Banque Central des Etats de I’Afrique de I’Ouest (Central Bank of West African States)
Development Assistance Committee (of the OECD)
deep deepening index
Enhanced Structural Adjustment Facility
foreign direct investment
heavily indebted poor countries
International Development Research Centre
International Labour Organisation
net present value
National Resistance Movement
Program to Alleviate Poverty and the Social Costs of Adjustment
Policy Development and Review Department (of the IMF)
Policy Framework Paper
real external debt burden
National Resistance Movement of Mozambique
Structural Adjustment Credit
Structural Adjustment Facility
Structural Adjustment Loan
Uganda Exporters Association
West African Economic and Monetary Union
Unilateral Declaration of Independence
Uganda Manufacturers’ Association
Uganda National Chamber of Commerce and Industry
United Nations Development Program
Zambia Consolidated Copper Mines
We were appointed by a decision of the Executive Board of the IMF in October 1996, to conduct the first-ever independent external evaluation of the Enhanced Structural Adjustment Facility (ESAF). The evaluation group was originally made up of Dr. Kwesi Botchwey, Harvard Institute for International Development (Coordinator and Convenor); Professor Paul Collier, Centre for the Study of African Economies, Oxford University; Professor Jan Willem Gunning, The Free University of Amsterdam; and Professor Yusuke Onitsuka, University of Tokyo. Following the unfortunate and untimely death of Professor Onitsuka, Professor Koichi Hamada of the Economic Growth Center, Yale University, was appointed as the new fourth member of the group.
Our evaluation was intended to complement an extensive internal review of experience under ESAF-supported arrangements by the Fund’s Policy Development and Review Department (PDR), which was expected to be concluded in the spring of 1997.
We held our first meeting in Washington in April 1997 to discuss administrative arrangements, to conduct preliminary discussions with Fund staff and with members of an Evaluation Group of IMF Executive Directors charged with the general oversight of the successful conclusion of our work, and to make all necessary arrangements for the release to us of all relevant information in possession of the Fund.
We met again in July 1997 to finalize the administrative arrangements, reach final agreement among ourselves on country sampling and other procedural matters, and conduct further in-depth meetings with staff of the PDR and other departments working on the countries in our sample. Most important, at this time we attended and participated in the meeting of the Executive Board devoted to a discussion of the full report on the PDR review of experience under ESAF-supported programs.
From about August through much of September and early October 1997, we undertook a program of country visits, which we interrupted in the latter part of September to attend the World Bank/IMF Annual Meetings in Hong Kong, where we conducted further interviews with governors and senior officials from a large number of countries, including some countries outside our sample. We also met with a number of IMF Executive Directors and senior officials from the Bank.
We resumed and concluded our program of country visits after the Annual Meetings and finally congregated at the Centre for the Study of African Economies at Oxford to review our findings from the field trips, debate outstanding issues, and agree on the form and content of our final report, which we present here.
We would like to express our gratitude and appreciation to all those who assisted us in various ways. In particular, we would like to thank the members of the Evaluation Group of Executive Directors; the Executive Directors of our sample countries; the Director and staff of the Office of Internal Audit and Inspection, especially Elena Frolia, Melanie Brown, and Cathy Song; the Directors and staffs of the Policy Development and Review Department and the African Department; and the Fund and World Bank representatives in the various countries we visited. We also wish to acknowledge the invaluable support provided by a number of research assistants—Michael Fabricius, Elizabeth Su-Dale (who also assembled the first draft), Yasuhisa Ojime, Mototsugu Shintani, and Alex Maynard. Above all, we would like to thank the ministers, government officials, private sector and trade union representatives, nongovernmental organizations (NGOs); and all others who gave us so much of their time during our country visits.
The terms of reference for this evaluation study distinguished three components: social impact, external viability, and ownership.
Although ESAF programs are often criticized for the uniformity of their design and policy focus, their social impact is highly diverse. There are instances of generalized initial losses in incomes and of generalized initial gains, of prolonged reductions in social expenditures and of sustained increases. The same social group may suffer severely in some ESAF programs and benefit in others: for example, ESAF radically lowered civil service real wages in Côte d’Ivoire while raising them in Uganda. Thus, the average experience across all ESAF programs conceals as much as it reveals, while a case study of an individual country is liable to be highly particular. Our approach has been to apply a simple analytic taxonomy, distinguishing a few important mechanisms whereby the poor might be affected.
