capital controls: country experiences with their use … paper aims to develop a deeper...

This paper aims to develop a deeper understand-ing of the role that capital controls may play in

coping with volatile movements of capital, and ofcomplex issues surrounding capital account liberal-ization. It provides a detailed analysis of specificcountry cases to shed light on the potential benefits orcosts of capital controls, including those used in crisissituations. It also considers the important link be-tween prudential policies and capital controls, includ-ing the improvement of prudential practices and ac-celerated financial sector reform to address the risksinvolved in cross-border transactions. Chapter II re-views the experience of selected countries with theuse or removal of capital controls based on a detailedreview and comparison of the experience of a groupof 14 countries that used various types of capital con-trols, often to manage episodes of unsustainable capi-tal flows. Chapter III examines the prudential ap-proach to managing the risks associated with capitalflows, and Chapter IVprovides some conclusions.

The review of country experiences in Chapter II isorganized around five key themes. These themes in-clude the use and effectiveness of controls on capitalinflows in limiting the potentially destabilizing ef-fects of short-term capital flows and preservingmonetary policy autonomy under tightly managedexchange rate systems (involving formal or de factopeg arrangements); the potential benefits and costsof reimposing selective controls on capital outflowsto reduce pressures on the exchange rate, includingin the context of currency or banking crises, as wellas of extensive exchange controls that may entail re-strictions on both capital and current internationaltransactions; long-standing and extensive capitalcontrols and their role in reducing financial vulnera-bility; and the benefits and costs of rapid and wide-ranging liberalization of previously restrictive ex-change control regimes. For each group, two to fivecountries were selected for case studies that providerecent and diverse experiences.1

For the first four key themes above, the study ex-amines the motivations of countries to limit capitalflows; the role that the controls may have played incoping with particular situations; the nature and de-sign of the control measures; and their effectivenesswith respect to influencing targeted flows and activ-ities and realizing their intended objectives. Thestudy also seeks to identify the factors that mayhave influenced the effectiveness of the controls, aswell as the potential costs that may have been asso-ciated with their use. Brief descriptions and assess-ments of each country’s experience can be found inChapters V–IX, and these form the basis for theanalyses in Chapter II. Appendices I–III provide amore detailed study of three countries that have re-ceived widespread attention in terms of their capitalcontrol measures: Chile, India, and Malaysia. In thecase of the benefits and costs of liberalization, thediscussion also focuses on the underlying reasonsthat have motivated countries to rapidly liberalizecapital flows, and the factors that may have im-pinged on the effectiveness of the liberalizationstrategies.

In analyzing the effects of capital controls, draw-ing conclusions from econometric and statisticalanalysis is inherently problematic, not least becauseof the difficulty in quantifying the capital controlmeasures, the quality of capital account data, and theconfluence of policy and the external environmentinfluencing the volume of capital flows. This paperadopts a descriptive approach, and concentrates onthe effectiveness of capital controls and the costs as-sociated with their use. While every effort has beenmade to provide an objective account and analysis ofthe developments, the country episodes may be opento different interpretations.

The prudential approach to managing the risks in-volved in cross-border transactions is described inChapter III. This area has only recently received

I Overview

3

1The choice of countries as well as the number of the countrycases for a group was based on ready availability of adequate in-formation to make an informed analysis. Conditions in worldgoods and financial markets have changed profoundly during the

last three decades, so the paper focuses on the experience of(mainly developing) countries that have used or liberalized capi-tal controls during the last 5 to 10 years. Most advanced countrieshad liberalized their capital accounts completely by the beginningof this decade.

I OVERVIEW

more widespread attention, and the prudential stan-dards themselves are under development. The chap-ter reviews progress in establishing prudential stan-dards for cross-border flows and issues in theirimplementation, and discusses their limitations andthe conditions for their effectiveness. The chapteralso examines the link between capital controls andprudential policies.

The analysis throughout the main paper takes asits starting point the observation that economic per-formance—and the volume, composition, andvolatility of international capital flows—will dependto a large extent on the mix of policies. The effec-

tiveness of particular measures or institutions is usu-ally gauged in the first instance with respect to theobjectives of a country’s economic policy. These ob-jectives may differ across countries, and so will theappropriate policy mix. Limiting macroeconomicand financial instability is among the most widelyshared objectives, and macroeconomic policies, cap-ital controls, and prudential measures may all have arole to play in achieving this goal. Although there isno unique best approach, the analysis in this paperunderscores that some types and combinations ofpolicies tend to be more effective than others, andhave fewer undesirable side effects.

4

General Considerations on the Use ofCapital Controls

This section provides a brief summary of thegeneral considerations involved in the use of capitalcontrols, including the objectives they have beenset to achieve, the ways in which their effectivenesshas been assessed, the forms they have taken, andthe potential costs that may be associated with theiruse. Country experiences presented in the subse-quent sections are assessed in light of these generalconsiderations.

Objectives of Capital Controls

Many arguments have been advanced in the eco-nomic literature to justify the use of capital controls.Among these, second-best arguments identify situa-tions in which capital account restrictions improveeconomic welfare by compensating for financialmarket imperfections, including those resulting frominformational asymmetries. Proposals to addressthese imperfections range from improved disclosureand stronger prudential standards to the impositionof controls on international capital flows.

Policy implementationarguments hold that capitalcontrols may help to reconcile conflicting policy ob-jectives when the exchange rate is fixed or heavilymanaged. These arguments include preserving mone-tary policy autonomy to direct monetary policy to-ward domestic objectives and reducing pressures onthe exchange rate. An additional, related, motivationfor capital controls has been to protect monetary andfinancial stability in the face of persistent capitalflows, particularly when there are concerns about(1) the inflationary consequences of large inflows, or(2) inadequate assessment of risks by banks or thecorporate sector in the context of a heavily managedexchange rate that, by providing an implicit exchangerate guarantee, encourages a buildup of unhedgedforeign currency positions. Finally, capital controlshave also been used to support policies of financialrepression to provide cheap financing for governmentbudgets and priority sectors. Other political economyarguments are outside the scope of this review.

Effectiveness and Potential Costs of Capital Controls

The effectiveness of capital controls has fre-quently been assessed on the basis of their impacton capital flows and policy objectives, such asmaintaining exchange rate stability, providinggreater monetary policy autonomy, or preservingdomestic macroeconomic and financial stability.Much attention has been given in the literature todifferentials between domestic and international in-terest rates, as capital controls tend to create awedge between domestic and external financialmarkets. This wedge, however, may itself create in-centives for circumvention; the effectiveness of con-trols will then depend on the size of this incentiverelative to the cost of circumvention. If the controlsare effective, capital flows would become less sensi-tive to domestic interest rates, which the authoritiescould then orient toward domestic economic objec-tives. These and other issues are considered in thecountry case studies, with an emphasis that variesaccording to the circumstances of the individualcountry and the availability of data and previousstudies.

Econometric and statistical studies of these is-sues have several methodological shortcomings. Inparticular, no generally accepted and reliable mea-sures of the intensity of capital controls are avail-able, and many studies simply use dummy variablesfor their presence or absence. Also, it is often diffi -cult to ascertain whether differences in the vari-ables to be explained are attributable to capital con-trols or other factors, some of which are alsodifficult to measure (e.g., the effectiveness of pru-dential supervision). Moreover, it has proven diffi -cult to distinguish in an economically meaningfulway between long-term and short-term capitalflows. Short-term loans are often rolled over repeat-edly, while long-term instruments can be often soldat short notice in secondary markets. This applieseven to foreign direct investment when the investorcan borrow against his collateral and short the cur-rency. Derivatives markets, including those forswaps and options, open up many additional av-

II Country Experiences

5

II COUNTRY EXPERIENCES

enues for changing the effective maturity of invest-ments. The extent to which the distinction betweenshort-term and long-term flows is erased dependsprimarily on the level of development of financialmarkets, and in particular on their depth and liquid-ity. These attributes of financial markets will in turnbe affected by government regulation, includingcapital controls.

Regardless of whether capital controls are effec-tive, their use (or reimposition) may entail somecosts. (See Bakker, 1996.) First, restrictions on capi-tal flows, particularly when they are comprehensiveor wide-ranging, may interfere with desirable capitaland current transactions along with less desirableones. Second, controls may entail nontrivial admin-istrative costs for effective implementation, particu-larly when the measures have to be broadened toclose potential loopholes for circumvention. Third,there is also the risk that shielding domestic finan-cial markets by controls may postpone necessary ad-justments in policies or hamper private-sector adap-tation to changing international circumstances.Finally, controls may give rise to negative marketperceptions, which in turn can make it costlier andmore difficult for the country to access foreignfunds.2

Types of Capital Controls

Controls on cross-border capital flows encom-pass a wide range of diversified, and often country-specific, measures. These restrictions on and im-pediments to capital movements have in generaltaken two broad forms: (1) “administrative” or di-rect controls and (2) “market-based” or indirectcontrols. In many cases, capital controls to dealwith episodes of heavy capital flows have been ap-plied in tandem with other policy measures, ratherthan in isolation.

Administrative or direct controls usually involveeither outright prohibitions on, or an (often discre-tionary) approval procedure for, cross-border capitaltransactions (Box 1). Market-based or indirect con-trols, on the other hand, attempt to discourage par-ticular capital movements by making them morecostly. Such controls may take various forms, in-cluding explicit or implicit taxation of cross-borderfinancial flows and dual or multiple exchange ratesystems. Market-based controls may affect theprice, or both the price and the volume, of a giventransaction.

Capital Controls to Limit Short-Term Inflows

Brazil (1993–97), Chile (1991–98), Colombia(1993–98), Malaysia (1994), and Thailand(1995–97) have all used capital controls to limitshort-term capital inflows. Short-term capital flows,though typically seen as less risky from the perspec-tive of individual banks and other investors,haveoften been regarded as speculative and destabilizingat the aggregate level. Long-term flows, by contrast,are usually considered to be more closely related tothe real economy and hence more stable and desir-able. It is not always straightforward to distinguishbetween short-term and long-term flows in an eco-nomically meaningful way. Figures 1–9 illustrate de-velopments in key economic indicators during theseepisodes. Part II, Chapter V, provides further detailsof the country experiences.

Motivations for Capital Controls on Short-Term Inflows

In all five countries, capital controls to limitshort-term inflows were imposed in response toconcerns about the macroeconomic implications ofthe increasing size and volatility of capital inflows,within the broader context of abundant capital flowsto emerging economies during the 1990s. Longer-term inflows generally reflected structural factors,notably wide-ranging economic reform (Chile,Colombia, and Malaysia) or the liberalization of ex-ternal transactions (Brazil, Colombia, and Thai-land). Short-term inflows reflected high domesticinterest differentials in the context of pegged (Thai-land) or heavily managed exchange rate regimes(Brazil, Chile, Colombia, and Malaysia), which hadoften given markets a false sense of security. Thelarge and persistent inflows complicated the imple-mentation of monetary policy, at times owing to alack of adequate monetary instruments (Thailand).In most cases, sterilization operations were the firstpolicy response to the inflows. However, such oper-ations typically entailed costs to the central bankowing to differentials between the cost of issuingsecurities and the return on foreign assets. Further-more, sterilization operations may have attractedfurther inflows as they tended to keep interest rateshigh.

Controls on capital inflows were imposed to re-duce reliance on sterilization, and in some cases topostpone other adjustment. These controls were typ-ically accompanied by other policies, including aliberalization of outflow controls (Chile and Colom-bia), an adjustment or progressive increase in theflexibility of the exchange rate (Chile and Colom-bia), and a further strengthening of the prudential

6

2Another issue, which is not addressed in this paper, is the ef-fect on other countries and the international economy at largewhen a country, or group of countries, resorts to capital controls.


framework for the financial system (Chile, Colom-bia, and Malaysia). In some countries, fiscal policyremained tight (Chile and Malaysia); in others, fur-ther tightening was limited (Brazil and Thailand);

and in some it remained loose, putting even greaterpressure on monetary policy (Colombia).

All five countries used inflow controls to preserveor enhance monetary policy autonomy. The controls

7

Capital controls have generally taken two mainforms: direct or administrative controls, and indirect ormarket-based controls.

Direct or administrative capital controls restrict capi-tal transactions and/or the associated payments andtransfers of funds through outright prohibitions, ex-plicit quantitative limits, or an approval procedure(which may be rule-based or discretionary). Adminis-trative controls typically seek to directly affect the vol-ume of the relevant cross-border financial transactions.A common characteristic of such controls is that theyimpose administrative obligations on the banking sys-tem to control flows.

Indirect or market-based controls discourage capitalmovements and the associated transactions by makingthem more costly to undertake. Such controls may takevarious forms, including dual or multiple exchange ratesystems, explicit or implicit taxation of cross-border fi-nancial flows (e.g., a Tobin tax), and other predomi-nantly price-based measures. Depending on their specifictype, market-based controls may affect only the price orboth the price and volume of a given transaction.

