Journal Issue

Policy: Learning by design: assessing IMF-supported programs

International Monetary Fund. External Relations Dept.
Published Date:
June 2005
  • ShareShare
Show Summary Details

What do IMF-supported programs aim to achieve? How are they formulated? How should their success be judged? And have they been successful? At the behest of the IMF’s Executive Board, staff took a detailed look at experience with IMF-supported programs during 1995-2000. The resulting papers suggest two broad challenges for future program design. In low-income countries, the key will be to build on the success to date in achieving macroeconomic stability and higher sustained output growth rates while still maintaining external viability and avoiding future debt-servicing problems. In middle-income countries, especially during capital account crises, programs need to restore market confidence rapidly to help avoid excessively abrupt—and economically disruptive—adjustment of the current account.

The past 15 years have seen important developments in the challenges facing member countries, and therefore in the objectives of the economic programs for which national authorities have sought the IMF’s support. Yet the very responsiveness of the IMF to these evolving needs has inevitably complicated program design and the evaluation of program success. As a first step in this review, staff classified programs by their main purposes. Despite variations in details, most programs could be placed in one of four broad categories:

  • Unsustainable current account deficits. In classic instances of external adjustment, countries having difficulty financing their current account deficit commit to reducing it to a sustainable level, while IMF financing helps the country reconstitute its reserves. This attenuates the necessary adjustment and allows for part of it to come through a positive supply response rather than through demand management alone.

  • Capital account crises. When sizable capital outflows force an abrupt external adjustment and, typically, a collapse of the exchange rate and economic activity, monetary and fiscal policies are geared more toward restoring confidence and limiting the adverse effects on activity than to promoting external adjustment (since the withdrawal of private financing is, in effect, achieving this).

  • Transition economies and low-income countries. Although there are many differences among them, these programs share a common emphasis on macroeconomic stabilization and structural transformation to enhance economic efficiency and promote sustained growth—while maintaining external viability.

  • Policy credibility and public debt sustainability. Where external accounts are largely in balance, programs can nevertheless help lower interest rates and spreads, putting public-debt dynamics on a more sustainable footing by enhancing the credibility of the authorities’ policies.

What constitutes success?

All countries should emerge from their IMF-supported programs with sustainable external positions. External adjustment thus affords the first measure of success. Over the short term, however, such adjustment can take strikingly different courses. A country with little recourse to financing can rapidly reduce its external indebtedness and lower the likelihood of a future crisis, but too abrupt an adjustment may take a significant toll on economic activity, employment, and the exchange rate. By contrast, a country with access to additional financing may postpone the necessary adjustment but later face a crisis that forces a much more painful adjustment. The key challenge is to strike the proper balance between adjustment and financing.

Medium-term debt sustainability provides a useful metric for determining that balance, on the grounds that countries should adjust sufficiently—but by no more than required—to ensure that their external debt position becomes sustainable. The IMF provides sufficient direct financing, or helps generate it indirectly through catalytic effects, to enable the member country to adjust at the appropriate pace, while its advice helps the member design a policy program that minimizes the economic and social disruption.

Against these criteria, how has external adjustment fared? The findings suggest a sharp demarcation between the experience of middle- and low-income countries. In middle-income countries, programs have generally targeted and achieved current account adjustments in line with debt sustainability considerations. But in a number of cases—especially, though not exclusively, capital account crises—large capital outflows meant that external adjustment was more abrupt than programmed or larger than would be indicated by debt sustainability considerations. This did not reflect tight macroeconomic policies—fiscal consolidation generally fell short of program targets—but rather a lack of financing owing to private capital outflows. At the same time, there is evidence that IMF-supported programs helped achieve improvements in the current account at lower cost in terms of lost output growth, perhaps because of more efficient policies.

For the low-income countries, however, the adjustment story is almost diametrically opposite: Programs generally did not target sufficient external adjustment to ensure debt sustainability, and the actual improvement in the current account balance was even less than programmed. These programs, therefore, did not aim at, or achieve, external viability through external adjustment but instead implicitly relied on debt relief. Indeed, the Heavily Indebted Poor Country (HIPC) and Enhanced HIPC initiatives were instituted during this period (in 1996 and 1999, respectively), but current account deficits—programmed and actual—would have been too large to stabilize debt ratios even following this debt relief.

External viability, of course, is just one program objective. Programs may also seek to reduce inflation and stabilize the economy, raise output growth, and reduce poverty. Middle-income countries with IMF-supported programs saw durable reductions in inflation and a dip in real GDP growth during the program period, followed by a recovery of growth rates but not an increase relative to preprogram performance. By contrast, programs in low-income countries saw sustained improvements in growth performance during and following the program, driven by a combination of better macroeconomic policies (lower inflation and smaller after-grants fiscal deficits) and a more benign external environment.

