the financial reforms of the1990s focused on freeing interest rates and creditallocations, and made much less effort to improvethe institutional basis of finance—a much harder,longer task.

2. Outcomes in the Financial Sector during the 1990s

Private sector credit is a key factor in growth.10

Banks can intermediate funds and take risks only ifprivate credit is not crowded out by governmentdebt. Over the 1990s, deposits grew faster than inthe previous decade, but in many countries bankcredit to the private sector from domestic sourcesgrew only slowly.The increase in loanable fundswas largely absorbed by the public sector.

Deposit GrowthBank deposits grew as a share of the gross domesticproduct (GDP) in the 1990s, unlike in the 1980s(figure 7.1 above and Hanson 2003b).Thus, mostmajor countries and most regions achieved a majorobjective of financial liberalization. And in Indiaand some East Asian and Latin American countries,nonbank deposits supplemented the rapid growthin bank deposits.

Box 7.1 discusses the resumption of depositgrowth in India as it gradually liberalized, as well asthe growth of India’s capital market.

Many factors contributed to the deposit growth,including the slowdown in inflation in the 1990s,11

the positive real deposit rates, and new depositinstruments.Another factor was the legalization offoreign currency deposits. Deposits in foreign cur-rency grew as a share of total deposits in manycountries in the 1990s, and in some cases they sup-plied more than half the total by the end of thedecade (Honohan and Shi 2003).12 Not surpris-ingly, the foreign currency deposits were popular


that had freer rates and were subject to lowerdirected credit requirements. Some countries alsobegan admitting foreign currency deposits, to attractoffshore funds and foreign currency holdings intothe financial system as well as to allow residents legalaccess to foreign currency assets (Hanson 2002;Honohan and Shi 2003; Savastano 1992, 1996).

Partial interest rate liberalizations soon generatedpressures for more general freeing of interest rates(albeit in some cases after reversals of liberalization).As borrowers directed funds into deregulated instru-ments and sectors, demand for low-cost loansincreased and repayments on them deteriorated.7

Unfortunately, when the macroeconomic situationwas unstable and interest rates were freed, very highreal interest rates developed, creating corporate andbanking problems that added to the overhang ofweak credits that were exposed by liberalization.

At very different speeds in different countries,interest rate liberalization came to be supplementedby other changes:8

• Central banks were made more independent.They abandoned their earlier developmental roleto focus on limiting inflation, often in the con-text of stabilization programs.

• Reserve requirements and directed credit wereeased.

• Capital accounts were liberalized, even in coun-tries where domestic foreign currency instru-ments remained banned. Foreigners wereallowed to participate in capital markets9 andprivate corporations were allowed to raise fundsoffshore.

• Markets were set up for central bank debt andgovernment debt. Equity markets were set up inthe transition countries and liberalized wherethey already existed.

• In some countries, pension systems addeddefined contribution/defined benefits elements,often operated by private intermediaries.

• Gradually, state banks were privatized. Bankingcompetition increased, as a result of the entry of


BOX 7.1

India—A Successful Liberalizer with Strong Capital Markets

India liberalized its financial sector gradually overthe 1990s, with particular success in capital mar-kets, while avoiding any major crisis. In the

1980s, India had a repressed financial system (Han-son 2001, 2003a). This, plus increasing macroeco-nomic instability, slowed deposit growth. Financialliberalization was part of greater reliance on the pri-vate sector after the 1991 foreign exchange crisis.Interest rates were raised and gradually freed, bankregulations and supervision were strengthened, andnonbank financial corporations (NBFCs) were allowedunder easier regulations (Hanson 2003a). After a 1991capital market crisis, regulations were strengthened,listings were liberalized, foreign investors wereallowed in, and infrastructure was substantiallyimproved (Shah and Thomas 1999; Nayak 1999).

Bank deposits of nationals and nonresident Indiansresumed their growth and NBFC deposits grew sharplyafter 1992. The stock, bond, and commercial paper mar-kets became among the most vibrant in developing

countries, providing nearly one-fourth of India’s corpo-rate funding from 1992 to 1996 (Reserve Bank of India1998) with listings more than doubling from 2,000 in1991 to over 5,000 (Standard and Poor’s 2003). Thepost-1997 economic slowdown led to a stock marketfall, problems in the NBFCs (which were wound down),and crises in the government development banks andmutual fund, though public sector commercial banksperformed surprisingly well. Recently, large capitalinflows and higher growth have led to low interest ratesand better bank performance.

Although India’s approach to financial liberalizationserved it well, three major issues remain: (1) crowdingout, with government debt now absorbing more than37 percent of bank deposits compared to about 24 per-cent at the end of the 1980s; (2) a weak informationand legal framework, which, despite efforts at improve-ment, still contributes to nonperforming loans and lim-its access to credit; and (3) the still-dominant role ofpublic sector banks.

India: Money (M3 in Sept.)/GDP,Deposit Rate and Inflation

Inte

rest

rat

e an

d in

flat

ion

(%)

M3

(% o

f GD

P)

M3/GDP M3+NBFCs % of GDP

Avg. M3/GDP, 1987–92

75

60

45

30

15

0

–15

75

1971

1973

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

60

45

30

15

0

–15

Inflation (CPI)

Freerate

Interest rate1 year deposit

Source: Reserve Bank of India data.CPI. Consumer price index.GDP. Gross domestic product.

Access to credit expanded less than manyobservers hoped after the financial reforms, thoughit improved toward the end of the1990s.Panel stud-ies had suggested that financial liberalization wouldmake more credit available to a wider group of bor-rowers. (See, for example, Schiantarelli et al. 1994,and works cited there.) After liberalization therewas some growth, but in practice government andcentral bank debt crowded out many borrowers. Insome countries where nonbank intermediaries(henceforth, nonbanks) grew, they did increaselending to nontraditional borrowers. But bothbanks and nonbanks were hindered by the lack ofinformation on borrowers and weaknesses in thelegal and judicial systems in the areas of collateraland creditors’ rights.

Instead of increasing private credit, the rise inbank deposits over the 1990s tended to be absorbedby government and central bank debt, and by banksstrengthening their offshore positions. In particular,in the 25 developing and transition countries withthe largest banking systems, the average ratio of netgovernment debt to bank deposits rose by more than60 percent, from about 13 percent in 1993 to about21 percent in 2000 (Hanson 2003b).15 Similar pat-terns prevailed in the larger African countries.16

The main reason for the rise in government debtwas postcrisis bank restructurings, involvingreplacement of weak private credits, particularly inBrazil, Indonesia, Jamaica, Mexico, and someAfrican countries.But growing government deficitsalso played a key role in some cases, notably Indiaand Turkey. In general, the increases in banks’ hold-ings of government debt reflected rises in the stockof government debt, rather than any increasedattractiveness of government debt to banks, ordecreased willingness of banks to take risks.17

Banks also increased their net holdings of cen-tral bank debt—substantially in some countries—despite falling reserve requirements. On average inthe 25 developing countries with the largest finan-cial systems,banks’net holdings of central bank debtrose by nearly 5 percentage points of GDP over the1990s (Hanson 2003b). As a monetary policyinstrument, central bank debt had advantages over


with members of the public, many of whom hadlost their savings and pensions in inflation andrepressed financial systems.13 But foreign currencyloans were also popular with borrowers.14

The reforms reduced the burden on banks,widening their discretion over the allocation ofresources and lowering required reserves. Now thatgovernments could raise resources from newlydeveloping debt markets, they had less need torequire banks to invest in government debt or tohold low-return reserves with the central bank thatwere invested in government debt. In many of the25 largest developing countries, the average ratio ofreserves to deposits fell over the 1990s (Hanson2003b).Directed credit requirements were reduced,interest rates were raised on remaining directedcredits, and nominal market rates fell.

Credit:Absorption of Deposits by the Public SectorBank credit to the private sector grew much lessthan bank deposits and other bank liabilities in the1990s (figure 7.2).

20

15

10

5

1990–2000Latin America7 countries

Increased privatesector credit

Increased net publicsector and centralbank credit

Increased net foreign assets

4 countries 3 countries 10 countriesEast Asia Transition Africa

1993–2000

Liabilities + capital

1993–2000

0

–5

Chan

ges

as a

perc

enta

ge o

f GD

P

1990–2000

FIGURE 7.2

Changes in the Ratios of Bank Assets and Liabilities PlusCapital to GDP, 1990s(averages by country group)

Source: IMF, International Financial Statistics.

Note: Countries covered are the same as in Figure 1. South Asia is excluded because coun-tries in that region do not report private credit or capital separately.

the previous instruments of credit controls on indi-vidual banks, changes in reserve requirements, andvariations in central bank lending, which hadtended to limit competition, to affect banks bluntly,and to affect weak banks heavily. But the use ofcentral bank debt had costs, in that it crowded outwould-be borrowers.Central banks often sold theirdebt to sterilize capital inflows as well as to tightenmoney when capital flowed out.18 Central banksmay also have sold debt to mop up some of the liq-uidity that arose from lowered reserve require-ments, or when they needed to fund their ownquasi-fiscal deficits that arose from negative spreadsbetween their assets and liabilities.

Another reason for the slow growth of privatecredit was that banks themselves increased their netholdings of foreign assets for hedging purposes (seefigure 7.2 above).Those in the largest 25 develop-ing-country markets went from essentially a bal-anced foreign position in 1990, on average, to netborrowing of nearly 1 percent of GDP in 1993 andnearly 3 percent of GDP in 1997, before revertingto being net holders of foreign assets in 2000 (Han-son 2003b).After 1997, external lenders cut creditlines, banks wound down their external borrow-ings, and banks increased their external assets.19

Given the limited growth of private sectorcredit, a variable that has been shown to be linkedwith economic growth,20 it is not surprising thatthe rise in GDP growth associated with the finan-cial (and general) liberalization of the 1990s was lessthan hoped for.

However, the story is more complex than theslow growth of private credit. In the major develop-ing countries, especially in East Asia, the averagegrowth of private sector credit (especially includingexternal credits) and of GDP was much faster before1996 than after.About 1995, some countries beganto experience financial crises. Much of the privatecredit extended by banks and nonbanks proved tobe unproductive, in the sense that it became non-performing before or during the crises. Duringbank restructurings, these credits were replaced withgovernment debt (to ensure depositors were paid);when eventually executed, the associated collateral


was usually worth less than 30 percent of the facevalue of the loans, suggesting how unproductive thegrowth in private sector credit had been.In the tran-sition countries and African countries the quality ofcredit issues was typically more related to the publicsector’s use of the credits.21

Bond and Equity MarketsGovernment and central bank bond market devel-opment was fairly successful in the 1990s. By 2000,more than 40 developing countries, including allbut one of the 25 with the largest financial systems,had government bond markets (Del Valle andUgolini 2003), and more than 20 had central bankdebt markets. The government bond marketsallowed governments to reduce their reliance onforeign borrowing. The supply of this debt wasinelastic, but it was attractive to banks for severalreasons: its interest rates had been freed, it carried alow capital requirement, it was less risky than pri-vate debt, and it had liquidity once the marketsbecame active.

The growth of domestic equity and bond mar-kets contributed to private sector financing in EastAsia and India during the 1990s. In the major EastAsian equity markets, market capitalizationexceeded $20 billion in 2000, having risen 80 per-cent or more (except in Thailand) since 1990.Turnover averaged more than 50 percent and list-ings in the individual countries rose at least 40 per-cent between 1990 and 2000, to the point wherethey all exceeded listings in every Latin Americancountry except Brazil (Standard and Poor’s 2003).The Indian market was even more successful in pro-viding resources to a wide group of firms after list-ing regulations were eased (see box 7.1 above).Chile’s market also did reasonably well, though itsturnover was low because of the pension funds’buy-and-hold policies. However, even in thesecountries, banks remain the main source of finance.

Elsewhere, equity markets were less successful.On the seven largest Latin American stockexchanges, listings have declined since 1997, and onfive of those seven they have declined ever since

economic performance than from any obvious insti-tutional advantage.This suggests that in the shortrun, equity market growth mainly reflects generaleconomic performance, which attracts foreigninvestors willing to risk sums that are small to thembut large relative to the market. Simply setting up amarket may not add much to growth or allow firmsto raise funding. Over time, however, as institutionsimprove, equity markets do seem to contribute toeconomic growth (Levine 2003; Levine and Zervos1998).Another important element in equity marketperformance seems to be foreign investor participa-tion (Bekaert, Harvey, and Lundblad 2003).