• In dual (two-tier) or multiple exchange rate sys-tems, different exchange rates apply to differenttypes of transactions. Two-tier foreign exchangemarkets have typically been established in situa-tions in which the authorities have regarded highshort-term interest rates as imposing an unaccept-able burden on domestic residents, and have at-tempted to split the market for domestic currencyby either requesting or instructing domestic finan-cial institutions not to lend to those borrowers en-gaged in speculative activity. Foreign exchangetransactions associated with trade flows, foreigndirect investment, and usually equity investmentare excluded from the restrictions. In essence, thetwo-tier market attempts to raise the cost to spec-ulators of the domestic credit needed to establisha net short domestic currency position, while al-lowing nonspeculative domestic credit demand tobe satisfied at normal market rates. Two-tier sys-tems can also accommodate excessive inflowsand thus prevent an overshooting exchange ratefor current account transactions. Such systems at-tempt to influence both the quantity and the priceof capital transactions. Like administrative con-trols, they need to be enforced by compliancerules and thus imply administration of foreign ex-change transactions of residents and domesticcurrency transactions of nonresidents to separatecurrent and capital transactions.

• Explicit taxation of cross-border flows involvesimposition of taxes or levies on external financialtransactions, thus limiting their attractiveness, oron income resulting from the holding by residentsof foreign financial assets or the holding by non-residents of domestic financial assets, thereby dis-couraging such investments by reducing their rateof return or raising their cost. Tax rates can be dif-ferentiated to discourage certain transaction typesor maturities. Such taxation could be considered arestriction on cross-border activities if it discrimi-nates between domestic and external assets or be-tween nonresidents and residents.

• Indirect taxation of cross-border flows, in the formof non-interest-bearing compulsory reserve/de-posit requirements (hereafter referred to as unre-munerated reserve requirement (URR)) has beenone of the most frequently used market-based con-trols. Under such schemes, banks and nonbanksdealing on their own account are required to de-posit at zero interest with the central bank anamount of domestic or foreign currency equivalentto a proportion of the inflows or net positions inforeign currency. URRs may seek to limit capitaloutflows by making them more sensitive to domes-tic rates. For example, when there is downwardpressure on the domestic currency, a 100 percentURR imposed on banks would double the interestincome forgone by switching from domestic to for-eign currency. URRs may also be used to limit cap-ital inflows by reducing their effective return, andthey may be differentiated to discourage particulartypes of transactions.

• Other indirect regulatory controls have the charac-teristics of both price- and quantity-based mea-sures and involve discrimination between differenttypes of transactions or investors. Though theymay influence the volume and nature of capitalflows, such regulations may at times be motivatedby domestic monetary control considerations orprudential concerns. Such controls include provi-sions for the net external position of commercialbanks, asymmetric open position limits that dis-criminate between long and short currency posi-tions or between residents and nonresidents, andcertain credit rating requirements to borrowabroad. While not a regulatory control in the strictsense, reporting requirements for specific transac-tions have also been used to monitor and controlcapital movements (e.g., derivative transactions,non-trade-related transactions with nonresidents).

Box 1. Types of Capital Controls


were seen as a means of resolving the classic policydilemma that results from having more objectivesthan independent policy instruments. Typically,monetary policy was oriented toward reducing infla-tion while also attempting to stabilize the exchangerate under relatively free capital movements that

made it difficult to set monetary and exchange ratepolicies independently.

Prudential concerns also motivated the adoptionof controls on capital inflows, though in most cases,macroeconomic considerations appeared to bedominant. The controls were intended to alter the

8

0

1

2

3

4

5

9796959493929119902

4

6

8

10

12

14

16

18

959493921991

0

2

4

6

8

10

12

9795939189871985–10

–5

0

5

10

15

9796959493921991

–2

0

2

4

6

8

98979695949392911990

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 1. Countries with Controls on Short-Term Capital Inflows:Net Private Capital Flows(In percent of GDP; episodes examined in the paper are shaded)

Source: IMF’s World Economic Outlook database.


maturity composition of the inflows toward lessvolatile flows, in addition to reducing their overallvolume. Short-term flows were seen to have poten-tial adverse effects on macroeconomic and finan-cial system stability, particularly as the ability of fi-nancial institutions to safely intermediate the

inflows was uncertain (Colombia, Malaysia, andThailand). The case has also been made that thesecountries faced a “systemic” shock (owing to theabundance of capital flows to emerging economies)that could not be addressed by conventional poli-cies (Chile).

9

0

10000

20000

30000

40000

50000

60000

70000

5000

10000

15000

20000

25000

30000

35000

40000

0

5000

10000

15000

20000

10000

15000

20000

25000

30000

35000

40000

3000

4000

5000

6000

7000

8000

9000

10000

1990

1985

1990 92 94 96 98

1991 9392 94 9695 9787 89 91 93 95 97

1991 92 93 94 9591 92 93 94 95 96 97

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 2. Countries with Controls on Short-Term Capital Inflows:Foreign Exchange Reserves(In millions of U.S. dollars; episodes examined in the paper are shaded)

Source: IMF’s International Financial Statistics database.


Design of the Short-Term Capital Inflow Controls

Although in all cases the controls were adoptedfor broadly similar reasons, the design of the mea-sures varied. All five countries used some form of

market-based controls (mainly in the form of director indirect taxation of inflows and other regulatorymeasures, such as asymmetric open position limitsand reporting requirements). In some cases, thesecontrols were supplemented by administrative or di-rect controls (Brazil, Chile, and Malaysia).

10

–5

–4

–3

–2

–1

0

1

979695949392911990–11

–10

–9

–8

–7

–6

–5

–4

–3

959493921991

–10

–8

–6

–4

–2

0

9795939189871985–9

–8

–7

–6

–5

–4

–3

–2

–1

9796959493921991

–8

–6

–4

–2

0

2

4

6

8

98979695949392911990

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 3. Countries with Controls on Short-Term Capital Inflows:Current Account Balance(In percent of GDP; episodes examined in the paper are shaded)



Brazil adopted an explicit tax on capital flows (the“entrance tax” on certain foreign exchange transac-tions and foreign loans),3 in combination with a

number of administrative controls (outright prohibi-tions of or minimum maturity requirements on cer-tain types of inflows). The coverage of the measureswas extended as the market adopted derivativesstrategies based on exempted inflows to circumventthe controls; and the tax rates were successivelyraised or differentiated by maturity to target short-

11

3This tax resembles the “Tobin tax,” which proposes a uniformlevy on all foreign exchange transactions to discourage short-term speculative position-taking in foreign currency.

50

60

70

80

90

100

110

120

130

94

96

98

100

102

104

106

108

90

100

110

120

130

140

150

160

70

80

90

100

110

120

90

100

110

120

130

140

150

160

170

1990

1985

1990 91 92 93 94 95 96 97 98

1991 92 93 94 95 96 9787 89 91 93 95 97

1991 92 93 94 9591 92 93 94 95 96 97

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 4. Countries with Controls on Short-Term Capital Inflows:Real Effective Exchange Rate(Index, 1990 = 100; episodes examined in the paper are shaded)

Source: IMF’s Information Notice System database. Based on relative CPIs. Increase means an appreciation.


term inflows. The regulations were also adjusted attimes of downward pressures on exchange rates, toreduce pressure on the capital account (for example,during the Mexican and Asian crises).

Chile combined market-based controls (indirecttaxation of inflows through an unremunerated re-serve requirement (URR)) with direct (minimum

stay requirement for direct and portfolio investment)and other regulatory measures (minimum rating re-quirement for domestic corporations borrowingabroad and extensive reporting requirements onbanks for all capital account transactions). The URRwas initially imposed on foreign loans (except fortrade credits), but subsequently rates were raised and

12

–0.2

0

0.2

0.4

0.6

0.8

1.0

1.2

2.4

2.5

2.6

2.7

2.8

2.9

20

25

30

35

40

45

50

0

100

200

300

400

500

600

200

400

600

800

1000

1200

1400

1600

1800

1990

1985

1990 91 92 93 94 95 96 97 98

1991 92 93 94 95 96 9787 89 91 93 95 97

199191 92 9293 9394 9495 9596 97

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 5. Countries with Controls on Short-Term Capital Inflows:Nominal Exchange Rate(In domestic currency units per U.S. dollar; episodes examined in the paper are shaded)



coverage extended to those inflows that became po-tential channels for short-term inflows, includingforeign direct investment of a potentially speculativenature. Similarly, in Colombia, the URR was im-posed on external borrowing with a maturity of lessthan 18 months (including certain trade credits), but

was later adjusted by imposing higher rates forshorter maturities, changing the deposit term, andextending the coverage of inflows subject to theURR. Malaysia adopted a combination of adminis-trative (prohibition of nonresident purchases ofmoney market securities and non-trade-related swap

13

–2000

0

2000

4000

6000

8000

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

0

10

20

30

40

50

2

3

4

5

6

7

8

15

20

25

30

35

1990

1985

1990 92 94 96 98

1991 92 93 94 95 96 9787 89 91 93 95 97

1991 92 93 94 9591 92 93 94 95 96 97

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 6. Countries with Controls on Short-Term Capital Inflows:Inflation(In percent, 12-month rate; episodes examined in the paper are shaded)



transactions with nonresidents) and regulatory mea-sures (asymmetric limits on banks’external liabilitypositions for nontrade purposes and reserve require-ments on ringgit funds of foreign banks). And Thai-land adopted a number of indirect, market-based

measures (asymmetric open position limits, detailedinformation requirements, and reserve requirementson nonresident bank accounts and baht borrowing,finance company promissory notes, and banks’off-shore short-term borrowing).

14

–5000

0

5000

10000

15000

20000

0

10

20

30

40

50

–50

0

50

100

150

5

10

15

20

25

30

35

–20

0

20

40

60

80

1990

1985

1990 92 94 96 98

1991 92 93 94 95 96 9787 89 91 93 95 97

1991 92 93 94 9591 92 93 94 95 96 97

Brazil

Chile

Malaysia

Thailand

Colombia

Reserve money Broad money

Figure 7. Countries with Controls on Short-Term Capital Inflows:Monetary Aggregates(In percent, 12-month percentage change; episodes examined in the paper are shaded)



Effectiveness and Costs of Controls on Short-Term Capital Inflows

The effectiveness of the controls in achievingtheir intended objectives was mixed.4 The principal

macroeconomic motivation for inflow controls wasto maintain a suitable wedge between domestic andforeign interest rates while reducing pressures on theexchange rate. Controls seem to have had some ef-fect initially, but in none of the five countries do theyappear to have achieved both objectives. Most coun-tries were able to maintain a large interest rate differ-ential, but some had to adjust their exchange ratesgradually under sustained upward market pressures(Brazil, Chile, and Colombia). Real exchange rates

15

4This assessment is complicated by large differences in the ex-tent of previous research. The Chilean experience with capitalcontrols has by far received the greatest attention in the economicliterature; Part II, Chapter V, and Appendix I review these studies.

–30

–20

–10

0

10

20

30

40

–1

0

1

2

3

4

5

6

–100

0

100

200

300

400

–5

0

5

10

15

20

–150

–100

–50

0

50

100

–20

0

20

40

60

80

–60

–40

–20

0

20

40

60

80

0

10

20

30

40

50

–20000

0

20000

40000

60000

80000

100000

–20000

2000400060008000

100001200014000

Brazil1

Chile

Malaysia

Thailand

Colombia

1990 1991 92 93 94 95

1985 1991 93 95 97

1990 92 94 96 98

87 89 91 93 95 97

92 94 96

Depreciation-adjusted (left scale) Simple (right scale)

Figure 8. Countries with Controls on Short-Term Capital Inflows:Short-Term Interest Rate Differentials(In percent; episodes examined in the paper are shaded)

Sources: Various, including IMF’s International Financial Statistics database and country authorities.The charts showthe simple and depreciation-adjusted short-term interest rate differentials with the United States. Owing to dataavailability, the type of domestic currency–denominated financial instrument varies by country.

1Q1–Q2 1994 observations removed to limit chart scale.


appreciated significantly in all five countries (to alesser extent in Thailand and Malaysia), with moreor less deterioration in the external current balances.The controls did not seem to be effective in reducingthe total level ofnet inflows(except in Malaysia andThailand), but seemed to be at least partly successful

in reducing short-term capital inflows. Sterilizationoperationsalso had to continue in some countries toabsorb the continuing inflows, with their associatedcosts to the central bank (Brazil and Chile). In sum,there is some evidence that the inflow controls werepartly effective (1) in Malaysia and Thailand, in re-

16

–20000

0

20000

40000

60000

80000

0

1000

2000

3000

4000

5000

6000

7000

0

200

400

600

800

1000

1200

1400

1600

400

600

800

1000

1200

1400

200

400

600

800

1000

1200

1400

1600

1800

1990

1985

1990 92 94 96 98

1991 92 93 94 95 96 9787 89 91 93 95 97

1991 92 93 94 9592 93 94 95 96 9791

Brazil

Chile

Malaysia

Thailand

Colombia

Figure 9. Countries with Controls on Short-Term Capital Inflows:Local Stock Exchange Index(Episodes examined in the paper are shaded)

Source: International Finance Corporation’s Emerging Markets database.


ducing the level and affecting the maturity of the in-flows while curtailing sterilization operations, and(2) in Colombia and possibly in Chile, in maintain-ing a wedge between domestic and foreign interestrates and affecting somewhat the composition of theinflows.5 The controls maintained by Brazil appearto have been largely ineffective in achieving theirstated objectives (as detailed in Part II, Chapter V).

