The evolution of capital account liberalization has varied across developing countries, reflecting, among other things, their differing degrees of macroeconomic imbalance and balance of payments strength and the extent of general economic liberalization. Prior to the mid-1980s, liberalization of capital account transactions was almost always undertaken from a relatively strong balance of payments position (Indonesia,45 Malaysia, and Singapore), but more recently a number of developing countries have adopted full convertibility following balance of payments deficits (Mauritius and Trinidad and Tobago), including in some cases large external payments arrears (in a number of Latin American and Caribbean countries). In the Baltic countries, where the elimination of capital controls followed the breakup of the former U.S.S.R., the balance of payments was in surplus owing in part to large reserve-related transfers. Current account positions tended to be less problematic; where countries experienced current account deficits in the period leading up to liberalization, with a few exceptions (Costa Rica, The Gambia, and Jamaica) the deficits were relatively small, usually well below 5 percent of GDP.
Outstanding external debt was substantial in many cases, particularly among Latin American and Caribbean countries, where it typically exceeded 500 percent of export earnings, and in Indonesia (250 percent). Exceptions were Singapore, where the external debt burden was almost nonexistent, and the Baltic countries and Malaysia, where it was moderate. Gross reserve holdings varied considerably across the liberalizing countries, from a low of less than one month of import cover in Jamaica to a high of over one year of cover in Malaysia.46 Reserve holdings averaged four to five months of import cover when full convertibility was adopted, although holdings that were freely usable for purposes other than debt servicing were often lower. Domestic macroeconomic imbalances at the inception of the capital account reforms were large in a number of countries, with inflation rates ranging from 5 percent in Indonesia and Singapore to over 1,000 percent in the Baltic countries before liberalization.
Process of Capital Account Reform
The experiences of a growing number of countries in adopting capital account convertibility offer an opportunity to examine the sequencing and speed with which reforms have been undertaken, with a view to determining if particular approaches have been associated with successful liberalization. In particular, lessons can be drawn on whether the chances of success are enhanced by other reforms introduced before the opening of the capital account. There is also the issue of measures that need to be adopted concurrently with the elimination of capital controls, so-called necessary conditions.
General Issues of Sequencing
Historically, capital account convertibility tended to be part of a gradual, phased approach to economic reform, occurring after current account convertibility in some countries, usually in the context of financial sector reforms.47 An example of such an approach is Malaysia, where a system of administered interest rates was in effect when liberalization of capital commenced. Malaysian financial institutions were therefore reformed simultaneously, in order to provide scope for market-based interest rates and to encourage the development of a stock market. Mexico is another example. Although it did not achieve full convertibility, it had a tradition of a relatively open capital account. The debt crisis in the early 1980s led to the reimposition of exchange controls that were soon lifted in the context of a newly introduced free foreign exchange market. Significant trade and fiscal reforms were instituted over a five-year period starting in 1982. Financial reforms, however, were instituted relatively late in the process, when it became apparent that to enhance the efficiency of the financial system and to improve the effectiveness of monetary policy, greater reliance on market forces was needed. Indonesia, in contrast, is a longstanding example of a successful and sustained opening of the capital account in which capital liberalization took place, in 1971, before the financial sector was reformed. Trade was not substantially liberalized until 1980, while interest rates began to be determined by the market only in 1983. Certain constraints on the financial system were maintained until 1988, but, although administered, real interest rates were positive from the outset of capital convertibility.
More recently, the process of liberalizing various categories of capital has tended to be less gradual, while the sequencing has continued to vary. A number of countries have moved to capital convertibility in a one-step process (Costa Rica, Hong Kong, Jamaica, Kyrgyz Republic, Mauritius,48 Singapore, Trinidad and Tobago, and Venezuela49). In the Baltic countries, exchange controls were liberalized over a very short time, with both Latvia and Lithuania achieving convertibility at around the time of the introduction of their national currencies. Similarly, Estonia eliminated all remaining controls on capital account transactions in the year following the establishment of a currency board in 1992. In Latin America, Argentina began to liberalize current account transactions in 1989 and completed the transition to full convertibility in 1991 in conjunction with the introduction of a currency board arrangement. El Salvador started its move to capital convertibility in 1991 and implemented most of the measures within the following year. Similarly, convertibility of the Paraguayan guaraní was established over a two-year period (1991–92).
Roles of Monetary and Exchange Regimes
Freedom from controls on capital movements heightens the role of domestic interest rates in avoiding destabilizing capital flows. It is reasonable to expect that a successful opening of the capital account would require that domestic financial markets be competitive either prior to or concurrently with the adoption of full convertibility in order to achieve market-related interest rates. Real interest rates were positive in most, but not all, cases at the time that full convertibility was adopted. Negative real rates in Costa Rica, Estonia, Guyana, Latvia, Lithuania, and Peru were soon raised to internationally competitive levels. Not surprisingly, most developing countries that abolished controls on capital transactions freed their domestic financial markets either prior to, in conjunction with, or soon after the capital account reform. For example, in most countries in Latin America that have recently liberalized and where financial markets were already fairly well developed, interest rates on loans and deposits were freed, and indirect monetary instruments such as treasury bills were introduced well before the establishment of full convertibility. In some cases, credit ceilings were eliminated as a first step in the reform program, before the establishment of capital convertibility (Costa Rica, Jamaica, and Trinidad and Tobago).
Rapid credit expansion by the central bank had also been a major problem in many countries prior to liberalization. Measures to reduce the extension of central bank credit, including credit to government through tighter fiscal policies, were taken at the time of reform in almost all countries (e.g., El Salvador, Estonia, Jamaica, Lithuania, Mauritius, Peru, Trinidad and Tobago, and Venezuela). Reserve requirements received increased emphasis as a tool to manage liquidity after the elimination of the credit ceilings in a number of countries (Argentina, El Salvador, Jamaica, Peru, Trinidad and Tobago, and Venezuela). At first, there was a tendency to raise reserve requirements, but, with subsequent interest rate deregulation, such requirements were often lowered.