We distinguish between two main channels by which ESAF programs can affect the poor: via private incomes and via social expenditures.
Private incomes are inevitably affected by ESAF programs; indeed, the ultimate purpose of such programs is that incomes should be increased on a sustainable basis. However, any policy change that seeks to raise incomes in the aggregate is likely to inflict some income losses. A useful distinction in analyzing these losses is between those that arise as a result of a fall in the aggregate income of society and those that arise as a result of redistributions within society. In a well-designed program, any aggregate losses will be temporary: policy reform is intended to raise aggregate income in the medium and long term. In some societies, a temporary decline in aggregate income is unavoidable, the main instance of this being the transition economies. These economies have a large high-cost sector, which reduces the potential for other sectors to expand and which must therefore itself contract, while lacking the organizations of private enterprise so that growth elsewhere is initially slow. Because the temporary losses in transition economies are widespread, the targeting of safety net responses may not be difficult; rather, it is the sheer scale of need that may pose the challenge. National food-for-work types of intervention may be both simple and sufficiently self-targeting. The transition economies are, however, not the main focus of this study.
Most ESAF countries are not transition economies, although a few of them have one or other of the transition economy-type features. Hence, in most ESAF programs unavoidable income losses arise mainly or exclusively from redistributions between socioeconomic groups. Whereas the aggregate income losses in transition economies are inherently temporary in a well-designed program, the losses arising from redistributions are likely to be long lasting because they are caused by relative price changes that are intrinsic to policy reform. Such income losses are thus both more particular, hitting pockets of people in the midst of more generalized improvements, and longer lasting.
Safety net interventions are thus likely to need more targeting, and to last for longer periods. However, by no means do all losing groups warrant safety net intervention. The central issue is whether the losers are concentrated among the initially poor. Overall, we concur with the main academic study on this question, which concludes that reform will “generally have positive effects on growth and income distribution” (Sahn, 1996, p. 22). That is, on the whole, the groups that lose from reform are concentrated among the initially better-off rather than among the poor. However, this does not mean that there is no overlap between the poor and those who lose from the reforms. In some contexts, income redistributions work against important subgroups of the poor: we found this for maize growers in the remote regions of Zambia, estate workers in Malawi, and urban informal sector workers in Côte d’Ivoire.
The very particularity of these groups indicates that it is not possible to devise, a priori, safety net interventions that will work across ESAF programs. There is no substitute for detailed country-level work using socioeconomic survey data. Usually, however, there is sufficient information available prior to ESAF programs for safety nets to be built into the program design. To date this has not been done. We recommend that the Fund draw formally upon the household poverty expertise of the World Bank, integrating projections of social impact into program design and monitoring the outcomes.
In some nontransition ESAF economies there have been large, temporary contractions in aggregate income, Zambia and Zimbabwe being the main cases in our sample. While there were complicating circumstances due to drought, we conclude that a significant part of the decline in income was avoidable, being due to errors in the sequencing of the reform program. In both countries financial liberalization was, in our view, premature, in that it preceded fiscal stabilization and considerably delayed its attainment. We are also concerned about the sequencing of some structural reforms. Some reforms that would have enhanced the ability of the poor to benefit from the ESAF program should have been given earlier priority. We recommend that sequencing issues be explicitly analyzed at the stage of program design.
The main scope for poverty reduction through rising incomes occurs in the poststabilization phase of ESAF programs. Whereas in the stabilization phase it is usually appropriate and indeed essential that fiscal and external deficits be reduced, in the poststabilization stage the most important objective may be to increase the investment rate so as to achieve sustainable rapid growth. In low-income ESAF countries both the scope for, and the desirability of, financing increased investment through domestic savings is limited. A successful poststabilization ESAF economy is likely to have a phase in which deficits temporarily increase again as investment is financed by increased aid and private foreign capital. Since it now appears that aid effectiveness requires that aid be targeted to countries with good macroeconomic policy environments, there is potentially a serious conflict between the common Fund practice of planning for a rapid tapering off of total aid in poststabilization situations and the donor objective of increased aid effectiveness. We recommend that ESAF have a continuing role in stabilized low-income countries, but that a sharper distinction be drawn between the phase in which policy is oriented toward stabilization and the phase in which it is oriented toward growth. In the latter, it may be appropriate for the investment rate to become a monitorable program objective. The original purpose of ESAF as stated by the IMF’s Executive Board on December 15, 1987, was “to promote in a balanced manner” the objectives of payments viability and growth. This balance should, in our view, be achieved through a changing emphasis over the duration of programs.