Private bond markets grew even less than equitymarkets in the 1990s.They share the problems ofequity markets as well as having some of their own.Concerns about potential future macroeconomicinstability have led bond buyers to demand highreturns for committing funds for the long term, anddeterred issues of long-term bonds. Often onlypublic sector firms issued long-term bonds, andthen only in a few countries, notably in South Asia.Potential buyers of private bonds were also deterredby lack of protection in law and in fact for bond-holders’ rights, and by the lack of good bankruptcylegislation and enforcement. Nonetheless, someprivate bond markets have developed, for examplein India, and in Mexico, recently, for securitizedhousing finance.

External Finance for the Private SectorWithin the private sector in developing countries,external borrowing grew faster than domestic bor-rowing in the first part of the 1990s, as large privatecompanies increasingly drew on external credits. Forexample, in 17 of the countries with the largestfinancial markets, the ratio of private sector foreignborrowing (of more than one year’s maturity) to bor-rowings from domestic banks increased fairly steadily,from 16.5 percent in 1990 to 27 percent in 1997.23

Short-term borrowing also grew substantially.How-ever, after 1997 these credits slowed in dollar terms.In the same 17 countries, external credit to the pri-vate sector changed little in dollar terms. However,


1990; turnover in all seven is less than 50 percent ofmarket capitalization (Standard and Poor’s 2003). Intransition countries, equity markets were created aspart of the privatization process,22 often with onlybelated recognition of the importance of regulatoryframeworks. Listings declined in most of these mar-kets as some privatized companies were taken offthe market. Market capitalization is less than $20billion, except in Poland, and the Russian Federa-tion, and turnover exceeds 50 percent only in Hun-gary (Standard and Poor’s 2003).

Several factors lie behind the slow developmentof equity markets in developing and transitioncountries.The first is that potential investors aredeterred by the low turnover in these markets (usu-ally much less than 75 percent compared with 85percent in even the smaller deciles of traded com-panies on the U.S. NASDAQ) and by low liquidity,which reflects the small sizes of the listed compa-nies as well as the low turnover (Shah and Thomas2003). Second, listings on stock exchanges havebeen reduced by takeovers of firms by multinationalcorporations, and trading has been reduced by themigration of major firms’ listings to industrial-country markets. In 2000, companies listed offshoreaccounted for about 55 percent of the market cap-italization in 15 middle-income countries, and for27 percent of the market capitalization in 25 low-income countries; much of the trading in thesestocks also takes place offshore (Claessens, Klinge-biel, and Shmuckler 2003). Family firms that couldlist often do not, partly because the benefits are notgreat, partly because these firms often have privi-leged access to credit through related banks, andpartly because they fear that dilution of ownershipwill reduce their control.Third and more funda-mentally,weak institutional factors—poor informa-tion, poor treatment of minority shareholders, andweak regulation of market participants—weakenthe interest of investors, both domestic and foreign,in many equity markets (Glaesner, Johnson, andSchleifer 2001; LaPorta, López de Silenes, andSchleifer 2002b; Black 2001).

The better performance of the East Asian andIndian markets seems to result more from superior

the ratio of these credits to domestic credit rose by50 percent by 2000, reflecting crisis-related devalua-tions and the removal of private sector credits frombanks in restructurings in these countries.

The external credits to the private sector werenarrowly distributed.They went only to interna-tionally creditworthy borrowers, and four countriesaccounted for the bulk of private sector externalborrowing (in dollars) in 2000: Brazil (27 percent),Mexico (12 percent), Indonesia (9 percent), andThailand (7 percent).

Offshore equity sales were another source of cap-ital for many large private companies in the 1990s.The numerous developing-country companies thatwere listed offshore in 2000 largely reflected issues ofglobal depository rights and American depositoryrights during the 1990s. Of course, this source ofcapital was also narrowly distributed.

While financial liberalization benefited large,well-run companies and, indirectly, other borrow-ers in developing countries, it raised banks’ risks.The best firms obtained loans and equity financeoffshore at less cost than in the domestic market,albeit with currency risk.24 This left a larger por-tion of the limited domestic private credit availableto other borrowers, but it also increased the averagerisk in the banks’ loan portfolios. Moreover, banksin developing and transition countries increasedtheir net intermediation of external loans up to1997, especially in East Asia, and they also guaran-teed some direct external borrowings by the cor-porate sector, typically off their balance sheets.

The external borrowings were a major factor inthe East Asian external payments crises and werealso important in other crises of the 1990s. Asexternal borrowings grew, lenders shortened matu-rities, creating maturity mismatches for borrowers.Further, loans made by financial intermediariesbased on their own external borrowing, thoughtypically matched in terms of currency, entailedsubstantial risks when the borrowers lacked anassured source of foreign exchange. Eventually,lenders refused to roll over their credits becausethey considered the risks too high.This generatedboth a banking crisis and a foreign exchange crisis.


The resulting sharp devaluations increased debt-servicing problems on many foreign currency loansand led to calls on the guarantees, worsening thedifficulties of firms and banks.

Financial IntermediariesMost of the impact of financial reforms on the insti-tutional structure of the financial sector was not feltuntil the latter half of the 1990s.This limited thegains from liberalization during the decade andcontributed to crises.

State banks, with their well-known problems(LaPorta, López de Silanes, and Schleifer 2002a),decreased in importance only after 1995, andindeed still dominated many financial systems in2000 (figure 7.3).The continued large state bankpresence meant that credit allocations changed onlyslowly, despite liberalization.The problems of statebanks after liberalization were most obvious in thetransition countries,25 where the banks often sim-ply continued to lend to traditional clients or werecaptured by politically powerful groups; as a result,their loans were unproductive and their alreadylarge portfolios of nonperforming loans increased.State banks in other countries had similar problems.The continued dominance of these banks, the asso-ciated weakness in credit allocation,and the impliedstate guarantees that allowed them to raise increas-ing deposits despite their high incidence of nonper-forming loans, all limited the gains fromliberalization and accounted for a substantial part ofthe cost of crises in the 1990s.

Private banks changed gradually with liberaliza-tion, entry of foreign banks, and fiercer competi-tion, but their deficiencies also contributed tounproductive lending and crises.Their credit man-agement skills did not keep pace with changes inthe environment such as the growth of foreign cur-rency operations and the greater competition thattheir traditional borrowers were facing in the realsector. Moreover, many private banks in East Asiaand some Latin American countries were parts ofindustrial-financial conglomerates and continuedto provide funding to their increasingly unprofitable

aries also often suffered from problems of risky andconnected lending and were often the first to failwhen credit tightened.

Greater competition can also create problems forbanks by cutting their profits (Caprio and Summers1996;Dooley 2003).Although this problem seems tobe mainly one of adjusting to competition(Demirgüç-Kunt and Detragiache 1998), it doesforce owners to decide whether they should con-tinue costly competition,try to exit,or loot the bank.Thus regulation and supervision, particularly withregard to bank intervention and exit, are importantissues when liberalization increases competition.

Regulation, Supervision, and Deposit Insurance

Banking Regulation and SupervisionImprovements in the prudential regulation andsupervision of banks lagged behind the liberaliza-tions of the 1990s and contributed to crises.The


industrial partners. State banks that were privatizedto local buyers in weak institutional environmentsoften suffered similar problems and had to be rena-tionalized.

Foreign banks enlarged their role as new policieseased restrictions on their entry in the latter half ofthe 1990s,particularly in transition countries but alsoin Latin America and Africa.26 Their entry increasedcompetition in banking and cut costs for bankclients.They competed fiercely for the best clientsand drove down profits on business with them, andthey also competed in lending to small firms.27

A second approach to increasing competition,taken by a few countries, was to simply allow morebanks, by lowering entry requirements.28 Unfortu-nately, many of these new banks were “pocketbanks,” capturing deposits to lend to their owners’businesses and often suffering problems. A thirdapproach was to allow the growth of nonbankfinancial corporations (box 7.3).These intermedi-

FIGURE 7.3

State Ownership in Banking, 1998–2000

Source: Map Design Unit, World Bank.

[Pls. pro-vide vec-torimage ofthismap.]

oversight of banks in developing countries startedfrom a low base in the 1990s because, during theperiod of financial repression, bank supervisors hadfocused on compliance with directed credit rules.

International standards for supervision—the 25Basel Core Principles—were not agreed upon untilSeptember 1997. Countries did enact their own


prudential regulations and upgraded supervision,but implementation—a political as well as a techni-cal issue—often lagged, even after costly crises.Enforcement was patchy, even of weak regulationson income recognition, provisioning, capital, andconnected lending, and weak banks continued inoperation. International standards on minimumbank capital were not set until 1988, in the Baselagreements between industrial countries for inter-nationally active banks.

The issues in improving regulation and supervi-sion were not just technical but also political.Thecrises of the 1990s did engender attempts toimprove regulation and supervision. But even then,regulations were often not strengthened immedi-ately and forbearance was used to limit the capitalinjections that governments otherwise would havehad to make. For example, in East Asia, regulationson capital, income recognition, and provisioninglagged behind international standards after the 1997crisis (Barth, Caprio, and Levine 2001), and actualcapital in many Indonesian and Thai banks was stillwell below the Basel standard in 2000.

By letting weak banks overexpand, the poor over-sight contributed to the crises of the 1990s. In devel-oping countries, weak banks that were allowed tocontinue operations often opted for a high-risk/high-return lending strategy or, in the worst case, werelooted, as has also occurred in industrial countries.Market discipline, which might have restrained theexpansion of weak banks, was limited by poor infor-mation and implicit or explicit deposit insurance.

Deposit InsuranceDeposit insurance and, in crises, blanket guarantees,were standard recommendations of many financialadvisors, and formal deposit insurance was initiatedor improved in nearly 40 countries in the 1990s,mostly in transition and West African countries(Demirgüç-Kunt and Kane 2002; Demirgüç-Kuntand Sobaci 2001). In addition, countries such asEcuador, Indonesia, Korea, Malaysia, Mexico,Thai-land, and Turkey introduced blanket guarantees ofbank liabilities.Deposit insurance and blanket guar-antees are mainly attempts to reduce the risk of

BOX 7.3

Nonbank Financial Intermediaries(NBFIs) in the 1990s

NBFIs such as finance companies,co-op banks, and nonbank finan-cial corporations exist in many

countries and in the 1990s some of themwere an important factor in private sectorcredit and deposit mobilization. For example,India eased restrictions on nonbank financialcorporations in 1992 and by 1996 theirdeposits were equal to more than 5 percentof broad money (box 7.1 above). In Thailandand Malaysia, finance companies’ deposit andcredit growth picked up in the early 1990s.In Latin America, co-op banks and housingbanks in Colombia have been important forsome time. NBFIs usually offered higherdeposit rates and credit in different formsand to different clients than banks—forexample loans for construction, consumercredit, and small borrowers. NBFIs also wereoften subject to easier regulations on inter-est rates, reserves, and capital than werebanks, as well as less supervision. However,NBFIs had a history of periodic crises in LatinAmerica and East Asia, as for example inThailand in the 1980s (Sundararajan andBalino 1991, 47–48). After 1997, many NBFIsin India, Malaysia, and Thailand went bank-rupt, depositors shifted to banks, and, tosome degree, banks increased their loans tothe former NBFI borrowers.

smaller than the differentials in bank risk (Laeven2002a, 2000b, 2000c). This probably reflects thepolitical power of the local bank owners who ben-efit most from deposit insurance.

To sum up, the schemes that were introducedfor the support of depositors tended to create largecontingent liabilities and to increase moral hazardwhile reducing market discipline.They contributedto crises and volatility by encouraging the fundingof weak institutions after liberalization. Depositorsand external lenders, expecting to be bailed out ofany problems by a government guarantee, tended tosupply too much funding,particularly to state banksand well-connected financial-industrial conglom-erates. Market discipline, which might have limitedthis funding, was negligible, not just because ofweak information but also because of the implicitand explicit guarantees.