In other countries, reform of the financial sector took place together with broader reforms that included capital account liberalization. For example, in the Baltic countries, a financial infrastructure did not exist, and there was little choice but to undertake reforms on all fronts simultaneously. Virtually all countries emphasized the use of open market operations (sale or purchase of government securities by the monetary authorities)50 in order to be able to sterilize bank liquidity. Such operations were introduced following, or simultaneously with, the elimination of credit ceilings and a reduction in reserve requirements. Instruments used were central bank securities or government bonds, or both.51 Given that secondary markets were not developed in most countries, primary issuance was often used as the initial form of operations.
In countries with limited foreign exchange markets, approaches to the choice of exchange rate regime were mixed, although most had floating regimes.52 A fixed rate regime backed by a currency board was adopted in Argentina, Estonia, and Lithuania, providing a strong institutional commitment to exchange rate stability and low inflation. In contrast, floating exchange regimes were adopted along with full convertibility in El Salvador, The Gambia, Guyana, Jamaica, Latvia, Peru, and Venezuela, where dual official and free foreign exchange markets were unified on the basis of the arrangements in the free market. Mauritius and Trinidad and Tobago also adopted full convertibility in the context of freely floating exchange rate regimes, replacing their earlier exchange rate pegs (to a currency basket in Mauritius and to the U.S. dollar in Trinidad and Tobago). Exchange rate flexibility in Costa Rica was gradually increased in 1990 with the widening of the spread between official buying and selling rates, while a freely floating exchange rate regime was adopted in 1992 at the time that controls on capital transactions were eliminated. Other countries (Indonesia, Malaysia, and Singapore) adopted managed floating regimes, with an eye to ensuring the competitiveness of export industries.
Prudential Supervision and Market Information
The liberalization of capital movements has in most cases been accompanied or followed by strengthened prudential supervision and regulation, especially in the area of foreign exchange risk assumption by financial institutions. Opening the capital account could increase the risks for banks, through the impact of increased volumes of capital flows on the deposit base and through a possible increase in exchange rate volatility on banks’ open foreign currency positions. Capital account liberalization therefore required strengthened supervision related to foreign exchange risks, generally undertaken as part of ongoing, broad financial sector reforms. In most of the countries surveyed, the bulk of the reforms to improve prudential standards was under way prior to and during liberalization, while in some others (Costa Rica, Guyana, Indonesia, Jamaica, and Peru), the reforms took place mainly during and after adoption of capital convertibility. Prudential reforms have focused on improvements in the supervisory framework, particularly adopting new regulations and reporting requirements, and increasing the ability of the supervisory authority to enforce the regulations. Most of the reforms have been directed toward achieving international coordination and harmonization of supervisory practices and have used the Basle Accord on capital adequacy as well as other standards in the major industrial countries as yardsticks.
Internationally accepted accounting principles and disclosure norms become highly important when capital movements are free.53 This has been recognized in a number of countries. For instance, a new accounting system for banks in China became effective in July 1993, and accounting standards that meet EU requirements are soon to be implemented in the Baltic countries. In addition, as a member of the International Accounting Standards Committee (IASC), Hong Kong issues statements of standard accounting practices that in all respects are in accordance with IASC requirements. However, significant improvements in accounting and disclosure practices remain necessary in a number of the liberalizing countries. In some countries, in the context of inadequate accounting and information systems, sudden information disclosures may have contributed to the instability of capital flows.
Credit rating institutions promote market scrutiny of the prudential position of financial institutions and also enhance the ability of investors to protect their interests. For example, to assess banks better. Bank Indonesia has recently introduced a sophisticated rating system that gives weights to capital, asset quality, management, earnings, and liquidity in line with international practice. Malaysia set up a credit rating agency in 1990 to rate all issued private debt securities and to disseminate, widely and on a timely basis, information for the rated companies to potential investors. Since 1992, Argentina has required that all publicly offered securities be rated by at least two authorized private agencies.
It is common practice for financial institutions and other enterprises to disclose to the public information on their financial health on a regular basis, usually quarterly. In 1993, Malaysia introduced new guidelines and minimum standards governing the disclosure of information by stockbrokerage and securities firms. The Hong Kong Securities and Futures Commission oversees information disclosure guidelines that were revised in 1993, setting out in detail standards and procedures expected of institutions in their recordkeeping and procedures for customer identification and reporting. Since 1992, the government of Argentina has also begun implementing stronger reporting, disclosure, and auditing requirements.
Internal Phasing of Capital Account Decontrol
The sequencing of capital account liberalization would appear to have received less attention in developing countries, partly because it has become increasingly difficult to institute a phased approach to decontrolling specific categories of the capital account, except perhaps for those controls that can be enforced outside the exchange system.54 Financial asset flows have become increasingly fungible.55 Rapid financial innovations and increasingly attractive investment opportunities abroad appear to have made it more difficult for those countries that had not liberalized earlier to maintain the effectiveness of their controls. It has also proved difficult to resist accelerating liberalization measures once a certain critical mass has been achieved and close substitutes for controlled capital operations have become available.56
It may be possible to maintain controls on those foreign capital transactions that relate to underlying transactions (e.g., inward direct investment flows and nonresidents’ real estate transactions) even while other controls are removed, because they can be monitored and enforced outside the exchange system. Indeed, many countries do maintain such controls. The structure of controls presently maintained by developing countries (which can be seen in Table 1 in Section VI)57 may also shed some light on this issue. As in industrial countries, controls on direct investment are far more common than other controls in developing countries, with controls on inward investment being more common than those on outward investment. This suggests that these controls are generally not intended to support the balance of payments. Other categories of control are also often maintained, with a slight majority of controls on outflows, possibly indicating that countries consider a mutually supportive set of controls important.
Capital Controls in Developing Countries
A total of 155 countries were surveyed.