Partly as a result of the success of ESAF, there are now several low-income African economies that have recently achieved a satisfactory policy environment. These countries (Uganda in our sample) are currently growing rapidly. However, their investment rates remain low: current high growth is the temporary pay off to policy reform. This conjunction of high growth and low investment is not sustainable: either investment must rise or growth will decelerate. The increase in investment cannot be financed predominantly from domestic savings because incomes are so low: both enhanced private and public capital inflows will be needed until incomes have risen.
Private investment is currently deterred because these environments are rated as highly risky. The risk ratings for the newly reformed economies are improving, but from a very low base, and it will take another decade before the ratings reach the level of the current newly industrializing countries, at which major investment inflows become likely. Recent research has shown that in the reformed policy environments, aid acts as a catalyst for private investment: each dollar of aid induces almost two dollars of investment. Hence, investment inflows can be increased both by increased aid inflows and by a reduction in the perceived risks. In both of these, the Fund has a key role.
The Fund can reduce perceived risks by signaling that a country has reached the phase in which the macroeconomic policy environment is satisfactory for private investment. In this phase, the key role of the Fund is the surveillance of policy maintenance through the monitoring of a few key variables, rather than the negotiation of further promises of policy change. However, surveillance in itself may not be sufficient to achieve credible certification. To be fully credible, the Fund should put its own resources at stake and so have a program. However, it is essential that such a program be clearly distinguished from those that are designed to cope with crisis recovery. Countries would be seen to graduate out of a crisis period into a second phase of rising investment, before they graduated completely out of Fund programs. The graduation into this second phase would constitute a powerful signal to the investment community. It would also constitute a signal to the donor community. There is now compelling evidence that aid is effective in, and only in, satisfactory macroeconomic policy environments. The Fund has a key role in certifying that such an environment has been attained. Clearly, maintained Fund financing in these environments adds credibility to the message that donor funds should appropriately increase. The Fund itself is not a development finance institution and should not become one. Rather, its new role for this group of graduated, but investment-scarce, countries is temporary, in the initial phase of a reformed environment. The recent wave of reforms and temporary high growth in Africa has thus created a window of opportunity. The Fund is instrumental in whether this opportunity is seized.
The provision of social services can be affected by an ESAF program in four distinct ways. First, if ESAF raises or lowers GDP, then, if all other ratios stay unaltered, real per capita social service provision will change accordingly. Second, ESAF may change the share of government expenditure in GDP. Third, ESAF may change the composition of public expenditure. Finally, relative price changes may change the amount of service provision that a given value of public expenditure will purchase. Each of these four routes for change has been important in one or other of the countries in our sample, and for each of the routes there are examples both of improvements and deteriorations in social provision. We are, however, concerned that there has been excessive focus on one of the channels, namely, the composition of public expenditure. In our sample, the most important changes in social provision often arose from changes in its relative price, yet this appears to have gone unnoticed. We recommend that the Fund present data on the provision of social services in a standardized format that decomposes changes into the four components set out above, so that governments can more readily see what is happening and why.
A further factor affecting social provision is the volatility imposed on certain components of public expenditure by the operation of the cash budget. There is no easy solution to this problem. However, in some countries there is a clear trade-off between the procedures that have proved effective in controlling expenditure in aggregate and the efficiency of public expenditure. A mechanism such as the cash budget, which substantially reduces the efficiency of expenditure, should be regarded as a temporary stopgap rather than as constituting a solution. Better forecasting of the intrayear pattern of revenues and a more appropriate pattern of expenditure smoothing will be necessary as countries graduate from the circumstances of fiscal crisis.