Equity and Bond Market RegulationImprovements in equity and bond market regula-tion began in the 1990s and also proved difficult toimplement. Even improving trading rules was diffi-cult because of the difference between the interestsof buyers and sellers,on the one hand,and the short-run interests of market operators, on the other.Alsodifficult to resolve has been the conflict between theinterests of majority and minority shareholders.Attempts to create markets overnight have had onlylimited success not only in cases of limited regula-tion (Czech Republic), but also where investor pro-tection rules appeared to be reasonably good(Russia). Regulation in Poland seems to have beenrelatively successful, however (Black, Kraakmen, andTarassova 2000; Glaesner, Johnson, and Schleifer2001).As with bank regulation and supervision, theissues are not merely technical but also political.

PensionsAs described in chapter 6, a major change in pen-sion systems occurred in the 1990s, with manycountries shifting from pay-as-you-go systems tosystems in which at least part of pension income isbased on full funding for individual accounts. Chile


bank runs. Deposit insurance also has the second-ary, consumer protection benefit of protectingunsophisticated depositors. Governments likeddeposit insurance as it appeared to give benefits yethad no costs, at least until a crisis arrived. Local pri-vate banks, often politically important, liked itbecause it improved their competitiveness withstate and foreign banks.

The actual impact of deposit insurance and guar-antees has been mixed.The statistical evidence sug-gests that the gains from deposit insurance dependon its particular features, its credibility, and the insti-tutional environment (Demirgüç-Kunt and Kane2002; Demirgüç-Kunt and Detragiache 2002). Inmany cases, the insurance created large contingentguarantees, increased moral hazard, and reducedmarket discipline (Demirgüç-Kunt and Kane 2002;Demirgüç-Kunt and Sobaci 2001; Demirgüç-Kuntand Huizinga 1999).Large lender-of-last-resort sup-port and blanket guarantees in effect providedunlimited insurance not only for depositors but forowners, many of whom looted their banks.Theywere particularly ineffective in the context of opencapital markets and political and economic turmoil:for example in Ecuador (IMF 2004b) and inIndonesia, liquidity support to banks was almost aslarge after the introduction of blanket guarantees asit was before. One reason may be that as the likeli-hood increases that the deposit insurance or a blan-ket guarantee will be used, its cost and credibilitycomes into question,and runs on banks and the cur-rency may increase (Dooley 2000).

Various attempts have been made to adjustdeposit insurance so as to reduce moral hazard andincrease market discipline ex ante, but usuallydepositor losses have been socialized ex post. Forexample, insurance limits have been placed on largedeposits and on deposits carrying the high rates thatare often offered by weak banks, but often the lim-its have not prevented the insurance from extend-ing to all depositors in a crisis. Another approachhas been to use risk-based deposit insurance premi-ums, in an attempt to offset the moral hazard andmarket discipline problem, but in practice the dif-ferentials in premiums have been substantially

was the first developing country to adopt thisapproach, in 1981.Among the countries with largefinancial systems, Argentina, Colombia, Mexico,and Peru in Latin America, plus Hungary, Poland,and Thailand, all adopted variants of this systemafter 1994 (Fox and Palmer 2001).The new systemsgave individuals much better access to their pen-sions and held the promise of generating demandfor long-term financial instruments and therebystimulating capital markets.

The results were not as good as anticipated.First, all systems had to cope with the change-overproblem of paying existing pensioners while invest-ing the contributions to the new system into assets.Without large fiscal surpluses, the change-over gen-erated a large increase in government debt that thenew pension system had to hold, as occurred inArgentina. As a result, the demand for long-termprivate instruments did not rise much.Thus the ini-tial impact of pension reforms was simply to makethe government’s liability transparent. A secondissue is that because capital markets typically aresmall, pension funds either generated price rises, ashappened even in Chile, and/or had to invest inbank debt, as happened in Peru.Third, costs havebeen high in many of the private pension funds,reflecting set-up costs, insurance linked to the pen-sions, and a response to advertising that encouragedexcessive shifts between funds. Some of these prob-lems could have been reduced and country riskdecreased for the individual accounts by allowingthe funds to invest externally, but countries haveusually tried to retain pension contributions andavoid possible balance of payments pressures.

Financial Sector CrisesFinancial sector and external payments crises werefeatures of the 1990s.29 Costly crises occurred inMexico, the East Asian “Miracle” countries, Russia,Brazil, and some Eastern European and Africancountries.The new millennium began with crisesstarting in Argentina and Turkey and high nonper-forming loans in China. Africa also suffered costlyfinancial crises (figure 7.4).


What role did financial liberalization play in thefinancial and currency crises of the 1990s, dubbedthe “twin crises” by Kaminsky and Reinhart(1999)? The discussion below first assesses the rolesplayed by macroeconomic problems, financial liber-alization, and weak lending by state banks andfinancial-industrial conglomerates and then out-lines the difficult tradeoffs that policy makers facedin responding to crises over a short time horizon.

Macroeconomic ProblemsMost crisis countries had high debt and larger thanusual current account deficits and were pursuingexchange rate–based stabilization policies.30 Manyalso had open capital accounts, but a causal associa-tion with crises is not clear: not all countries withopen capital accounts experienced crises and, in the1980s, crises had developed even in countrieswhose capital accounts were nominally closed.

The combination of high debt and exchangerate–based stabilization seems to be associated withunsustainable booms in capital inflows, imports, andGDP and shifts in relative prices, followed by rever-sals in these variables as financing slows and matu-rities shorten, while interest rates rise.The slowingof inflows reflects both the inherent characteristics

Estimated starting year and restructuring cost

Cost

of

rest

ruct

urin

gas

a p

erce

ntag

e of

GDP

60

1975 1980

AR

UR

PH SG

Cdl

IM

BRVEN MX

CZ MY

BUTK

TH

KO

EC

INDO

CHK

1985 1990 1995 2000

50

40

30

20

10

0

FIGURE 7.4

Selected Financial Crises, 1980–99

Source: Caprio et al. 2003.

Note: The figure illustrates the estimated financial costs of restructuring after selectedcrises, but not the losses in GDP.

growth declined, and real interest rates rose.35Theseproblems were often connected to unsustainablefiscal policy and the defense of unsustainable cur-rency pegs with long periods of high interest rates.Problems in the timing and sequencing of liberal-ization, sometimes related to political issues, alsocontributed to the crises (box 7.4).

In assessing the role of financial liberalization inthe 1990s crises, an important question is whyinternational lenders and domestic depositors sup-plied so much funding.Part of the large increases inloanable funds may have reflected a natural over-shooting tendency in financial markets (Kindle-berger 2000; Minsky 1992). But any such tendencywas certainly exaggerated by the explicit andimplicit guarantees that governments provided tolenders. Government debt was directly guaranteed(although after crises it was sometimes restruc-tured). Growing private external debt, funneledthrough banks or guaranteed by them, and growingdeposits carried at least an implicit guarantee,whichex post often became explicit. Moreover, when lib-eralization led banks to lose franchise value and cap-ital, weak regulation and supervision did notprevent bank owners from engaging in high-risk/high-return lending or even looting. Nor didit limit banks’ overexposure to related borrowers.Thus market discipline was eroded by governmentguarantees, implicit or explicit, while weak regula-tion and supervision did not limit moral hazard.

Guarantees and their credibility may also explainwhy the crises in the 1990s seem to have happenedrelatively quickly (Dooley 2000).According to thisexplanation, avoiding a crisis depends on maintain-ing foreign investors’ and depositors’ perceptionsthat the guarantees (and the exchange rate peg) arecredible. Events, including fears of political change,can quickly change these perceptions, leading toshifts into foreign exchange, curtailment of short-term credits, and rollovers of maturing loans, trig-gering banking and exchange rate crises.36

Weak LendingA third factor in the 1990s crises was the weak lend-ing,old and new,by the old financial intermediaries,


of portfolio adjustment and the growth of investorconcerns regarding debt-servicing capacity andexchange rate pegs.31The rises in interest rates mayreflect a combination of smaller inflows, growingconcerns about the sustainability of the exchangerate peg, and attempts to defend the peg with tightmoney, often for long periods. Eventually, theexchange rate depreciates and debts need to berestructured. Not surprisingly, the financial sectorsuffers a crisis in the downward phase of suchcycles, reflecting liquidity squeezes on banks thathave borrowed externally; problems with borrow-ers, especially those indebted in foreign currency;and runs on the banks to speculate on the currency.

Various events may trigger a crisis. Externalshocks include deteriorating terms of trade,increases in international interest rates, and increasesin risk premiums in industrial-country markets thatautomatically affect developing-country debt.32

Contagion in financial markets has also beencited.33 Domestically, unstable or inconsistentmacroeconomic policies sooner or later lead topressures against banks and the currency. Politicaldevelopments, such as the ouster of presidents Mar-cos in the Philippines in 1986 and Soeharto inIndonesia in 1998, lead connected parties to liqui-date their assets, putting pressure on banks andlenders with whom they did business.

Financial Sector LiberalizationFinancial sector liberalization seems to have been afactor in crises (Demirgüç-Kunt and Detragiache1998, 2001; Kaminsky and Reinhart 1999). Itincreased capital inflows and deposits, whichallowed rapid growth in credit to weak public andprivate enterprises and the government, as well asto real estate. Over time, the quality of the lendingdeteriorated.This may be one explanation for thelags between liberalization and financial crises(Demirgüç-Kunt and Detragiache 2001), andbetween financial crises and currency crises(Kaminsky and Reinhart 1999).34 Eventually, cor-porate bankruptcies,banking problems, and runs onbanks and currencies developed, particularly whenthe rapid credit growth and inflows slowed, real

notably state banks and industrial-financial con-glomerates. Before liberalization these intermedi-aries had large overhangs of bad debt, which hadbeen rolled over several times to favored borrowers.Financial liberalization made these debts worsebecause of higher real interest rates and lower infla-tion. Moreover, lower protection and increasedcompetition reduced traditional borrowers’ abilityto service their debts. However, the increaseddeposits and capital inflows associated with financialreform provided new funds that enabled the banks


to roll over their loans again, adding to the ultimatevolume of nonperforming loans.For example, in theearly stages of liberalization in the transition coun-tries, “most state banks continued to lend asinstructed or for patronage purposes” (Sherif, Bor-ish, and Gross 2003, 21). In East Asia, banksexpanded their lending to related industrial con-glomerates, which were increasingly overleveraged(Claessens, Djankov, and Lang 1998). In addition,crises tended to generate a shift of deposits to statebanks, because of expectations of government guar-

BOX 7.4

Problems with the Process of Financial Liberalization

In the late 1980s Nigeria liberalized interest ratesand bank entry but retained a multiple exchangerate regime that was accessible only through banks.

This raised the demand for bank licenses, many of whichwent to well-connected parties who were interested inarbitraging foreign exchange between the multiplerates, not in banking. Though the number of bankstripled, the ratio of deposits and credit to GDP fell, and,by the 1990s, banks were experiencing significant dis-tress (Lewis and Stein 2002).

In Korea, the de facto rapid liberalization of short-term borrowings in the early 1990s, both internationallyand internally, led the heavily leveraged corporations tobe increasingly financed by short-term inflows andthrough less regulated intermediaries. In the run-up tojoining OECD, Korea had opened its capital account byfreeing short-term foreign borrowings, but left longer-term borrowings subject to restrictions in an attempt tolimit total capital inflows (Cho 2001). This policyencouraged a maturity mismatch in lending and a cur-rency mismatch on the part of borrowers, especiallysince rates were much lower on foreign currency loansthan domestic ones. Although deposit rates were for-mally liberalized in 1993, their rise was limited by moralsuasion, government guidance, and high reserve require-ments until 1996. New intermediaries (finance compa-nies converted to merchant banks) sprang up to meet

demands for funds by intermediating external inflows.Bank trust accounts were liberalized and grew relativeto bank deposits; they also were allowed to take short-term commercial paper, which had relatively free inter-est rates. Finally, the freeing of interest rates onconsumer loans contributed to a shift of loanable fundsto these activities and may have dampened Korea’s sav-ing rate, augmenting the country’s increased reliance on(short-term) external borrowing.

