The IMF treated the problem as a liquidity rather than a solvency crisis.

When the crisis hit in 1982, the IMF and creditor governments sought to contain it through a “case-by-case” strategy aimed at providing enough additional financing to cover the time required for the indebted countries to implement adjustment programs and generate enough growth to restore normal financial relations. The additional financing, however, was mostly in the form of debt obligations. As Eichengreen and Kenen (1994) recently summarized the implications, “In imprecise but helpful terms, an insolvent debtor must pursue a debt-reducing strategy, but an illiquid debtor should pursue a debt-raising strategy so as to make its interest payments and defend its creditworthiness.”

The primary difficulty with this argument, as Eichengreen and Kenen acknowledge, is the ambiguity of the distinction between a liquidity shortage and insolvency. ^1/ If a country has enough real resources to generate the foreign exchange to service its debts but faces a temporary inability to convert resources into foreign exchange, then it faces a liquidity crisis; without sufficient real resources, it faces a solvency crisis. In that strict sense, none of the heavily indebted Latin American countries ever faced a solvency crisis in the 1980s. ^2/ Mexico—to take the most readily quantifiable example—had petroleum and natural gas reserves totaling some 72 billion barrels, valued at more than $2,000 billion in 1982, compared with outstanding external public - sector debts totaling just over $60 billion. Merchandise exports (mostly by the public sector) totaled $21 billion, compared with scheduled debt service payments (amortization plus interest) on external public-sector debts of $16 billion. ^3/ The problem was not a shortage of real resources; it was rather that the degree of required adjustment in domestic expenditure for the Mexican authorities to mobilize those resources was simply not feasible.

A more accurate phrasing of this criticism would be to charge that insufficient attention was paid to the feasibility of the required adjustment or to the real costs of the increased indebtedness engendered by the additional financing that characterized the first few years of the debt crisis (see, for example, Dornbusch (1993), pp. 53-55). To continue with the example of Mexico, imports in dollar terms fell by two thirds from 1981 to 1983 and remained depressed through 1987, and real wages declined by a similar magnitude; meanwhile, the stock of external debt rose by more than a third. The magnitude of these changes and the effects that they had on Mexico’s economic stability were far greater that anyone foresaw in 1982.

Was the IMF excessively optimistic in its assessments of the growth prospects of the major Latin American countries after the debt crisis, as many critics have charged? Throughout Latin America, the adjustment programs developed in 1982 and 1983 were predicated on forecasts of a rapid resumption of economic growth that would gradually bring down debt service ratios to sustainable levels. In Mexico, the late-1982 staff projection for real GDP showed zero growth in 1983, 3 percent growth in 1984, and 6 percent in 1985. For Brazil, the program initially assumed a 3 ½ percent decline in real GDP in 1983, to be followed by growth of 2 percent in 1984 and 4 percent in 1985. Not surprisingly, output growth proved to be difficult to forecast in the prevailing crisis conditions; perhaps more surprisingly, however, the forecasts on average were only modestly optimistic. At one extreme, in Mexico, three-year growth totaled 2 percent, compared with the initial projection of 9 percent; but in Brazil, growth totaled 10 percent through 1985, compared with a forecast 2 percent. For the whole region, the one-year-ahead forecasts published in the IMF’s World Economic Outlook added up to just over 6 percent growth, and the outturn was closer to 4 percent; ^1/ excluding Brazil, of course, the gap would have been wider.

The explanation for the slowness of growth to resume in some but not all heavily indebted countries after 1982 is complex and cannot be encapsulated in the liquidity-solvency dichotomy. First, a drop in real wages was essential to restore financial balance, but wage cuts also depressed aggregate demand; hence, a few countries that resisted adjustment were able to maintain short-run growth better than those that took early action—but that growth was unsustainable without the needed adjustment. Second, both the Fund staff and officials in indebted countries were slow to recognize the breadth of structural reform that was necessary if macro-economic stabilization was to be implemented without stunting growth; not until the mid-1980s did market liberalization and reform become a full partner with stabilization. Third, and largely for these reasons, political resistance to effective adjustment was strong and prevented some countries from fully implementing the programs to which they had agreed. These factors were in principle foreseeable; if they were not fully incorporated in the projections, bias would result. ^2/

