November 1, 2003

The IEO provides objective and independent evaluation on issues related to the IMF. The Office operates independently of IMF management and at arm's length from the IMF's Executive Board.

Statement by Montek S. Ahluwalia, Director, Independent Evaluation Office
of the International Monetary Fund
at the General Assembly of the Club de Madrid
November 1, 2003

IMF Operations and Democratic Governance: Some Issues¹

This conference focuses on the relationship between IMF policies and programs and democratic governance in the countries in which the IMF operates. While the purely technical aspects of IMF policies, have been extensively studied, their relationship with democratic processes has received much less attention, and yet it is extremely important because most emerging market countries function in a democratic environment and this has implications for how the IMF should operate. In my remarks I propose to present some issues relevant to this theme which emerge from the work of the Independent Evaluation Office (IEO) thus far.² Since the three case studies prepared for this conference relate to emerging market countries, I will focus on issues that are especially relevant for this group. The IMF also plays an important role in low-income countries that have somewhat different problems and governance issues also arise in the IMF's relationship with these countries. I will not deal with these issues in this presentation except to mention that the IEO is currently engaged in a major study of the experience of six low-income countries (Guinea, Mozambique, Nicaragua, Tajikistan, Tanzania, and Vietnam) in negotiating programs under the IMF's Poverty Reduction and Growth Facility (PRGF). The results of this evaluation will be available in 2004.

Crisis management in a democratic environment

Emerging market countries that are integrated with global financial markets have proved to be highly vulnerable to what are called capital account crises caused by a sudden loss of confidence leading to a massive reversal of capital flows. The key to managing such crises is for governments to restore confidence and this calls for quickly outlining adjustment policies and appearing credible in their capacity to implement them, which includes the ability to obtain legislative support if necessary.

This poses special challenges in a democratic environment because democracies are characterized by an adversarial style of politics in which it is the function of the opposition to criticize and oppose the government in power. Since a crisis situation is usually also one in which large numbers of people experience hardship, the political system naturally seeks to assign blame. This makes it difficult for an incumbent government to mobilize broad support for corrective policies. Even a new government may find it difficult since corrective policies take time to yield results, and usually involve some pain in the short run. Capital account crises explode so quickly that adjustment programs have to be put in place in a matter of days rather than weeks and there is little time to explain the need for corrective steps, build a consensus, and carry people along. Even when the need for corrective steps is accepted, the issue of which groups should bear most of the burden of adjustment is inherently controversial and a consensus can prove elusive.

The involvement of an external agency such as the IMF in an otherwise charged atmosphere only adds to the difficulty. It raises concerns about unwarranted external interference and loss of sovereignty and the suspicion that the policies being adopted reflect ideological biases or agendas of outsiders instead of being tailored to the needs and circumstances of the country.

I should emphasize that I am not suggesting that crisis management is necessarily easier in non-democratic states. Such states face different challenges at time of crisis-including questions about their legitimacy and sustainability-that can make crises even more severe by exacerbating political instability. My only point is that managing crises within a democratic governance framework poses special challenges and the IMF has to find ways of operating in such an environment in a manner which can help achieve the economic objectives which are its primary concern, while respecting the functioning of democratic institutions, and strengthening good governance.

The importance of domestic ownership

Perhaps the most important issue relevant for democratic governance is the degree and quality of domestic ownership of the adjustment policies to be followed. Ownership is not easy to define. There are difficult questions about the extent to which ownership should extend-should it cover government, or major political groups or civil society generally? It is also not clear whether the primary concern should be with process issues i.e. the extent of public debate and participation, or with the substantive result in terms of the extent of consensus achieved. At the very least, however, democratic governance requires that the conditionality in IMF-supported programs should not appear to be imposed from outside but be based on policies emerging from domestic policy formulation process, which should also command a wide range of support. This perception of ownership is now widely shared in the Fund and the development community generally.