Capital Controls in Developing Countries
| Category | Number of Countries Maintaining Controls1 |
|---|---|
| Any form of capital control | 119 |
| Comprehensive controls | 67 |
| On outflows | 67 |
| On inflows | 17 |
| Foreign direct investments | 107 |
| Of nonresidents | 84 |
| Of residents | 35 |
| Profit repatriation and capital liquidation | 34 |
| Taxes on capital transactions | 9 |
| Nonresident-controlled enterprises | 6 |
| Portfolio investments | 61 |
| Of nonresidents | 30 |
| Of residents | 33 |
| Security issuance by nonresidents | 15 |
| Security issuance abroad by residents | 6 |
| Debt-to-equity conversion | 2 |
| Financial transactions | 78 |
| Of nonresidents | 41 |
| Of residents | 66 |
| Trade-related financial transactions | 7 |
| Deposit requirements for borrowing from abroad by residents | 2 |
| Deposit accounts | 83 |
| Of nonresidents in foreign exchange | 37 |
| Of nonresidents in local currency | 52 |
| Of residents abroad | 29 |
| Of residents in foreign currency with domestic banks | 23 |
| Other capital transfers | 70 |
| Personal capital transfers | 34 |
| Blocked accounts | 24 |
| Real estate transactions | |
| Of nonresidents | 23 |
| Of residents | 30 |
A total of 155 countries were surveyed.
Capital Controls in Developing Countries
| Category | Number of Countries Maintaining Controls1 |
|---|---|
| Any form of capital control | 119 |
| Comprehensive controls | 67 |
| On outflows | 67 |
| On inflows | 17 |
| Foreign direct investments | 107 |
| Of nonresidents | 84 |
| Of residents | 35 |
| Profit repatriation and capital liquidation | 34 |
| Taxes on capital transactions | 9 |
| Nonresident-controlled enterprises | 6 |
| Portfolio investments | 61 |
| Of nonresidents | 30 |
| Of residents | 33 |
| Security issuance by nonresidents | 15 |
| Security issuance abroad by residents | 6 |
| Debt-to-equity conversion | 2 |
| Financial transactions | 78 |
| Of nonresidents | 41 |
| Of residents | 66 |
| Trade-related financial transactions | 7 |
| Deposit requirements for borrowing from abroad by residents | 2 |
| Deposit accounts | 83 |
| Of nonresidents in foreign exchange | 37 |
| Of nonresidents in local currency | 52 |
| Of residents abroad | 29 |
| Of residents in foreign currency with domestic banks | 23 |
| Other capital transfers | 70 |
| Personal capital transfers | 34 |
| Blocked accounts | 24 |
| Real estate transactions | |
| Of nonresidents | 23 |
| Of residents | 30 |
A total of 155 countries were surveyed.
Developments Following Liberalization of Capital Accounts
Initially, the balance of payments responded strongly to the liberalization of capital accounts in most of the developing countries reviewed. Rather than resulting in capital flight, the elimination of capital controls on both outflows and inflows, accompanied in most instances by a tightening of financial policies, would appear to have generated confidence and resulted in new inflows. For example, Argentina’s overall balance of payments position moved from a deficit equivalent to 7 percent of GDP in 1989 to a surplus of over 2 percent of GDP by 1993, while Venezuela’s balance of payments recovered from an overall deficit of 8 percent of GDP in 1988 to a surplus of 3 percent of GDP in 1991. In the other Latin American countries, overall balance of payments positions also improved sharply, although in some cases they deteriorated in later years (Costa Rica, El Salvador, Guyana, and Mexico). Because the inflows have reflected the confidence of the markets in the liberalized systems, as well as the general course of stabilization and restructuring, they have remained sensitive to macroeconomic policy reversals or failures.
Current account performance following liberalization was uneven, with deficits decreasing in some countries (El Salvador, Jamaica, and Malaysia) and increasing in others (Argentina, Estonia, The Gambia, Lithuania, Peru, and Singapore). To a certain extent, a larger current account deficit in the balance of payments would be expected, as credible reforms lead to larger capital inflows. International reserves tended to increase, and official reserves holdings grew in all countries except Costa Rica. Moreover, many countries that had accumulated substantial external payments arrears were now able to reduce or eliminate them altogether through cash payments or rescheduling and, more important, to avoid accumulating new arrears (except Costa Rica).58
Domestic policies were usually tightened as capital controls were eliminated, as part of a broad stabilization effort in the context of IMF-supported economic programs. Credit expansion was curbed in most countries (Argentina, Costa Rica, Estonia (initially), Guyana, Indonesia, Latvia, and Venezuela), although in some credit growth remained relatively high following the liberalization (Jamaica, Peru, and Singapore), and some countries experienced a rapid depreciation of the exchange rate (Jamaica and Peru).59 Only in El Salvador and Latvia did the nominal exchange rate appreciate. Moreover, fiscal policies were significantly tightened in most countries (Argentina, El Salvador, Estonia, The Gambia, Indonesia, Latvia, Lithuania, and Venezuela).60 Inflation generally declined immediately following convertibility (exceptions were Jamaica and Singapore), owing partly to the freeing up of import constraints associated with capital inflows and, in certain cases, to the effects of exchange rate appreciation or a slower rate of depreciation.
As noted earlier in this section, the environment for prudential supervision and regulation changes with the liberalization of the capital account. Reforms of market standards have therefore been undertaken to enhance the ability of the financial system to handle risk in an international setting while strengthening domestic banks to compete with foreign banks and reducing the overall need for government intervention. This process has involved the privatization of some banks and the licensing of new ones in El Salvador, Estonia, and Peru; and recapitalization, mergers, and closures in Costa Rica, El Salvador, and Peru. For this purpose, the responsibility for bank supervision and regulation has been clarified in a number of countries, including Costa Rica, El Salvador, Jamaica, and Venezuela. New bodies have also been established to regulate capital markets, particularly with regard to reporting, disclosure, and insider trading, and to streamline supervision.
