Abstract

The sufficiency of financing is key to the viability of any IMF-supported program. The Asian crisis countries' estimated financing needs were heavily dominated by the capital account and in particular the assumed rollover rate on short-term foreign debt.1 The size of the short-term liabilities was such that it was essential that creditors roll over at least a good part of their positions. Inducing them to do so required persuading them that the programs would work, snowing that there was enough official money available to make them work, and suggesting that pulling money out unilaterally would not be in their longer-term interests—a particularly difficult task when dealing with short-term, fixed-value credits.

The sufficiency of financing is key to the viability of any IMF-supported program. The Asian crisis countries' estimated financing needs were heavily dominated by the capital account and in particular the assumed rollover rate on short-term foreign debt.1 The size of the short-term liabilities was such that it was essential that creditors roll over at least a good part of their positions. Inducing them to do so required persuading them that the programs would work, snowing that there was enough official money available to make them work, and suggesting that pulling money out unilaterally would not be in their longer-term interests—a particularly difficult task when dealing with short-term, fixed-value credits.

In each of the programs (but particularly in Indonesia and Korea) very large official financing packages, together with sound economic policies, were intended to restore confidence and limit private capital outflows. However, the programs were not initially successful in restoring confidence, and private capital outflows far exceeded program projections. Several factors contributed to weak confidence, including hesitant program implementation, political uncertainties, and other factors casting doubt on the authorities' ownership of the programs, the revelation of market-sensitive information, problems with the coverage of government guarantees, and uncertainties surrounding the financing packages.

In the event, the financing available was inadequate to protect the programs from a failure to restore market confidence quickly. This suggests two main alternatives. One would have been a larger official financing package (or greater front-loading of the packages), although this was limited by resource constraints and moral hazard concerns. A second would have been earlier concerted involvement of the private sector; such action could have been considered at an earlier stage in these countries, but there are no straightforward mechanisms to assure such involvement, and the attempt, at a moment of nervousness across the emerging market countries, could have increased the risk of contagion. The experience, which is discussed below, underscores the importance of ongoing work on international financial architecture, including more effective ways of involving the private sector in the resolution of financial crisis.

Official Financing and Program Projections

The approach taken in these cases involved trying to strike a delicate balance, with a promise of official financing that, although large, was far from sufficient to constitute a guarantee of external liabilities. It was hoped that this commitment, together with firm policy implementation in line with the programs, would elicit a spontaneous response from private market participants such that the official financing package would not be needed, at least not in full. The alternatives would have been to withhold support or initiate a more formal approach to private creditors (to the extent that they could be identified and organized) to keep their money in place.

Given the size and openness of the countries involved and thus the enormous potential volume of capital outflows, the amount of financing provided in these packages had few precedents (Table 4.1).2 In addition to the financing provided by the IMF itself, large amounts of bilateral and other multilateral (the World Bank and the Asian Development Bank) support were pledged, which exceeded the IMF's support. In Thailand, bilateral funds amounting to $10.5 billion were part of the package and have been disbursed in step with the IMF's resources. In Indonesia and Korea, a “second line of defense” was pledged by bilateral creditors, which, however, had not been disbursed as of October 1998.

Table 4.1.

Official Financing

article image

Duration of original arrangements was 36 months for Indonesia and Korea and 34 months for Thailand.

Original financing package, not including augmentations since July 1998.

Large as this official financing was, it would have been sufficient to support the programs in Indonesia and Korea only on the assumption that they would elicit a broadly positive response on hte part of private markets, especially in the initial phase of the programs.3 It was known, however, that there was a substantial risk that private capital outflows would turn out larger than assumed. If this risk materialized and the program financing thus turned out to be inadequate, the perceptions that had led investors to panic in the first place would be confirmed, leading into a vicious circle. This risk was compounded by the fact that much of the external debt outstanding was of private corporations and banks: this meant that external creditors were less easily reassured by the IMF's and other commitments of official resources—which left these creditors with significant uncertainties about individual debtors' solvency and ability to finance the purchases of foreign currency needed to service their debts.

