This paper has presented a preliminary assessment of some key aspects of the programs in Asian crisis countries. Clearly, any conclusions drawn at this stage, while events are still unfolding, must necessarily be tentative. Moreover, the experiences of the countries examined contain many important differences as well as similarities. It is nonetheless useful to try to distill some lessons from the experience so far.

The programs adopted by the authorities and endorsed by the IMF were based on the assumption that policies, together with the commitment of official financing, would restore confidence in the markets and attract private capital flows; official financing and current account adjustment would then need to be sufficient to satisfy the external financing constraint. The more successful the strategy in restoring confidence, the more limited would be the need for disbursement of the official commitment of funds. However, in the event, particularly in Indonesia, but also in Thailand and in Korea, the programs and their initial implementation did not restore confidence rapidly enough, capital accounts were much less favorable than assumed and so the reverse happened: given the climate of economic and political uncertainty, investors (including domestic ones) were not reassured, so a vicious circle of capital outflows and depreciation resulted. The vicious circle was exacerbated as deepening insolvency of financial institutions and corporations created counterparty risk that added to pressure on the foreign exchange market. The depreciations and the severe recessions that ensued took most of the burden of closing the financing gap through a massive current account adjustment. More recently, markets have been stabilized with the restoration of confidence, and exchange rates have appreciated toward precrisis levels.

The expenditure-switching effect of exchange rate depreciation in these economies was attenuated by the concurrent depreciations in several countries in the region. At the same time, the depreciations had strong expenditure-reducing effects via their impact on the balance sheets of financial institutions and corporations. Further balance sheet effects came from sharp drops in asset prices and the disclosure of existing problems in portfolio quality. The resulting wealth effects and disruptions to financing, along with adverse effects on confidence, were reflected in a collapse of domestic investment and a severe decline in consumption associated with the sharp economic downturn. The downturn was also exacerbated by other shocks: internal economic and social disruption (whose seriousness differed across countries) had adverse effects on aggregate supply, while external demand was further weakened by other factors such as the deepening slump in Japan. However, the magnitude of the downturn was largely forced on these economies by the substantial current account adjustment dictated by capital outflows for which it was impossible to compensate through even larger official financing.

The program projections badly misgauged the severity of the downturn. In part, this reflected the fact that the IMF's projections were somewhat more sanguine than the consensus, partly reflecting pressures to agree with the authorities on a common set of program projections and, perhaps, partly a concern to avoid damaging confidence through gloomy forecasts. Erring on the side of optimism in this way was probably detrimental to the programs' credibility. However, it should also be noted that very few foresaw the severity of the downturn—neither the authorities, the private sector, nor academic observers. Failing to foresee the depth of the recession meant that the monetary programs were originally set to allow more rapid growth of money and credit; and fiscal targets were originally more restrictive than they would otherwise have been;¹ it also meant underestimating the magnitude of financial sector restructuring needed.

A variety of factors conspired to make it more difficult to restore confidence. These factors included political uncertainties, the appearance of irresolution in policies, difficulties in communicating the logic and features of the programs to the markets and the public, problems in the coverage of government guarantees, some lack of public support for the programs, and the public debate that took place regarding certain aspects of the IMF-supported programs. Moreover, these factors were operating in a setting where programs were vulnerable to such shifts in market sentiments: the programs' financing had been based on the assumption of a high rollover rate for private short-term debt—in effect, assuming that a virtuous circle would materialize. The phased, contingent nature of official commitments, and uncertainty over the disbursement of the second lines of defense, may also have been factors weakening efforts to restore confidence.

It is thus clear, in hindsight, that the programs were not adequately financed to be carried out in an environment where the crucial effort to restore confidence failed. There would have been two obvious hypothetical alternatives: more official financing or greater private sector bail-in. More official financing would have been difficult, given limited resources and moral hazard considerations. Earlier concerted involvement of the private sector could have been pursued, but if done too aggressively there could have been adverse consequences for emerging markets more generally if the private sector concluded that there had been a change in the rules of the game. Here, the main lesson is that such avenues should be explored in preparation for the next crisis; that is, instruments and mechanisms need to be found to elicit the maintenance of private sector exposure to a country facing a potential loss of market access, without inducing adverse contagion.