A number of factors may have played a role in theeffectiveness (or lack thereof) of the controls in real-izing their intended objectives, though it is not possi-ble to be certain. In Brazil, well-developed and so-phisticated financial markets (with active trading ofcurrency futures and other derivatives) seem to havereduced the cost of circumventing the continuouslywidening coverage of the regulations. Incentives todo so were strong owing to large interest rate differ-entials and expectations of a stable exchange rate.Similarly, in Chile, the dynamic response of opti-mizing agents in a sophisticated financial systemseems to have reduced the effectiveness of the initialset of regulations and facilitated the exploitation ofloopholes in the system. The Chilean authorities inturn were obliged to continuously extend the cover-age of the regulations to the extent permitted bylegal and political considerations. While strong en-forcement capacity through a comprehensive infor-mation and disclosure system between the centralbank of Chile and the commercial banks may havebeen instrumental in identifying the loopholes, theexemption of trade credits from the controls and po-litical constraints on closing all potential loopholesseem to have weakened the effectiveness of the con-trols over time. In Colombia, subjecting certain tradecredits to the URR may have eliminated a significantchannel for circumvention, but the shift from debtcreating inflows to other financing sources (e.g., for-eign direct investment) opened another potentialchannel for circumvention.

Factors other than controls may also have played arole in reducing the volume of inflows or changingtheir maturity composition in some cases. These in-cluded (1) adjustments in monetary policy to narrowinterest rate differentials and curtail sterilization op-erations (Malaysia); (2) a somewhat more flexibleexchange rate arrangement to discourage speculativeinflows (Chile and Colombia); (3) further strength-ening of prudential regulations and supervision(Chile, Colombia, and Malaysia); and (4) a deterio-ration in investor confidence (Thailand). In addition,potential data problems—as well as an increase in

short-term inflows channeled through exempted in-flows and thus not recorded as such (trade credit inChile and foreign direct investment flows in the caseof Colombia)—may potentially hide the magnitudeof short-term inflows and give a misleading picturein terms of the effectiveness of the controls.

Conclusions

The foregoing suggests the following tentativeconclusions. First, to be effective, the coverage ofthe controls needs to be comprehensive, and thecontrols need to be forcefully implemented. Consid-erable administrative costs are incurred in continu-ously extending, amending, and monitoring compli-ance with the regulations. Even then, controls maylose effectiveness over time as markets exploit thepotential loopholes in the system to channel the “un-desired” inflows through the exempted ones. The ef-fectiveness of the controls seems to be limited by so-phisticated financial markets, which reduce the costof circumvention relative to the incentives.6 Second,although capital controls appeared to be effective insome countries, it is difficult to be certain of theirrole given the problems involved in disentanglingthe impact of the controls from that of the accompa-nying policies, which included the strengthening ofprudential regulations, greater exchange rate flexi-bility, and adjustment in monetary policies. Third,inflow controls may not be ideally suited as instru-ments of prudential policy, as they are often imposedand modified for macroeconomic rather than micro-economic reasons, for example at times of down-ward pressure on exchange rates (Brazil during theMexican and Asian crises and Chile and Colombiaduring the Asian crisis).

The experience of a number of countries (e.g.,Brazil and Thailand) also suggests that the use ofcontrols on inflows may not provide lasting protec-tion against reversals in capital flows if they are notaccompanied by necessary adjustments in macro-economic policies and strengthening of the financialsystem. In these cases (as well as in Colombia), re-sorting to capital controls may actually have delayedthe necessary policy adjustments, making the even-tual adjustment more severe. Moreover, in countries

17

5National data support this conclusion. However, the evidenceis mixed in the case of Chile, where more detailed examinationsof the data have cast some doubt on the proposition that the con-trols affected the composition of flows. (See also Appendix I.)

6The experience here closely parallels earlier episodes in in-dustrialized countries under an adjustable peg regime. For exam-ple, Germany during 1968–73 attempted to resist episodes ofstrong capital inflows by measures including minimum reserverequirements on the growth of liabilities to nonresidents. Thesemeasures contributed to disintermediation from the banking sys-tem, and obliged the Bundesbank to introduce an ever-broadeningrange of indirect and quantitative controls. Nonetheless, the con-trols were largely ineffective in preventing short-term capital in-flows and ultimately the appreciation of the currency.


with weak prudential and supervisory frameworks,banking systems took on excessive risks despite thecontrols (Thailand).

Capital Outflow Controls DuringFinancial Crises

This section examines the experiences ofMalaysia (1998–present), Spain (1992), and Thai-land (1997–98) with the use and effectiveness of se-lective controls on capital outflows, with a focus onthe role that the controls may have played in copingwith crisis situations. Figures 10–18 illustrate devel-opments in key economic indicators during theseepisodes, and Part II, Chapter VI, provides furtherdetails of the country experiences.

Motivations for Imposing Capital OutflowControls During Financial Crises

The desire of the authorities to limit downwardpressure on their currencies has been one of the mostfrequent motives in imposing controls on capitaloutflows. Earlier reviews of country experiences in-dicated that such restrictions have mainly been ap-plied to short-term capital transactions to countervolatile speculative flows that threatened to under-mine the stability of the exchange rate and depleteforeign exchange reserves. These restrictions havealso served at times as an alternative to the promptadjustment of economic policies and thus helped theauthorities “buy time.” They have also been em-ployed to insulate the real economy from volatilityin the international financial markets (see Bakker,1996, p. 20).

All three countries reimposed controls on capitaloutflows in the context of significant downwardpressure on the exchange rate: Spain during the Eu-ropean currency turmoil of the fall of 1992, andMalaysia and Thailand in the context of the Asian fi-nancial crisis of 1997–99. Spain was a member ofthe European Monetary System’s ERM (exchangerate mechanism), where decisions on exchange raterealignments were subject to agreement with theother members of the system; Thailand was main-taining a pegged exchange rate regime when thecontrols were imposed; and Malaysia had been fol-lowing a managed float, before fixing the ringgit vis-à-vis the U.S. dollar along with the imposition of thecontrols in September 1998. In all three countries,the controls aimed at containing speculation againstthe currencies and stabilizing the foreign exchangemarkets against a backdrop of sharply declining offi -cial foreign exchange reserves. Room to use interestrates in defense of the exchange rate was limited inall three countries—in Spain, by market concerns

about adverse macroeconomic fundamentals, includ-ing a large fiscal burden, and in Malaysia and Thai-land, by concerns about the adverse impact of highinterest rates on fragile domestic economies andbanking systems. In Spain, the peseta had been de-valued by 5 percent before the imposition of the con-trols, but market pressures had not subsided. A fur-ther realignment of the exchange rate appearednecessary but could not be carried out immediatelygiven high tensions within the ERM, which alsoruled out interest rate increases, while the authori-ties’ strong commitment to European MonetaryUnion (EMU) precluded an exit from the ERM.

In all three countries, the controls were imposedin an environment where capital account transac-tions had already been largely liberalized. Spain’scapital account had been completely liberalizedseven months before the reintroduction of the capitalcontrols, while those of the other two countries hadbeen fairly open (mainly on the inflow side in Thai-land). Malaysia had liberalized most portfolio out-flows, except for corporations with domestic bor-rowing, and had adopted a liberal approach toportfolio inflows. Malaysia had also liberalizedcross-border transactions in ringgit, including fortrade-related transactions, and financial transactionswith nonresidents; offshore trading of ringgit securi-ties was tolerated. In Thailand, nonresidents werefree to obtain baht credit from domestic banks andoperate in well-developed spot and forward markets.As a result, an active offshore market had developedfor both the ringgit and the baht.

Design of Capital Outflow Controls DuringFinancial Crises

While the design of the controls imposed by thethree countries varied significantly, in all cases theymainly targeted the activities of nonresidents (identi-fied as “speculators”), by restricting their access todomestic currency funds that could be used to takespeculative positions against the domestic curren-cies. The controls explicitly exempted current inter-national transactions, foreign direct investmentflows, and certain portfolio investments. In Spain,the controls took the form of a compulsory, non-interest-bearing 100 percent deposit requirement ondomestic banks, to discourage speculation by mak-ing it costly for banks to engage in certain transac-tions with nonresidents. The requirement initiallyapplied to increases in banks’long foreign currencypositions, peseta-denominated deposits and loans tononresidents, and peseta-denominated liabilities ofdomestic banks with their branches and subsidiaries.These requirements were subsequently limited to asingle deposit requirement on increases in banks’swap transactions with nonresidents (seen as the

18

Capital Outflow Controls During Financial Crises

most likely avenue for speculation). In Thailand, atwo-tier currency market was created, with the goalof segmenting the onshore market from its offshorecounterpart through a mix of direct and market-based measures. In particular, the Thai banks wererequired to suspend all transactions with nonresi-

dents that could facilitate a buildup of baht positionsin the offshore market (involving spot and forwardsales, and lending via swaps); the repatriation of pro-ceeds from asset sales in baht were prohibited andtheir conversion had to be on the basis of onshoreexchange rates.

19

–6

–4

–2

0

2

4

6

8

989796951994–6

–4

–2

0

2

4

6

98979695949392911990

–15

–10

–5

0

5

10

9493929190891988–6

–4

–2

0

2

4

6

98979695949392911990

–20

–15

–10

–5

0

5

10

15

989796959493921991–12

–10

–8

–6

–4

–2

0

2

4

98979695949392911990

Malaysia Romania

Spain Russian Federation

Thailand Venezuela

September 1998 measures

Figure 10. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Net Private Capital Flows(In percent of GDP; episodes examined in the paper are shaded)



In Malaysia, the controls were more wide-rang-ing and combined capital controls with exchangecontrols, but without restricting payments andtransfers for current international transactions andforeign direct investment. After an initial (and ineffect unsuccessful) attempt in August 1997 to iso-

late the domestic market from the offshoremarket,7 a number of direct controls were adopted

20

7The control took the form of swap limits on banks’non-trade-related offer-side swap transactions with nonresidents.

15000

20000

25000

30000

35000

40000

0

1000

2000

3000

4000

20000

30000

40000

50000

60000

70000

80000

0

5000

10000

15000

20000

25000

10000

15000

20000

25000

30000

35000

40000

2000

4000

6000

8000

10000

12000

14000

16000

1994

1988

1991 1990 92 94 96 989392 9594 9796 9998

1994 95 96 97 98 9989 90 91 92 93 94

1990 92 94 96 9895 96 97 98 99

Malaysia Romania


Thailand Venezuela

September 1998measures

February 1999measures

Figure 11. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Foreign Exchange Reserves(In millions of U.S. dollars; episodes examined in the paper are shaded)



to stabilize the onshore ringgit market by eliminat-ing its offshore counterpart, where speculativepressures on the ringgit had been putting pressureon domestic interest rates. Practically all legalchannels for a possible buildup of ringgit funds off-shore were eliminated. Offshore ringgit were re-

quired to return onshore, limits were imposed onimports and exports of ringgit currency, the use ofringgit in trade payments and offshore trading ofringgit assets were prohibited, and transfers be-tween external accounts of nonresidents and ringgitcredit facilities between residents and nonresidents

21

–15

–10

–5

0

5

10

15

989796951994–9

–8

–7

–6

–5

–4

–3

–2

–1

98979695949392911990

–4.0

–3.5

–3.0

–2.5

–2.0

–1.5

–1.0

–0.5

9493929190891988–2

–1

0

1

2

3

4

98979695949392911990

–10

–5

0

5

10

15

989796959493921991–10

–5

0

5

10

15

20

98979695949392911990

Malaysia

Romania

Spain

Russian Federation

Thailand Venezuela


Figure 12. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Current Account Balance(In percent of GDP; episodes examined in the paper are shaded)



were prohibited. To contain the outflows, transfersof capital by residents were also limited, and repa-triation of nonresident portfolio capital wasblocked for 12 months. In February 1999, the lattermeasure was replaced with exit levies on the repa-

triation of portfolio capital that decline with theholding period of the investment. The controlswere supported by additional measures to eliminatepotential loopholes, including an amendment of theCompany Act to limit distribution of dividends

22

70

80

90

100

110

120

130

20

40

60

80

100

120

140

160

80

85

90

95

100

105

60

70

80

90

100

110

120

130

140

60

70

80

90

100

110

120

80

100

120

140

160

180

200

220

1994

1988 1994 95 96 97 98 99

1991 1990 92 94 96 9892 93 94 95 96 97 98 99

89 90 91 92 93 94

1990 92 94 96 9895 96 97 98 99

Malaysia Romania


Thailand Venezuela


February 1999 measures

Figure 13. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Real Effective Exchange Rate(Index, 1990 = 100; episodes examined in the paper are shaded)

Source: IMF’s Information Notice System database.


while still complying with Malaysia’s obligationsunder Article VIII of the IMF’s Articles of Agree-ment, and the demonetization of large denomina-tions of ringgit notes to limit the outflow of ringgitfunds.