The importance of ownership is illustrated by the contrasting experience of Korea and Indonesia in 1997. The conditionality associated with the November 1997 program in Indonesia was owned by the technocrats who were the principal interlocutors of the IMF, but was not owned by the Government at the highest political level. As a result, critical parts of the policy package, such as the commitment to maintain the tight stance of monetary policy, were not implemented and others such as the closure of some banks were suddenly reversed. This led to the failure of the November program and prolongation of the crisis. The crisis worsened in December, following rumors about the health of the President and the resulting political uncertainty connected with succession in a regime that had held power uninterruptedly for thirty years. The IMF negotiated a revised program in January 1998, which included extensive structural conditionality, much of it related to macro-economic efficiency. Although the new letter of intent was publicly signed by President Suharto to signal ownership, other actions of the government suggested that real ownership at the political level was lacking. There were deficiencies in the program also, including in our view an excessive overload of structural conditionality with insufficient attention to banking sector problems. The political climate also deteriorated with ethnic riots against the Chinese community which had a dominant role in the private corporate sector. As is well known, Indonesia experienced the most severe currency collapse of the region, as well as the sharpest and most prolonged downturn in output.

The Korean experience was quite different. When the Korean crisis surfaced in October 1997, a program was negotiated with an outgoing administration in an environment of political uncertainty since Presidential elections were scheduled to be held in mid-December. The President elect, Mr. Kim Dae Jung, on assuming office on December 19, indicated his firm resolve to implement the program. Despite this commitment, the Korean program initially failed to stabilize the situation because the financing made available was inadequate without parallel action to lock in private creditors. This problem was resolved subsequently by an agreement among the G-7 governments to encourage a coordinated restructuring of Korea's commercial bank debt (and I will return to the issue in a minute). Thereafter the program was able to restore stability relatively quickly because it enjoyed strong domestic ownership and was implemented firmly. Korea experienced a quick V-shaped recovery in output.

Two clarifications about the role of ownership are appropriate. First, it is worth noting that the high level of ownership in Korea did not mean there was no criticism in the early stages. On the contrary, the initial downturn in the economy and the loss of employment it entailed, generated severe criticism at the time, both internally and externally, with many critics ascribing this outcome to the pursuit of wrong policies. It is only after the program was seen to have worked well that perceptions changed, and there is today fairly broad agreement that the adjustment program in Korea, including some of the structural reform measures taken, served the economy well. Ownership does not therefore mean absence of dissent and criticism. It only requires that government must show a sufficient degree of commitment to implement the program in substance and in a democracy this is unlikely to be sustained unless sufficiently broad support can be mobilized, if not immediately then in a reasonably short time.

Second, it is important to clarify that ownership may be necessary for success, but it is by no means sufficient. A program may be fully owned but may prove unviable as with the initial Brazil program of 1998 which was designed to protect the currency peg. The IEO study shows that the IMF had reservations about the viability of the currency peg and had been arguing in favor of greater exchange rate flexibility, but it went along with the program primarily because the Brazilian authorities were determined to maintain the peg. Brazil's economic management performance had been good, and it was decided to give the benefit of doubt to the authorities.³ In the event, the effort failed and the peg collapsed in January 1999. The program was renegotiated immediately thereafter and the revised program was based on a flexible exchange rate and a shift to inflation targeting, combined with institutional reforms to increase fiscal discipline. The new program was also strongly owned and was effectively implemented and succeeded in stabilizing the economy. It is interesting to note that despite tight monetary and fiscal policy-which are often the subjects of criticism-Brazil's GDP growth performance was better than projected. Growth was only marginally positive, but there was no sharp output contraction as in East Asia.

Current IMF policy requires that programs must be fully owned and the IMF must also be satisfied that they have a reasonable chance of restoring external viability in a manner consistent with achieving sustainable growth. Since there is obviously no guarantee that this convergence will take place, there is room for disagreement between the authorities and the IMF. The nature of these differences and the manner in which they are resolved are relevant for democratic governance. The studies conducted by the IEO offer some insights on these issues.