Given that unpredictable private capital outflows were central to program financing, an obvious question is whether more direct action should have been taken at an earlier stage to limit these outflows by attempting a rescheduling of private external debt (Box 4.1 reviews attempts to involve the private sector in the three countries). Programs in Korea and Indonesia were formulated without any advance agreement to restructure debt. Such agreements were discussed with major private external creditors only at a later stage, Korea brought about an effective standstill on bank debt in late December 1997 and made a provisional agreement with its private bank creditors in late lanuary 1998. Indonesia concluded an agreement with private creditors in late June. In Thailand, however, the authorities at the outset received certain assurances and indications regarding maintenance of credit lines of foreign banks resident in the country.4 Would it have been desirable—if it had been possible in the limited time available—for the IMF to have exerted pressure to bring about a refinancing or rescheduling beforehand in the other countries? This would likely have lengthened and complicated program negotiations by turning them into (at least) a three-way process, and (particularly in Korea) there was not much time before a moratorium and/or exchange controls would have had to be activated. Greater assurances that the countries would be adequately financed might have been worth some additional delay. On the other hand, there was concern that an approach that was perceived as “heavy-handed” could both precipitate greater capital flight from the countries immediately concerned and unsettle conditions for market access by other countries both in and beyond the region—countries that were themselves already under some pressure from the market uncertainties created by the Asian crisis. These issues are among the thorniest being addressed in current discussions of the international financial architecture.

Involving Private Sector Creditors

In all three countries, uncertainty about the rollover of short-term foreign debt presented a major risk to the programs. The initial focus was on the restoration of confidence through convincing packages of policies and official financing to induce private creditors to maintain their exposure voluntarily. At different stages, more direct action was also taken to involve private creditors in the closing of financing gaps. The form and timing of this involvement reflected the specific circumstances of each country.

Thailand

In Thailand, steps were taken at the start of the program to encourage the rollover of a significant part of maturing short-term debt. These steps were facilitated by the fact that some two-thirds of total short-term debt outstanding prior to the program was owed by foreign bank branches and subsidiaries, mainly of Japanese banks.

In August 1997, shortly before the Stand-By Arrangement was approved by the IMF's Executive Board, the Thai authorities received assurances and indications that credit lines of foreign banks resident in Thailand, the bulk of which involved Japanese banks on the creditor and debtor side, would be maintained. As a result, rollover rates for short-term obligations of foreign banks in Thailand remained high through April 1998, but subsequently declined, mainly reflecting problems at home of Japanese creditor banks. The rollover of short-term obligations of Thai banks and corporations meanwhile declined sharply in early 1998, but recovered by midyear.

Korea

In Korea, the financial sector accounted for the bulk of short-term foreign liabilities, but unlike in Thailand, most of the short-term debt (over half at the end of November 1997) was owed by domestic financial institutions (including their overseas branches and subsidiaries) and the geographical distribution of creditor banks was more dispersed. Efforts to involve private sector creditors were thus likely to be more complicated.

When the original Stand-By Arrangement with Korea was approved by the IMF's Executive Board on December 4, 1997, it was expected that the program, combined with the announcement of a large financing package, would turn around market sentiment. Talks with private sector creditors were not envisaged.

In late December, however, with rollover of short-term debt down sharply and usable official reserves effectively depleted notwithstanding the injection of about $10 billion from the IMF, discussions with creditor banks became critical. Talks in Japan, the United States, and Europe led to voluntary cooperative understandings on the maintenance of interbank credit lines to Korea through end-March 1998. At the same time, discussions on a framework for voluntary restructuring of short-term debt were initiated. A detailed debt-monitoring system was set up to track daily rollover rates. In early January, rollover rates rose significantly.

On January 28, 1998, the Korean authorities reached an agreement in principle with a committee of foreign banks on a voluntary restructuring of the short-term debt of 33 commercial and specialized banks (including their overseas branches) as well as certain merchant banks. The eligible debt, amounting to some $24 billion, covered interbank obligations and short-term loans maturing during 1998.

The debt-restructuring agreement was signed on March 31, 1998, with 134 creditor banks from 32 countries tendering loans and deposits amounting to $21.8 billion. The original obligations were exchanged for government-guaranteed debt of one-year maturity at 225 basis points over the London interbank offered rate (LIBOR) (17 percent of total), two-year maturity at 250 basis points over LIBOR (45 percent of total), and three-year maturity at 275 basis points over LIBOR (38 percent of total). As a result of the debt restructuring, Korea's short-term debt declined from $61 billion at end-March to $42 billion at end-April 1998.