Effectiveness and Costs of Controls on CapitalOutflows During Financial Crises

The effectiveness of the controls in realizing theirintended objectives was mixed. In Malaysia, elimi-

23

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

–5000

0

5000

10000

15000

60

65

70

75

80

85

90

–5

0

5

10

15

20

25

20

25

30

35

40

45

50

55

0

100

200

300

400

500

600

700

Malaysia Romania


Thailand Venezuela

1994

1988

1991 1990 92 94 96 9892 93 94 95 96 97 98 99

89 90 91 92 93 94 1994 95 96 97 98 99

95 96 97 98 99 1990 92 94 96 98


February 1999measures

Exchange rate band

Figure 14. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Nominal Exchange Rate(In domestic currency units per U.S. dollar; episodes examined in the paper are shaded)

Source: IMF’s International Financial Statistics database. For Spain, the exchange rate is computed with respect tothe deutsche mark.


nation of most potential sources of access to ringgitby nonresidents effectively eliminated the offshoreringgit market, and, together with the restrictions onnonresidents’repatriation of portfolio capital and onresidents’outward investments, contributed to the

containment of capital outflows. In conjunction withother macroeconomic and financial policies, thecontrols helped to stabilize the exchange rate. Sincethe introduction of the controls, there have been nosigns of speculative pressures on the exchange rate,

24

1

2

3

4

5

6

7

8

0

50

100

150

200

250

300

350

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

–200

0

200

400

600

800

1000

1200

–2

0

2

4

6

8

10

12

0

20

40

60

80

100

120

Malaysia Romania


Thailand Venezuela

1994

1988

1991 92 93 94 95 96 97 98 1990 92 94 96 9899

89 90 91 92 93 94 1994 95 96 97 98 99

95 96 97 98 99 1991 92 93 94 95 96 97 98 99



Figure 15. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Inflation(In percent, 12-month rate; episodes examined in the paper are shaded)



despite the marked relaxation of fiscal and monetarypolicies to support weak economic activity. Norhave there been signs that a parallel or nondeliver-able forward market is emerging; and no significantcircumvention efforts have been reported. In Spain,

initially the large deviation of onshore from offshoreinterest rates and the stabilization of the pesetawithin its ERM bands suggested that the controlshad succeeded in curtailing access to peseta funds byspeculators, in segregating the onshore and offshore

25

–50

–30

–10

10

30

50

70

0

50

100

150

200

–40

–30

–20

–10

0

10

20

30

–50

0

50

100

150

200

250

–20

–10

0

10

20

30

40

–50

0

50

100

150

200

Malaysia Romania


Thailand Venezuela

1994

1988 1994 95 96 97 98 99

1991 92 93 94 95 96 97 98 99 1990 92 94 96 98

89 90 91 92 93 94

1990 92 94 96 9895 96 97 98 99

September 1998 measures


Reserve moneyBroad money Reserve money

Broad money

Reserve moneyBroad money


Narrow moneyBroad money


Figure 16. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Monetary Aggregates(In percent, 12-month percentage change; episodes examined in the paper are shaded)



markets, and thus in limiting speculation against thepeseta. However, once the scope of controls was re-duced and clarified, the differentials narrowed. Thepeseta again came under pressure in November,which recurred as weekends approached until thepeseta was devalued in a negotiated realignment ofthe ERM in late November 1992, after which thecontrols were lifted and the authorities moved toraise interest rates. In Thailand, large differentialsinitially emerged between offshore and onshore in-

terest rates, trading in the swap market virtuallystopped, and speculative attacks temporarily ceased.However, the controls soon began to develop leaks,pressure on the baht resumed, and within twomonths after the controls were imposed, the authori-ties floated the baht. Most of the controls were abol-ished or substantially modified at the end of January1998; the prohibition of banks’noncommercialtransactions with nonresidents was replaced withlimits; and the two-tier market was unified.

26

–200

0

200

400

600

800

0

50

100

150

200

250

300

350

–200

0

200

400

600

800

1000

1200

–100

0

100

200

300

400

–60

–40

–20

0

20

40

60

2

3

4

5

6

7

8

9

–100

0

100

200

300

400

–5

0

5

10

15

20

–100

–50

0

50

100

150

–2

0

2

4

6

8

–100–80–60–40–20

0204060

0

10

20

30

40

50

60

Malaysia Romania


Thailand Venezuela

1994

1988

1991 1990 92 94 96 98 92 93 94 95 96 97 98 99

1994 95 96 97 98 9989 90 91 92 93 94

1990 92 94 96 9895 96 97 98 99

Depreciation-adjusted (left scale) Simple (right scale)



Figure 17. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Short-Term Interest Rate Differentials(In percent; episodes examined in the paper are shaded)

Sources: Various, including IMF’s International Financial Statistics database and country authorities.The charts showthe simple and depreciation-adjusted short-term interest rate differentials with the reference country (Germany forSpain and United States for all others). Due to data availability, the type of domestic currency–denominated financialinstrument varies by country.


A number of factors may have played a role in therelative effectiveness of the measures. In Spain, thewide-ranging and restrictive measures as first intro-duced effectively curbed not only speculative activi-ties, but a much broader range of transactions, in-

cluding financial operations associated with thehedging of exchange rate risk by nonresident im-porters and exporters. Initial uncertainty about theprecise scope of the measures may also have damp-ened activity in the market. The authorities subse-

27

200

400

600

800

1000

1200

1400

1600

1800

200

400

600

800

1000

1200

1400

1600

150

200

250

300

350

400

0

500

1000

1500

2000

2500

0

500

1000

1500

2000

–5000

0

5000

10000

15000

20000

25000

30000

35000

1994

1988

1991 1990 92 94 96 9892 93 94 95 96 97 98 99

1994 95 96 97 98 9989 90 91 92 93 94

1990 92 94 96 9895 96 97 98 99

Malaysia Romania


Thailand Venezuela



Figure 18. Countries with Selective Controls on Outflows (left column)and with Extensive Controls (right column):Local Stock Exchange Index(Episodes examined in the paper are shaded)

Sources: International Finance Corporation’s Emerging Markets database, Reuters, and country authorities.


quently clarified the regulation and narrowed thecoverage of the deposit requirement to swap opera-tions, which were identified as the preferred methodof speculative financing. Market participants tookadvantage of additional loopholes, given the expec-tations of further peseta depreciation. Similarly, thecontrols seem to have been initially effective inThailand, reflecting the absence of extensive salesby domestic holders of baht assets and the strict ap-plication of the controls by the central bank andcommercial banks. The effectiveness of the mea-sures was eventually undermined by the persistentlylarge return differentials in the still active offshoremarket and expectations of baht depreciation. Thecontrols may have delayed the implementation of acomprehensive structural reform and stabilizationpackage, and thus worsened the crisis.

In Malaysia, the wide-ranging nature of the mea-sures and their strict and effective enforcement bythe authorities and the commercial banks seem tohave been instrumental in effectively eliminating theoffshore ringgit market and thus in contributing tothe containment of the speculative pressures. Therelatively favorable economic fundamentals ofMalaysia at the outset, the authorities’efforts to dis-seminate information to increase the transparency ofthe controls, and their efforts to accelerate thestrengthening of the financial sector also seem tohave played an important role in improving the ac-ceptability of the measures both domestically and in-ternationally. The general return of confidence in theregion, the sharp improvement in Malaysia’s exter-nal balance, and the ex post undervaluation of theringgit following its peg to the U.S. dollar (whileother currencies in the region started to appreciate)also seem to have reduced the incentives for circum-vention. It is too early to judge at this stage whetherthe controls will have long-run adverse effects on in-vestor sentiment.

Conclusions

In short, the reimposition of controls on capitaloutflows during episodes of financial crisis seems tohave provided only a temporary respite of varyingduration to the authorities.The controls gave theMalaysian authorities some breathing space to ad-dress the macroeconomic imbalances and implementbanking system reforms. In Spain, the measures didnot avoid a second realignment of the peseta, thoughthey may have provided some additional time in ne-gotiating the realignment within the ERM. The ex-periences of the three countries suggest that (1) to beeffective, the controls must be comprehensive,strongly enforced, and accompanied by necessaryreforms and policy adjustments; (2) controls do notprovide lasting protection in the face of sufficient in-

centives for circumvention, in particular attractivereturn differentials in the offshore markets andstrong market expectations of exchange rate depreci-ation; (3) the ability to control offshore market activ-ity may have been instrumental in containing out-flows and stemming speculative pressures; and (4)effective measures risk discouraging legitimatetransactions, including foreign direct investment(Malaysia) and trade-related hedging transactions(Spain), and may raise the cost of accessing interna-tional capital markets (as indicated by the rise inMalaysia’s relative risk premium following the con-trols). The effectiveness of Malaysia’s controls wasprobably further enhanced by the strengthening ofcontrols over residents’outward investment.

Extensive Exchange Controls DuringFinancial Crises

This section draws on the experiences of threecountries, Romania (1996–97), Russia (1998–pre-sent), and Venezuela (1994–96), that resorted to ex-tensive systems of controls on both current and capi-tal transactions in connection with crises. Thesecrises entailed severe downward pressure on ex-change rates and a sharp reduction in foreign ex-change reserves owing to extensive official foreignexchange intervention. Part II, Chapter VII, providesfurther details of the country experiences.

Motivations and Design of Extensive ExchangeControls During Financial Crises

All three countries resorted to extensive exchangecontrols to stabilize their foreign exchange markets.The controls involved administrative measures toclose or significantly restrict access to the foreign ex-change market for both current international pay-ments and transfers, and capital movements. Roma-nia and Russia had been maintaining fairly restrictivecapital account regimes prior to the imposition of thecontrols. By contrast, Venezuela’s capital accountwas highly liberalized; and the imposition of con-trols was thus a sharp reversal of policy. Both Russiaand Venezuela had achieved current account convert-ibility prior to the episodes under consideration.

In Romania, the measures took the form of a sus-pension of the foreign exchange dealer licenses ofall but four state-controlled banks, and limits on theovernight cash position of foreign exchange bureaus,which severely constrained the operation of the for-eign exchange market. In Venezuela, the measuresconsisted of restrictions on the availability of foreignexchange for import and export payments, and forinvisible transactions; surrender requirements on ex-port receipts and certain capital inflows; and a prohi-

28

Long-Standing and Extensive Controls and Their Liberalization

bition on the repatriation of nonresident investmentsand all other capital transactions, except for the re-payment of external debt.

In Russia, the measures were even more exten-sive, and combined a reintensification of capital con-trols (tightening of existing restrictions and reimpo-sition of inflow controls) with restrictions on currentinternational transactions (temporary closing of theinterbank foreign exchange market and creation of adual market), and debt default. The combination ofcapital controls with debt default may have been mo-tivated by concerns about a massive capital outflow,the accompanying depreciation of the ruble, and afragile banking system with large unhedged foreignexchange positions. The authorities abandoned theexchange rate band as downward pressures persistedafter the adoption of the measures. By contrast, inVenezuela and Romania, the controls were accompa-nied by a temporary fixing of the exchange rate.

Effectiveness and Costs of Extensive ExchangeControls During Financial Crises

It is not fully clear whether the controls achievedtheir intended objectives. In Romania andVenezuela, the controls seem to have containedsome of the initial pressures in the foreign exchangemarket, allowing the authorities to maintain rela-tively stable exchange rates for some time. Access toforeign exchange was severely constrained and par-allel foreign exchange markets emerged. These mar-kets were characterized by substantial premiumsover the official rate, reflecting continued macroeco-nomic imbalances and problems in the financial sys-tem. In Russia, the full impact of the control mea-sures remains to be seen, as the economic situationhas not yet been durably stabilized. Despite the per-vasiveness of the measures, foreign exchange mar-ket pressures did not subside until the first quarter of1999, reserves continued to decline, and outflows in-creased sharply. This deterioration took placeagainst the background of continued fiscal problemsand a further weakening of the banking system. Thesharp depreciation of the ruble contributed to a full-scale financial crisis, as banks’large unhedged for-eign currency positions, which had been accumu-lated under the tightly managed exchange rateregime, caused large foreign exchange losses. Innone of the three countries did the measures fullysucceed in stemming capital outflows.

It seems, however, that in Venezuela the controlsincreased the degree of monetary policy autonomyin the context of a fixed exchange rate system. Thelower interest rates associated with the controls mayhave enabled the government to reduce the immedi-ate cost of the banking crisis and improve its fiscalbalance, possibly at the expense of higher external

debt service and more limited access to internationalfinancial markets. In this regard, it is noteworthythat Venezuela’s share in total foreign direct invest-ment in Latin America fell while the controls were ineffect, and increased when they were lifted. How-ever, these developments may also reflect foreign in-vestors’views on the banking system and the politi-cal situation. Similarly, the interest rate differentialon Venezuela’s Brady bonds fell sharply followingthe lifting of the controls. Difficulties in accessingforeign capital were also observed in Romania, andforeign direct investment declined relative to othertransition economies. More severe problems of thistype were observed in Russia, where access to inter-national capital markets halted, foreign direct invest-ment inflows fell sharply, and the yield differentialon Russian securities rose significantly, though thesedevelopments may have largely reflected the chillingeffect of debt default.