In addition, however, there were important unforeseeable factors at work. For one, protectionism, especially through non-tariff barriers, reduced the markets for developing-country exports and affected different markets in quite different ways. ^1/ Even more importantly, export market conditions varied across commodities: oil prices (important for Mexico) drifted downward throughout the first half of the 1980s and collapsed in 1985-86; coffee prices (important for Brazil) were reasonably strong in the early eighties, skyrocketed in 1986 in response to a disastrous harvest, and then collapsed; copper prices (important for Chile) weakened at the beginning of the decade and rose sharply toward the end. On balance, commodity price declines weakened growth in Latin America during the critical adjustment period in the early 1980s; correcting for that effect would explain much of the apparent bias. ^2/

2. The IMF acted like the handmaiden of the commercial banks.

Sachs (1986a, p. 84) stated this argument clearly: “The basic strategy of the IMF and the creditor governments since 1982 … has been to ensure that the commercial banks receive their interest payments on time.” ^3/ More recently (and with more balance), Dooley (1993, p. 10) concluded “that for a long time the strategy that was intended to force the banks to continue to lend while the debtor countries embarked on reform programs worked in the narrow interests of the banks.”

The basis for this criticism is that the net effect of indebted countries reaching agreements with the Fund and the banks in the 1980s was to transfer resources from debtors to creditors. That is, the combination of the reduced present value of debt service under rescheduling agreements and the provision of “new money” through IMF-supported concerted lending was generally less than the originally scheduled debt service; hence indebted countries still had to make some net payments to banks and thus in a direct financial sense were worse off than if they had just defaulted.

The underlying premise is that in the absence of concerted lending, the indebted countries would have had little choice but to default, after which the discounted value of the banks’ assets would have been substantially lower. Moreover, because the exposure of a number of the major.money-center banks to heavily indebted countries exceeded their capital, a series of defaults–but not a series of negotiated reschedulings–could have led to a collapse of the international banking system. Clearly the IMF played a key role in preventing that catastrophe. The right question, however, is not whether the Fund helped the banks; the question is whether it did so at the expense of, rather than to the mutual benefit of, the indebted countries.

The case for mutuality of interests rests directly on two arguments: that default would result in a loss of access to international capital markets, and that the value to indebted countries of maintaining access exceeds the real economic and political cost of the adjustment that is required if the country is to be able to service its debts. Both the requirements for and the value of capital market access have been extensively debated, without a clear resolution. ^1/ The Fund position, however, was based on the logically prior proposition that the required adjustment would be beneficial for its own sake, regardless of its ability to crowd in private foreign capital.

To understand the IMF position on this issue requires a perspective broader than that of a game-theoretic model of financial relations. From a macroeconomic perspective, maintaining debt service is one element of a strategy to prevent a slide into autarchy. The heavily indebted countries in the early 1980s could not service their debts in the short term without obtaining external support, and they could not service them in the longer run without undertaking fundamental policy reforms. Even if these countries could have defaulted without rupturing future relations with creditors, they would likely have been worse off than if they had undertaken the required policy adjustment and stayed current on payments to creditors. The case for this proposition rests essentially on the value of the adjustment program, rather than on the expected penalties from default.

The record on policy reform in Latin America suggests that countries that adopt unilateral policies on debt service are likely to experience economic deterioration rather quickly, and that countries that delay macroeconomic adjustment are likely to become worse off after a spurt of short-term benefits. The first category would include Peru under Alan Garcia and Brazil under José Sarney; the second category would include (for much of the 1980s) Argentina as well as Brazil. In contrast, successful post-crisis adjusters such as Chile, Mexico, and Bolivia laid the basis for more balanced growth over the longer run; the achievement of that growth will provide the surest test of the case-by case strategy.

3. The IMF recognized much too late that debt relief was needed.

When the Brady Plan was introduced in March 1989, the IMF reacted quickly to support it and to play a key role in implementing it. For three years or so preceding that development, however, a variety of debt-relief proposals were floated by advocates including Bill Bradley, Peter Kenen, and Felix Rohatyn. During that period, the IMF kept a low profile on the issue, and a general perception arose that the institution was opposed, or at best indifferent. As I.G. Patel put it recently, “the Fund … was certainly too late in actively advocating debt relief - as indeed was the [World] Bank” (Patel (1994), p. 12).