Thailand set up its “offshore”/onshore BangkokInternational Banking Facility in 1993 with tax andregulatory advantages that were justified as an attemptto create a regional financial center operated bynational banks. The facility allowed locals to deposit inforeign currency and local borrowers to escape (albeitwith short-term loans) from the government’s tightmoney policy and foreign currency–denominated loans.Its operations became a major factor in the expansionof Thailand’s external debt (Bordo et al. 2001; Alba,Hernandez, and Klingebiel 1999). Pressure on the gov-ernment from these borrowers was probably a factor inthe government’s lengthy, costly attempt to defend thebaht even as it supported the borrowers, taking a mon-etary stance inconsistent with the fixed exchange rate.After the devaluation, these obligations were a majorfactor in the banks’ problems and in the closure of manyfinance companies.

sis; the support needed for their privatizationamounted to about half their assets (Clarke and Cull1999). In Eastern Europe, the cost of the public sec-tor banks’ bad debt overhang was enormous—forexample about 16 percent of GDP in Bulgaria andabout 18 percent of GDP in the Czech Republic(Sherif, Borish, and Gross 2003). In China, officialestimates of the nonperforming loans of the fourlargest state banks exceeded 20 percent of loans in2003; various private estimates were much higher.

Privatization is often considered as a remedy forthe weak lending of public sector banks, but it hasbeen costly in cases where the state has retained acontrolling interest or where sales have been madeto weak owners whose operations were poorly reg-ulated and supervised. Mexico’s 1991 privatizationis perhaps the best known example. Soon after pri-vatization, partly because of the currency crisis, the


antees.37This allowed further increases in lending tofavored clients who often used the loans to buy for-eign exchange and then defaulted on the loans.

The overhang and growth of state banks’ non-performing loans, and their cleanup,were substantialelements in the crises of the 1990s.A notable exam-ple is Indonesia (box 7.5). In Thailand,more than 80percent of Bank Krung Thai’s loans became nonper-forming. Brazil’s BANESPA (the state bank of SaoPaulo) was estimated to have more than 90 percentnonperforming loans; the estimated cost of the fed-eral government’s 1997 cleanup, prior to privatizingthe bank, was about $20 billion or nearly 3 percentof GDP.The Finance Ministry has estimated thatrestructuring Banco do Brasil and Caixa EconomicaFederal may cost $50 billion. In Argentina, the bank-rupt state of the smaller provincial banks wasexposed by the spillover of Mexico’s “Tequila” Cri-

BOX 7.5

Indonesia: Early Liberalization and Weaknesses Related to Political Connections

In Indonesia the freeing of interest rates and easingof capital and reserve requirements contributed tolarge deposit growth, as well as a doubling of the

number of commercial banks (Hanson 2001). By 1996,competition and the expansion of 10 private banks hadreduced state banks to about 45 percent of the system.However, all banks were very weak (World Bank 1996,1997a). Despite the rules, state banks were overexposedto well-connected borrowers, and private banks to theirowners. State banks reported low capital and theirreported nonperforming loans, though high, were under-stated, given the rollover of bad loans and other maneu-vers. At least two state banks were insolvent. Loanswere often inflated by “commissions” to loan officers.Private banks reported better figures but weak supervi-sion provided no check on them. Exposure limits werenot enforced and many small banks were bankrupt.

The spillover from the July 1997 Thai devaluationexposed these weaknesses and the dependence offinance on the political regime. Capital outflow devel-

oped and rollovers of the large amount of short-termexternal loans stopped as investor concerns mounted(despite the imposition of limits on currency specula-tion). Monetary policy was loosened to ease borrowers’problems. The November IMF program brought littlerelief—runs on private banks and the currency speededup with the closure of 16 banks (small depositors didnot begin to be paid until January 1998) and theDecember illness of Soeharto. State banks, which ben-efited from shifts in deposits, made loans to well-con-nected borrowers on the basis of projected exports thatdid not materialize. In January 1998, outflowsincreased with the poor reception of the 1998–99budget, panic buying of goods, riots that frightenedIndonesians of Chinese origin, and the possibility ofintroducing a currency board. The exchange rate fell toless than one-seventh of its precrisis level. Massivecentral bank liquidity support, often well in excess ofbanks’ capital and in some cases up to 75 percent oftheir assets, would have doubled the money base had

(Box continues on the following page.)

Mexican government was forced to renationalizethe banks; it then cleaned up their balance sheets atan estimated cost of more than US$70 billion andreprivatized them to international banks,beginningin 1998 (box 7.6).

In Eastern Europe, the initial bank privatizationswent poorly,particularly where governments retaineda controlling interest (Clarke,Cull and Shirley 2003).In Africa, too, the experience with bank privatizationwas often bad, with long delays and sales eventuallymade to undercapitalized owners who did notimprove credit management and abandoned thebanks when they lost their capital (box 7.7).

Difficulties in Policy Responses to CrisesThe crises presented difficult new policy problems.Traditional macroeconomic policies of tighter fiscaland monetary policy and devaluation were appro-


priate for reducing excess demand and currentaccount deficits to financeable levels. But the finan-cial sector problems, and their implications for thebalance of payments, raised a new set of more com-plicated issues and tradeoffs, for which no singlebest practice exists.

To deal with an individual bank’s problems, thestandard recommended response is to provide liq-uidity support at high interest rates and then inter-vene with protection for small depositors.But banksbecome insolvent well before they become illiquid,and owners of insolvent banks may then choose arisky lending strategy or even attempt to loot thebank (de Juan 2002).38 Moreover, problems in onebank typically indicate more widespread problems,and closing a bank without promptly compensatingdepositors may trigger runs on other banks andlooting by bank owners.

BOX 7.5

(continued)

reserves not fallen (Kenward 2002; World Bank 2000c).Imposition of a blanket guarantee at the end of Janu-ary temporarily slowed outflows. Soeharto’s reelectionin March was followed by severe riots, often directedagainst Indonesians of Chinese origin and Soehartocronies. Capital outflows rose once again, as did liquid-ity support. In May 1998 Soeharto resigned but pres-sures on banks continued.

In sum, liberalization encouraged deposit growthand foreign inflows, but credit access depended not onprofitability but on political connections, includingaccess to external loans from international banks (cor-porations did much of Indonesia’s external debt bor-rowing; the state banks’ external borrowing waslimited by policy). Lenders and depositors looked atconnections, not at risk and corporate leverage. Theeasing of bank licensing and the lack of enforcementof exposure limits worsened this problem. Then, whenpolitical concerns developed, the well-connected triedto withdraw their assets and an outflow developed,

exacerbated by the concerns of the middle-classIndonesian Chinese. As a result, the blanket guaranteestopped the bank runs only temporarily—total liquid-ity support was nearly as large after the blanket guar-antee was imposed as before, according to the figuresin Enoch et al. (2001). Of the US$20 billion liquiditysupport that went mostly to private banks, 96 percentwas unrecoverable and a substantial amount wasdiverted into foreign exchange speculation, accordingto an ex post study by the National Auditor. The costof the crisis is estimated at more than 50 percent ofGDP. Bank Mandiri, a merger of four state banks ofwhich at least two were bankrupt before the crisisbegan, accounted for about 30 percent of the cost ofthe crisis (more than 17 percent of GDP). More than 70percent of the losses in the state banks were in loansthat had to be taken off the books. The poor quality ofthese loans is shown by the eventual recovery rate ofless than 30 percent, most of which was realized fourto five years after the crisis.

loose money will help borrowers in local currency,but put additional pressure on the exchange ratethat will hurt borrowers in foreign currency. Use ofreserves delays this problem, but reserves are finiteand their decline can provoke a speculative attackon the currency.

Nontraditional policies have had only mixedsuccess. Capital controls have not been effective instopping currency runs.40 A blanket guarantee mayor may not halt bank or currency runs, dependingon how it affects concerns about the credibility ofthe guarantee and the burden of future costs (Doo-ley 2000).A few countries, including Argentina andEcuador, have tried to stop bank runs by freezingdeposits and devaluing, but the disruption to thepayments chain has led to massive recessions. Ifdeposits are to be written down, in parallel withloans, it is probably best to make a politically


As a bank problem becomes systemic, bank runsturn into currency runs and pose severe problemsfor which there is no standard answer.The govern-ment faces the unpleasant choice of either inter-vening in weak banks, thus possibly provoking runson other banks,39 or providing liquidity support—loose money—that will spill over into pressure onthe exchange rate and international reserves, espe-cially in open economies (World Bank 2000c).

Another choice is that of how much to supportthe exchange rate with reserves and tight money(offsetting the loose money from liquidity support)versus how much to allow a depreciation. Tightmoney helps to protect the exchange rate, as in thetraditional policy response, and thus helps borrow-ers in foreign currency, but it hurts borrowers inlocal currency and it hurts banks, particularly if it ismaintained for a long time. Liquidity support and

BOX 7.6

Bank Privatization in Mexico

Mexico nationalized its commercial banks in1982. It decided to privatize them in 1991,as part of liberalization and to raise fiscal

resources. At the time, the privatization was acclaimedas a resounding success, fetching higher prices thanpredicted. Although open only to domestic purchasers,the sale was considered technically well designed andexecuted. Bidders were first qualified, and the auctionswere transparent and quick, without scandal.

The macroeconomic (Tequila) crisis of 1995 took theshine off this success: loan defaults increased sharplywith the collapse of the peso and the rise in interestrates. Failing banks were found to have made poor loansunder the relaxed regulatory framework, often to polit-ically powerful groups connected to their controllingowners (Haber and Kantor 2003; LaPorta and López deSilanes 2003). The connected lending meant that thebanks had effectively financed much of their own pur-chase. Taking into account loans from developmentbanks, the buyers actually had minimal equity, but this

had not prevented their purchase of the banks. The gov-ernment renationalized the failed banks and protectedthe depositors but taxpayers were left with a huge bill,estimated at 18 percent of GDP.

The features of the sale that were praised earlier arenow often criticized: the privatization for being too hastyand the purchase prices as too high. Yet a sale was con-sidered necessary to raise fiscal resources and sustain thegovernment’s commitment to reform. The main mistakeswere the exclusion of foreigners and the acceptance ofpurchases by politically powerful but heavily leveragedbuyers. The exclusion of foreign bidders was partly a cal-culated risk to shore up domestic support in a nationalis-tic country. Even with foreign participation, highlyleveraged locals might have bid more for the banksthrough their access to loans. Thus, above all, the Mexicanexperience illustrates not just a flawed privatization butthe complicated issues that bank privatizers must juggle,including the difficult problems associated with dealingwith local elites in a sector as sensitive as banking.

unpalatable exchange of tradable bonds, asArgentina did in its January 1990 Bonex plan.Whatever is done,GDP growth is almost certain toslow if not decline (Frankel and Wei 2004).

In sum, the crises of the 1990s appear to berelated to macroeconomic problems, but also to


financial liberalization in the context of the over-hang of old political and economic relationships,manifested in state banks and politically powerfulfinancial-industrial conglomerates. Governmentguarantees encouraged a rise of funding for theseintermediaries, which they channeled into weak

BOX 7.7

Bank Restructuring and Privatization in Sub-Saharan Africa

At the end of the 1980s many African banks wereinsolvent and illiquid. Governments undertookmajor restructuring programs over the 1990s to

deal with these problems. A gradual return to macro-sta-bility and balanced government budgets—a prerequisitefor bank restructuring—occurred in programs supportedby the International Monetary Fund (IMF) and the WorldBank. Directed credits were abandoned and interest ratesliberalized. Government arrears to the banking sectorwere often securitized on various terms, with debt ser-vice often guaranteed. Money markets were established.Many countries issued new banking laws, overhauled reg-ulations, and set up supervisory authorities.

Bank restructurings were both organizational andfinancial and sometimes led to privatization, but theprocess also often required multiple restructurings andwas hesitant (World Bank 2001, box 3.3). Some banks,particularly public sector banks, were closed or weakbranches were turned into agencies. In Benin, theextreme case, all public sector banks were closed in1990, leaving the country without banks for somemonths until new private banks entered the market. Inother cases, bad assets were provisioned and losses wereabsorbed by existing shareholders (governments and theprivate sector); in a few cases, new capital was injectedby the private sector; and in others, bad loans wereremoved from banks. Asset recovery corporations wereset up to manage bank liquidations and/or to recoverloans and reimburse depositors/creditors (for example inCameroon, Côte d’Ivoire, Ghana, Uganda, and Senegal)with mixed results. Management was changed, staff

retrenched, internal controls were put in place, and newloan procedures were gradually developed.