Conclusions

Extensive controls on capital and current interna-tional transactions may temporarily relieve pressureson the balance of payments, but they do not providelasting protection when the fundamental causes ofthe imbalances remain unaddressed. As with themore targeted controls discussed in the section oncapital outflow controls during crises, the controls inRomania, Russia, and Venezuela may have reducedaccess to foreign capital. Difficulties of this sort seemto have motivated the authorities in Romania andVenezuela to remove the restrictions (with Venezuelaopting for a big bang approach), and to addressmacroeconomic and financial sector imbalances.

Long-Standing and Extensive Controlsand Their Liberalization

While China and India were not been immune tothe Asian crisis of 1997–98, they were less affectedby it than other countries in the region. The rela-tively closed capital account regimes of these twocountries have been credited with helping to limitvulnerability to financial contagion, though otherfactors may have played a role as well, includingmost notably large and relatively closed economiesand strong foreign exchange reserves positions.While the Asian countries most affected by the crisissuffered severe recessions and major banking prob-lems, both India and China experienced only a minorslowdown in their strong growth, and the impact ofthe crisis on their financial systems was limited.China was able to maintain the de facto peg of itscurrency to the U.S. dollar. India continued to followa flexible exchange rate policy, which appears to

29


have further reduced the impact of the crisis. Part II,Chapter VIII, provides further details of these coun-try experiences.

Initial Circumstances

Given the relatively early stages of financial mar-ket development and some structural shortcomingsin their financial systems, both countries have fol-lowed a gradual and cautious approach to liberaliz-ing their capital accounts. The restrictive capital ac-count regimes in India and China have historicallybeen just one facet of a generally closed and heavilystate-controlled economic system. Nevertheless,economic liberalization, including external liberal-ization, has become an important medium-term goalin both countries. India in particular has made largestrides toward reversing several decades of statedomination of the economy, though the financialsector is still largely publicly owned and directedlending remains extensive. Capital controls havebeen quantity based, rather than price based, andhave been administratively enforced. In both coun-tries, the capital control regimes in place during the1990s encouraged longer-term flows (in particular,foreign direct investment) over short-term ones, withthe controls oriented toward limiting reliance onshort-term and debt-creating flows. As a result, capi-tal controls in both countries have shifted the com-position of measured capital inflows toward longer-term flows and more creditworthy borrowers, partlyby curtailing access of noncreditworthy domesticborrowers to foreign financing.

Effectiveness and Costs of Controls

While the capital controls in both China and Indiaare believed to have been effective in limiting mea-sured capital flows, there also seems to be some evi-dence of evasion and avoidance, for example,through the misinvoicing of trade transactions orlarge errors and omissions in the balance of pay-ments statistics. In both countries, the extensive re-strictions gave rise to significant administrativecosts, burdened legitimate transactions, and mayhave reduced the efficiency of resource allocation.

While the effects on India of the Asian crisis in1997–98 have not been severe, the country hasnonetheless proven vulnerable to external shocksduring periods of large domestic imbalances (in1980 and 1990–91), though this vulnerability mayhave been lessened by the reorientation of capitalcontrols since 1991 to discourage volatile foreignfinancing. In China, the authorities have noted thatillegitimate current account transactions had facili-tated substantial capital flight (about $11 billionduring the first half of 1998), reflecting the outbreak

of the Asian crisis.Concerns about large outflows—driven in part by fears of an imminent devaluationof the renminbi, the falling interest rate differential,and increased evasion—prompted the authorities tointensify the enforcement of the existing controlsduring the second half of 1998.Administrativescreening of capital account transactions was en-hanced and documentation and verification require-ments for current international transactions weretightened. The authorities considered these mea-sures to be necessary in view of their commitmentto a stable exchange rate. In mid-1999, the authori-ties restricted overseas yuan transactions by pro-hibiting domestic banks from accepting inward re-mittances in domestic currency. The measure mayhave helped to prevent the illegal movement ofyuan out of China, and to clamp down on offshoretrading of the yuan by Chinese financial institu-tions. These measures were accompanied by otherinitiatives to facilitate the efficient operation of ex-change controls, and in particular reduce the delaysin approving legitimate transactions. The regulatoryframework was made more transparent, and newtechnology was adopted to facilitate screening andenforcement.

Conclusions

The experiences of China and India seem to sug-gest that the long-standing and extensive controls oncapital transactions may have had some role in re-ducing the vulnerability of these countries to the ef-fects of the recent regional crisis. In particular, theyhelped shift the composition of capital inflows to-ward longer-term flows. However, other factors mayhave played a role as well in reducing their financialvulnerability. These include, for both countries, astrong external position with ample foreign ex-change reserves; larger sizes of the domestic mar-kets; relatively weak trade and financial linkageswith the rest of the world compared with the othercountries in the region; relatively earlier stages of fi-nancial market development, with a lower level offinancial intermediation by the banking systems; anda flexible exchange rate policy in the case of India.In both India and China, enforcement of the controlswas facilitated by strong administrative capacity.

Rapid Liberalization of Capital Controls

This section draws on the experiences with capitalaccount liberalization of Argentina (1991), Kenya(1991–95), and Peru (1990–91)—all of which im-plemented a relatively rapid liberalization of thecapital account. Argentina and Peru liberalized all

30

Rapid Liberalization of Capital Controls

capital transactions using a big-bang approach, andin Kenya the liberalization was also relatively rapidbut spread over five years. Part II, Chapter IX, pro-vides further details on these country experiences.

Earlier reviews of country experiences with thecapital account liberalization have suggested that or-derly liberalization required not only a proper se-quencing and pace of changes in capital account reg-ulations, but also strong and consistent supportingpolicies. The speed and sequencing of capital ac-count liberalization have generally reflected a coun-try’s initial conditions and its broader economic de-velopment and restructuring. As a consequence,countries have followed diverse approaches. Big-bang approaches have usually been part of programsintended to signal a strong commitment to reform.

Initial Circumstances and Motivations forRapid Liberalization of Capital Controls

In all three countries, the liberalization of the capi-tal account was preceded by a period of severe imbal-ances in the domestic economy. In the case of Ar-gentina, liberalization took place following a periodof hyperinflation, an almost complete loss of policycredibility, and a collapse in demand for money andbanking services. These developments prompted theauthorities to adopt the Convertibility Plan in 1991,which involved the establishment of a currencyboard, and the elimination of all restrictions on cur-rent and capital account transactions. Similarly, inPeru, liberalization followed a period of hyperinfla-tion, depletion of foreign exchange reserves, and asharp decline in output and investment in the late1980s. In Kenya, liberalization followed a period oflarge fiscal deficits, a deteriorating balance of pay-ments, severe shortages of foreign exchange reserves,high inflation, and a slowdown in economic growth.

Capital account liberalization was intended to sig-nal a strong precommitment to reform and was moti-vated by a desire to create conditions that would at-tract foreign financing and achieve sustainedgrowth. In all three countries, the liberalization ofthe capital account was just one part of a wide-rang-ing liberalization program that included the deregu-lation of the financial system (in particular, of inter-est rate and credit controls), trade liberalization, andprivatization of public enterprises. In Argentina andPeru, attempts have also been made to strengthen thesupervisory and regulatory frameworks for the fi-nancial system, maintain tight monetary and fiscalpolicies to make further progress in reducing infla-tion, and enhance labor market flexibility.

Effects of Rapid Liberalization ofCapital Controls

Liberalization of the capital account was followedby an increase in foreign investment in Argentinaand Peru, but only to a lesser extent in Kenya, wheresome initial pickup in capital inflows was reversedsharply from 1992. In Argentina, foreign direct in-vestment and portfolio inflows reached 11 percent ofGDPin 1993, compared with less than 1 percent in1990. Subsequent capital outflows related to theMexican crisis of 1994–95 were managed withoutresort to capital controls, with the authorities insteadopting to tighten fiscal policy and provide some li-quidity assistance to the banking system within theconfines of the currency board. Further measureswere also taken to strengthen the banking systemand lengthen the maturity structure of the publicdebt.

In Peru, significant capital inflows were associ-ated with an appreciation of the exchange rate andsome deterioration in the current account. Currentaccount deficits were financed partly by an increas-ing share of short-term inflows in total inflows. Theincreased reliance on short-term financial credit bybanks made the financial system somewhat vulnera-ble, as evidenced by a weakening in the financialcondition of several institutions. However, the au-thorities’efforts to strengthen prudential regulationshelped increase the resilience of the banking system.In the period following liberalization, growth re-sumed and inflation was reduced sharply.

The liberalization of the capital account failed toprevent a sharp economic downturn in Kenya, withthe onset of an economic crisis, a significant rise inmoney supply and inflation, the emergence of exter-nal payment arrears, and a sharp depreciation of thecurrency. The crisis took place against the back-ground of inconsistent economic policies ahead ofthe election period in the early 1990s, with gover-nance problems in the financial system, weaknessesin prudential supervision, and delays in structuralreform.

Conclusions

The experiences with rapid liberalization of thecapital account highlight the importance of soundmacroeconomic policies combined with ongoing ef-forts to strengthen the financial system and imple-ment associated reforms. In the absence of adequatemacroeconomic and financial policies, capital ac-count liberalization may increase vulnerability to ex-ternal and domestic shocks.

31

One major objective of capital controls is tomanage the various risks associated with

capital flows.8 Capital control measures focus onspecific types of transactions and attempt to manageand reduce risk by influencing the composition ofparties to, and the volume of, such transactions.Chapter II examined a number of measures adoptedby various countries, ranging from quantitativerestrictions to a Tobin tax–like mechanism. Thepredominant motivations for the use of capital con-trols were macroeconomic, mainly to facilitate thepursuit of both monetary policy and exchange rateobjectives.

An alternative approach to managing the risks as-sociated with capital flows is not to attempt to con-trol the flows directly, but to limit the vulnerabilityof the economy to the risks associated with suchflows. Prudential policies applied to financial insti-tutions can contribute to this goal by influencingrisk-taking on the part of financial institutions andby improving the robustness of the financial systemto external shocks. As seen in Chapter II, a numberof countries have recognized this and taken steps tostrengthen bank supervision to cope with capitalflows (for example, Argentina, Chile, and India), butit is only recently that the potentially crucial role ofprudential policies in managing the risks associatedwith capital flows and financial intermediation gen-erally has been widely appreciated. The understand-ing of how prudential policies may affect macroeco-nomic performance is still at an early stage, andstatistical or econometric analysis of such links is alargely uncharted field. This chapter therefore high-lights a number of important issues rather thanreaching definitive conclusions.

Financial institutions are major parties to interna-tional capital transactions. They accept cross-borderand foreign currency deposits, initiate external bor-rowings, make foreign loans and investments, and

generally intermediate cross-border transactions. Ithas sometimes been observed that financial institu-tions are prone to excessive risk-taking. When thishappens in connection with cross-border and foreigncurrency transactions, sudden and large reversals ofcapital flows or large currency movements can havedamaging consequences on the health of individualfinancial institutions. Moreover, shifts in sentiment,leverage, and liquidity problems can multiply andtransmit shocks throughout the financial system, andin extreme cases they result in financial panics. Byrequiring more effective risk management in indi-vidual institutions, prudential policies can helpdampen transmission and contagion, and contributeto stemming the development of a major financialcrisis. The experience of the Asian economies since1997 has underscored the role that a weak financialsystem can have in accentuating a crisis.9

The distinction between prudential policies andcapital controls is not always clear-cut. A rapid andlarge buildup of foreign assets and liabilities by fi-nancial institutions driven by periods of “irrationalexuberance” followed by pessimism and a retrench-ment of positions can itself be a source of excessvolatility of currency prices and capital flows. If pru-dential regulation and supervision can make individ-ual institutions manage the risks associated with ex-ternal assets and liabilities more prudently, then thevolatility of capital flows may be reduced or the con-sequences of volatility limited. A targeted prudentialmeasure that seeks to limit a particular risk, for ex-ample, banks’foreign currency exposure, may influ-ence specific types of capital transactions. More-over, if banks’activities dominate capital flows inand out of a country, as is often the case, then con-straining the risks that can be taken by banks may ef-fectively limit the overall size of capital flows, aswell as their riskiness. Thus, measures that are typi-cally considered as prudential may in fact be usedfor capital control purposes.

III Prudential Framework for ManagingRisk in Cross-Border Capital Flows

32

8Capital controls in many instances may be regarded by the au-thorities as serving other important purposes, including strength-ening national sovereignty, protecting national security, andachieving specific social objectives.

9Baliño and others (1999) reviews how inadequate prudentialpolicies and weak banking systems contributed to and deepenedthe crisis in countries such as Indonesia, Korea, and Thailand.