While in our view there is no systematic tendency of the poor to lose disproportionately from ESAF programs, there are important subgroups of the poor that are so weakly integrated into the economy that they are left behind during growth. In Uganda, growth has been predominantly urban based, and subsistence farmers have not been in a position to benefit. In the medium term, reintegration of these groups into the market economy will raise their incomes. In some countries, there is a conflict between this medium-term objective and the short-run objective of revenue maximization. In Uganda, both the current high level of petroleum taxes and the previous windfall coffee tax work to slow integration of rural households into the market. There are again no easy solutions to such problems, but government choices should be made in the context of a more informed appreciation of the costs of taxation. In the short term, the best way of improving the living standards of these economically detached households is through increased provision of basic social services. In the context of rapid growth, rising rural social provision is likely to be affordable and could reasonably have been included as monitorable program objectives.
Our recommendations on social impact are as follows.
First, in our view the Fund should not invest in building up expertise in poverty analysis. Rather, we recommend that at the stage of program design the Fund formally ask the World Bank to identify ex ante which groups among the poor are likely to lose from the proposed reforms. The Bank would then provide the Fund with projected time paths of the real incomes of the main groups of the poor and also with projected outputs of social services. The output projections for social services would take into account the relative price changes that we have identified as so substantial that they can radically change the conclusion from social indicators. The projected time paths would be incorporated in program documents, along with the traditional fiscal and monetary monitoring variables. Whether a program would be considered in need of revision would be decided in part on the basis of a comparison of outcomes with these projections. Clearly, the time lag for income data is much greater than that for financial targets. However, the use of income data in assessing whether the outcome of a program was consistent with the initial projections would be salutary since major deviations would require explanation.
Second, in program design, trade-offs between the short and the long run should be explicitly analyzed. This analysis would address sequencing issues, the efficiency costs of revenue measures, the need for front-loading of slowly maturing structural reforms, and the appropriateness of cash budgets.
Third, in the area of fiscal policy, where the macroeconomic concerns of the Fund and the microeconomic concerns of the Bank currently overlap, collaboration between the Bank and the Fund should be increased. Specifically, boundaries need to be more clearly delineated and, where overlaps are accepted, a more formalized requirement for joint analysis and decision should be negotiated so that country-level staff members are clear about their respective powers and duties.
Finally, in already stabilized economies the Fund should shift from ex ante negotiation of short-term targets and policies to an ex post evaluation over a longer period. This would help reforming governments to build their reputations and would enable the Fund to play a useful role in potential ESAF countries that now reject the instrument. In postcrisis economies the Fund would focus on encouraging and managing increased external inflows, public and private. Except in the transition economies, ESAF funds would taper in with adjustment rather than taper out. Conversely, the Fund would be more circumspect in providing support in stabilization contexts where the commitment of the government is in serious doubt.
ESAF as an instrument is unusual in that it is not used as budget support (except in the CFA franc zone). It thereby directly affects only private incomes and not social expenditures. Its impact on private incomes is through the exchange rate. Inadvertently, this has the effect of implicitly taxing exports. While this is an inevitable feature of all public transfers, it is usually offset by the beneficial effects of increased public service delivery. We recommend that ESAF funding be provided as budget support, as indeed already applies in the CFA franc zone.
Since an ESAF program mainly consists of the provision of a concessional loan, it should be analyzed from an intertemporal perspective. The current account of the balance of payments is the difference between savings and investment, as well as the difference between exports and imports. It can also be regarded as the difference between savings and investment, which is an intertemporal concept. As Yusuke Onitsuka demonstrated, a typical nation goes through a dynamic pattern that starts from a debtor position and then approaches a creditor stage in its process of economic development. In this dynamic path of growth, or in the optimal growth path, such indicators as the debt-service ratio and the debt-GDP ratio do not necessarily stay constant. According to modern macroeconomics, a current account surplus emerges for a nation that expects a decline in the future (permanent) income.
There are many indicators of the external viability of a nation. As the numerators, we take debt service (a flow variable) at one time, and debt outstanding, or the net present value, at another. For the denominator, we use exports and GDP. The combination of these creates the debt-service ratio, for example, which is equal to the debt service divided by exports. All the combinations are static in nature, except the concept of net present value, which is the discounted sum of all future obligations.