In some cases, deficiencies in financial sector reforms (particularly in the areas of supervision and intervention) have created problems. In those cases in which banking sector problems intensified after rapid liberalization of the capital account (Argentina, Costa Rica, Latvia, and Venezuela), they appear to have reflected mainly magnified effects of preexisting weaknesses in the structure of banks’ balance sheets, including large volumes of nonper-forming loans and insufficient capital, and institutional weaknesses. Inadequate or delayed implementation of prudential reforms has spilled over into exchange markets in some instances (Indonesia and Venezuela). Bad banking practices and asset stripping were largely responsible for the recent problems in Latvia. In Venezuela, resort to exchange controls in the 1970s and 1980s contributed to extensive disintermediation and to the consequent weakness in the banking system, eventually culminating in the banking crises of early 1994. The resulting crisis of confidence and pressures on the exchange rate of the bolivar led to the reimposition of exchange controls. (For further discussion, see Section VII.)
Effectiveness of Maintaining and Reintroducing Control Mechanisms
Studies of the effectiveness of capital control mechanisms have suffered to some extent from a lack of agreement on what constitutes effectiveness.61 Controls have often had some impact in the sense that they have created differentials between domestic and international interest rates. The evidence accumulated, nevertheless, points to the general inefficacy of such controls, particularly on outflows, in maintaining an unsustainable exchange rate beyond a brief period.62 Although controls may contribute to some differential between domestic and foreign yields, the differentials tend to be small and relatively unimportant compared with the gains from successful speculative attacks on an exchange rate under pressure.
Continued Maintenance of Controls
Although the available estimates of capital flight show substantial variation—underlining the difficulty with inherently disguised flows—they suggest that capital controls have generally not been all that effective in curbing capital flight, partly because significant linkages have continued to exist between domestic and foreign markets. Different ways to define flight capital have been put forward in the literature.63 Some studies have defined capital flight as large domestic capital outflows, while others have argued that foreign capital precipitously pulled out of a country should also be included. The channels through which capital flight takes place are numerous and may also include legal channels that are used to make international payments and transfers (e.g., transfers through the banking system). Further, misinvoicing trade transactions and changing the terms and conditions under which trade financing is extended to importers and exporters are the main forms of large-scale capital flight in developing countries with extensive capital controls.
For a group of developing countries, mostly in Latin America, a variety of studies by IMF staff and others provide estimates of capital flight when these countries maintained capital controls in the early 1980s (Argentina, Chile, Mexico, Philippines, and Venezuela), supporting the view that the controls were not very effective. For instance, during 1982 in Argentina capital flight was estimated to have continued at a brisk rate, despite the reintroduction of exchange controls on capital transfers. Capital flight in Chile is estimated to have ranged from US$800 million to US$900 million during 1982 despite capital controls. In Mexico, several estimates show capital flight throughout 1976–84, although the estimated peak year varies. Similarly, capital flight from the Philippines is estimated to have been sizable in the period following substantial tightening of capital controls in 1993. In Venezuela, capital controls were reintroduced in 1983, but capital flight continued, with estimates ranging from US$1 billion to as high as US$5 billion. Most available estimates suggest that the capital flight occurred despite the maintenance of controls on capital flows.
In the context of IMF technical assistance to member countries that has focused on the elimination of capital controls, the IMF staff has often provided estimates of the magnitude of capital flight, emphasizing that controls on capital outflows have been largely ineffective and that their elimination would recognize the de facto situation. In most of these cases, estimates of capital flight ranged between 20 percent and 30 percent of export receipts (Nigeria, 1981–84; Guatemala, 1982–88; Honduras, 1980–82; Jamaica, 1983–89; and Venezuela, 1984–86). In a few cases the magnitudes were even larger, for example in Egypt, where the estimate was close to two-thirds of export receipts; in Trinidad and Tobago, where it amounted to more than one-half of export receipts; and in Somalia, where such estimates equaled annual export receipts. It also appears that the scale of capital flight often intensified prior to liberalization (e.g., in Nigeria and Venezuela). In many countries, the main channel for capital flight was identified as trade misinvoicing (Egypt, Guatemala, Honduras, Jamaica, Nigeria, Somalia, and Trinidad and Tobago), although in some cases capital flight took the form of short-term outflows (El Salvador, Nigeria, and Venezuela) or private transfers (Fiji).
Reimposition of Controls
Reimposing controls on capital outflows would appear to reintroduce the disadvantages of maintaining controls on outflows discussed earlier. In particular, the recent experience of Venezuela (reviewed in Section VII) has been consistent with the general findings regarding their effectiveness. However, the phenomenon of large capital inflows associated with renewed access to international financial markets in the context of stabilization and liberalization has raised the question of whether temporary capital controls on inflows may be useful to deal with short-term shocks to the balance of payments or complications to monetary management. The broad conclusion drawn here is that controls on capital inflows cannot substitute for adjustments to fundamental macroeconomic policies, such as fiscal policy and exchange rate policy, and can contribute to distortions and inefficiency. The temporary use of such controls may have provided a brief breathing space to the authorities, allowing them to take necessary macroeconomic policy measures, although the available evidence is not conclusive.
Both industrial and developing countries that have adopted full convertibility have occasionally reintroduced controls on capital transactions, sometimes in conjunction with controls affecting current transactions. One rationale put forward for reintroducing the controls has been that they can help protect the balance of payments position. Alternatively, the controls might serve the longer-term and more systemic purpose of ensuring greater independence of financial policies and exchange rate stability. The success of such policy reversals is examined here for a group of countries that had reintroduced exchange controls on capital transactions after the onset of the debt crises in the early 1980s. The experiences in the 1990s of Chile, Colombia, Malaysia, and Venezuela are the subject of individual case studies in Section VII.
Controls on Outflows
Argentina, Chile, Mexico, and Venezuela reintroduced controls on capital transactions in the early stages of the 1980s debt crisis, as capital outflows mounted. Although these countries introduced extensive controls on international capital transactions (in isolation or in conjunction with exchange controls on current international transactions), they were generally unable to avert balance of payments crises or to sustain overvalued nominal exchange rates. The large-scale capital flight that continued through this period of capital controls is well documented.