Design of Prudential Policies for Managing Risks Associated with Capital Flows

Conversely, capital control measures may haveprudential effects, for example by restricting banks’short-term borrowing and thus limiting liquidity andother risks to banks associated with such borrow-ings. As was seen in Chapter II, countries includingChile, China, and India have on occasion used capi-tal control measures to pursue prudential objectives.The effective use of such measures rests on the exis-tence of an adequate administrative capacity.

Design of Prudential Policies forManaging Risks Associated withCapital Flows

Cross-border capital flows involve the same fun-damental categories of risks as do purely domestictransactions, but with added dimensions in each cat-egory. Box 2 provides an overview of the principalrisks that arise in the context of an open capital ac-count. Many of these relate to the use of foreign ex-change, but some also arise from differences in otherinstitutional arrangements.

The basic similarity between the risks of interna-tional capital flows and the risks of purely domestictransactions means that they can be addressed withinthe overall prudential framework by adapting andextending regulations and supervision used in thedomestic financial market. Best practice prudentialregulations would seek to manage the additionalrisks inherent in international capital flows by limit-ing the institution’s risk exposure relative to its risk-taking and management capacity. Thus, while pru-dential regulations do not target capital flowsdirectly, they can influence their volume, composi-tion, and volatility. Such regulations would includeenhanced monitoring, disclosure, and reporting; pru-dential limits (in the form of certain balance sheetratios); rules for loan classification, asset valuation,provisioning, and income recognition; norms requir-ing strong internal risk management procedures; andaccounting and control systems.

Prudential standards need to give particular atten-tion to banks, given their large (though somewhat di-minished) role in the provision of credit, their centralposition in the payments system, the systemic impli-cations of their high leverage, and the mismatch inthe liquidity of their assets and liabilities. Also, capi-tal inflows are often intermediated by the bankingsystem, and their reversal may be associated with abanking crisis if banks are not adequately prepared.Even when financial crises were triggered by eventsin nonbank financial institutions, the eventual sever-ity and duration of the crisis was largely determinedby the ability of the banking system to withstandshocks. The public good aspect of the banking sys-tem’s services and the potential cost of resolving a

banking crisis provide a further rationale for focus-ing on banks.

Recent experience in Asia has highlighted how vul-nerabilities can increase under a weak prudentialregime (Baliño and others, 1999). Capital inflows intothe banking sector helped fuel rapid credit expansion,with banks being increasingly exposed to credit andforeign exchange risks and to maturity mismatches inforeign currencies. Banks’foreign currency lending tocorporate borrowers that did not have secure foreignexchange revenue streams created major problemsonce the currencies started to depreciate. More gener-ally, rapid growth of assets also strained banks’capac-ity to assess risk adequately. Prudential regulation andsupervision that might have mitigated these problemshad serious deficiencies, including with respect to for-eign currency mismatches.10

Considerable work has been undertaken in inter-national forums to develop principles for prudentialregulation and supervision.11 Reflecting the height-ened interest and concern about international capitalflows, prudential standards and procedures are beingupdated to reflect the risks in bank intermediation ofcross-border capital flows. The Basel Committee hasproposed revisions to the capital adequacy frame-work, the development of methodologies for creditrisk and interest rate risk management and modeling,banks’interactions with highly leveraged institutions(notably, hedge funds), sound practices for loan ac-counting and credit risk disclosure, bank trans-parency and internal control systems, and opera-tional risk management. The Basel Committee hasalso issued papers on authorization procedures andprinciples for the supervision of banks’foreign es-tablishments, the information flow between bankingsupervisory authorities, and the relationship betweenbank supervisors and external auditors.

Ensuring an adequate capitalization of banks iscentral to limiting banking system risks, including

33

10In Korea, for example, the crises in the banking and corporatesectors, and the related external payments crisis, were to a largeextent rooted in excessive lending of foreign currency to corpo-rate borrowers with inadequate foreign exchange earnings. Theseexposures were not adequately monitored and controlled, eitherby the banks themselves or by the supervisory authorities.

11A comprehensive review of work in this area was provided inAnnex IV to the October 1999 International Capital Marketsre-port, “Proposals for Improved Risk Management, Transparency,and Regulatory and Supervisory Reforms.” The Basel Committeeon Banking Supervision has played a central role in this area.Work is also under way in the context of the Financial StabilityForum, which has established working groups on capital flows,off-shore financial centers, and highly leveraged institutions. TheJoint Forum on Financial Conglomerates—which comprises theBasel Committee, the International Organization of SecuritiesCommissions (IOSCO), and the International Association of In-surance Supervisors (IAIS)—has also issued a report on the su-pervision of financial conglomerates.

III PRUDENTIAL FRAMEWORK FOR MANAGING RISK IN CROSS-BORDER CAPITAL FLOWS

those arising from international capital flows. Thecentral objective of the new draft capital adequacyframework (which would replace the 1988 Accord)is to promote safety and soundness in the interna-tional financial system, and it gives even more atten-tion than in the past to the activities of large, interna-tionally active banks (see Basel Committee, 1999a).The proposed framework would maintain a modifiedversion of the existing accord as the standard ap-proach to minimum capital requirements, but alsoconsiders the option of providing greater scope forthe use of internal credit ratings and portfolio mod-els in establishing minimum capital. The coverage ofthe framework would also be extended to fully cap-

ture the risks in a banking group, with a view to ac-curately representing an institution’s risk profile. Su-pervisory evaluation and market discipline throughincreased disclosure are additional pillars of the cap-ital adequacy regime. The revised framework doesnot propose to change the minimum capital ade-quacy ratio of 8 percent. This level is not likely to besufficient for those institutions that are systemati-cally exposed to greater risk, such as those in emerg-ing markets, where the authorities are already inmany cases requiring or encouraging banks to main-tain higher capital. The advantage of higher capitaladequacy ratios would be to make financial systemfailures less likely; and when failures do occur, a

34

The opening up of the capital account and cross-border activities of banks introduce additional risks thatmay increase the magnitude, or complicate the manage-ment, of the risks that banks typically face in their do-mestic activities. The key risks with an open capital ac-count and how to cope with these risks are discussedbelow.1

1. Credit risk isthe failure of a counterparty to per-form according to a contractual arrangement. It appliesnot only to loans but also to other on- and off-balance-sheet exposures such as guarantees, acceptances, andsecurity investments. Additional dimensions of creditrisk in cross-border transactions include

• transfer risk:when the currency of obligation be-comes unavailable to the borrower;

• settlement risk:risk in the settlement of foreign ex-change operations that arise because of time zonedifferences; and

• country risk: risk associated with the economic,social, and political environment of the borrower’scountry.

2. Market risk is therisk of losses in banks’on- oroff-balance-sheet positions arising from movements inmarket prices that change the market value of an assetor a commitment. This type of risk is inherent in banks’holdings of tradable securities, financial derivatives,open foreign exchange positions, and interest-sensitivebank assets and liabilities.

Foreign exchange riskrefers to the risk of losses inon- or off-balance-sheet positions arising from adversemovements in exchange rates. It tends to be mostclosely identified with cross-border capital flows.Banks are exposed to this risk in acting as market mak-ers in foreign exchange by quoting rates to their cus-tomers and by taking unhedged open positions in for-eign currencies.

Interest rate riskrefers to the exposure of a bank’s fi-nancial condition to adverse movements in interestrates; it arises as a result of a mismatch (gap) between abank’s interest-sensitive assets and liabilities and af-fects both the earnings of a bank and the economicvalue of its assets, liabilities, and off-balance-sheet in-struments. Excessive interest rate risk may erode abank’s earnings and capital base. The primary forms ofinterest rate risk are

• repricing risk, whicharises from timing differ-ences in the maturity and repricing of bank assets,liabilities, and off-balance sheet positions;

• yield curve risk, whicharises from changes in theslope and shape of the yield curve; and

• basis risk, whicharises from imperfect correlationin the adjustment of the rates earned and paid ondifferent instruments with otherwise similar repric-ing characteristics.

Risk also exists in derivatives transactions. Deriva-tives are an increasingly common method of taking orhedging risks. The actual cost of replacing a derivativecontract at current market prices is one measure of aderivative position’s exposure to market risk. Sincemany of these transactions are registered off-balance-sheet, supervisors need to ensure that banks active inthese transactions are adequately measuring, recogniz-ing, and managing the risks involved.

3. Liquidity risk arises from the inability of a bank toaccommodate decreases in its liabilities or to fund an in-crease in its assets at a reasonable cost or liquidate itsassets at a reasonable price in a timely fashion. Inade-quate liquidity, then, affects profitability and, in ex-treme cases, can lead to insolvency. There are no inter-nationally agreed prudential standards on bank liquidity,but supervisors require banks to have adequate systemsto monitor and control their liquidity needs and estab-lish contingency plans for periods of liquidity stress.Regulation of liquidity risk focuses on the degree ofmismatch between maturities of assets and liabilitiesand dependability of access to funds in future periods.

Box 2. Risks in Banks’ Cross-Border Transactions

1This draws on Johnston and Ötker-Robe (1999).

Design of Prudential Policies for Managing Risks Associated with Capital Flows

higher portion of the cost would be borne by the pri-vate sector. The incentives of banks, as leveraged in-stitutions, to “gamble for resurrection” would alsodecrease. The principal disadvantage of higher mini-mum capital ratios is that they raise banks’costs andthus encourage further disintermediation. Differ-ences in minimum capital ratios across countries canalso distort competition among banks and influencedecisions on where to incorporate, and thereby alsopossibly affect cross-border capital flows. The newframework proposes to address this issue by giving alarger role to supervisory review in setting capitalrequirements for individual financial institutions thatappropriately reflects risks borne by the institution.Specific proposals affecting banks’international ac-tivities include the following:12

• External risk assessments prepared by ratingagencies would be used to establish risk weightsfor sovereign borrowers.13 Under the 1988 Ac-cord and its amendments, sovereign riskweights were based mainly on whether or not acountry is a member of the OECD.14 The draftproposes that only those countries rated mosthighly would be eligible for a zero risk weight(a minimum Standard & Poor’s rating of AA–),with the risk weight rising in stages to 150 per-cent for claims on countries with a rating of B–or below. Furthermore, sovereign risk could beweighted at less than 100 percent only if thecountry has subscribed to the Fund’s SpecialData Dissemination Standard (SDDS).

• Subject to some limitations, external risk as-sessments would also be used to establish riskweights for exposures to banks, other govern-mental entities, securities firms, and corporates.Under the existing procedures, all claims onbanks incorporated in OECD countries andshort-term claims (i.e., up to one year) on banksincorporated in non-OECD countries have re-ceived a 20 percent risk weight, while longer-term claims on non-OECD banks were risk-weighted at 100 percent. This standard has been

criticized on the grounds that it may have biasedcredit flows to emerging markets towardshorter-term maturities.

Although the use of external risk assessments inassigning risk weights aims to better reflect the actualrisk of assets than the current uniform and somewhatarbitrary risk weights, there is still considerable de-bate about the quality of external assessments. Creditratings have come under close scrutiny since the out-break of the Asian crisis, when a number of assess-ments proved to be far too favorable in retrospect.

Consideration is also being given to the need forprudential oversight of highly leveraged institutions(including hedge funds). There are concerns that theoperations of highly leveraged institutions have con-tributed to the volatility of capital flows to emergingmarkets. Ongoing discussion has focused onwhether the activities of highly leveraged institu-tions should be directly regulated, or whether theircreditors, particularly bank creditors, need to be heldto tighter prudential standards in their dealings withsuch institutions. The Basel Committee has empha-sized the latter approach:

• Before conducting business with highly lever-aged institutions, a bank should establish clearpolicies governing its involvement with theseinstitutions consistent with its overall credit riskstrategy. Sufficiently sophisticated risk manage-ment procedures need to be in place to identifyand measure the specific risks associated withhighly leveraged institutions, particularly therisks associated with derivatives. A preemptiveapproach rather than one informed mainly bynet asset values is essential.

• Overall credit limits need to be established, andcollateral and early termination provisionsshould take into account the ability of highlyleveraged institutions to rapidly change tradingstrategies, risk profile, and leverage.

Sound practices for loan accounting, credit riskdisclosure, bank transparency, and related matterswill also help to mitigate the risks associated with in-ternational capital flows. The Basel Committee re-cently issued a paper listing 26 sound practices forloan accounting and recognizing credit risk exposure(Basel Committee, 1999b). The suggested practicesreflect the judgment that capital adequacy standardslack meaning, and risk management is seriously im-paired, if loans are not properly valued and loanlosses are not adequately recognized and provisionedfor in banks’balance sheets. Generally good prac-tices in this field are found in a number of advancedeconomies, but these may not be easily transferableto countries with less developed financial and regu-latory infrastructures. In such countries, more me-chanical approaches relying on simple quantitativecriteria may be more appropriate. Box 3 discusses

35

12Other specific proposals that could influence banks’interna-tional activities include those on risk weights for over-the-counterderivatives and securitized assets.