The official public stance of the Fund, as articulated by the Managing Director and in keeping with the Fund’s role as an intergovernmental institution, tended to reflect the development of political support for debt relief. As the leaders of major industrial countries proposed various debt-relief schemes in 1987 and 1988, the IMF responded positively; when the Brady plan culminated this process in March 1989, the Fund acted immediately to implement it. Whether a more aggressive public posture might have accelerated the process could be debated, but not with more than purely conjectural content. In any event, the case for generalized debt relief could not have gained credence until enough time had elapsed to allow both creditors and debtors to pass from crisis to quagmire: that is, until 1987 at the earliest. By that point, the management of the Fund was actively encouraging–publicly as well as privately–an expansion of options for reducing debt stocks more aggressively. ^1/

The substantive questions are whether the staff were leaders or followers in the intellectual field, and whether the Fund was prepared to accept debt relief once the countries and the banks agreed on a deal. The staff, of course, were not always of one mind on this issue; concerns about encouraging unilateral debt repudiation or discouraging a return to voluntary bank lending kept the bandwagon in the garage. Nonetheless, the staff position increasingly favored recognition of the need for debt relief once the initial systemic crisis began to fade. As is well known, Mike Dooley–then in the Fund’s Research Department–was an early proponent; his string of studies on the issue began in 1985 with a paper arguing that the existence of a stock of bank debt selling at deeply discounted prices implied that a bank making a new loan to that borrower would face an immediate capital loss; only by eliminating the debt overhang could the borrower expect to regain normal access to such credits. ^1/

Operationally, the Fund implicitly accepted debt relief in the 1986 stand-by arrangement with Bolivia, in that the program was predicated on the assumption that the official interest obligations would be met only in part; the buybacks that gave Bolivia its initial debt relief in 1987 were implemented while that program was active and were administered by the IMF through a contribution account. Shortly thereafter, the December 1987 deal in which a portion of Mexico’s bank debt was voluntarily replaced by collateralized bonds with a lower face value was also undertaken with the implicit support of the Fund, with which Mexico had an active stand-by arrangement at the time. Hence, while the Fund displayed no less than its usual caution, it cannot be said to have impeded or even delayed the development of debt-relief plans in the second half of the 1980s.

4. The IMF departed from its basic surveillance mandate.

The view is often expressed that the IMF’s role of overseeing the international monetary system and helping countries overcome short-term balance of payments problems is incompatible with–or at least orthogonal to–assisting and advising developing countries on longer-term structural problems. This general criticism was stated recently by the Bretton Woods Commission (1994, p. 6), which concluded that “in developing [countries] the IMF should focus on short-term macroeconomic stabilization.” With specific respect to the debt crisis, Edwards (1989, p. 8) characterized the criticism as being that “the Fund has ceased to operate as a financial institution” and “is acting more and more as a development aid-granting agency.”

What specifically did the IMF do in response to the debt crisis, and how did those activities relate to the Articles of Agreement and the Fund’s previous work? In general, the Fund assisted countries in designing macroeconomic adjustment programs and supported those programs financially through stand-by and extended arrangements. The only relevant question in that regard is whether these loans were consistent with the Fund’s mandate to provide temporary financing for balance of payments purposes. The creation of the Extended Fund Facility (EFF) in 1974 had provided a mechanism for making the IMF’s general resources available for longer periods than under stand-by arrangements (SBAs). The EFF, which was designed to cover situations where a country’s payments imbalance resulted from deep-seated structural problems, was activated for several Latin American countries, including Peru in 1982 and 1993, Mexico and Brazil in 1983, Chile in 1985, Mexico and Venezuela in 1989, and Argentina in 1992. ^1/

How temporary was this financing? A number of these countries became prolonged users of Fund resources as they dug out from the initial crisis, but only Peru fell even temporarily into arrears with the IMF. Most borrowers gradually but punctually repaid their debts. Mexico’s indebtedness peaked in 1990 at SDR 4.8 billion and fell to 3 billion by mid-1994; Argentina reduced its obligations from SDR 4 billion in 1988 to 1.6 billion in early 1992 before undertaking new drawings under the EFF; and Brazil reduced its indebtedness from a peak of SDR 4.3 billion in 1984 to less than 200 million by mid-1994. ^1/ Maturity profiles were elongated in the 1980s, but they did not vitiate the requirement of temporariness.