The role of financing

A successful strategy for crisis management requires some combination of adjustment and financing (either official or unofficial) and the appropriate balance between these elements depends on the underlying cause of the crisis. If the crisis has been caused primarily by temporary nervousness or panic, there is a case for arguing that the problem could be quickly solved by a large injection of liquidity to quell panic, with a relatively modest component of corrective measures.⁴ If however the capital flow reversals are viewed as being caused by genuine weakness in fundamentals, which may not have been evident earlier but have now surfaced, the need for corrective policies to address the root cause of the problem is much greater.

The current exceptional access policy makes it possible in principle for member countries to obtain IMF resources in excess of the normal limits but in practice several constraints may apply. The financing need itself can be very large, even allowing for the flexibility provided under exceptional access. The total resources available with the IMF could be constrained if the IMF faced multiple crises simultaneously. More importantly, even when total resource availability is not a constraint, there is concern in many quarters that easy recourse to IMF financing creates the wrong incentives for lenders and generates moral hazard.

These considerations shift the balance towards greater reliance upon strong adjustment efforts aimed at bringing about a restoration of confidence and resumption of normal private capital flows as quickly as possible. However, in practice, countries may find themselves in a situation where even strong adjustment may not suffice if financing is inadequate, leading to avoidable economic contraction. The case of Korea illustrates the problem. The Korean authorities had publicly announced that they would request support ranging from $50 to $60 billion including loans from G-7 countries. Limitations of finance meant that the actual amount that was ultimately provided by the IMF package (which included amounts from the World Bank and ADB) was only $21 billion, with a further $20 billion from bilateral government loans as a "second line of defense," which would be made available if necessary.

The IEO study shows that the IMF's own assessment at the time was that this amount would not suffice and if more financing was not available, Korea's commercial bank debt would have to be restructured for the program to work. Both options were raised by the IMF with the major shareholders, but additional financing was not forthcoming and there was no support for a debt restructuring initiative at the time. IMF staff responded by arbitrarily increasing the assumed rate of rollover of the commercial debt falling due, in order to make the program appear viable with the financing available, while referring to a second line of defense in the form of bilateral financing as a fall back possibility if needed. However, the conditions under which these resources would become available were unclear. The inclusion of the second line of defense in the "headline" figure for the financing package, without clear indication of the terms on which these resources would be available, only focused attention on the inadequacy of the package. The won appreciated briefly for one day but then resumed its slide and it is only after the restructuring arrangement was put in place that the situation stabilized.

The main lesson from this experience is that the size of financing can be critical to the success of the program. The financing package needed should therefore be carefully and realistically assessed and only resources that are transparently available on an assured and predictable basis should be included in the official financing package. The volume of access in each case will have to be determined on the basis of the specific circumstances, but considerations of democratic governance in the borrowing countries require that members in similar circumstances must be seen to be treated similarly. This is essential if the institution is to be seen as fair in its dealings with borrowing countries.

Program design issues

The conditionality associated with adjustment programs is often the focus of public criticism and democratic governance requires that these issues be addressed transparently so that the reasons for the IMF's insistence on particular conditions is better understood. Our studies throw light on a number of issues that are relevant in this context.

One of the reasons IMF-supported programs meet with political opposition in borrowing countries is the perception that they rely upon a mechanical "one size fits all approach", routinely prescribing fiscal austerity which may not be appropriate in many cases and at times may even be counter-productive. The IEO study of fiscal adjustment, using a data base of 116 programs, comes to some interestingly nuanced conclusions on this issue. Contrary to the general impression, there was considerable variation in the fiscal stance prescribed in IMF-supported programs, with one third of the programs involving an increase in the fiscal deficit. However, the study also found that the rationale of fiscal adjustment proposed in individual programs was not adequately explained in program documents. The lack of transparency on this important issue contributes to the perception that the prescriptions are arbitrary and prevents a serious debate on the issues involved.