13National supervisory authorities would need to be satisfiedthat the risk assessment institutions meet minimum standards fortransparency, objectivity, independence, credibility, and accuracy.Also, banks would be expected to follow a consistent approach inusing such assessments (that is, cherry-picking ratings would notbe permitted).

14For the purpose of the Accord, OECD countries include fullmembers of the OECD and those countries that that have con-cluded special lending arrangements with the IMF associatedwith the Fund’s General Arrangements to Borrow, but excludeany country that has rescheduled its sovereign debt during theprevious five years.


the experience of the United States, with a particularemphasis on the treatment of cross-border loans.

International capital flows also generate addi-tional risks for financial institutions in terms of mar-ket risk and liquidity risk. For market risk, the 1996Amendment to the Capital Accord to incorporatemarket risks required internationally active banks tohold capital against risks related to exchange ratechanges and movements in the price of assets heldfor trading purposes. For liquidity risk, the Commit-tee’s 1992 paper “Framework for Measuring andManaging Liquidity” provides a summary of prac-tices and techniques employed by major interna-tional banks in measuring and managing liquidity,

and provides a benchmark for liquidity managementby banks. Management of foreign currency liquidityis particularly important because, unlike in domesticcurrency, there are limits on the ability of centralbanks to provide assistance to banks to tide overtemporary liquidity problems.

Implementation of Prudential Standards

To effectively manage the risks from interna-tional capital flows, authorities must establish ade-quate prudential standards in all markets. Recent fi-

36

In determining the adequacy of loan valuation andloss provisioning, U.S. supervisors evaluate eachbank’s risk management capacity and internal policiesand procedures (internal credit review procedures,loss evaluation techniques, and the adequacy of loan-loss provisions) relative to its individual portfoliocomposition and risk profile. Evaluations are per-formed according to general guidelines, which eschewthe application of mechanical rules. The same princi-ples and methodology are applied to both domesticand cross-border credit exposures but additional re-quirements are applied to a bank’s internal policiesand procedures for managing material cross-bordertransfer risk.2

Loan Grading

Supervisory guidelines establish five categories(pass, special mention, and classified credits, consist-ing of substandard, doubtful, and loss), based on a de-fined set of key factors, for grading the risk quality ofloans.3 Delinquent (overdue) credits are also distin-guished but do not automatically determine the riskcategory, and performing (nondelinquent) loans withwell-defined credit weaknesses may be adversely clas-sified. Additional factors are taken into account inevaluating partially charged off or formally restruc-tured credits, as well as guarantees and off-balance-sheet items.

Loan-Loss Provisioning

Banks are required to establish a loan-loss reserve(“allowance for loan and lease losses” or ALLL),charged against current income. All loans, or portionsof loans, recognized (classified) as “loss” must becharged off immediately.4 For all loans not classifiedas loss, the ALLL must be increased by (1) losses esti-mated over the remaining effective lives of loans andleases classified as substandard or doubtful, (2) alllosses estimated for the forthcoming 12 months oncredits not classified, and (3) estimated losses fromtransfer risk exposures. A bank’s management is re-sponsible for grading the loan portfolio and making thenecessary provisions or charge-offs at least quarterly.Estimated losses on individual credits should meet thecriteria for accrual of a loss contingency set forth inU.S. generally accepted accounting principles(GAAP). In addition, supervisors evaluate the ade-quacy of the overall level of the ALLL based on ananalysis of the bank’s policies, practices, and proce-dures, its historical credit-loss experience and the rea-sonableness of the management’s overall methodology.Reasonableness is assessed by comparing the actuallevel of the ALLL against the sum of 50 percent ofdoubtful and 15 percent of the substandard loans5 plusestimated losses on other credit exposures (excludingthose classified as loss) over the forthcoming 12months. Shortfalls from this alternative calculationtrigger a more intensive review of management’s

Box 3. The U.S. Supervisory Approach to Loan Classification and Provisioning1

Source: Board of Governors of the Federal Reserve System(1994), and various supplements updated through May 1999.

1As set forth in the “Interagency Policy Statement on theAllowance for Loan and Lease Losses” issued December 21,1993.

2Transfer risk is a subset of country credit risk and refers tothe borrower’s capacity to obtain the foreign exchange re-quired to service its cross-border debt.

3Banks’own loan grading systems, when different, must bereconcilable with the regulatory framework.

4The value of credit (net of the realization of the net liqui-dated value of collateral or realization of guarantees) and theALLL account must both be reduced by the amount of theloss. All applicable unpaid interest accrued during the currentyear should be charged against current income and unpaid in-terest accrued in prior years should be charged off to theALLL.

5These weights are based on the industry average loss expe-rience over time for these classifications.

Implementation of Prudential Standards

nancial crises have shown that prudential standardsfell well short of best practices in many countries,even when judged against norms that do not take ac-count of the more recent advances in this area. Al -though conforming to prudential standards is oftenin the self-interest of banks and other financial insti-tutions, prudential standards are also designed tocombat moral hazard and related incentives for ex-cessively risky behavior. When prudential regula-tions and practices differ markedly across countries,this will create opportunities for regulatory arbi-trage, reduce the effectiveness of the regulations,and increase systemic risks. In these circumstances,prudential standards in one country will need to

take account of the adequacy of prudential stan-dards in other countries. As discussed previously,the Basel Committee’s draft framework for banks’capital adequacy takes account of countries’imple-mentation of the Core Principles in setting riskweights; but there may be scope for more systemati-cally using information on countries’prudentialstandards. Prudential regulators and supervisors inboth advanced economies and emerging marketsmight have been able to mitigate the Asian financialcrisis—the former by taking adequate steps to dis-courage financial institutions and other investorsfrom exposing themselves to excessive risks in theemerging markets; the latter by putting in place

37

analysis, but management’s estimates are usually ac-cepted when it has (1) maintained effective systemsand controls for identifying, monitoring, and address-ing asset-quality problems in a timely manner, (2) rea-sonably analyzed all significant factors affecting thecollectibility of the portfolio, and (3) established an ac-ceptable ALLL evaluation process.

Credit Exposures Involving Cross-Border Transfer Risk

Banks are required to report quarterly on individualcountry exposures that are significant relative to theircapital and the country’s economic performance and tohave in place additional procedures for managing asso-ciated transfer risk as well as for grading and provision-ing against these exposures. An official InteragencyCountry Exposure Review Committee (ICERC) evalu-ates and classifies transfer risk exposures to specificcountries based on criteria as set forth in the Interna-tional Lending Supervision Act of 1983.6 Banks are in-formed about classifications of only those loans spe-cific to their portfolios and adequate safeguards arerequired to prevent such information being divulged to

unauthorized personnel.7 An ICERC classificationtakes precedence over the general classification guide-lines only when it is more severe. The ICERC frame-work also includes a nonclassified category of expo-sures that supervisors incorporate into their generalassessment of a bank’s asset quality and adequacy of itsreserves and capital. This category includes exposuresto countries taking economic adjustment measures,generally as part of an IMF program, to restore debtservice, or to countries where recent debt service per-formance indicates that an earlier classification is nolonger warranted. The ICERC generally accords morefavorable treatment to performing trade credits andbank credits. Measurement of country exposure is alsoadjusted for guarantees and collateral and the risk is re-allocated to the country where the guarantor resides, thecollateral is held, or the issuer of stocks or bonds usedas collateral resides. The International Lending Super-vision Act of 1983 requires banks to set up a separateloss reserve “Allocated Transfer Risk Reserves(ATRR)” for loss provisions on transfer risk exposuresclassified as “value impaired.” Required provisions arebased on mandated percentages unless the loss is di-rectly charged off. The ATRR must be segregated fromthe ALLL and cannot be considered as part of capital.Allocations to ATRR are not initially required whennew loans are made in the context of an IMF-supportedor other appropriate economic adjustment program thatgenerally enhances the debt service capability of thecountry concerned. However, such allocations could berequired subsequently on the basis of performance U.S.branches of foreign banks are not subject to the regula-tions establishing the ATRR but are expected to have inplace policies for recognizing and writing off losses ontransfer risk exposures. U.S. supervisors evaluate theirtransfer risk and related procedures.

6“Substandard” is applied to countries that (1) are not com-plying with external obligations; (2) are not in the process ofadopting an IMF or other suitable economic adjustment pro-gram or adhering to such a program; and (3) the country andits bank creditors have not negotiated a viable reschedulingand are unlikely to do so in the near future. “Value impaired”is applied to a country with protracted arrearages as indicatedby more than one of the following factors: (1) it has not fullypaid interest for six months; (2) it has not complied, nor arethere immediate prospects for complying, with an IMF pro-gram; (3) it has not met rescheduling terms for over one year;and (4) prospects for an orderly restoration of debt service inthe near future are indefinite. “Loss” applies when the loan isconsidered uncollectible. This classification would apply, forexample, if a country were to repudiate its obligations tobanks, the IMF, or other lenders.

7The approach for exposures to transfer risk parallels an ap-proach used for credits to large domestic borrowers, the“Shared National Credit Program.”


sound and transparent prudential standards thatwould have limited and made clearer the risks fac-ing international investors.

All countries (especially emerging markets thatwould benefit directly from reduced risk to theirown financial systems) have an interest in establish-ing adequate prudential standards, but countrieswith large and advanced markets have a particularresponsibility to ensure that their investors and fi-nancial institutions are heedful of the risks involvedin cross-border transactions. The size of institu-tional portfolios in the major advanced economies issuch that modest changes in their asset distributioncan have a significant macroeconomic impact onsmaller open economies. The magnitude and rapid-ity of such portfolio adjustment can be substantial,which leads to excessive inflows into smaller coun-tries and to their later reversal. Rapid portfolioshifts have a number of causes, which fall into twomajor categories: a failure to draw adequate distinc-tions between different countries in a region, or dif-ferent assets in a country; and a “run” on a countryor region similar to a bank run.15 Prudential authori-ties in the large advanced countries will have an in-terest in limiting these problems to enhance sys-temic stability. Establishing a regulatory andsupervisory framework that obliges investors tomore accurately analyze and differentiate countriesand assets reduces the first type of shortcoming,generally helps to improve risk management, andmay thereby also contribute to limiting the risk ofruns.

The establishment and maintenance of prudentialstandards rest on three pillars: public regulation andsupervision, internal practices and controls, andmarket discipline. Moreover, the prudential supervi-sion and regulation framework must continuallyadapt to the evolving state of market developmentand internal governance in individual institutions.Especially in developing countries, one or more ofthese pillars may be weak. In mature markets, rapidinnovations in financial technology pose particularchallenges, in that management and supervision can-not sufficiently keep pace with these developmentsand fail to identify risks in the financial institutionsand in the financial system. Large losses incurred byeven the most sophisticated institutions in their de-rivatives and global trading activity point to the seri-ousness of the risks.

Countries with weak supervisory agencies often,but not always, also suffer from relatively weakskills in the private financial sector, and thus fromserious shortcomings in the ability and incentives offinancial institutions to adequately manage risk. Di-rected and connected lending, evergreening of loans,and excessive credit concentration are known tohave been the immediate cause of major bankingproblems in countries, and they may be compoundedby weaknesses in the legal system and other gover-nance problems that impede effective monitoring bycounterparties and shareholders as well as loan col-lection efforts. Such financial systems are some-times said to suffer from a weak “credit culture.”The absence of satisfactory disclosure rules and theinability of the supervisory authorities to enforcethem may further weaken the operation of marketdiscipline in such systems. Resolution of these prob-lems is usually not quick, and it must be viewed aspart of the overall process of long-term economicdevelopment.

Effectiveness of Prudential Measures inLimiting Risks Associated with CapitalFlows and the Role of Capital Controls

Prudential policies could contribute importantly toreducing the risks associated with international capi-tal flows, by strengthening the ability of the financialsystem to withstand volatile market conditions. Theymay also be useful in reducing the volatility of flowsinvolving financial institutions, which may actuallybe a principal element of destabilizing capital flows.As discussed in Chapter II, countries that have madesubstantial progress in this field—for example, Ar-gentina and Chile—have been quite successful incontaining the risks from international capitalflows.16 On the other hand, a number of the most so-phisticated financial institutions have exposed them-selves to excessive risks in their cross-border trans-actions, which underscores the need to adaptprudential policies to innovation in the marketplace.Shortcomings in internal risk management proce-

38

15In such a run, a rapid and large-scale sell-off of a country’sassets (and by implication its currency) has adverse effects on thereal economy, further depressing asset values. Investors, expect-ing other investors to sell off, seek to be the first through theexits.

16Chile’s prudential policies are discussed in Appendix I. Ar-gentina’s financial sector reforms are discussed in detail in vari-ous issues of the IMF Staff Country Reports on Argentina. It isimportant to note that efforts to improve prudential policies in Ar-gentina have been ongoing, and additional elements of best prac-tice have been implemented almost continuously. In 1996–99, forexample, minimum capital requirements were tightened throughthe introduction of more stringent criteria for calculating risk-weighted assets and by making minimum capital requirements afunction of the degree of maturity mismatch between banks’as-sets and liabilities.