The IMF did, however, introduce some important innovations in the course of managing the debt crisis. First, it assumed a much more active role than heretofore in arranging the total financing packages for adjustment programs. Prior to the crisis, the standard practice for the staff had been to estimate the financing that would be provided by other private and official creditors under normal conditions and with good policies in place; the Fund would then negotiate a program that would make such financing possible, and if necessary it would provide financing (within the established access limits) to close any remaining gap. For Mexico and Argentina in 1982, the initial working assumption had to be that the commercial bank creditors would make every effort to reduce their exposure as rapidly as possible, in which case (a) the access limits would be inadequate to finance the program and (b) IMF financing would do little more than to enable the country to service its dwindling bank debts.

The solution was to require concerted lending by the banks so as to stabilize their aggregate financing. That solution worked because it was in the collective interest of the banks but not in the individual interests of those who might otherwise have become free riders. If the procedure had been generalized, it would have represented a major extension of the IMF’s activities; in practice, however, it was used only in cases where the program could not otherwise have been financed, and usually only where there was a systemic threat in case of a program failure. By the late 1980s, concerted lending had evolved into a more general “menu” approach in which the role of the IMF was generally the traditional one of approving an adjustment program that could serve as the basis for the use of Fund resources and serve as a catalyst for outside financing.

The second major innovation was the development of procedures for monitoring adjustment programs that were not supported by IMF financing. Under the heading of “enhanced surveillance,” the IMF in 1984 began authorizing the release to private creditors of staff reports that evaluated programs that were financed by creditors other than the Fund; in most cases, the programs were undertaken in conjunction with multi-year rescheduling agreements (MYRAs) that extended beyond the conclusion of IMF financing. As with concerted lending, it quickly became apparent that the practice had merits but that it would be a mistake to generalize it. In particular, care had to be taken to ensure that private creditors did not come to think of the IMF as a credit-rating agency, and that the institution’s credibility would not be undermined by de-linking policy advice from financial commitment. The practice therefore was limited primarily to cases where the Fund had provided financing and maintained close contact with the authorities through the subsequent completion of the adjustment process; in addition, it was usually applied by making information and analysis available to private creditors without implying an official seal of approval. Since 1990, there have been few new cases of enhanced surveillance.

Finally, the practices of the IMF were extended in 1989 to provide for the application of IMF resources to support reduction of bank debt under the Brady Plan. Fund resources were lent to member countries but were earmarked for the repurchase of bank loans at prices approximating those prevailing in secondary markets. To ensure that benefits accrued to the borrower and not just to creditors, buybacks under this program initially were restricted to cases where banks were increasing their exposure; and, in selected cases, programs were approved and drawings were permitted before the borrowing country had reached agreement with commercial creditors to settle arrears. As the plan evolved, banks took advantage of an increasingly sophisticated menu of options for participating.

The temporary effect of these innovations was to make bankers into the bedfellows of the IMF: for nearly a decade starting in 1982, bankers working on Latin American loans had to maintain close and frequent contacts with IMF staff and management. By the early 1990s, however, after the Brady deals were in place, the traditional arms-length relations had been restored, and the Fund had largely avoided the risks of moral hazard and bureaucratic overreach that could have ensued if crisis management had ossified into standard operating procedure.

5. The resolution of the debt crisis was delayed by poor coordination between the IMF and development institutions, notably the World Bank.

The IMF and the World Bank have been criticized both for stepping on each other’s toes and for dancing to different drums. Feinberg and Bacha (1988) expressed the allegation of coordination failure as follows: “The Fund’s financial reaction to the Latin American debt crisis was both swift and deep–but it was not lasting. [Meanwhile, ] … Initially perceiving the debt crisis to be a temporary phenomenon, the World Bank sat back and watched the IMF take the lead” (pp. 377, 383).