Treatment of depositors in failed banks varied fromcountry to country. In some, the government left thedeposits in the restructured bank or reimbursed alldepositors. In others, repayment of depositorsdepended on the asset recovery of the failed institu-tions. Priority was given to compensation for smalldepositors. Depositors incurred substantial losses inCameroon and the Republic of Congo, for example.Some countries introduced deposit insurance in the late1990s, but it is unclear whether these systems couldhandle banking crises as large as those of the 1990s.

Privatizations generally went to a major institu-tional partner, often foreign. In some cases the foreignbanks were large and well known, with a reputation toprotect. African private banks that operate in severalcountries have developed (Ecobank, Bank of Africa,Financial Bank, CBC, Stanbic) as a result of privatiza-tion involving foreign partners. However, in somecases, the foreign banks provided little improvement.

As a result of the restructurings, African commercialbanks have become more solvent, liquid, and profitableand a safer haven for deposits, but many problemsremain. In many countries, banks are still weak in theirlending and operations. Bank deposits have declined insome countries, probably reflecting a mix of bank clo-sures, discouragement of small deposits, and civilstrife. Commercial bank lending has generally been lim-ited, reflecting crowding out of government debt andbankers’ selectivity in lending.

Source: World Bank 2001, box 3.3.

• Privatizations and restrictions on foreign entryin the financial sector often allowed local elitesto retain or increase their economic and politicalpower.

• Implicit and explicit guarantees of deposits andinternational loans supported local elites’ abilityto raise resources.

• Liberalization of bank licensing led to “pocket”banks that mainly engaged in connected lendingor regulatory arbitrage, not expansion of access(the record of nonbank intermediaries is some-what better but they too were often linked toindustrial-financial conglomerates).

• Development of the framework for capital mar-kets—such as reasonable information, legal andjudicial treatment of bankruptcy, treatment ofminority shareholders, conduct rules for marketparticipants—was slow, compounding the prob-lems that capital markets in developing coun-tries face in terms of concerns aboutmacroeconomic stability,high costs, and low liq-uidity.

The process of liberalization and the limitednature of the results in the 1990s suggest four majorlessons, discussed next.

Finance Depends on InstitutionsPerhaps the most important lesson of the 1990s forfinance is that the financial sector’s contribution todevelopment depends not just on resource mobi-lization but also on attention to institutions: inter-mediaries, markets, and the informational,regulatory, legal, and judicial framework. Resourcesneed to be allocated to those that offer the best com-bination of return and risk, and this depends on thequality of institutions. Building up these institutionsis not easy, takes time, and requires political support.

In the 1990s, the traditionally weak loans of statebanks and financial institutions linked to industrialconglomerates were further weakened by thehigher interest rates that followed liberalization, aswell as by increased import competition and realappreciations that cut the profitability of traditional


loans. Eventually crises developed and the deposi-tors and external creditors were bailed out by gov-ernments. Privatization in these environments didnot solve the problems, and often required costlyrenationalizations.

3. Lessons of the 1990s

In the 1980s and 1990s the approach to financeshifted from the repression of prices and marketsand the use of government credit allocations to amore market-based, internationally open system.Yet this shift, along with the other reforms, had lessthan the expected effects on growth. Access tofinancial services does not seem to have improvedsubstantially in the 1990s, though there are indica-tions of improvements recently. Expectations maywell have been too high. Another reason was anapparent “boom in bust[s]” (Caprio 1997), relatedto macroeconomic policies but also to financial lib-eralization in the context of an overhang of weakinstitutions—financial intermediaries, financialmarkets, and informational, legal, and judicialframeworks. Problems in these areas reduced theimpact of liberalization and in some cases led toperverse results.

The weaknesses of institutions were not just atechnical issue: they reflected the difficulty of chang-ing the previous state-led development system and,more fundamentally, its underlying political-eco-nomic basis within a short period, while restraintson markets could be and were quickly lifted.Theoverhang of these factors during the 1990s was animportant reason behind the following:

• Credit allocation was weak and continued to goto the public sector, well-connected individuals,financial-industrial conglomerates, and tradi-tional state bank clients.

• Bank privatization was slow and partial privatiza-tions left control of intermediaries in the hands ofgovernment in many transition and Africaneconomies, leading to continued preferential treat-ment of the traditional borrowers from state banks.

borrowers. Explicit and implicit guarantees allowedthese financial institutions to obtain much of theliberalization-induced increase in deposits and cap-ital inflows, and to substantially expand lending totheir traditional borrowers,private and public.Reg-ulation and supervision did not prevent this; theirweaknesses reflected not just technical but politicalissues. Market discipline was eroded by poor infor-mation and, more important, by implicit andexplicit guarantees. Better capital market develop-ment could have relieved some risks and absorbedpart of the shocks. However, the capital marketsfaced competition from implicitly or explicitlyguaranteed deposits and external loans. Marketdevelopment also was hindered by the inherentproblems of capital markets in developing countriesand the difficulties of building up a reasonable insti-tutional framework quickly.

By the end of the 1990s, it became clear thatmuch of the increased deposits and capital inflowshad gone into (1) unproductive private borrowingor state enterprise debt that had to be replaced bygovernment debt in order to bail out depositorsand lenders, (2) deficit finance, and (3) central bankdebt to stabilize the economy.Thus it is not sur-prising that the financial liberalizations of the 1990sdid not live up to the high expectations regardingsustained increases in growth or credit access.

Focusing on the poor quality of credits exposesa common thread in the slow growth and financialcrises of the 1990s: the continuation of preferentialaccess, related to the overhang of old institutions,that was changed only slowly by the financialreforms. In many countries in the 1980s and 1990s,public sector borrowing,with its implicit guaranteefrom future tax revenues,was excessive and eventu-ally led to crises and slow growth. But even incountries with smaller public sectors and relativelylimited fiscal problems, such as Chile in the late1970s and East Asian countries in the 1990s, loansto industrial conglomerates—made from the guar-anteed deposits in the private financial intermedi-aries that they controlled or from state banks andinternational lenders, to which they had preferen-tial access because of the institutional set-up—


eventually became nonperforming and contributedto crises.As noted earlier, the poor contribution ofsuch loans to sustained growth is shown by the lowvalue of the associated collateral when it was even-tually sold.

Delaying Needed Policies Is CostlyLimiting the incidence and cost of financial crisesdepends on resisting political pressures to prolongunsustainable booms and to delay action on weakbanks,41 as well as on avoiding socializing theirlosses. In the 1990s,governments often tried to pro-long booms and did not limit the expansion ofweak banks.42 Unfortunately, such policiesincreased the ultimate volume of bad loans and thesize of the crises.Then, after crises occurred, gov-ernments typically responded by bailing out depos-itors and external investors through liquiditysupport, expansion of whatever deposit insuranceexisted, and blanket guarantees, all of which gener-ated large increases in government debt and contin-gent liabilities.

Expectations that losses would be socialized,through explicit and implicit guarantees, also con-tributed to crises and volatility by encouragingweak institutions to mobilize funds after liberaliza-tion. Depositors and external lenders, expecting tobe bailed out of problems by a government guaran-tee, supplied funding to state banks and financialintermediaries that were part of financial-industrialconglomerates.The funding was well in excess ofwhat could be used productively. Market discipline,which might have limited this funding, was weak-ened by the implicit and explicit guarantees.43 Theprocess was unstable,however.When a rise occurredin the subjective probability that the guaranteeswould be called, net capital outflows developed, asdepositors and investors became concerned abouthow the guarantees would be paid and funded.44

The capital flight was facilitated by the liquiditysupport to weak banks and the support of theexchange rate by reserve sales.The combination ofhigh initial returns, limited losses on the funds thatwere taken out of the countries just before thecrises, and the ultimate provision of government

external borrowing and offshore equity issues didprovide lower-cost funding but only to larger cor-porations in a few countries, and the loans were sub-ject to currency and rollover risk.The slowdown innet private-to-private disbursements and short-termloans was a major factor in the crises.Though it willbe difficult, better development of domestic capitalmarkets, even in the countries that have done rela-tively well,would reduce the impact of future crises.

Successful Finance Depends on MacroeconomicStabilityAnother old lesson is that successful financial liber-alization and successful finance depend on macro-economic stability (World Bank, World DevelopmentReport 1989). If anything,open capital accounts andvolatile international capital flows place a larger pre-mium on sound macroeconomic management.However, financial reforms, or at least more mar-ket-based interest rates, were often put in place inthe 1990s in the midst of macroeconomic imbal-ances, complicating what was already a technicallydifficult problem.46 For example, countries withunsustainable fiscal policies often used financial lib-eralization to continue their debt buildup and delayadjustment.47 Even when fiscal deficits were smallerthan in the 1980s, the countries that liberalizedfinance often had large external and internal debtoverhangs that contributed to volatility.

Even a strong financial system has difficulty pro-tecting itself against default by an overindebtedgovernment, as the recent Argentine crisis illus-trates.48 Also, many countries that liberalized werepursuing exchange rate–based stabilization, or hadrelatively fixed exchange rates.These macroeco-nomic policies, and the tight monetary policy andthe credibility issues associated with them, oftenmeant extended periods of high real interest ratesand burdensome external borrowing, which even-tually contributed to countries’ inability to servicedebt and to financial crises.Thus, the problems withfinancial liberalization, the crises, and the limitedresults from financial liberalization in the 1990soften reflected macroeconomic policy deficienciesand the overhang of large external debts.


guarantees left the depositors and investors withgood returns during the 1990s.45

Improvements in these policies will depend notjust on new measures but also on strong implemen-tation, which has been difficult even in industrialcountries.

Financial Liberalization Increases FinancialResourcesA financially liberalized economy tends to generatemore financial resources than a repressed economy.This is an old lesson (McKinnon 1973; Shaw 1973)that had been forgotten during the financiallyrepressed 1980s. During the 1990s, the growth inbank deposits (relative to GDP) speeded up in manycountries.This acceleration reflected lower infla-tion, more realistic interest rates, and a wider menuof financial instruments, including foreignexchange–denominated instruments. In addition,domestic capital markets were started and devel-oped and private firms increased their external bor-rowing and external equity issues.

Deposits and domestic capital markets per-formed best where growth was already rapid, wherethere was a history of high deposit mobilization, andwhere investors were willing to take risks to getequity shares in rapidly growing corporations: EastAsia and India. Elsewhere, deposit growth was lessand capital market performance was less good.Deposit growth picked up much less in Latin Amer-ica, reflecting the region’s history of inflation andgovernment intervention.Also, much of the growthwas in foreign currency deposits that complicatedpolicy making.The decline in listings in equity mar-kets in Latin American and transition economiessuggests that access to finance through equity issuesdid not widen much. Even where capital marketperformance was better,access suffered from the lackof scale and liquidity in the markets; multinationaltakeovers of major firms; migration of listings to lesscostly, more liquid industrial country markets; and,more fundamentally, weak institutional frame-works—in particular the lack of information, regu-latory protection of minority shareholders, andbankruptcy protection for bond holders. Private

4. The Future of FinanceLooking ahead, the general pattern seems likely toremain one of more market-based finance. In mostcountries the financial liberalizations of the 1990sare unlikely to be reversed in their broad aspects,barring large macroeconomic policy errors. Awidespread return to financial repression is proba-bly now untenable for two reasons. One reason ispolitical: lower inflation and a more market-based,more open financial system became politicalimperatives in the 1990s. Repressed finance hadhigh costs and regressive distributional effects.Overtime, increasingly politically active households havedemanded protection for their savings and access tothe investment opportunities that were once avail-able only to political and economic elites.A secondreason, noted earlier, is that the increased access toexternal financial markets brought about by theenormous growth in trade, travel, and migrationand by improvements in communications has madefinancial repression difficult.49

Although macroeconomic stability, on whichgood finance depends, seems to exist in many coun-tries, macroeconomic issues remain. First, in today’sopen economies, slow policy responses or policyerrors quickly translate into macroeconomic insta-bility. Second, large government debt overhangsand/or large unfunded pension liabilities are prob-lems in many countries.The burden of these prob-lems has been eased by low world interest rates, butrising world interest rates, as well as other shocks,may lead macroeconomic policy astray.As the 1980sand 1990s show, excessive government debt caninteract with inconsistent exchange rate and mone-tary policy to lead to massive capital flight, large cur-rency depreciations, and costly financial sectorcollapses.When the government goes bankrupt, thefinancial system and the whole economy suffer.