Indonesia

The original Stand-By Arrangement with Indonesia, which was approved by the IMF's Executive Board on November 5, 1997, assumed that the official financing package, supplemented by part of Indonesia's own reserves, would be sufficient to cope even with a relatively large decline in the rollover of short-term debt. At the time, steps to restructure external obligations did not seem pressing. Moreover, with nearly half of total external debt (three-fourths of private external debt) owed by private corporations, efforts to involve private sector creditors were likely to be particularly complicated.

With the deepening of the crisis and continued depreciation of the rupiah in late 1997 and early 1998, how-ever, the external debt of the private sector became an issue that had to be addressed. Talks with a steering committee of private bank creditors began in February 1998, followed by meetings in April (New York), May (Tokyo), and June (Frankfurt). A private external debt team set up by the authorities prepared and coordinated the negotiations with assistance from outside consultants and in collaboration with the IMF, the World Bank, and the Asian Development Bank. In addition, a system was established to monitor daily the rollover of short-term interbank credit lines.

On June 4, 1998, the Indonesian authorities reached an agreement with the steering committee of creditor banks on a multifaceted deal to support the restructuring of the external debt of the banking and corporate sectors. The agreement on interbank debt involved an offer to exchange short-, medium-, and long-term obligations maturing by end-March 1999 against new loans carrying a full dollar guarantee from Bank Indonesia and maturities of one year (not more than 15 percent of the new loans) to four years (at least 10 percent of the new loans), with interest rates ranging form 275 to 350 basis points over LIBOR. Regarding trade credit, participating banks agreed to use their best effort to maintain, for the period of one year, aggregate credit to Indonesian banks at the level outstanding at end-April 1998.

The agreement on corporate debt provided a frame-work for the voluntary restructuring of external obligations of the corporate sector. It offered a government exchange guarantee to creditors and debtors who agreed to restructure their debt on the basis of certain minimum conditions (a three-year grace period and an eight-year maturity). A new government entity, the Indonesian Debt Restructuring Agency (INDRA), was to be established to operate the scheme, which was similar to the FICORCA scheme in Mexico. INDRA would not take on commercial risk but would ensure foreign payments to the creditor on the basis of rupiah payments received from the debtor, the latter being determined based on the most appreciated real exchange rate during a specified period.

Experience with the implementation of the June agreement has, so far, been mixed. By mid-October 1998, interbank obligations amounting to $2.9 billion had been exchanged; while this represented a very large proportion of the identified eligible debt, it was considerably lower than originally anticipated on the basis of preliminary estimates of eligible debt. Regarding trade credit, assurances equivalent to $2.7 billion have been received. Implementing the framework for corporate debt restructuring has proved time-consuming. How-ever, with the establishment of INDRA in August 1998, complemented by a set of guidelines for debt workouts (Jakarta Initiative) and subsequent steps to set up the necessary legal framework, preparatory work is now largely complete. INDRA is promoting the program among creditor and debtor groups and intends to circulate the documentation shortly.

Market Reactions

In any case, the programs were vulnerable to adverse market reactions, and those reactions turned out to be far less favorable than hoped—especially in Indonesia and Korea, with the situation in both cases sliding into funding crises.5 As a result, private capital outflows were much larger (Box 4.2) and exchange rates much weaker than originally envisaged (Table 4.2).

Table 4.2.

Exchange Rate Assumptions

article image

Dates when programs were negotiated.

End-period exchange rate for month preceding date shown for program or review.

Year average exchange rates.

April 1997-March 1998 for Indonesia; calendar year for Korea.

April 1998-March 1999 for Indonesia; October 1997-September 1998 for Thailand; and calendar year for Korea.

Market reactions were less favorable than anticipated in the initial programs for several reasons: it took longer than expected to establish the credibility of economic policies, including in the structural area; most of the external debt was private, so that creditors needed to be assured not just of the country's but also of the individual debtor's ability to pay; and as the crises unfolded, investors became increasingly aware of these countries' vulnerabilities, in particular the depth of problems in the banking and corporate sectors.

Establishing credibility—including reassuring foreign investors that private sector creditworthiness would be restored—was intrinsically difficult. For example, plans to recapitalize banks took considerable time to design in detail, let alone implement, and announced programs could only specify the broad outlines and discrete measures used as performance criteria. The outcome of such financial sector reform plans also depended on the authorities' commitment to implementation, and some early developments cast doubt on ownership of the programs. A degree of market skepticism was thus understandable.