In this report, we propose the use of more dynamic concepts: the real external debt burden (REDB) developed by Obstfeld and Rogoff (1996), and the debt deepening index (DDI). REDB indicates the rate of the change in the debt-GDP ratio in the absence of the current account deficit (or surplus), and DDI indicates the rate of change in the debt-GDP ratio under the actual current account balance.
In particular, we consider that the ratios with respect to exports are overemphasized in the practices of the ESAF program and the HIPC (Heavily Indebted Poor Countries) Initiative. For example, the debt-service ratio (i.e., the ratio of debt service to exports) and the ratio of debt service to GDP could produce quite a different ranking in the viability of nations because of the difference in the degree of openness of nations. We suggest that at least equal weight be given to the indicators related to GDP, compared with those related to exports.
Our recommendations on the issue of external viability are as follows.
One should rely more on ratios of debt service to GDP and debt to GDP as indicators of external viability of a nation since they are less affected arbitrarily by the degree of openness of a nation than are ratios of debt service to exports and debt to exports, because the latter ratios are overly sensitive to the openness of the economy.
To supplement the static nature of the above ratios, one should also refer to the REDB and to the DDI. The net present value of debt is indeed a dynamic concept, but it should be matched not only with the current GDP, which is a static concept, but also with the net present value of the national income of a nation or the net present value of the savings-investment balance.
Ownership and Governance Issues
The one common theme that runs through perceptions of ESAF at the country level is a feeling of a loss of control over the policy content and the pace of implementation of reform programs. On the one hand, there is broad agreement that ownership is a necessary condition of successful policy reform and program implementation. This much is acknowledged in official declarations of both donor and recipient governments, and by most multilateral institutions, including the IMF, the World Bank, and the regional development banks. Academic writers on the subject also find, predictably in our view, that high conditionality programs do not generally do well, and the recent review by the Fund’s Policy Development and Review Department implicitly confirms this view in its finding that a substantial proportion of program interruptions are attributable to policy disagreements between governments and the Fund staff.
Our review of country experiences also shows a correlation between the degree of ownership and successful program implementation. Vietnam, especially before the ESAF program, Uganda, Ghana, and to a lesser extent, Côte d’Ivoire and Bangladesh before the instability of the Ershad period, are good examples.
On the other hand, in spite of the apparent consensus on the importance of ownership, it has not been possible to move matters beyond mere theory. On the donor side, development cooperation ministries and offices point to the need to explain to taxpayers how their money is used abroad. This is a real enough political problem, although it must be noted that it is in practice often aggravated by public misconceptions about the size of aid in relation to national budgets, the role of aid in domestic economic crises, and the extent of true charity in aid flows. On the part of the multilateral institutions, there is the sheer weight and convenience of established practice and the commanding authority that comes with controlling the purse, to say nothing of the genuine difficulties that complicate attempts at giving operational meaning to the concept of ownership.
The challenge therefore is how to foster strong country ownership and at the same time provide adequate assurances to both multilateral and bilateral sources of financial assistance that their resources will not be wasted. The solution, in our view, lies not in reducing ownership to simply persuading the country to adopt what others want, but in finding a middle ground that enables the country to express its will and build consensus behind a program capable of achieving sustainable growth. This requires actions both at the country level to improve the decision-making and consensus-building processes, and by the Fund to make the negotiation process and conditionality regime more supportive of country ownership.
Accordingly, we make recommendations for country initiatives, as well as for modifications in the Fund’s operating procedures. Without prejudice to our general and country-specific analysis on ownership and to the suggestions we make by necessary implication, we propose the following specific recommendations.
Initiatives at the Country Level
The country itself should, first and foremost, take steps to define its medium- to long-term vision and the policy agenda that goes with it before it begins formal negotiations with the Fund, the Bank, and other agencies. For this purpose, the country should avail itself of all possible sources of technical assistance, including those of its nationals abroad and technical assistance from the Fund, which many countries acknowledge is particularly helpful in such circumstances. The country then should take steps to build a body of national consensus behind such a national program.