During the first phase of the crisis, the Bank played a relatively limited role while it concentrated instead on longer-run development problems, especially of lower middle-income countries. ^1/ The Bank made substantial adjustment loans to Brazil, but it lent less to Mexico and little to Argentina. In 1983-84, the IMF took the lead and extended net credit totaling approximately $9 billion to the eleven most heavily indebted Latin American countries, compared with a net flow of just over $3 billion from the World Bank. By 1985, the picture began to change: Brazil’s and Mexico’s Fund-supported adjustment programs had faltered, and the Bank had begun to step up its lending throughout the region; overall, net flows from the Bank were slightly larger than those from the Fund (more than $1 ¾ billion, compared with $1 ½ billion).

One purpose of the Baker strategy, introduced in October 1985 at the Annual Fund-Bank Meetings in Seoul, Korea, was to bring the Bank group more fully into the picture. Secretary Baker called for the IBRD and the Inter-American Development Bank to increase their lending to the “principal debtor countries” and for other agencies in the World Bank group (IFC and MIGA) to work toward attracting equity capital flows to those countries. This initiative significantly strengthened the Bank’s role in the debt strategy: not only did its own lending rise sharply, but the example of IBRD disbursements to some extent supplemented IMF credits as a trigger for commercial bank rescheduling agreements. Over the next three years (1986-88), Bank lending to the eleven major Latin American borrowers totaled nearly $6 1/2 billion, at a time when net credit from the IMF was just around $500 million, as the normal flow of repayments to the Fund nearly matched new lending. ^1/

Neither before nor after the Baker plan was lack of coordination between the two Bretton Woods institutions a systemic problem. In the early eighties, the Fund took the lead, and in the second half of the decade the Bank played a larger role; total net flows from the two institutions were reasonably stable throughout. Rather than reflecting a coordination problem, this passing of the baton reflected the long-standing differences in mandate and priorities of the Fund and the Bank. Both institutions did occasionally tread on each other’s toes, most notably in 1988 when they differed in their assessment of the viability of Argentina’s proposed reforms. While the IMF was still negotiating the terms for an SBA to replace the one that was about to expire, the Bank announced (at the Annual Meetings in Berlin) the approval of four loans (three sectoral loans plus one for trade policy) totaling $1 1/4 billion. That embarrassing contretemps forced the institutions to develop more detailed understandings of their respective responsibilities, and the incident was not repeated. ^2/

6. The IMF imposed inappropriate “Washington consensus” conditions.

In an influential paper, John Williamson (1990) characterized the IMF’s approach to adjustment programs as part of a “Washington consensus” founded on “prudent macroeconomic policies, outward orientation, and free-market capitalism,” an orientation that implicitly dismisses “the ideas spawned by the development literature” (pp. 18-20). With respect to Latin America, he concluded that it “is not at all clear that the policy reforms currently sought by Washington adequately address all of the critical current problems” (p. 18). As examples, he cited the need for price controls as a component of inflation-reducing strategies, the need to allow for entrepreneurial skepticism in projecting the ability of adjustment programs to generate growth, and the need to allow for the likely persistence of capital flight following implementation of a stabilization program. ^1/

No one would argue seriously against the “charge” that the IMF has insisted that financial responsibility and stability require market-oriented policies, low fiscal deficits, and limits on the growth of domestic credit financed by the monetary authorities. The issue is whether these prescriptions are applied rigidly in cases where they are not strictly appropriate. In practice, the specific elements of adjustment programs vary greatly according to the circumstances facing each country, but the real issue is whether a different approach altogether might have been required in some cases. In Latin America, the alleged exceptions include, in addition to the specific examples cited by Williamson:

• countries where the state plays a large role in promoting development, for example through the operation of state enterprises or by directing capital flows toward favored sectors, may not be susceptible to market-oriented stabilization policies;

• countries where inflation arises primarily from structural or inertial forces may suffer especially high real costs from conventional stabilization programs; and

• countries where confidence in the domestic currency or in domestic financial institutions is low may be destabilized by a dismantling of capital controls.