The financial liberalizations of the 1990s havecreated a sounder basis for finance in at least sixways:

• Crises cleared away the “debris” of past nonper-forming loans, although they left large holdingsof government debt that created problems.


• Intermediaries and capital markets haveimproved. In Eastern Europe, Latin America,Africa, and even East Asia, many financial inter-mediaries were gradually replaced by reputableforeign banks. Such banks have better lendingskills, are more able to engage in arms-lengthlending and resist government pressures, and,potentially, impose fewer demands on govern-ment for bailouts than the intermediaries theyreplaced. Capital markets have also been set upor improved, but they still face many structuraland institutional challenges.

• Government and central bank debt markets havedeveloped.They allow central banks to carry outmonetary policy more efficiently, increase banks’liquidity, and allow less inflationary finance offiscal deficits.The growth of government debtmarkets also helps provide a benchmark that canmake private debt markets more efficient.

• Access to credit is growing in some countries,fromforeign banks (Clarke et al. 2004), new domesticbanks, and bank-like intermediaries.With the clo-sure of the old intermediaries,bad credit no longerdrove out good credit. New intermediaries thathold the promise of a sustainable increase in small-scale lending were able to grow. In Ecuador, forexample, the collapse of the public sector interme-diaries has left room for dramatic growth in pri-vate banks’ small credits in the last two years.

• Information is improving. The accounting andauditing of intermediaries and borrowers isimproving.So is information on small borrowers—public credit bureaus have been established in 23(mostly transition countries) since 1994 and theprivate credit bureaus that already existed in manycountries are improving (World Bank 2003d).

• Prudential regulation and supervision seem tobe improving and, in a few cases, the combina-tion of regulation, supervision,and a better safetynet have limited the impact of crises in individ-ual banks, for instance in Peru, although supervi-sion has also missed major weaknesses in somecountries, such as the Dominican Republic.

often said that private credit is currently limited bythe unwillingness of banks and markets to takerisks. In fact, it is limited by the large volume ofinelastically supplied government debt. This isbecause, to ensure that all government debt is held(either by financial intermediaries or by individu-als), the spread between interest rates on govern-ment debt and private debt has to be big enough tocrowd out enough private debt. Hence the way toexpand the supply of private credit is not to try tomake government debt less attractive but to leavemore space for funds for the private sector in thefinancial system,or to make private debt and equitymore attractive, so that more financial resources canbe raised in total.

Reducing the Volatility of Flows and Its ImpactGovernments have made various efforts to reducethe volatility of flows, especially on the upside of aboom, and to ease the impact of volatility, particu-larly by building up international reserves to offsetshocks and, within banks, by externally hedgingforeign currency liabilities.

But much remains to be done. Some analystshave argued for reducing incentives to excessivecapital inflows that can easily turn into excessiveoutflows.They argue, for example, that India’s suc-cess in avoiding the 1997 crisis was related to itslimits on banks’ (and firms’) offshore borrowing,even as it allowed inflows into the stock market andliberalized direct foreign investment regulations.Chile’s implicit taxes on short-term inflows alsoappear to have had some success in reducinginflows, extending their maturities, and in limitingthe impact of shocks, but at the cost of reducingcredit availability to the private sector (Edwards1999; Forbes 2003).

Another approach would be to reduce theincentives to banks for increasing their net offshoreborrowings.This would involve at least leveling theplaying field through application of the samereserve, liquidity, directed credit requirements, andpremiums for “deposit insurance” as on domesticdeposits. Here, too, little has been done. In the areaof international bond issues, some countries have


Improving FinanceFurther improvements in the contribution offinance to development depend on improving thekey tradeoff between safe and sound finance andrisk taking in the financial sector’s intermediationbetween savers and investors.The crises of the 1990snaturally have raised concerns about financial insta-bility that can lead to poor growth. Governments,attempting to reduce the future costs of crises, haveoften tended to emphasize prudence. But there is atension between stability and the ability of thefinancial system to carry out the key intermediaryroles for development—mobilizing funds fromsavers, allocating these funds to investors that willyield the best combination of return and risk,reducing risk, and shifting risk to those most will-ing to bear it.A financial system that does these taskswell will contribute greatly to development.

Improving the tradeoff between stability andintermediation in finance depends not just onmaintaining the systems of market-based interestrates and credit allocations that arose during the1990s, but also on the following:

• Reducing the crowding out of private credit bythe current large overhang of government debt;

• Reducing the volatility of resource flows, partic-ularly on the upside of cycles and to weak insti-tutions;

• Improving intermediaries and markets; and

• Widening access to credit.

The discussion below addresses each in turn.Progress will depend heavily on countries’ successin building institutions, improving their informa-tional and legal frameworks, and, ultimately, achiev-ing more competitive political systems that willreduce the power of political-economic elites.

Reducing the Crowding out of Private CreditPerhaps the most immediate obstacle to the abilityof the financial system to carry out its intermediaryrole, as well as a threat to stability, is the large over-hang of government debt in many countries.50 It is

begun to try to reduce the bias in bond buyers’beliefs that any restructuring will favor them, bymaking restructurings easier in terms of loweringthe percentage of bond holders that is needed toaccept a restructuring offer (the Collective ActionClause). However, it remains to be seen how thischange will operate—U.S. courts have oftenallowed individual creditors to seek preferentialtreatment. In the case of Elliot Associates vs. Peru,settled in 2000,Elliot Associates obtained a restrain-ing order on the payments on the restructured debtto which Peru had agreed with other creditor rep-resentatives. Peru eventually settled by paying ElliotAssociates $56 million for the debt that they hadbought for $11 million in 1996. Ultimately, allattempts at limiting excessive inflows depend onpolitical will to limit a boom, while in practice,countries often have eased restrictions on capitalinflows in order to prolong a boom.

Internally, governments have tried to developcapital markets as a shock absorber for the volatilityof external and internal flows. Funds invested inequity or long-term domestic government and pri-vate debt represent much less of a threat to the econ-omy than do volatile short-term external capitalflows.51 Thus, capital market development couldcontribute to stability as well as assisting the alloca-tion of funds to promising activities. One problem,of course, is that investors in such instrumentsdemand high returns under the current environ-ment in developing countries, so such instrumentsare often unattractive to potential issuers.This prob-lem adds to the structural problems of small size,lack of liquidity, and high costs that limit capitalmarket development.Domestic capital markets,par-ticularly in the larger countries, could be stimulatedby improvements in institutional factors, such as bet-ter information on firms, better rules on marketconduct and corporate treatment of minority share-holders, and better legal and judicial treatment ofbankruptcy. Generally, such improvements requiresubstantial time and effort.

Better market discipline is another approach toenhancing both intermediation and stability. Mar-ket discipline means ensuring that depositors and


international lenders have appropriate incentives tolimit their funding to weak intermediaries, byensuring that they stand to receive lower returns ondeposits and investments if problems occur. Marketdiscipline complements government regulation andsupervision and evidence exists that it can work indeveloping countries (Martinez-Peria andShmuckler 2001; Calomaris and Powell 2001).Unfortunately, market discipline depends on goodinformation.Though accounting and auditing areimproving, much remains to be done. For example,regulations could encourage prompt disseminationof accurate information and impose stiff penaltiesfor failure to do so.

Perhaps more important, market discipline isblunted by widespread implicit and explicit govern-ment guarantees that developed in the 1990s.Tomake market discipline work, governments face thedifficult task of establishing credible limits on liquid-ity support, blanket guarantees, and deposit insur-ance, so that at least the holders of banks’ largeobligations consider themselves at risk. One way tobegin improving market discipline might be to limitpayoffs to large providers of funds, especially sincethe latter can be expected to have relatively goodinformation about the strength of individual banks.It would also help to prevent problems in one bankfrom contaminating the rest of the system.However,the policy would immediately pass the problems ofa weak bank on to the central bank as lender of lastresort—a role that also would need to be limited, tocontain costs. Deposit insurance would also comeinto play,and would need to be truly limited to smalldeposits. Premiums for deposit insurance wouldneed to reflect differences in risk in different classesof banks. Unfortunately, the systems of risk-basedpremiums that have been adopted have largelycopied the pricing from industrial economies and,though better than flat, premiums still provide sub-stantial subsidies to domestic private banks, probablybecause of the banks’ political power.

When banks’ problems have become more sys-temic, the past responses—large lender of last resortsupport and blanket guarantees—have underminedfuture market discipline and been costly to future

hesitant to raise issues.52 Protecting supervisors com-pletely from legal action raises another issue—therisk that they will engage in malfeasance. Hence atribunal separate from the court system is needed todeal with accusations of malfeasance by supervisors.

More fundamentally, improvements in regulationand supervision face substantial political roadblocks,which have arisen in industrial as well as developingcountries.For example,from time to time,U.S.finan-cial economists have raised concerns about someU.S. banks being too big to fail. Also in the U.S.,political forces and regulatory forbearance are oftencited as a contributory factor in the U.S. savings andloan crisis. In many developing countries a few largebanks dominate the system, and bankers and majorborrowers are often one and the same. In this con-text, regulatory capital does not have even the mini-mal incentives that it does in arms-lengthtransactions between intermediaries and borrowers.The industrial-financial groups are the principalentrepreneurs in many countries, even large ones, solimits on connected lending are not feasible. If prob-lems of loan quality develop, the strength of the eco-nomic and political elite is likely to lead to regulatoryforbearance. Even if supervisors can identify capitalinsufficiencies and other regulatory violations, itwould be difficult for them to stand up to mono-lithic political elites,particularly when the alternativeis simply to ignore a problem in return for a supple-ment to their small salaries. Finally, the potentialstrength of regulation and supervision is limited bythe still-important role of large state banks that carryout government policies and are nontransparentalmost by design. In sum, regulation may not be suc-cessful unless it also empowers the market to moni-tor banks better, by encouraging market discipline.

Improving IntermediariesThe entry of reputable foreign banks is one way toimprove intermediation as well as to limit the costof crises. Reputable foreign banks bring bettertrained staff to the country and generally have bet-ter systems for evaluating and managing credit riskthan local banks.These advantages often spill overinto the local banking system, from competitive


generations. In effect, they have provided nearlyunlimited insurance not only for depositors but alsofor owners who can loot their banks. Alternativeoptions for dealing with crises would need to beginwith a different approach to dividing the costs of thecrises between current holders of liabilities andfuture generations. Lengthy suspensions of depositwithdrawals have proved to be undesirable: theybreak down the payments chain and have con-tributed to large declines in output, as has happenedin Argentina and Ecuador. But brief suspensions ofdeposit withdrawals,while term deposits are replacedby long-term, marketable instruments that involve asubstantial discount (in present value terms), are apossible alternative that would make current deposi-tors bear part of the cost of the crisis (Beckerman1995; IMF 2004b).Of course, such policies are polit-ically difficult to implement. But they would notonly limit the burden of crises that future generationswould have to pay, they might also reduce the size offuture crises, by strengthening market discipline.

To limit weak lending and crises, governmentshave also improved their banking laws and prudentialregulation and supervision.Since the strengthening ofprudential regulation and supervision only began inthe later 1990s, not much evidence has accumulatedon how well it can work to prevent crises.At the sim-plest level, regulators and supervisors in developingcountries may lack the technical skills even to dealwith loan quality and provisioning, not to speak ofmore complicated aspects of banking, such as evaluat-ing complex operations in capital markets and foreignexchange,swaps and derivatives that are poorly valuedin imperfect markets, and risk management models.Deficiencies exist in the consolidated supervision offinancial-industrial conglomerates and in the supervi-sion of offshore activities—important areas in devel-oping countries that will not be improved simply bygiving supervisors more power.

Partly these problems reflect incentives: typicallysupervisors are poorly paid and have an incentive toshift into banking, especially once they have beentrained to handle tasks well. Often supervisors aresubject to lawsuits by bankers, even for actions inperformance of their duties—which makes them

pressures and the movement of personnel. In addi-tion, reputable foreign banks also are likely to coverany losses on their loans or operations withoutdemanding government support, so as to avoiddamaging their reputations.