The problem arises in part because IMF programs typically make over-optimistic assumptions about the pace of economic recovery in the period following the crisis, with performance often falling short of projections. This over-optimism may reflect the fact that program targets have to be negotiated with the authorities and there are political compulsions to project rosy outcomes, but this can lead to flaws in program design. Unrealistic assumptions about the revival of GDP growth lead to optimistic revenue forecasts, which in turn lead to over-ambitious fiscal targets. Failure to meet these targets in turn feeds the perception that the IMF routinely prescribes excessively tight fiscal policies.

The criticism that the IMF prescribes excessively tight fiscal policies was commonly heard at the time of the East Asian crisis and is worth revisiting. The IEO study of Korea came to the conclusion that the fiscal correction initially proposed was probably unnecessary because Korea did not have a fiscal problem nor was public debt high. However, it did not support the view that the fiscal policy stance was responsible for the collapse in output because the fiscal correction originally proposed was very mild to begin with, and in any case was very quickly corrected. Exactly the same thing happened in Indonesia: over-optimistic growth assumptions validated a more optimistic fiscal stance which was however corrected in a sequence of steps. The real problem in both Korea and Indonesia was that the program did not take into account the negative balance sheet effects of the exchange rate collapse upon the financial and corporate sectors and, therefore, on private investment. These effects were themselves the consequence of balance sheet mismatches in the banking or corporate sectors, a factor which was not adequately recognized earlier in IMF analyses but which now receives much more attention. In the case of Indonesia, investment was further badly hurt by ethnic riots against the Chinese community which had severe adverse effects on the investment climate and on private investment.

Our studies suggest that IMF programs tend to make over-optimistic assumptions about the pace at which private investment is expected to recover and this in turn leads to the adoption of restrictive fiscal policy in order to make room for the expected growth in investment. A more realistic assessment of the pace of recovery of private investment could, in certain circumstances, justify a more counter-cyclical stance of fiscal policy provided the debt situation and debt profile are not a source of worry. As it is, IMF programs are generally "reluctant to call a downturn", as documented in the fiscal adjustment study, with the result that the need for counter-cyclical fiscal policy is rarely discussed ex ante, even if programs are often readily modified ex post once the extent of growth decline becomes apparent.

Another issue related to program design, which has generated a great deal of criticism, is the proliferation of conditionality in IMF programs. Overloading programs with multiple conditions, exemplified by the Indonesia program of January 1998, diffuses attention making it difficult to ensure implementation of what is really critical. It is also often resented as being excessively intrusive, especially in a democratic environment. The IMF has recognized this problem and responded through its current streamlining initiative, which provides that the Fund should limit conditionality (whether in the form of prior actions, performance criteria or structural benchmarks) to those elements that are felt to be critical for achieving the objectives of the program. The problem in practice is how to define what is critical. Policies related to macroeconomic variables and stability in the banking sector are indisputably critical for macro stability. However, the economic objectives of IMF supported programs typically also include restoration of sustainable growth, and the list of potentially relevant policies is obviously much larger if growth objectives are included. Furthermore, as the IEO study on capital account crises points out, in dealing with a capital account crisis case, the most important objective of the program must be to restore confidence. This raises the possibility that policies not directly linked to macro-economic stability, may be viewed as critical in certain circumstances if they relate to areas where corrective action is needed to restore market confidence.

A case is sometimes made for broad conditionality on the ground that crises are also opportunities when there may be a political willingness to make changes on a broad front and reformist constituencies within the government may actually find it advantageous to piggy back a broader reform agenda on IMF-supported programs. This may well be true, and governments must be free to push for such changes on their own, and even to include their intentions in the letter of intent. However such policy measures need not be incorporated formally into Fund conditionality if they are not judged to be critical. Broad- based conditionality, covering many areas not directly linked to crisis resolution, can undermine ownership and lead to unnecessary politicization, which is best avoided.