Without attempting to resolve such issues in a short discourse, it seems fair to conclude that there is grounds for a debate here. Throughout Latin America, the adjustment programs supported by the IMF in the wake of the debt crisis were predicated on a model in which the development role of the state, inertial inflation, and the autonomous role of financial weakness played little part. Rather, the basis for IMF policy advice was (and is) that structural reforms should aim at reducing market distortions and giving full play to market incentives; that price stability is essential for growth and can be promoted effectively only through appropriate macroeconomic policies; and that capital controls are ineffective at best and are usually counterproductive. The fact that–in some cases, especially in the early 1980s —negotiators on each side of the table were arguing on the basis of such different models was doubtless an important contributor to the failure of the governments concerned to assume “ownership” of their own programs and implement them firmly in the face of domestic political opposition.

As a result of the intensive and extensive dialogues that took place between the IMF and the authorities of indebted countries in the 1980s, these debates were partially resolved. On the side of the Latin American countries, a “silent revolution” gradually weakened the belief in state dominated economic development and in the need for capital controls. ^1/ For its part, the IMF showed a degree of flexibility throughout the decade. For example, program conditions (beginning with Brazil in 1984) gave increasing attention to the role of inertial inflation through acceptance of the operational deficit as a measure of fiscal policy. In a few cases where the threat of external shocks made the success of a program especially risky (most notably in the EFF arrangement negotiated with Mexico in 1986), the Fund incorporated contingency clauses in the program terms. More generally, the Fund endorsed a wide range of innovative and heterodox programs, such as Argentina’s 1985 Austral Plan. Nonetheless, room remains for further dialogue and research on the linkages between the literature on macroeconomic stabilization and on structural reform and development (see Corbo et al. (1987) and Khan and Montiel (1989)).

7. Debt relief was hampered by the IMF’s refusal to write down its claims.

The Brady Plan provided for a coordinated approach to debt reduction in which commercial banks, bilateral official creditors, and international financial institutions (IFIs) would all play a role. The new role for the IMF was to allow borrowers to apply a portion of their drawings from the Fund to reduce their outstanding bank claims on the basis of market, rather than face, values. Subsequently, however, criticism arose because the debt owed to the IMF and other IFIs was spared from any rescheduling, much less writing down in value. Helleiner (1994) summarizes the views of many in developing countries: “The possibilities and modalities for writing-down some IMF/Bank debt need to be discussed directly and openly rather than remaining unacceptable topics for discussion” (p. 13).

The IMF Articles of Agreement do not allow for the writing down or rescheduling of member countries’ obligations to the Fund; to do so would be inconsistent with the institution’s mandate to make its resources available on a temporary basis. Instead, countries with debt-servicing difficulties have been encouraged to implement Fund-supported adjustment programs that can serve as the basis for rescheduling or other debt-relief agreements with private and official bilateral creditors. That process, which has underpinned the debt strategy from the beginning, would be seriously compromised by any effort to weaken member countries’ commitments to it. More fundamentally, this issue is irrelevant from a financial perspective. The debt strategy since 1989 has recognized that countries cannot extricate themselves from a depressed-growth path unless their debt overhang is eliminated and debt service obligations are reduced to a sustainable level in relation to anticipated export receipts. Although there are nominally three tranches to external debt (obligations to private creditors, bilateral official creditors, and IFIs), practically speaking there are only two: private and official. Since official creditors already have numerous mechanisms for taking joint action in this field, including the Paris Club and the G-7 summit process, there is no free-rider problem that would require intervention by the IFIs. Whether creditor governments choose to take the full “hit” on bilateral credits or allocate some portion of it to IFI credits has little financial or even political relevance to the debtor. ^1/

What is relevant is the total size of the write-down on official credits and the IMF’s role in supporting debt reduction. For the six years from the Bolivian buybacks of 1987 through mid-1993, eleven Latin American countries with IMF-supported adjustment programs were granted rescheduling agreements through the Paris Club; those agreements covered more than $35 billion in debts. With only a few exceptions (notably for Bolivia and Nicaragua), official bilateral debt cancellations were not applied to this region. ^1/ Private-sector debt relief, supported both by official bilateral creditors and by the IMF and other IFIs, was much larger: over the same period, eight Latin American countries used a variety of operations to reduce their debt and debt-service obligations by $42 billion (on an initial stock of $104 billion), at a cost of just over $14 billion. ^2/ Whatever the political or economic constraints may have been on these numbers, there is no basis for thinking that a more direct participation by the IMF would have been either necessary or sufficient for increasing them.