The common factor behind the problems facing these countries was poor economic policies. Loose financial policies generally led to an appreciation of real exchange rates, eroded external competitiveness, and contributed to widening current account deficits. In the early stages, the current account deficit was usually financed by capital inflows so that the overall balance of payments remained in surplus. As it became clear that policies were inconsistent with a sustainable balance of payments position, expectations of devaluation started to mount, contributing to private capital outflows. Capital flight continued despite exchange controls on capital transactions, as evidenced by large negative errors and omissions in the balance of payments and estimates of misinvoicing from partner country trade data. Further, as a result of restrictions on current and capital account transactions, parallel market premiums emerged. Therefore, despite the policy autonomy that exchange controls were intended to provide, capital flows remained sensitive to changes in monetary and fiscal policies, and controls were increasingly circumvented through deepening parallel markets. Subsequent tightening of fiscal and monetary policies, together with exchange rate adjustments, generally led to a rapid improvement in the overall balance of payments as the direction of capital flows was reversed. Overall, controls on capital outflows did not succeed in isolating the economy from external developments or in protecting the balance of payments position. Further, the controls often gave rise to distortions in financial markets, inhibiting intermediation and development of those markets, as well as promoting tax evasion. Venezuela, which reintroduced exchange controls in 1994, has had a similar experience.
An alternative perspective on the effectiveness of capital controls is provided by the experience of the group of countries that varied the extent of controls on capital account transactions over an extended period (Egypt, Hungary, India, Korea, Nigeria, Turkey, and ZaÏre). Despite their structural differences, these countries show some similarity in the composition of capital account transactions, as capital flows appeared to be dominated by official transactions. Recorded private sector capital flows, including portfolio flows and foreign direct investments, appeared rather modest. However, other private flows that may be similar to capital flows were generally significant (e.g., trade-servicing transactions or workers’ remittances). Moreover, it would seem that these flows were in fact relatively sensitive to changes in the authorities’ policies, both fiscal and monetary. In several instances, as confidence in the authorities’ policies weakened, outflows associated with workers’ remittances intensified, while balances held in domestic foreign currency accounts were quickly remitted offshore. Capital flight, usually in the form of over- and underinvoicing of imports and exports, accelerated in response to inconsistent policies. Official reserve positions deteriorated as the authorities attempted to defend overvalued nominal exchange rates, and external payments difficulties emerged, leading to the accumulation of external payments arrears (Egypt, Nigeria, Turkey, and Zaire).
Controls on Inflows
Some countries that have experienced destabilizing surges of capital inflows have resorted to exchange controls or related incentives to help cope with them.64 The inflows have generally resulted from a combination of successful economic stabilization, liberalization, improved confidence in economic policies and performance, renewed access to international markets, distortions in domestic financial markets that resulted in high real interest rates, and speculation regarding exchange rate appreciation. In some cases, costless forward cover facilities provided by central banks have also played a role.65
Countries have generally relied on a mix of quantitative restrictions and indirect measures aimed at influencing effective interest rate differentials or incentives for banks to accept nonresident funds. Indirect measures have included differential reserve requirements on nonresident deposits in domestic banks, unremunerated reserve requirements on foreign borrowing that raised the effective interest rate paid, interest rate caps, and the assessment of taxes on the transfer of marketable instruments.
The evidence on the effectiveness of the relatively few countries with controls targeting surges of capital inflows is not conclusive. Often, countries that have applied controls on capital inflows have taken other policy measures concurrently to counter the inflows, making it difficult to dissociate the impact of the controls. While there is some evidence that short-term inflows may have been reduced by the controls targeted at them, they may often have been replaced by capital flows with nominally longer-term maturities. Thus, the potentially expansionary effects of the inflows on liquidity may not have been effectively curbed.66
Malaysia recorded increasing capital inflows in the early 1990s in response to widening interest rate differentials in its favor, heightened foreign interest in Malaysia’s stock market, and speculation of a pending appreciation of the ringgit. Tightening monetary policy to limit the inflationary consequences of these inflows encouraged further inflows, and in September 1991, statutory reserve and liquidity requirements were imposed on ringgit borrowing in order to equalize the regulatory costs between offshore and domestic sources. Despite these and other measures, as well as a depreciation of the ringgit, the inflows persisted, although they slowed as domestic liquidity conditions were eased in 1994 and the currency was allowed to appreciate in nominal terms. Further controls on inflows were introduced between January and August 1994.
Chile took steps to limit temporary capital inflows while maintaining an attractive environment for longer-term foreign capital. In early 1992, the 20 percent reserve requirement on external loans was extended to foreign exchange deposits, while some controls on outflows were removed. The reserve requirement, which applied for one year regardless of the maturity of the loan or deposit, was designed to discourage temporary inflows. Increased commissions were applied to swap operations. At the same time, the central bank conducted open market operations aimed at sterilizing the capital inflows and also gave greater flexibility to the exchange rate. The reserve requirement on bank borrowing abroad was subsequently increased to 30 percent in May 1992, as was the requirement on foreign borrowings by Chilean companies in August 1992. Despite these measures and the relaxation of controls on capital outflows, the capital account surplus continued to increase in 1992 and remained high in 1993. The controls may have led to a change in the composition of inflows in 1993, as short-term inflows slowed while long-term inflows strengthened.
Overall, the effectiveness of capital controls, particularly on outflows, is likely to be limited.67 Countries that have introduced effective multiple exchange rates to deal with the problem of capital inflows have had similar experiences. Nevertheless, there are circumstances in which controls or incentives can provide a breathing space during which an appropriate macroeconomic policy response can be formulated and policy measures can be taken. The OECD Code of Liberalization of Capital Movements provides explicitly for the introduction of such emergency measures, although subject to strict time limitations.68 However, even with such time limits, the usefulness of a reprieve from inflows is likely to be limited. The ability to circumvent controls will make any breathing space very short, and the amount of policy adjustment that can be introduced in such a time frame (say, a few months) will be correspondingly lessened.
IMF Policy Treatment
The IMF has generally welcomed the liberalization of restrictions on capital account transactions in developing countries. This support, however, has been selective, with particular prominence being given to issues related to capital controls, for example, where capital movements were a significant factor in macroeconomic developments and exchange rate management. The IMF’s policy advice has focused on appropriate adjustments in fiscal, monetary, and exchange rate policies in response to large capital inflows while generally discouraging the tightening of controls over capital movements. An acceleration of the liberalization process has been sought when conditions were deemed appropriate, for example, if it was felt that further liberalization of capital outflows could help mitigate the macroeconomic complications associated with strong inflows. In other cases, however, a more measured and gradual approach to liberalization of inflows was supported.