The decision to float exchange rates—in the absence of any clear domestic policy anchor to focus expectations—opened the door to continued market depreciation. But there was no practical alternative under the IMF-supported programs, especially in Thailand and Korea where the initial efforts of the authorities to defend their exchange rates resulted in the exhaustion of reserves and removed much of their room for maneuver. Credible step devaluations may have been possible and less disruptive, but only when the countries still had resources and the resolve to defend a new peg. If rates had been repegged based on expectations of capital flows at the time of the programs, they would soon have had to give way in the face of the capital outflows, unless a punishingly tight interest rate policy had been attempted—and that with no guarantee of success.

Although no targets were announced for exchange rates, the exchange rate was the central focus of monetary policy, and interest rates the operating target. The monetary performance criteria specified in the programs were not so much policy targets as secondary tools for monitoring policy outcomes. Policy itself concentrated on leaning against the wind with regard to exchange rate movements. This approach—which put the emphasis on adapting policy to changing conditions—may have been the only viable option in the crisis. The exchange rate was the best available guide to policy, as no other nominal variable was immediately observable.

The basic objective of monetary policy in these programs was to avoid an inflation-depreciation spiral. As suggested by the experience of Indonesia, the possibility of such a spiral was genuine, even in countries with a track record of relatively low inflation. Given concern that excessive monetary tightening could severely depress economic activity, though, the policy followed in the programs was intended as a middle course, leaning against the wind in the foreign exchange market rather than an all-out pursuit of any exchange rate target.

During 1997, the authorities in all three program countries showed some reluctance to tighten monetary policies, both before and after exchange rate pegs were abandoned. This initial vacillation made the task of stabilizing more difficult later on. By early 1998, nevertheless, significant tightening had occurred in Korea and Thailand, but this tightening was not extreme when set against the benchmark of previous crises elsewhere, and is unlikely to have been a major factor behind the output decline. At the same time, persistent reports of credit crunches are of concern. These may be attributable largely to dislocations in the microeconomic allocation of credit, attributable to problems of credit risk rather than tightening of aggregate liquidity, and pointing to the need to move ahead with financial and corporate restructuring.

In Korea and Thailand, the authorities were able to ride the hurricane, and have succeeded in averting an inflation-depreciation spiral. In Indonesia, in contrast, monetary policy went widely off track and inflation became a serious concern, reflecting deeper structural as well as political and social problems and the weakness of the central bank. More recently, the situation in Indonesia has stabilized, with a substantial recovery of the rupiah.

Given unhedged foreign currency exposures, currency depreciation might have had a greater impact on corporations than the higher interest rates needed to stop it—although both are likely to have hurt, particularly given the high leverage ratios of corporations in these countries. Although monetary tightening, if carried to extremes, could in principle result in depreciation rather than appreciation, there is no evidence that this perverse response occurred in these countries.

The original programs in all three countries included some element of fiscal adjustment, in the face of an expected deterioration due to the economic environment, to make room for part of the prospective costs of bank restructuring and to support the external adjustment and thus bolster confidence. The fiscal measures presented in the initial programs were fairly modest in Korea and Indonesia (and even smaller when compared with the expected outcome for the previous fiscal year rather than to the authorities' original plans for the program period). In Thailand, where initial fiscal and external current account imbalances were larger, the fiscal measures in the initial program were more substantial.