Finally, an important issue from the perspective of democratic governance is how to demonstrate ownership. Governments can make declarations of ownership, but since programs must be acceptable to the IMF to qualify for support, critics can always claim that such declarations of ownership are only an acceptance of the inevitable. The IEO study on Prolonged Use of IMF Resources had recommended a procedural change that might help. Instead of the existing procedure in which the letter of intent requesting IMF support is the outcome of a negotiating process, the procedure could be changed so that governments initially propose their own adjustment program, which could then be accepted by the IMF or modified through negotiations to make it acceptable. Governments would be free to publish their initial proposal if they wished, enabling clear identification of what was initially proposed by the government itself. It could still be argued that governments would anticipate the requirements for IMF approval and these considerations would be reflected in their initial proposal, but the two stage procedure would increase transparency, help to identify critical areas of difference and contribute to a more open domestic debate on controversial issues.

An implication of this approach is that the Fund, on its part, should acknowledge the importance of ownership by moderating its own expectations and being willing to settle for "second best programs", which may not be ideal in its view but which have the advantage of full ownership and also meet minimum standards. The minimum standards used for this purpose should obviously be determined in as transparent a manner as possible and surveillance has an important role to play in this context.

Surveillance

Surveillance is a core activity of the IMF and can have an impact on democratic governance in several ways. First, it has a critical role to play in crisis prevention in emerging market countries and a number of steps have been taken in recent years to strengthen surveillance for this purpose. Surveillance has been broadened to include assessments of debt sustainability, reports on the observances of standards and codes, and assessments of financial sector stability.⁵ To the extent that these steps encourage adherence to internationally accepted good practices in the fiscal, monetary, and financial areas, they will increase transparency and thereby strengthen democratic governance.

More generally, surveillance could have an impact on democratic governance if surveillance reports are candid and also made public. A number of difficult issues arise, which have been explored in the studies done by the IEO. Surveillance can only be effective if it is candid. However, our studies have found that IMF surveillance has often been lacking in candor. In Brazil for example, the IMF staff had significant reservations on the viability of the exchange rate regime in the second half of the 1990s, but these reservations did not surface sufficiently candidly in presentations to the Executive Board. Similarly, weaknesses in the banking system in Indonesia, many arising from governance related issues, were known to the staff, but were understated in documents that went to the Board.

Increased candor should not present problems if surveillance is viewed purely as a confidential advisory activity. However, the effectiveness of surveillance in crisis prevention does not depend only on the advice provided by the Fund. It also depends on its role of informing the Executive Board and generating peer pressure in that forum and possibly also on the even wider role of informing markets, and creating a structure of incentives that would encourage countries to take preventive steps. These wider objectives can only be achieved if Fund surveillance is made more transparent. In this context, the IEO study on capital account crises had recommended that there should be a presumption of publication of Article IV reports. This would obviously have implications for democratic governance in borrowing countries since it would release into the public domain the IMF's assessment of the macro-economic situation, which in turn could generate a more informed debate on these issues.

Publication of Article IV reports also raises some difficult problems. Country authorities may worry that surveillance reports highlighting risks might precipitate the very crises they seek to prevent. This is a legislative concern, but it can be argued that it may be overdone. If greater transparency were followed as a general policy for all countries, it would probably create fewer problems for individual countries than the authorities currently imagine. The system would soon get used to the fact that systematic surveillance will identify possible problems in some countries at some times and this does not necessarily portend crises, especially if corrective action is also being taken. This is especially so if the signals themselves are more continuous rather than abrupt. To avoid the disruptive effect of abrupt negative signals, the IEO study on capital account crises recommended that the IMF should evolve a system of "escalated signaling" in which signals can be progressively intensified to deal with situations where vulnerabilities identified over several rounds of surveillance are not addressed. The resort to a gradual escalation of signals relating to underlying problems would reduce the danger of sharply adverse market reactions, except in circumstances where such a reaction is perhaps really needed.