The IMF also supported capital account liberalization in countries where significant private and public capital inflows were accompanied by large current account surpluses. In one such case, the authorities worked very closely with the private sector to abolish eventually the existing capital controls and to unify the financial and commercial exchange rates. The IMF took the view that a timely removal of the remaining capital controls would be welcomed while noting that their abolition might require supporting measures to stimulate the development of a full-fledged capital market in the country.
On several occasions, concerns have been expressed about the imposition of capital controls. In one case, an IMF member introduced controls on short-term capital flows to stabilize market conditions, and the authorities believed that they had succeeded in halting speculative flows. Nonetheless, the IMF felt that such measures introduced distortions and were not desirable in the longer term and supported subsequent decisions to remove the controls. In another case, extensive exchange controls were imposed in response to instability created by a banking crisis. This resulted in substantial delays in processing foreign exchange transactions and contributed to the accumulation of large external arrears and to the curtailment of access to private capital markets. Concerns about the persistent accumulation of external arrears and the serious distortions created by exchange controls led the IMF to support the elimination of exchange controls while recognizing that this could be achieved only in the context of strong fiscal and credit policies and might need to be implemented in a phased manner.
Notwithstanding this general approach, however, it is recognized that, in individual cases, special considerations might justify a more favorable view of measures restraining capital inflows. In one case where the member was faced with a high external debt burden, the authorities introduced controls on public sector borrowing by all public sector entities, including those with no more than indirect links to the state, although controls were not imposed on private sector borrowing. The IMF endorsed the authorities’ determination to maintain control over publicly related borrowing.69 More generally, on the part of the authorities, a more measured approach to capital account convertibility has reflected their desire to protect independence in conducting monetary policy. Chile, for instance, used additional instruments to create disincentives to short-term capital inflows.70
IMF-supported programs have played a role in external sector liberalization in the transition economies of Eastern and Central Europe.71 A rapid liberalization of certain capital account transactions has been largely motivated by the expectation of significant private capital inflows and, in particular, of private foreign direct investment and the associated managerial and technological resources that are considered critical for the transformation process.
Virtually all IMF-supported programs in Eastern and Central Europe included measures to facilitate the inflow of foreign capital, such as the passage of liberal foreign investment codes.72 These measures were further strengthened by opening the privatization process to foreign investors and often granting them treatment equal to domestic investors.73 When the liberalization of foreign investment regimes was not proceeding as rapidly as originally agreed, the IMF recognizing the contribution that foreign investment could make in the transformation process, has at times encouraged the authorities to take a more positive and consistent view of the role that foreign capital could play in the economy. In the assessment of progress in external liberalization, the IMF has generally endorsed up-front measures to liberalize the external trade and payments system. However, it has also been recognized in some instances that restrictions on capital outflows can be maintained until financial imbalances are reduced and the external reserve position improved.
The IMF programs have frequently included provisions regarding certain restrictions pertaining to current account transactions but that may also have significant implications for the capital account. Measures that require foreign exchange repatriation or surrender requirements can often be intended to restrict residents’ ability to engage in capital account transactions. Without such requirements, effective restrictions on capital outflows may be difficult to enforce regardless of the intensity of restrictions on current and capital account transactions, where the monitoring and control of capital transactions are relatively weak. However, the same weakness also tends to affect the enforcement of the surrender and repatriation requirements.
In programs in Eastern and Central Europe, the IMF has pursued a case-by-case approach with regard to exchange repatriation, surrender requirements, and internal convertibility. In some cases, the IMF has argued against the introduction of surrender requirements, stressing that in a proper policy environment a surrender requirement would not be necessary and, therefore, could not be positively recommended. In other cases, IMF programs have included commitments to abolish surrender requirements, with several countries also permitting banks to offer foreign exchange accounts.
In Czechoslovakia and its successor states, the Czech Republic and the Slovak Republic, the IMF encouraged the liberalization of transactions on current and capital accounts. In the early 1990s, Czechoslovakia pledged to introduce a very liberal foreign investment regime under which foreign companies were subject to less restrictive regulations than those applicable to domestic enterprises as regards repatriation of profits, transfers of capital participation, and borrowing from foreign banks; it also permitted the establishment of majority- or wholly foreign-owned companies. Following the breakup, the Czech Republic maintained its commitment to further liberalize capital account transactions. The Slovak Republic, however, introduced some restrictions relating largely to external trade financing in response to the virtual depletion of its foreign exchange reserves. In that context, the IMF pointed to the high costs of a strategy that had included reliance on external controls. The restrictions were removed in December 1993.
On occasion, the IMF has encouraged an accelerated transition to full capital account convertibility. In one case, where such an objective was explicitly included as one of the principal elements of the program, it was motivated by the openness of the economy in the context of limited administrative capacity. Rapid liberalization of capital outflows has also at times been considered as a possible measure that could help alleviate the impact of large capital inflows experienced by some countries. Such a measure can be particularly useful when the authorities’ ability to sterilize inflows is limited and because of the likely adverse impact of exchange rate appreciation on competitiveness.
Technical Assistance
One focus of the IMF’s technical assistance in the area of foreign exchange systems has traditionally been its efforts to facilitate current account convertibility in its member countries. However, since the mid-1980s, technical assistance has also involved advice on the adoption of full current and capital account convertibility, and exchange system reforms have been linked frequently to the establishment of full convertibility.74
In view of extensive capital flight and disintermediation of foreign exchange identified in the context of technical assistance, one approach that has on occasion been adopted to achieve full convertibility has included the legalization of a parallel (illegal) market, thereby generalizing a free market for foreign exchange. One of the first cases in which convertibility for both current and capital account transactions was proposed by the IMF staff as part of the reform program was in 1986, when a transitional two-tier exchange system was introduced with the IMF’s technical assistance.75 A move to full convertibility was generally indicated by the lack of transparency and inefficiencies associated with controls, which had often already been undermined by extensive parallel market trading. Formal elimination of controls was seen as bringing realism to the exchange system and providing proper incentives for the private sector to rationalize the allocation of foreign exchange resources.