Projected Private Capital Flows in the Three Programs

Thailand

Excessively optimistic projections of private capital flows do not appear to have been a problem in Thailand at the outset; in fact, the outturn for 1997 was somewhat stronger than original program projections. This was in large part due to an informal understanding with foreign (mainly Japanese) banks with subsidiaries resident in Thailand that their lines of credit would be maintained. Average rollover rates were still very high in October 1997, but declined significantly in December 1997 and January 1998, reflecting a sharp drop in the rollover of obligations of Thai banks and corporations. Since then, average rollover rates have recovered, but capital flow projections for 1998 have been revised downward substantially since the first review.

article image

Excluding errors and omissions and official financing.

Preliminary outcome.

Korea

In Korea, private capital outflows in late 1997 (mainly through the domestic banking system) turned out to be considerably larger than projected in the original program, with attendant strong pressures on the exchange rate. The original program assumed that the “bulk of the short-term debt will be rolled over.” In the event, rollover rates declined sharply in December 1997, prompting discussions with private creditors at the end of the year, which were concluded on January 28, 1998. Nevertheless, further downward revisions to the capital account projections for 1998 were necessary at the time of the second review, mainly on account of trade financing difficulties experienced by small and medium-sized enterprises.

article image

Including errors and omissions and excluding offical financing.

Preliminary outcome.

Indonesia

In Indonesia, the original program severely underestimated the extent of capital outflows in the initial phase of the program. Outflows from the stock market, which amounted to $5.5 billion in the last quarter of 1997, played a key role, together with low rollover rates of short-term debt. The need for an agreement with creditor banks became increasingly apparent in late January-early February 1998. Negotiations were protracted amidst uncertainty about the possibility of a unilateral pause in debt payments. An agreement was concluded in early June 1998, covering the restructuring of interbank debt, trade financing, and a framework for voluntary restructuring of corporate debt involving a government exchange guarantee scheme (INDRA).

article image

Including errors and omissions and official capital flows.

Preliminary outcome.

Including errors and omissions, official capintal, and effect of resheduling.

A lack of firm resolution in the implementation of macroeconomic policies also undermined credibility. The Indonesian authorities, for example, initially raised interest rates in line with the program but then rolled back the increase a week later. The Korean authorities likewise were reluctant to raise interest rates at the outset. More generally, the credibility of monetary policies was impaired in all three countries by the financial sector weaknesses that were seen by many to limit the authorities' scope to raise interest rates. Political uncertainties also played a major role—notably, in Korea, the presidential elections and the initial disavowal of the program by the presidential candidates immediately after its acceptance by the government; and in Indonesia, conflicting signals from the regime regarding its commitment to the program. In Thailand, there were also some uncertainties as the government that negotiated the original program was a fragile coalition that eventually fell; a new, more stable coalition government was established only in November 1997. While political uncertainties were by no means unique to these countries—indeed, they are a salient element in many IMF-supported programs—they had a particular impact given the countries' vulnerabilities to international capital outflows.

Some other country-specific factors may also have contributed to the failure to reverse capital out-flows. For instance, prior to the approval of Korea's initial program with the IMF, the Bank of Korea established a facility to provide foreign currency refinancing to commercial banks. The interest rate was initially set at a small spread over the London interbank offered rate (LIBOR). One of the measures introduced in connection with the IMF-supported program was to widen this spread to 400 basis points over LIBOR (on December 4, 1997), seen at the time as a penalty rate. In the event, market interest rates facing Korean banks at the time turned out to be much higher, so the rate charged on this facility continued to entail a subsidy. The banks made extensive use of this facility, channeling the funds to their offshore subsidiaries that were having difficulties rolling over their foreign currency liabilities. Such flows to offshore subsidiaries accounted for a large share of Korea's capital outflows during December 1997. The rate charged on refinance was subsequently raised (to 1,000 basis points on December 23, 1997) to bring it into line with market rates.