Reputable international banks have enteredmany countries in recent years, but losses and, insome cases, their own lack of capital have limitedtheir interest in further expansion. Some banks thatexpanded in Eastern Europe, in hopes of establish-ing a presence before countries acceded to theEuropean Union, suffered losses as competitiondeveloped. Some that expanded in Latin Americahave suffered losses from operations and from thedevelopments in Argentina. In the recent re-priva-tizations of Indonesian banks, only one bid camefrom a well-known global bank. Lesser-knownbanks have been expanding internationally, butsuch banks can generate more supervision prob-lems than local banks,because of the problems withinternational supervision. Moreover, without repu-tations to lose, such banks may pull out when thingsgo bad in the country or in their home market,leaving governments to bear the costs.

Improving Access to FinanceIncreasing small clients’ access to finance is a criti-cal issue for the financial sector in its support ofdevelopment. It involves the tradeoff between mak-ing banks safe and sound and making sure theycontinue to intermediate.A prerequisite to increas-ing access is to reduce the absorption of loanableresources by the government and the central bank.

Pressures remain great to direct low-cost creditto small borrowers. Historically, however, theseefforts have usually been unsuccessful, undermin-ing sustainable finance for rural and small andmedium-sized enterprises, just as occurred underfinancial repression.

A few intermediaries have successfully sus-tained loans to small borrowers (box 7.8). Themore traditional banking operations among themhave common features that explain their success:interest rates that cover costs, good deposit mobi-lization, containment of administrative costs, and a


high rate of loan collection, all backed by appro-priate internal incentives for good staff perform-ance (Yaron, Benjamin, and Piprek 1997).Theirexample needs to be followed.The informationalinfrastructure for small lending also improvedtoward the end of the 1990s with the founding andimprovement of credit bureaus.

Greater competition in banking services,through greater entry of banks and nonbanks andlooser regulations and supervision, is sometimesrecommended to improve access and lending ingeneral. Certainly, regulations should provide roomfor intermediaries that take funds from groups ofwell-informed investors/depositors and “nip at theheels” of banks, by offering better returns to depos-itors (though with greater risk), along with betterservice and innovation in products and lending.

Exactly how these entities should function—forexample as venture capital funds or deposit takers—and where the lines should be drawn between themand “banks,”are country-specific details.Such inter-mediaries operated in some East Asian and LatinAmerican countries and in India in the 1990s, and

BOX 7.8

Extending Credit for Small Borrowers

In addition to the well-known examples ofthe Grameen Bank (begun in 1976) andBank Rakyat Indonesia after its 1983

reform (Robinson 2002), other successfullenders began to expand toward the end of the1990s. These included CrediFe in Ecuador,MiBanco in Peru, CrediAmigo in Brazil, and, inIndia, SEWAH (which uses a Grameen-typeapproach) and self-help groups that use a mix-ture of the Grameen approach and traditionalbanking. Some of these intermediaries receivedsupport from donors. The Grameen approachrelies on the social responsibility of borrowerswho belong to a narrow group—an approachthat has also been used by some banks.

improvement of credit bureaus will be an importantdevelopment in improving access to credit as well asthe quality of loans. Important issues that need tobe addressed in this process are banking secrecy;how to make banks comply promptly and accu-rately with the requirement to provide informa-tion; whether the credit bureau is to be private orpublic; the inclusion of related information such asinstallment purchases; and consumers’ rights tochallenge and amend the information.

Improvements in the legal and judicial frame-work, notably the definition and execution of col-lateral and bankruptcy laws, are also important inimproving credit access and lending in general.Financial intermediaries prefer not to execute col-lateral—they are mobilizers and allocators of funds,not managers of firms—but the threat of executingcollateral gives an incentive for prompt debt ser-vice. Good bankruptcy laws make the survival ofviable firms easier and allow shifts of physical capi-tal from nonviable firms to others, with creditorsreceiving the maximum settlement.The potentialto improve credit access through better informa-tion, contract enforcement, and technology is great:in the United States, the cost of processing a smallloan is now below the price of a modest lunch.

Good access to financial services also involvesefficient deposit and payments services—importantfacilities given the increase in domestic and inter-national migration. In Africa, unfortunately, thestrengthening of the banking system has in somecases reduced access to deposit and payments ser-vices for small transactions. In other parts of theworld, payments services are often limited anduncompetitive. Post office banks—narrow banks,holding only government debt—with better tech-nology, and banks providing only these services (forexample in Tanzania and Mongolia) are examples ofinnovative ways to serve these needs.

5. Conclusion

While financial liberalization delivered in someaspects during the 1990s, its benefits are likely to lie


the outcomes illustrate their positive and negativesides. Before they fell victim to crises in 1997, thenonbank intermediaries increased finance forunderbanked sectors such as consumer durables andconstruction. But to some extent their success wasnot in competition and innovation but based onregulatory arbitrage relative to banks, which wereconstrained by interest rate controls (in India) ortight money policy (in Thailand). A critical issuewith such intermediaries is whether politically theycan be denied access to the bank safety net, orwhether they should be regulated and supervised inthe same manner as banks to protect taxpayers aswell as depositors. India appropriately resisted bail-ing out depositors, but Thailand’s attempt to offerthese intermediaries access to liquidity fundingcontributed to an easing of monetary policy thatwas inconsistent with the pegged exchange rate.

Another related issue is the size of theinvestor/depositor base: as it widens, the distinctionblurs between these institutions and banks, the pres-sures for claims on the safety net increase, and thegovernment may be drawn into supervision andregulation. Such problems have occurred in co-opbanks in India and in countries such as Indonesia,Nigeria, and Russia, where banks were allowed toset up with negligible capital. In Indonesia, 48 ofthese banks, many run by the politically well-con-nected, borrowed from the lender-of-last-resortfacility well in excess of their capital during the cri-sis (Kenward 2002), and used the funds to supportforeign exchange purchases and related businesses(see box 7.5 above).

Improving access, as well as the quality of creditallocation in general, depends heavily on improvingthe informational, legal, and judicial framework.The poor supply of information about borrowers,though improving, limits lending to smaller clients.In some countries, this problem has been circum-vented by lending through third parties that ineffect guarantee the loans.53 More generally, how-ever, better information would enhance competi-tion for sound borrowers while giving borrowersan incentive to service their loans to maintain goodcredit records. Thus, the continued spread and

in the future and to depend on further institutionalreforms.The crises of the 1990s, and the limitedcontributions of liberalization to growth and accessto finance, reflect to a large degree the continuationof the weak institutional framework related to theoverhang of the old financial system and,more fun-damentally, the persistence of old political and eco-nomic power centers.The freeing of interest ratesand credit allocation increased resource mobiliza-tion. But the persistence of the former institutionalframework meant that resource allocationimproved less rapidly. Implicit and explicit guaran-tees, by removing market discipline, contributed toexcessive expansion of lending for the low-produc-tivity projects of well-connected borrowers.Weakregulation and supervision reflected not just tech-nical problems but also political pressures for regu-latory forbearance. Large, generalized liquiditysupport during the crises often went to favored par-ties that bought foreign exchange with it. Informa-tion, which might have helped market disciplineand limited excessive lending had guarantees beenless, was not a focus of regulation, and it sufferedfrom the lack of transparency typical of manydeveloping countries. Limited credit accessreflected the crowding out of public sector andcentral bank borrowing. In addition, it reflected alack of information related not only to technicalissues but also to the unwillingness of establishedintermediaries to share information on their bor-rowers. Weak legal and judicial frameworks,designed to protect borrowers and often responsiveto economic and political elites, reduced the incen-tives to service debts and made it difficult for newborrowers to gain access to finance by pledging col-lateral effectively. Capital markets, which mighthave absorbed some of the shocks, grew slowlybecause of the weak institutional framework andunderlying structural problems.

The lessons of the 1990s are that improving thecontribution of finance to growth depends heavilyon macroeconomic stability, governments that arewilling to take steps to limit unsustainable booms, amarket-based approach, and the quality of institu-tions (financial intermediaries, information, and the


quality of the legal and regulatory framework).Thequality of institutions was not changed much by thestroke-of-the-pen liberalizations of interest ratesand credit allocations. Improving these institutions,and thereby improving financial intermediation,will depend on institution building,better informa-tional and legal frameworks, and, ultimately, morecompetitive political systems. Success will dependon a mix of increased market discipline and limitingguarantees, better regulation and supervision thatincludes encouraging greater market discipline ofintermediaries, greater participation of reputableforeign banks, and capital market development.Government is needed to support better markets,without intervening excessively in them, backed byan open political process that limits the distortionsof finance in favor of well-connected parties.

Notes

1. As Lenin cogently put it, “The big banks are the stateapparatus which we need to bring about socialism andwhich we take readymade from capitalism” (quoted inLaPorta,López de Silanes,and Shleifer 2002a,266).Thuscommunist, socialist, and planned economies national-ized domestic and foreign commercial banks. Ger-schenkron (1962) was among the first to provideacademic support for the provision by government andstate banks of funds for industrialization and long-termcredit. In addition to state banks, specialized develop-ment finance intermediaries, generally public, were setup to provide credits for small-scale industry, agricul-ture, housing, and long-term industrial credit. Theywere financed by government-guaranteed external bor-rowing, including bilateral and multilateral loans; bylow-cost directed credits from banks and other inter-mediaries; and by government revenues. Often theseintermediaries went bankrupt, reflecting failures to col-lect debt service and dependence on unhedged externalborrowing.

2. For example, Brownbridge and Harvey (1998) describesuch financial repression in Africa.

3. Dornbusch and Edwards (1991); Alesina, Grilli, andMilesi-Ferreti (1994); and Garrett (1995, 2000).

4. Estimates of aggregate subsidies range from 3 to 8 per-cent of GDP annually (World Bank, World DevelopmentReport 1989; Hanson 2001). Regarding allocations, inCosta Rica in the mid-1970s for example, the publicBanco Nacional’s interest rate subsidy on agricultural

13. As an example of the popularity of these measures, inPeru after hyperinflation at the end of the 1980s, the1993 Constitution (Article 64) guaranteed citizens theright to hold and use foreign exchange. More than 50percent of deposits are in dollars, even in the non-Limasavings banks.

14. The interest rates on foreign currency credits avoid thehigh, up-front cost of an expected depreciation thatmay not occur for some time—the “pesoproblem”(Hanson 2002). This improves cash flows(lower deficits for governments using cash accounting)and increases a loan’s effective maturity. Moreover,when a depreciation does come, the cost is spread outin the amortization period. Not surprisingly, govern-ments borrow externally and, many countries, forexample Mexico in 1994 and Brazil and Turkeyrecently, have indexed some domestic debt to foreigncurrency. For private firms, there is also the hope that adepreciation may lead to a government bailout, eitherby a favorable takeover of their foreign loans or anasymmetric conversion of domestic foreign currencydebts and deposits to local currency, as occurred inMexico (1982) and Argentina (2002).However, foreigncurrency loans do increase bank risks, even whenmatched with foreign currency deposits, since the bor-rowers may not have easy access to foreign currencyearnings. Banks could have adjusted the foreign anddomestic currency proportions of their balance sheetsby varying interest rate differentials, but, given thedemand for foreign currency deposits, the spread prob-ably would have been high,creating moral hazard prob-lems in loans in domestic currency.

15. These figures understate the relative growth of publicsector debt because they include China, where depositgrowth was large and banks’ accumulation of govern-ment debt was relatively small, but the accumulation ofstate enterprise debt was large. In those transition coun-tries for which relevant data are available, privatizationreduced borrowing by public enterprises, thereby off-setting the rise in government debt, but deposits grewonly slowly and were largely absorbed by increased cen-tral bank debt.

16. Note that these figures understate the growth of privatecredit in India and East Asia before 1997 and overstate itafter 1997, because of the growth and decline of thenonbank sector.

17. Government debt was either injected into the banks aspart of restructurings or, in the case of deficit finance,sold at whatever rates would ensure its purchase.Thus,as a first-order approximation, the supply was inelastic(except for changes in the proportions sold internallyand externally).The liquidity, low risk, and low capitalrequirements on government debt affected only the rate


credits was equal to about 4 percent of GDP and 20percent of agricultural value added.About 80 percent ofthe credit went to 10 percent of the borrowers; the aver-age subsidy on these loans alone would have put eachrecipient into the upper 10 percent of the income dis-tribution (World Bank,World Development Report 1989).The situation in other countries was similar. See Adamsand Vogel (1986); Adams, Graham, and Von Pischke(1984); Gonzalez-Vega (1984); and Yaron, Benjamin,and Piprek (1997). Larger firms often accessed directedcredit and on-lent it to their suppliers, capturing thespread between repressed and free rates. Directed cred-its were also diverted into loans with free rates, forexample through curb markets, or, when some depositrates were freed, into deposits that paid higher rates thanthe loan rates on directed credits.