Because the free markets were to be built on existing parallel markets, the exchange rate regimes proposed in the context of technical assistance were usually floating ones (Egypt, El Salvador, Guatemala, Honduras, Jamaica, Korea, Nigeria, and Venezuela). However, in each case, country-specific circumstances had to be considered. In some cases, there was a secondary market free of exchange restrictions, and foreign exchange purchases for both current and capital account transactions could be effected freely in this market. The official exchange rate, however, was generally significantly overvalued and foreign exchange at the subsidized rate was rationed, subject to extensive leakages and exemptions. The unification of official and parallel market rates was therefore seen as playing an important role in ending the implicit subsidies and thus in overall fiscal adjustment supporting convertibility.
In one case, most banks were state owned, and collusion in the banking sector was a concern; arrangements in the free market were seen as being able to provide a benchmark to establish the exchange regime for the banks. In another, where a unified exchange system had been introduced earlier, the spread between the official and illegal market rates was negligible. This suggested that de facto convertibility was already in effect and that formal abolition of the controls would make the exchange system more transparent and improve confidence. In several countries, it was recommended that a free market be established on the basis of the parallel market, because most foreign exchange transactions were already channeled through this market. The IMF staff recommended full convertibility in one case as a means to ensure a unified exchange system after the introduction of an auction market for the distribution of foreign aid-related funds in 1986, because there was concern that multiple exchange rates could emerge without the elimination of all exchange controls.
Other technical assistance in the area of capital convertibility has placed less emphasis on the existing role of the informal market (Czech Republic, Cyprus, Fiji, Korea, Lithuania, Malta, Sri Lanka, and Trinidad and Tobago). Lithuania is an example of a transition economy in which market-based systems did not exist initially and had to be developed. It was recommended that a fully convertible exchange system be established to facilitate integration into international markets and to make resource allocation efficient. Technical assistance was provided to the authorities in drafting a currency law that included capital convertibility provisions.
In some other cases, the elimination of remaining controls on capital account transactions was perceived by the IMF staff as a logical extension of earlier reforms involving the elimination of controls affecting current account transactions. Requests for technical assistance in eliminating controls on capital transactions in the Czech Republic, Korea, and Malta were motivated in part by these countries’ desire for membership in another international or regional organization (OECD in the case of the Czech Republic and Korea, and the EU in the case of Malta). The IMF staff’s advice in these instances focused on the possible sequencing of reforms leading to full convertibility.
While recommendations for full convertibility in the context of technical assistance reflected the diverse circumstances of countries, some common themes can be found. In virtually all countries, the ineffectiveness of existing controls was evidenced by estimates of capital flight. The improved transparency associated with a free exchange system and the recognition of an active informal market were also key considerations in the recommendations. Furthermore, the promotion of a competitive and efficient exchange system was also frequently a factor.
Specific rationales for promoting capital convertibility in individual countries have included the removal of exchange controls to facilitate investment and growth as the mismatch between import and production capacity and the underutilization of installed capital was eliminated; unification and decontrol of interest rates to provide incentives for re-flows; elimination of the implicit subsidy resulting from an overvalued exchange rate that distorts income distribution and creates a bias toward imported goods away from the use of domestic labor; decontrol and better monetary and fiscal policies for more efficient allocation of resources and to facilitate import and export substitution; improved transparency as regulations are brought into line with reality; capital decontrol to remove incentives for massive capital flight; most transactions are already handled through parallel markets; convertibility facilitates integration into the world economy; controls prevent efficient pricing and allocation; capital liberalization would be a natural extension of the reform process; elimination of capital controls would help establish the same degree of liberalization as in neighboring countries; macroeconomic conditions are conducive for the move to full convertibility; liberalization is likely to encourage competition in the financial sector; to promote policy discipline; and to release pressure on monetary policy.
The IMF staff has also provided technical assistance in support of convertibility involving related operational aspects, such as linkages with monetary management, development of instruments and markets for securities, reserves management, reporting systems, and the development of interbank market operations in foreign exchange. Technical assistance on exchange systems since 1988 has also included recommendations regarding mechanisms for managing liquidity in anticipation of capital inflows resulting from the liberalizations, including analysis of money demand and management of sterilization operations. Recent technical assistance provided to the Czech Republic, which was already experiencing strong capital inflows, emphasizes the need to continue the liberalization of exchange controls and the domestic financial markets in order to minimize arbitrage opportunities and ensure optimal portfolio allocation.
In countries with floating exchange rates, the emergence of a forward market for foreign currency was a necessary underpinning for convertibility. Such markets have proved sensitive to the overall state of openness of financial systems and have therefore been slow to emerge in developing countries, although the benefits associated with forward cover in offsetting exchange rate risks are widely understood. The IMF staff has generally noted that the development of a forward market arrangement by the private sector is helped by market-based exchange rate regimes and flexible interest rates. At the same time, the staff has cautioned strongly against the provision of forward cover by official entities (such as the central bank), because such guarantees have often resulted in substantial quasi-fiscal losses.
A special problem has emerged in some developing countries resulting from large backlogs of exchange applications or accumulations of external payments arrears. Technical advice on convertibility in such cases has often involved sequestering the overhang in order to ensure that the freely convertible foreign currency market functions properly.
Indonesia’s current account was in deficit when it liberalized in 1971, but the deficit was small and the overall balance of payments was in surplus.
Almost all liberalizing countries had floating exchange rate regimes.
For a discussion of sequencing issues, see S. Edwards, “On the Sequencing of Structural Reforms,” NBER Working Paper No. 3138 (Cambridge, Massachusetts: National Bureau of Economic Research, 1989); and P.J. Quirk, “Capital Account Convertibility: A New Model for Developing Countries,” in Frameworks for Monetary Stability, ed. by T. Balino and C. Cottarelli (Washington: International Monetary Fund, 1994).