Effectiveness of Prudential Measures in Limiting Risks Associated with Capital Flows

dures and the failure of supervisors to detect andcorrect these problems were partly at fault. Theremay also have been an element of moral hazard inthe actions of some financial institutions, brought onby expectations of a bailout. An important potentialbenefit of improved prudential standards and prac-tices is that supervisors can more easily recognizeand prevent financial institutions from engaging inbehavior that may in the end necessitate a bailout bythe public sector.

While prudential policies are intended to promotesoundness, it must be recognized that prudentialstandards, if not carefully designed and applied, mayhave unintended and undesirable consequences byproviding distorted incentives that result in exces-sive risk-taking in specific areas, as well as facilitat-ing contagion. Notably, risk-weighting schemes incapital adequacy regulations that do not adequatelyreflect the riskiness of different borrowers could en-courage banks to take on greater than warranted ex-posure to high-risk borrowers. Similarly, when in-vestments made by institutional investors arerequired to carry a minimum credit rating, largeamounts of capital may be pulled out from a countrywhose credit rating has been downgraded, thus gen-erating a self-fulfilling downturn in that country. So-phisticated, statistically based risk managementtechniques, if used to maximize trading profits byexploiting correlations between markets withoutgood judgment as to the limitation of such correla-tions, may prove to be quite vulnerable in periods ofstress when historical relationships between marketsbreak down. In such cases, a rush to close downloss-making positions may further accentuate mar-ket volatility.

Prudential policies must also strike an appropriatebalance between reducing the threat of excessiverisk-taking and constraining the freedom of institu-tions to take the normal risks inherent in financial in-termediation. In this connection, care must be takenso that regulations are not oriented toward control-ling capital flows at the expense of weakening therole of prudential policies in maintaining the safetyand soundness of financial institutions. Althoughcross-border transactions often entail added dimen-sions of risk (such as foreign exchange risk), thisdoes not necessarily mean that these transactions orassets are inherently riskier than domestic assets.Nor, for that matter, do the risks related to externaltransactions and assets usually comprise a major partof risks run by institutions. Indeed, prudential regu-lations based excessively on the foreign-domesticdistinction will not be effective in addressing finan-cial risks.

Although prudential measures and improved riskmanagement at individual institutions can help tolimit the risks associated with international capital

flows, they will not be able to discourage all unsus-tainable flows. Prudential measures cannot be sostrict as to kill off risk-taking activity altogether, andcarefully managed risk-taking strategies could un-ravel under unexpected shocks. Sentiments can alsooverride the best prudential measures. Moreover,prudential measures target financial intermediariesthat manage other people’s money, and are not in-tended to regulate the portfolio decisions of individ-uals and nonfinancial corporations that invest theirown funds.17 For example, cross-border portfolio in-vestment may lead to a speculative asset price bub-ble, just as a bubble may arise in domestic financialmarkets. The collapse of a bubble can have seriousmacroeconomic consequences, partly because de-clining asset prices affect wealth and private con-sumption. Prudential regulations may help to reducethe effects of an asset price deflation on financial in-stitutions, thereby mitigating its consequences forreal activity, but they may not be able to preventsuch an event from occurring.

When prudential standards and practices areweak, and possibly when institutions that are outsidethe scope of prudential policies (such as nonfinan-cial firms) are an issue, other measures, includingcapital controls, may prove useful in managing thespecific risks associated with international capitalflows.18 As discussed in Chapter II, capital controlsdiffer in how effectively they perform a prudentialfunction (when they are used for this purpose); andin how severely they distort resource allocation in fi-nancial and other markets. Capital controls also dif-fer in how difficult they are to administer and howeffectively they can be enforced. As with all types ofeconomic regulation, including prudential regula-tion, unintended side effects may arise, and individ-ual controls must be judged not in isolation but onlyin the context of a country’s overall regulatory andinstitutional framework.

The design of a well-functioning system of capitalcontrols to serve a prudential function is thus a com-plex task. Outright prohibitions of capital transac-tions may be easiest to administer and enforce, butonly when controls are fairly comprehensive. If cur-rent payments and some capital transactions havebeen liberalized, such controls may be circum-vented, for example by disguising controlled trans-

39

17Corporate governance and monitoring by creditors are ex-pected to provide oversight, but the basic presumption would bethat economic agents must be allowed to invest their ownmoneyas they see fit.

18Some countries with weak domestic prudential institutionshave limited their resort to capital controls by encouragingforeign bank ownership, with supervision by the banks’homesupervisors.


actions as uncontrolled ones; or by duplicating thepayoffs of a restricted contract with an unrestrictedone. The principal drawback of a blanket prohibitionis that it will preclude sound as well as risky transac-tions, and may therefore be highly distortionary. In-troducing elements of administrative discretion—such as a licensing system for capital flows—can al-leviate this problem somewhat but is administra-tively more costly and may raise governance issues.Price-based controls modeled on prudential regula-tions, such as the unremunerated reserve require-ment on inflows used by Chile, are less distor-tionary. However, such controls are generally moredifficult to administer and enforce than outright pro-hibitions, and possibly than quantitative restrictions.Loopholes in their coverage will need to be identi-fied and closed as they are progressively exploitedby arbitrageurs. The ultimate complexity and de-mands on a country’s administrative capacity ofsuch controls may thus be similar to that of pruden-tial regulation and supervision. They may nonethe-less prove useful in countries where other pillars of afunctioning prudential system (market discipline,transparency and internal controls in financial insti-tutions) are weak.

Conclusions

The use of prudential policies in coping with therisks associated with capital flows needs to be ana-lyzed further, in terms of both understanding howthey best function and studying actual implementa-tion by countries. Despite the relatively favorableexperiences of a number of countries that havestrengthened their supervisory regimes, country ex-periences still offer only limited evidence on howwell prudential measures can limit the risks associ-ated with capital flows, and additional work isneeded on this point. Nonetheless, the discussionabove highlights the need for a careful design of pol-icy, the risks of targeting capital flows at the expenseof the safety and soundness of institutions, and theimportance of implementation capacity—a particu-larly demanding challenge for emerging markets.

The use of capital controls in pursuing prudentialobjectives is more controversial. The positive rolethat controls may potentially play in an environmentof weak supervisory systems is tempered by the dif-ficulty of administering a sophisticated system ofcontrols and the distortionary effects that simpler di-rect controls may have.

40

This review of the use and liberalization of capi-tal controls in 14 countries cannot be considered

exhaustive. It illustrates the difficulty of preciselyassessing the effects of capital controls, which mayhave benefits as well as costs. The analysis of the re-lationship between prudential policies and capitalcontrols is a first step, and considerable further workwould be needed to fully clarify their respectiveroles, interdependencies, and limitations. This papernonetheless sheds some light on arguments previ-ously advanced in the literature, on some of the op-erational issues related to the design of capital con-trols, and on the relationship between capitalcontrols and prudential policies. Despite the diver-sity of the country experiences examined in thispaper and the absence of a single best approach tocapital account liberalization, a number of apparentregularities may prove useful in formulating policy.

The evidence presented in this paper supports theconclusion that capital controls cannot substitute forsound macroeconomic policies.Countries with seri-ous macroeconomic imbalances and no credibleprospect for improvement in the short run were reg-ularly unable to address large-scale capital flows, ortheir adverse economic effects, with capital controls.Not even comprehensive and strictly enforced ad-ministrative controls have always protected coun-tries from balance of payments or financial crises.

To what degree capital controls are effective in in-sulating a country from external shocks or in provid-ing a breathing space in which to adopt sound poli-cies is a more difficult question to answer from thecountry case studies. Countries have tended to em-ploy a number of policy instruments in unison to-ward a policy goal, so that it is difficult to disentan-gle the contribution of capital controls in achieving acertain objective. More flexible exchange rate poli-cies, prudential policies, and liberalization of out-flows (in case of excessive inflows) are some of thepolicies that have been employed in conjunctionwith capital controls. Some countries that have em-ployed capital controls appear to have been moresuccessful than others in achieving their policy ob-jectives, and one can draw a number of generallyuseful observations from the countries’experiences.

First, no single capital control measure is effec-tive across all countries at all times. Effectivenessdepends on a host of factors, including the serious-ness of macroeconomic imbalances, which may giverise to strong incentives for circumvention of thecontrols.

Second, selective controls on a targeted range oftransactions, while possibly effective in limitingthose specific transactions, tend to be quickly cir-cumvented as market participants find ways toachieve their desired ends through unrestricted chan-nels. To achieve their objective, controls need to bewidened as market participants find new ways of cir-cumventing the restrictions. The ease with which re-strictions are circumvented is mitigated somewhat incountries that have a strong monitoring and enforce-ment capacity and that are able to quickly adjustcontrols to close off avenues for circumvention. Inmost cases, however, regulators have encountereddifficulties in anticipating and countering the marketresponse to controls. This is particularly true for acountry with well-developed financial markets.Countries’experiences also show that even currenttransactions and foreign direct investment have beenvehicles for circumvention, which attests to the diffi -culty of targeting even at the broadest level. To beeffective in the somewhat longer run, controls inmost cases needed to be quite comprehensive.

Third, administrative capacity and the level of fi-nancial market development also have a bearing onthe choice of controls and their effectiveness. Prop-erly designed market-based controls are more likelyto be the less distortionary choice for a financialmarket that is substantially developed or liberalized.Nevertheless, measures such as the Chilean URR de-mand a degree of administrative sophistication ifthey are to be effective. Direct controls have beenapplied with some success in relatively closedeconomies at an earlier stage of financial market de-velopment. However, countries such as India andChina that took this course also possessed an effec-tive administrative apparatus. While direct controlsmay be somewhat less administratively demandingthan market-based controls, one cannot concludethat direct controls are, other things equal, more ef-

IV Conclusions

41

IV CONCLUSIONS

fective than market-based controls. Direct controlsmay also be circumvented when they are not suffi -ciently comprehensive, or when implementation ca-pacity is not sufficiently strong. Also, discretionarycontrols open up governance issues related to theirfair and transparent implementation.

The need for controls to be comprehensive inorder to be effective implies that more effective con-trols are also more distortionary and hence morecostly. The benefits of effective controls thus need tobe carefully weighed against their costs. Compre-hensive direct controls can allow a country with aless developed financial market to insulate itself tosome extent from external shocks and pressures, butsuch policies may impede financial market develop-ment, and may lead to a loss of the efficiencies thatderive from liberalized markets. In countries withmore sophisticated financial (and other) markets,very strong controls may be needed to ensure effec-tiveness. At some stage it may become difficult todesign a set of controls that can limit “undesirable”capital flows without unduly restricting “desirable”transactions, seriously disrupting financial markets,and reducing access to foreign capital. The unfavor-able trade-off has prompted many countries to dis-mantle comprehensive controls, including those thatwere introduced during periods of stress.

The evidence is mixed on whether capital controlscan be used to correct financial market imperfectionsand serve a prudential purpose. Capital controls, par-ticularly on short-term inflows, may temporarily andpartially substitute for full-fledged supervisory insti-tutions. In particular, it is clear that building effec-tive supervisory and regulatory institutions may takea long time. On the other hand, the experience of thecountries reviewed here suggests that while pruden-tial concerns sometimes played a role in the decisionto use capital controls, macroeconomic considera-tions were typically more important and indeed deci-sive in many cases. When governments adopt andmodify capital controls primarily in response tomacroeconomic factors, this may detract from theirusefulness in attaining prudential goals.

The converse of the previous question is whetherprudential regulation and supervision of financialinstitutions can help in managing the risks from in-ternational capital flows, by influencing the vol-ume, composition, and hence the volatility of suchflows. The evidence on this point seems more per-suasive. Strong prudential policies were found toplay an important role in orderly and successfulcapital account liberalization and in reducing thevulnerability to external shocks, and such policiesmay, to some extent, be an alternative to capitalcontrols, in addition to being an inherently valuablemeans of enhancing financial system stability. Ofcourse, prudential policies alone will not be able toeliminate the risks associated with internationalcapital flows. Properly used, however, they willcontribute to lessening such risks, in conjunctionwith appropriate macroeconomic policies. Withprudential policies, as with capital controls andother government intervention, there is a need toguard against misregulation and overregulation.Moreover, as countries differ in their ability to im-plement and enforce various types of policies, theappropriate mix of capital controls and prudentialpolicies to be used in moving toward capital ac-count convertibility will also need to be tailored to acountry’s specific circumstances.

Finally, with regard to sequencing, both capitalaccount liberalization and other financial sector re-forms are ongoing and interrelated processes, whichappear to be closely linked to the overall level ofeconomic development. The impetus for necessaryfinancial sector restructuring has often come from amore general opening of the economy, and improvedfinancial sector stability is in turn conducive to fur-ther external liberalization. These processes are alsocomplex, and involve changes in many dimensions,including market development, governance, pruden-tial regulation and supervision, monetary opera-tions—the entire infrastructure of finance. Againstthis background, it is difficult to prescribe in generalthe sequence in which capital controls on differenttypes of flows should be liberalized.

42

capital controls: country experiences with their use … paper aims to develop a deeper...

Documents