5. Abiad and Mody (2003) note the link between greateropenness to trade and financial liberalization.

6. Capital controls, particularly in the context of macro-economic imbalances, increase incentives for corrup-tion, worsen the income distribution, and, because theyfail, create disrespect for laws. Even in the 1970s, a highproportion of the massive capital inflows into LatinAmerica leaked out (Dooley et al. 1986).More recently,in China, net short-term outflows of capital and errorsand omissions in the balance of payments were verylarge (World Bank 1997a, 2000c).

7. For example, in Mexico after the post-1982 high infla-tion, the limits on interest rates on agricultural loanswere below the rates on some deposits for a period.Rural borrowers often simply took their loan proceedsand deposited them, earning a positive return on theloans with much less effort than by farming.

8. Abiad and Mody (2003).9. Stock markets were opened to foreign investors

between 1986 and 1993 in the major East Asian andLatin American countries and in India and Pakistan(Bekaert, Harvey, and Lundblad 2003).

10. Levine and Zervos (1998); Levine, Loayza, and Beck(2000).

11. The sharp fall in inflation in the 1990s made interestrates more realistic, even with declines in nominal rates;it also reduced other financial distortions associated withinflation.Among the 25 developing countries with thelargest financial systems, those with hyperinflation at thebeginning of the 1990s reduced inflation sharply (insome cases, such as Argentina, to single digits), whilemost of those with initial inflation of 10–50 percentannually reduced inflation to single digits by 2000. InAfrica, inflation also fell and in most transition coun-tries, inflation fell sharply from initial high levels.

12. Foreign currency holdings also were often large relativeto financial systems (Hanson 2002).

differential between the debt and private credit that wasneeded to crowd out the equivalent amount of privatecredit, rather than the amount of government debt held,which was determined by the inelastic supply.

18. In some cases, the central banks also temporarily actedas large lenders of last resort.

19. The increase in external assets probably reflected anattempt to hedge the risks from their foreign currencyliabilities, including deposits (Honahan and Shi 2002).Although banks’ net external positions were small in2000, gross external assets and liabilities were muchlarger than earlier (Hanson 2003b), suggesting thatfinancial liberalization had increased banks’ ability todiversify themselves.

20. For statistical evidence on the importance of privatesector credit in growth see Levine and Zervos (1998);Levine, Loayza, and Beck (2000).The evidence of thelink between savings/investment and financial sectorliberalization is mixed (see, for example, Bandiera et al.2000), but the investment ratio does seem to have risenin the 1990s in the larger Asian countries, though not inthe larger Latin American countries and it actuallydeclined in the larger African countries.The differencebetween saving and investment ratios may, of course,reflect differences in capital inflows.

21. Crises,unproductive credits, and their links to the unre-formed institutional and political framework thatremained after liberalization are discussed in the sectionbelow on financial crises.

22. Stock markets were reported as of 1991 in Hungaryand Poland; in 1994 in Croatia, the Czech Republic,Romania, Russia, the Slovak Republic, and Slovenia; in1995 in Bulgaria, Latvia, Lithuania, and Mongolia; in1996 in the Former Yugoslav Republic of Macedonia,Moldova, and Uzbekistan; and in 1997 in Estonia,Kazakhstan, and Ukraine (Standard and Poor’s 2003).

23. This average is for the 17 of the 25 largest financial mar-kets for which data are available on banks’ domesticcredit to the private sector. It excludes China, India, andKorea, which do not report separate data on privatesector credits.These three countries are large externalborrowers in absolute terms but are likely to havesmaller ratios of private external borrowings to bankcredit than the average for the 17 countries.

24. The additional currency risk of these funds was less thanit might seem, as domestic credit in many countries wasincreasingly denominated in foreign exchange.

25. “[The state banks’] commercialization as joint stockcompanies was not accompanied by sufficient commer-cialization of their credit management, product devel-opment, service levels, operational efficiency, or riskmanagement.All this meant poor loan performance andeventually insolvency. Many factors worked against


early detection of such problems—poor accounting andauditing standards, inexperienced supervisory person-nel, inadequate prudential regulations,decentralized andincomplete information systems (often branch accountsnot consolidated with headquarters accounts) and thetraditional reliance on the government for additionalfunding when liquidity became short.… Managementinformation systems were weak. All these factorsworked against timely and effective scrutiny of manage-ment behavior” (Sherif,Borish, and Gross 2003,21–22).

26. Western European banks entered Eastern Europe hop-ing to gain market shares before the European Unionexpanded.The shares of foreign banks in the number ofbanks and in total bank assets grew rapidly in Bulgaria,the Czech Republic, Hungary, and Poland. In Russiaand Ukraine, however, foreign banks represented only asmall fraction of the total number of banks,even in 2000(Sherif, Borish, and Gross 2003). In Latin America,Spanish banks became a major force by taking over stateand private banks. In Africa, foreign banks reentered andSouth African banks were playing an increasing role insouthern Africa at the end of the 1990s.

27. Research suggests that in Latin America foreign banksare at least as good as domestic banks at lending to smallfirms (Clarke et al. 2004), and in India foreign banks’lending to small and medium-size firms has grown fasterthan that of state banks.

28. Indonesia took liberalizing bank entry to an extreme,with almost “free banking” (box 7.5). Russia and Nige-ria later followed a similar approach.Most new banks inthese countries were “pocket” banks, capturing fundsfor their owners’ firms. In Indonesia, these banks werehit hard by the crisis and proved costly to the govern-ment when deposits were guaranteed.

29. Caprio and Klingebiel (2002) list 117 systemic financialcrises (in which most of banks’ capital was exhausted) in93 countries and 51 borderline crises in the period1970–99. See also Sundararajan and Balino (1991) andWorld Bank,World Development Report 1989.

30. Argentina, Russia, and some African countries had highpublic sector debt compared to public revenues (IMF2004a). Other countries, notably in East Asia, had highprivate debt, including high external private debt, rela-tive to GDP.Variants of exchange rate–based stabiliza-tion were being used by Argentina, Chile, and Uruguayin the late 1970s and by Argentina, Brazil, and Mexicoin the 1990s. Other countries, notably the East Asiancountries and Turkey in the 1990s, limited the flexibil-ity of their exchange rates.The relation between the1990s crises and the current account deficits is similar tobut not the same as “Generation I”models of balance ofpayments crises (Krugman 1979). In the 1990s crises,the problem was not just financing the current account

if bank failures lead to deposit insurance payments (Ben-ston and Kaufmann 1997). It is unclear how well thisapproach would work in developing countries.

39. Even if small depositors are promptly paid off, largedepositors may switch to foreign exchange.

40. Arioshi et al. (2000);Dooley (1996).The Malaysian con-trols are often cited as an example of effective controls,but they were put in place after the crisis was largelyover (World Bank 2000c).

41. To paraphrase William McChesney Martin, formerchairman of the U.S. Federal Reserve Board, the role ofgovernments is to take away the punchbowl before theparty gets too wild.

42. Such government behavior occurs not only in devel-oping countries but also in industrial countries, forexample in the U.S. savings and loan sector before itscrisis.

43. This weakness would have existed even if good infor-mation had been available.

44. Interestingly, additional deposits often flowed into statebanks during these periods. Despite the weakness oftheir lending, the public typically considered them tohave better guarantees.These banks, in turn, often madeadditional loans to weak borrowers.

45. See, for example, Klingen et al. (2004).46. Financial liberalizations, even gradual ones, are not easy

to manage. Errors in liberalization are not always tech-nical; they sometimes reflect pressures by influentialgroups.

47. Financial liberalization also tended to increase the fiscaldeficit and make it more costly to finance, as the gov-ernment lost seigniorage revenues and had to pay moremarket-based interest rates on its debt.

48. World Bank (1998a) describes the substantial strength-ening of Argentina’s financial system in the mid-1990s.

49. Some policies and some countries will of course devi-ate from the general trends. Some governments wheredemocracy is limited may attempt to impose capitalcontrols and return to the inflation tax as a means ofcapturing resources. And many countries remain con-cerned about the narrowness of credit access for theircitizens, and seek ways to provide funds for rural andsmall and medium-size enterprise lending at below-market rates through specialized intermediaries,notwithstanding the past failures of this approach andthe increases in access that are occurring.

50. Such debt is not completely bad—it can serve as a liq-uid asset to improve the payments system and as a wayfor individual banks to deal with limited runs.However,governments’ low revenue-generating capacities make itdifficult to service these debts, lead to cuts in publicsocial and infrastructure spending, and divert govern-ments from developmental issues by the day-to-day


deficit but net amortizations of long- and short-termloans, which could change suddenly.

31. Portfolio adjustments to improved investment opportu-nities generate rapid inflows initially, followed by a slow-down in inflows and net negative foreign exchangeflows (because of interest payments that require internaladjustment).

32. See, for example, Demirgüç-Kunt and Detragiache(2002); and Kaminsky and Reinhart (1999). In the crisesof the early 1980s, high U.S. interest rates, as well as thefall in petroleum prices, probably played a role, whilelower interest rates contributed to the large capitalinflows to developing countries in the early 1990s andagain recently.The rise in international interest rates thatstarted in 1993 probably contributed to a gradual tight-ening of credit conditions for developing countries.

33. A substantial literature has evolved over the possibilitythat the crises in the 1990s, particularly those in EastAsia, reflected contagion in financial markets, not fun-damentals; see Claessens and Forbes (2001) and workscited there. Contagion is one explanation of “Genera-tion II” models of crises in which there are multipleequilibria, associated with high and low rates of capitalinflow. No doubt international investors exhibit someherding behavior for various reasons. Another expla-nation is that events in one country could lead exter-nal investors to reevaluate the subjective risks in othersand reduce their exposures.This also would seem likecontagion.

34. The lag between liberalization and crises seems fairlylong (Demirgüç-Kunt and Detragiache 2001, 105).Thelag may also reflect the difficulty of pinpointing liberal-ization and crises, both of which occur over time, as dis-cussed in Eichengreen (2001). Demirgüç-Kunt andDetragiache (2001) date liberalization from the removalof some interest rate controls and note that the esti-mated lag may reflect the gradualness of interest rate lib-eralization. Of course, the initial rise in deposit interestrates may also reflect part of a defense against a run onthe currency, as for example in India in 1991.The lagbetween financial crises and currency crises may reflectliquidity support to weak banks at the start of financialcrises, as discussed below.

35. See Diaz Alejandro (1985); the capsule discussions ofcountry experience in Sundararajan and Balino (1991,40–49); and the descriptions of financial crises inKindleberger (2000).

36. Of course, this explanation is related to Generation IImodels of crises, discussed in footnote 33.

37. State banks have not been closed without paying offdepositors, except in a few African countries.

38. The United States has required intervention in weakbanks well before capital is exhausted, and explanations

problem of rolling over the debt, raising the risk of areturn to inflationary finance.These potential problemshave been eased by the fall in interest rates worldwide.However, when interest rates begin to rise again, andthe costs of debt service correspondingly increase, theproblems may reappear.

51. International equity market markets can also act as ashock absorber, but only the largest and most transpar-ent firms can list in these markets. Offshore bond mar-kets also are developing in private as well as publicbonds; they reduce the risk of credit crunches butincrease currency risk.


52. Protecting supervisors completely from legal actionraises another issue, the risk that they will engage inmalfeasance. Hence, a tribunal separate from the courtsystem is needed to deal with accusations of malfeasanceby supervisors.

53. For example,making loans for scooters, cars, and homesto workers in the formal sector who often cannot befired; making loans to farmers that are repaid by deduc-tions from the contracts the farmers have with cropbuyers; and lending to small and medium-size enter-prises either through larger firms or by discounting theirorders from such firms.

Financial Liberalization: What Went Right,What Went Wrong?

Documents

Transcript of Financial Liberalization: What Went Right,What Went Wrong?