It is obviously for each country to decide how national consensus can be built most effectively, and no ironclad laws can be laid down in these circumstances. Moreover, the idea is not to make this yet another conditionality, as it is tending to become with particular regard to donor demands for civil society participation. However, based on the experiences of countries that have managed to create space for sustained policy reform through national dialogue, we recommend, for the consideration of the countries themselves, the creation of economic management teams made up of the economic and key sector ministries and political leaders to oversee the reform process so that it does not become the exclusive business (and burden!) of the minister of finance, and the holding of national conferences where alternatives and trade-offs can be openly debated.
Initiatives by the Fund
Side by side with what we propose should be done by the country to develop ownership, we recommend that the following steps be taken by the Fund to enhance national ownership in negotiating agreements.
First, we recommend that, at the earliest opportunity and at a sufficiently high management level, the Fund engage in intensive and informal policy dialogue with the country’s political leadership to understand the country’s political constraints and possibilities and, in this way, be able to form the right political judgment for determining the mandate for formal negotiation with the country.
Second, we recommend that the timing and duration of IMF missions be so arranged as to allow adequate time for country preparation in advance of negotiation and consensus building during the negotiation process itself.
Third, and perhaps most important, we recommend that steps be taken to relieve widespread concerns about the Fund’s perceived inflexibility in negotiations through the introduction of an element of choice in the negotiation and conditionality regime. We accordingly recommend that some flexibility be built into the mandate for negotiations in the currently essentially ex ante negotiation and conditionality regime. One of the ways (but by no means the only one) this could be done is to formulate alternative program paths through the negotiation process, leaving it to the country to decide, with the advice of the staff, what best (or better) suits its particular circumstances. We recognize and discuss some of the operational problems and risks associated with such a flexible regime more fully in the text of our report. At the very minimum, each alternative program path would need to satisfy a minimum condition of viability, in order not to saddle the Fund with unacceptable risks of program failure and, more important, to guarantee the country sustainable growth.
Next, we recommend that the Fund develop a more systematic mechanism for providing ex post support for country-initiated, or home-grown, programs. We believe that this will enable the Fund to play an important role in countries that have balance of payments need but where agreement is thwarted or delayed even though the areas of convergence between the Fund and government are substantial; this may happen, for instance, where a government feels unable to accede to formal agreement with the Fund, mainly for political reasons. We believe that an essential element of such a mechanism of ex post support for homegrown programs would be the provision of technical assistance to the countries wishing to develop such programs, as this will improve program quality and, commensurately, the occasion for exiting by the Fund for reasons of program failure. While such a system will need to have entry and exit points to make it workable the point should be to strive to reduce the prospect of such programs breaking down and, therefore, to minimize the risk of exits, which can be disruptive and costly to the country in the end. The entry point could be triggered, for instance, by a major reform initiative by the government, such as a strong budget or a major adjustment to an overvalued exchange rate.
We also recommend that ways be found to both humanize and demystify the Fund’s image, so as to assuage the political hazard that countries perceive to be associated with dealing with the Fund. Here, we have in mind not so much the dramatic new initiative by the Bank to send staff members on familiarization missions to villages, but simpler ways in which Fund staff can have more systematic and interactive contact with a broad cross section of stakeholders in the countries. In our view, the policy framework process could provide a more convenient forum for these broad-based contacts than the Article IV consultative mechanism because of its forward-looking policy focus.
In the area of Fund/Bank cooperation, we note that against the background of increasing overlap in the work of the Fund and the Bank—with the Fund becoming more “structural” in focus, while the Bank becomes more “macro”—it is particularly important that Bank/Fund relations be better coordinated, because they have an obvious bearing on country ownership. Policy choice by the country is not helped when the Fund and the Bank (and donors) pull the country in different directions. We therefore recommend that urgent steps be taken to develop more effective and operational instruments of coordination through close consultation between the two institutions.
Finally, to reinforce these strategies aimed at fostering country ownership, we strongly recommend that resident missions be established in all ESAF countries. We recommend further that the missions be strengthened by a delegation of more authority to them, especially in matters that are dependent on knowledge of the concrete domestic situation, and also through the selection of high-flying, technically strong, and politically mature staff members to head them. We believe that a strengthened resident mission is a more effective way to address concerns about program interruptions than is intensified monitoring and frequent visits by headquarters-based staff.