Another factor that influenced market reactions to the programs, by both domestic and foreign investors, was the coverage of government guarantees. This was a particularly controversial factor in Indonesia: when the government closed 16 banks in November 1997, it announced guarantees on deposits in those banks that (in line with the IMF's advice) covered only deposits up to the equivalent of about $5,000. This signaled to large depositors who held the bulk of deposits that their funds—at least those in private banks—might not be repaid should any more banks be closed. More-over, no announcement was made regarding the protection that would be provided depositors—beyond that presumed to apply to the deposit liabilities of the state banks—in the event of any subsequent bank closures. This policy toward guarantees, at a time when banking weaknesses were widespread, with public knowledge that many other private banks were in as bad a condition as the 16 that had been closed, was an important factor contributing to bank runs. There is no doubt that some banks needed to be closed (see Section VIII below) and it is quite possible that closing a larger number of banks in this early phase of the program might have induced more confidence in the banking sector. However, the policy regarding guarantees now appears to have been ill-advised—notwithstanding good economic reasons in principle for limiting depositor protection. Indeed, in late January 1998, this policy was modified, with a full guarantee issued on all bank liabilities for two years. Anecdotal evidence that there was little public awareness of any deposit guarantee, and that many depositors who participated in the bank runs had deposits below the maximum covered, suggests, however, that other factors may also have contributed to the bank runs.

The markets also became more aware of the weaknesses of the authorities' financial positions as the programs unfolded. In part, this was a reflection of the fact that (especially in Korea and Thailand) the authorities turned to the IMF quite late, and only after exhausting their reserves. A related issue was the revelation of information associated with the programs themselves. Notably, in Korea, the staff report was leaked over the Internet, informing market participants that “usable reserves” were at a perilously low level in relation to maturing short-term debt (since a substantial portion of Korea's reported reserves were actually illiquid claims on overseas branches of Korean banks). Similarly, the Thai authorities, in the midst of the crisis, were required to release data on the central bank's forward foreign exchange positions that revealed the weakness of the country's reserve position. This was intended as a step toward greater transparency, in line with the IMF-supported program, but its timing weakened the impact on confidence that the announcement of the program could have had.

Another element of market uncertainty surrounded the official financing packages. Of particular relevance are the “second lines of defense,” which were pledged by bilateral creditors in Indonesia and Korea but had not been disbursed as of October 1998. The precise terms and conditions under which the second line would be disbursed were never clearly specified. If the “virtuous circle” assumed in the IMF-supported programs had materialized, the second line would not have been needed; however, the uncertainties about their availability may have influenced market participants in their decision to continue their exit—in effect testing the second lines of defense.

A related issue is that IMF and other official financial support was “phased”—that is, disbursed in tranches at the outset of the program and on completion of successive reviews, conditional on the program's remaining on track; although in the Asian crisis countries, total financing was unprecedently large and was heavily front-loaded compared with other programs. Such phasing is a standard feature of IMF programs, aimed at providing support to meet balance of payments needs while safeguarding the IMF's resources and maintaining the authorities' incentives for continued implementation of the program. Choosing the appropriate schedule of disbursement involves a trade-off between these considerations and considerations related to the market's uncertainty about the availability of such resources: phasing implies that the full amount of the financing package is not available to the authorities from the start and there always remains the possibility that later tranches will not be disbursed as scheduled, in the event that the programs go off track.

The communication of the rationale and substance of the programs may also have influenced the reaction to the programs, by both market participants and the general public. Several weaknesses in communication were apparent in the initial Asian crisis programs. One shortcoming in all three countries at the outset was the absence of an effective government economic spokesperson, available to explain the program to the public, underscore the government's support for it, and respond to public concerns as events unfolded. The many public statements of the IMF in support of the policies followed by Indonesia and Korea before (and right after) their currencies collapsed did little to restore confidence. Even if many of the miscommunications that occurred were beyond the IMF's control—and indeed, this element improved significantly, especially with the establishment of new governments in Korea and Thailand—they gave impetus to the efforts under way to improve communication about IMF-supported programs.

1

Current account imbalances were also important especially in Thailand, but in all three countries capital flows were a major element of variability in the external situation.

2

By way of comparison, the financial support package for Mexico in early 1995 amounted to $50 billion, including $18 billion from the IMF.

3

The assumption was that a virtuous circle would be generated by the knowledge that official financing would be available.

4

These assurances, involving credit lines of $19 billion, were received at a meeting with Japanese creditor banks in mid-August 1997. Some uncertainties remained, however, regarding short-term credit lines to Thai banks ($11 billion).

5

By way of comparison, during the Mexican crisis of 1994–95, the announcement of several international credit packages in January and February 1995 failed to restore confidence, and the peso continued to depreciate. However, the announcement of a strengthened fiscal plan on March 9, 1995, and the subsequent authorization from the United States to draw the first US$3 billion of a loan agreed a few days earlier, had a substantially favorable effect on confidence and the exchange rate appreciated by 20 percent against the U.S. dollar between then and end of April 1995.

Cited By