In recommending publication of surveillance reports, it is necessary to consider how to deal with situations where there are serious differences of view between the authorities and the IMF. It cannot be assumed that the IMF is always right that there is a case for broadening surveillance to incorporate fresh perspectives to this end, the IEO recommended that it should be possible for country authorities to request the IMF to seek second opinions from outside experts. These opinions would not be binding on the IMF, but they would provide an additional input into the surveillance reports submitted to the Board, increasing the transparency of surveillance activity. ⁶

Another aspect of surveillance which is relevant for democratic governance is its role in identifying the critical areas of weakness where corrective action may be needed if a program has to be negotiated in future. Advance identification of the areas where corrective action is likely to be required by the IMF would help create transparency about the minimum standards which a future program would have to meet.

Surveillance can also help in developing mechanisms that would enable emerging market countries to protect critical social programs in times of crisis. The IEO study of fiscal adjustment found that, contrary to the general impression that IMF supported programs lead to a squeeze on social expenditures, the expenditures are not lower, in countries with IMF programs, than in countries which have crises but don't have IMF programs. However, the study also reveals that aggregate expenditures are not the critical factor. Even if aggregate social expenditures are maintained in times of crisis management, critical expenditures relevant to the quality of service delivery for the most vulnerable groups, such as expenditure on basic medicines especially in rural facilities, are starved of funds while wage expenditures are maintained. The study showed that the resources needed to protect the quantity and quality of services relevant for the poor are relatively small, but there is a lack of public expenditure management systems which can ensure that such programs are identified and effectively protected.

This is clearly not an area where the IMF has a comparative advantage. However, the IEO report suggested that surveillance could be used to encourage countries to outline their own plans for protecting critical expenditures in crisis situations. These plans could be developed by the countries themselves, with assistance from agencies such as the World Bank and WHO. Anticipatory work along these lines would make it easier in times of crises to identify and effectively protect programs critical for the poor. This expanded approach to surveillance runs the risk of being viewed as excessively intrusive, especially in normal times. However, the IEO report argued that if the initiative is left to the authorities, on a purely voluntary basis, it is possible that surveillance could play a useful role at least in some circumstances.

These messages, emerging from our studies, are relevant for anyone interested in evaluating the impact of IMF policies and programs on democratic governance. The promotion of democratic norms and institutions, is not an explicit objective mentioned in the Articles-the Fund is technically neutral on political systems-but the promotion of good governance has been accepted as an explicit objective and measures to promote good governance will inevitably strengthen democratic institutions wherever they exist.

¹ The IEO was established in 2001 to conduct studies of various aspects of the functioning of the IMF. It is completely independent of management and report to the Executive Board. The views expressed in this statement are those of the author and do not necessarily represent the views of the Executive Board or management of the International Monetary Fund.

² The IEO has completed three studies in its first year of operations: "Evaluation of Prolonged Use of IMF Resources (2002)"; "The IMF and Recent Capital Account Crises (2003); and "Fiscal Adjustment in IMF-Supported Programs (2003)" (being printed). The studies have been discussed by the Executive Board of the IMF and are available at www.imf.org/ieo along with a summary of the Executive Board discussion.

³ The unsettled conditions in international financial markets at the time following the Russian default in August and the LTCM crisis was also a factor in favor of trying to avoid a currency crisis at that time.

⁴ Governments are especially likely to take this view if they have earlier viewed capital inflows as a rational response to the underlying strength of the economy and the economic policy environment.

⁵ The Independent Evaluation Office is considering an evaluation of the financial sector assessment program and the associated financial sector stability assessments for its work program next year.

⁶ The Executive Board discussion of the IEO report on capital account crises revealed a number of reservations on recommendations regarding escalated signaling and second opinions. These recommendations cannot therefore be viewed as being among the many others that have been unreservedly accepted.