Fiscal plans were revised substantially during the course of the programs in response to changing economic conditions—specifically, declining economic activity, a deteriorating external environment, unduly depreciated exchange rates, and (in Indonesia) falling oil prices. In the early program reviews, additional measures were introduced to offset part of the deterioration of the fiscal balance resulting from changing conditions. More recently, with growing concerns over the recession throughout the region, and lesser need for external adjustment due to the rapid adjustment in the current account, the balance of priorities has shifted toward supporting output and increasing the support available under social safety nets. From an early stage, fiscal deficits were allowed to expand to accommodate at least part of the automatic effect of declining activity and income and the exchange rate depreciations, providing support for economic activity from early 1998 on. In recent reviews, fiscal programs have been eased further to augment the automatic stabilizers.

The overall direction of fiscal policy measures can be seen by examining how the change in the fiscal balance was affected by policy changes. By this measure, fiscal policy actions are estimated to have significantly expansionary effects in both Korea and Thailand, relative to a policy of pure accommodation. In Indonesia, in contrast, fiscal policy changes offset up to one-third of the very large deterioration in the fiscal balance associated with changing economic conditions (where the latter includes the increase in food subsidies in response to the exchange rate depreciation). These results do not give credence to the view that fiscal stringency was a major factor accounting for the output decline in these countries.

Given the need for external adjustment forced by the capital outflows during the crisis, fiscal policy may have had more influence on the composition than on the magnitude of the output decline. An easier fiscal policy at the outset would likely have required more real exchange rate adjustment and/or higher interest rates in the face of capital outflows, depressing private domestic expenditure further via balance sheet effects. The net stimulus to economic activity might thus have been relatively small. By the same token, the recent shift of policies in the direction of supporting activity has become appropriate in light of the external adjustment by the private sector, and the easing of the external financing constraint with the abating crisis.

Although a complete understanding of these countries' structural problems emerged only as events unfolded, it was known from the start that structural reforms needed to be a central pillar of the programs. A strong package of structural reforms was essential, in light of major weaknesses, especially in the financial and corporate sectors, that underlay the crisis. Critics have argued that since many reforms, even if sensible in the medium term, have some adjustment costs, the large number of reforms entailed an excessive burden at a time of great economic weakness. While these concerns cannot be dismissed lightly, they ignore the real nature of the crisis—much more the result of cumulating structural weaknesses than of macroeconomic maladjustment. In that context, lasting recovery depended on comprehensive structural change. Attempting stabilization without strong structural reforms, especially in the financial and corporate sectors, would have been a costly effort to treat the symptoms without credibly addressing the causes of the disease. Moreover, in light of complementarities among different reform measures, dropping some of the reform measures from the packages on the grounds that they were too costly or too difficult to put in place would likely have impaired the effectiveness of others. Indeed, as the programs evolved, they revealed greater depths and complexities to the weaknesses that the reforms were to address. At this point, however, hindsight suggests that corporate restructuring should have been given higher priority at the outset—as indeed it was given increasing emphasis as the programs evolved.

Critics have also argued that closures of financial institutions at the outset of the programs undermined confidence (especially in Indonesia) and that any needed closures should have been delayed. How-ever, delays in closures may only have made their costs larger. The main problem with Indonesia's banks during November 1997-January 1998 was not early closure but closing a subset of problem institutions under inappropriate conditions. One factor accelerating bank runs was the initial treatment of deposit guarantees, which were very limited in amount, inadequately publicized, and covered only those institutions already closed. In contrast, the experience with financial restructuring in Korea and Thailand was much more favorable.

The record of implementation of policies shows that two of the countries—Korea and Thailand—have, on the whole, been rather successful in implementing the programs as agreed, whereas in Indonesia, in part because of the severity of the underlying political crisis, the program has repeatedly veered off course and has required substantial modification. While this period has been very difficult for all three countries, developments have been much more favorable in the two that have been able to keep to their programs. In Korea and Thailand, the challenge is to persevere with their adjustment, and get through the difficult phase where measures have begun to bite but their credibility has not yet been established, into the phase where they can start to reap the benefits. Indonesia, in contrast, still faces a more difficult task, owing to the need to repair repeated policy slippages and arrest a slide into an increasingly difficult social situation; its progress in this direction, however, has been encouraging.