In Mauritius, convertibility was preceded by freeing up export surrender requirements.
Venezuela subsequently (five years after liberalizing) reim-posed exchange controls (see Section VII).
In countries with currency board-like arrangements, less priority was accorded to open market operations. For instance, in Argentina, the purpose of open market operations was mainly to smooth out sharp fluctuations in short-term interest rates and international reserves.
Central banks in Venezuela and Lithuania offered deposit facilities to market participants. However, the facilities were discontinued in Lithuania with the introduction of the currency board and were replaced by sales of central bank certifications in Venezuela.
For a discussion of this issue, see H. Reisen, “On Liberalizing the Capital Account: Experiences with Different Exchange Rate Regimes,” Discussion Paper (Lugano: Organization for Economic Cooperation and Development and University of Basle, 1991).
For example, the 1994 banking crisis in Venezuela revealed shortcomings in the accounting rules covering such areas as capital adequacy, loan classification, provisioning, and lending to related parties.
For those industrial countries that liberalized controls over extended periods (e.g., Denmark, Japan, and the Netherlands), liberalization generally began with less volatile transactions and with those more directly necessary to normal business activities, so as to soften the short-term impact of liberalization of capital transactions on the balance of payments and the economy. Thus, outward direct investments were usually authorized sooner than portfolio investments abroad, and trade credits were liberalized before financial loans; transactions in shares were liberalized be-fore those in interest-bearing securities and, when these were liberalized, transactions in long-term bonds were permitted first, while controls on transactions in money market instruments were maintained as long as possible.
See Claessens, Dooley, and Warner, Portfolio Capital Flows, pp. 153–74.
These problems are not new: as noted above, Germany attempted to impose partial controls in the 1970s but was increasingly obligated to widen the coverage of the controls as loopholes developed, until the controls were ultimately removed.
The categories of capital transactions in the survey reported in the table are broadly those employed by both the OECD and the EU in their codes for capital transactions.
Costa Rica’s gross foreign reserves fell slightly in 1993 following the liberalization package introduced in 1992 and then deteriorated further in 1994 as the fiscal and current account imbalances widened, which contributed to a drop-off in private capital inflows.
Fiscal policy was already tight in Estonia and Latvia.
Argentina’s public sector balance switched to a surplus, while the fiscal improvement in Venezuela was only temporary as a significant turnaround took place in 1991.
Dooley, Survey of Academic Literature.
See, for example, Mathieson and Rojas-Suarez, Liberalization of the Capital Account; N. Fieleke, “International Capital Transactions: Should They Be Restricted?” IMF Paper on Policy Analysis and Assessment 93/20 (Washington: International Monetary Fund, December 1993); and R.B. Johnston and C. Ryan, “The Impact of Controls on Capital Movements on the Private Capital Accounts of Countries’ Balance of Payments: Empirical Estimates and Policy Implications,” IMF Working Paper 94/78 (Washington: International Monetary Fund, July 1994).
See, for example, R. Cumby and R. Levich, “On the Definition and Magnitude of Recent Capital Flight,” in Capital Flight and Third World Debt, ed. by D.R. Lessard and J. Williamson (Washington: Institute for International Economics, 1987); M.P. Dooley, “Capital Flight: A Response to Differences in Financial Risks,” Staff Papers, International Monetary Fund, Vol. 35 (September 1988); and J.T. Cuddington, “Capital Flight: Estimates, Is-sues, and Explanations,” Princeton Studies in International Finance No. 58 (Princeton, New Jersey: Princeton University, 1987).
Most recently, Brazil, Chile, Colombia, Korea, Malaysia, Mexico, and Spain.
A 1988 IMF study of practices in forward exchange markets concluded that assumption of exchange rate risks by governments had led to large fiscal or quasi-fiscal losses in many countries, in some representing multiples of the monetary base. See P.J. Quirk, Graham Hacche, Viktor Schoofs, and Lothar Weniger, Policies for Developing Forward Foreign Exchange Markets, IMF Occasional Paper 60 (Washington: International Monetary Fund, 1988), pp. 17 and 21–23. Since the mid-1980s, such practices have been far less prevalent, although losses by the Mexican government on issuance of dollar-denominated tesobonos in 1994 could be as high as 5 percent of GDP (at the present exchange rate).
Case studies of such controls in Chile, Colombia, and Malaysia are discussed in Section VII.
This view is supported by the experiences, discussed above, of Ireland and Portugal in 1992.
Article 7 allows for the introduction of temporary restrictions in response to balance of payments difficulties. See Section V for further details.
Controls imposed by government on its own operations do not have the same economic significance as capital controls.
See Section VII.
This review covers IMF-supported programs in Albania, Bulgaria, the Czech Republic and the Slovak Republic, Estonia, Hungary, Latvia, Lithuania, Poland, and Romania.
These typically included the right to unrestricted remittance of profits abroad and provisions for liberal repatriation of capital in the event of liquidation.
In some cases, large investments in strategic sectors and investment in large public sector enterprises required special approval.
Countries to which the IMF has provided technical assistance in the area of capital account liberalization include the following: Liberia (1985), Nigeria (1986), Somalia (1986), Dominican Republic (1987), Venezuela (1988), El Salvador (1989), Guatemala (1989), Korea (1989, 1992, and 1994), Honduras (1990), Jamaica (1990), Egypt (1990–91), Lithuania (1992), Trinidad and Tobago (1993), Fiji (1993), Cyprus (1994), Malta (1994), Sri Lanka (1994), and Czech Republic (1995).
Regional convertibility had been discussed in 1982 in the context of an intraregional monetary arrangement. Technical assistance in the area of capital account convertibility was provided by the Exchange and Trade Relations Department (ETR) until 1992, when this function was passed to MAE.
Note: This section was prepared by Przemyslaw Gajdeczka and Alexander Kyei. It is drawn in part from material in European Economy, Directorate-General for Economic and Financial Affairs, European Commission, Brussels, No. 26 (May 1988); and Liberalization of Capital Movements and Financial Services in the OECD Area (Paris: OECD 1990).