The findings leave two important sets of questions. For the middle-income countries, especially during capital account crises, the key question is how to engender a faster return of confidence and private capital flows, bringing external adjustment better into line with debt sustainability considerations and avoiding excessively abrupt and disruptive corrections of the current account balance. Would this require greater IMF financing? Capital controls? A stronger policy response? Or is it unavoidable, with prevention the only cure?

Challenges remain … in predicting and understanding the implications for policy effectiveness of large capital outflows in capital account crises, and in devising better models of medium-term growth.

For programs in low-income countries, the key challenge will be to continue to improve growth performance—including through lower inflation and smaller after-grants fiscal deficits, but also through structural policies—while maintaining external viability. Will this require further debt relief? A larger proportion of financing in the form of grants rather than loans? Or a fundamental rethinking of program design?

Frameworks and formulations

In formulating advice for national authorities on program design, IMF country teams use a wide variety of analytical methods. This eclectic approach allows staff to tailor program design to country circumstances. For the short-run macroeconomic framework, this approach worked surprisingly well: with the exception of capital account crises, neither inflation nor growth projections exhibit systematic biases. At longer horizons, however, growth projections have optimistic biases, especially in low-income countries. One notable consequence was overly rosy debt projections, which new debt sustainability templates are intended to help address.

Challenges remain, however, in predicting and understanding the implications for policy effectiveness of large capital outflows in capital account crises, and in devising better models of medium-term growth, including to improve debt-sustainability assessments.

Policy content

In tackling the question of policy content, the staff asked three questions: Were policies geared toward achieving program objectives? Were they carried out? What was the outcome?

Macroeconomic stabilization and external adjustment are usually cornerstones of IMF-supported programs, so it is, perhaps, surprising that the exchange rate regime is no more likely to be altered at the outset of a program than at other times, and upfront devaluations are extremely rare.

Nevertheless, there is some tendency of middle-income countries embarking on disinflation efforts to adopt an exchange rate peg, while low-income countries generally attempt disinflations under floating regimes.

Success rates under the two approaches were virtually identical. What appears to differentiate success from failure is whether the programmed fiscal consolidation was achieved. External adjustment was easier under floating exchange rate regimes, in the sense that a given improvement in the current account balance was associated with a smaller reduction in output growth, though the effect is not quantitatively large. While countries with pegged exchange rate regimes are thought to be more prone to excessive foreign currency borrowing—and therefore suffer sharper external adjustment when pegs collapse—the association did not hold generally, although it may have happened in some capital account crises.

As for monetary policies, programs generally targeted a reduction in broad money growth rates and in inflation rates. The targeted monetary tightening was greater the larger the programmed reduction in inflation and the improvement in the current account balance, and smaller the larger the initial output gap and when the exchange rate was floating. Programs did succeed in lowering inflation, though not always by as much as was targeted, in part because of broad money overruns. Notably, it does not seem to matter whether these overruns derived from balance of payments inflows or domestic credit creation, which raises concerns about the need to sterilize large donor inflows or capital inflows if inflation targets are to be achieved. And while monetary tightening in programs was typical, there is no evidence that this slowed output growth.

Depending on initial levels of government expenditure, the fiscal deficit, and the programmed improvement in the current account balance, programs on average envisaged a fiscal tightening of around 1-2 percent of GDP over a two-year period; controlling for these initial conditions, programs in low-income countries targeted about 1 percent of GDP less fiscal adjustment than middle-income country programs. While the fiscal tightening in the initial program year was generally achieved, there tended to be important slippages by the following year, particularly when growth turned out to be weaker than expected, and the envisaged adjustment was primarily on the revenue side or was particularly large. In turn, fiscal slippages contributed to failures at disinflation and to worse public-debt dynamics—though the largest source of errors in projections of public-debt dynamics came from valuation changes on foreign-currency-denominated debt and from the fiscal costs of banking crises.

Smaller budget deficits were associated with faster output growth— most likely because of the boost to confidence and lower interest rates, and greater availability of banking system credit for the private sector.

Fiscal tightening in IMF-supported programs is often controversial because of the possible contractionary effect on the economy, but the empirical evidence does not suggest that fiscal consolidation led to slower output growth. On the contrary, smaller budget deficits were associated with faster output growth—most likely because of the boost to confidence and lower interest rates, and greater availability of banking system credit for the private sector.

Finally, IMF-supported programs incorporate structural measures to underpin macroeconomic adjustment, enhance economic efficiency, and reduce vulnerability to future crises. Programs in transition and low-income countries have a relatively larger proportion of efficiency-enhancing reforms, while programs in middle-income countries, in particular during capital account crises, have a relatively larger share of measures directed at reducing vulnerabilities, especially in the financial sector. Assessing the effects of individual measures on program goals is difficult, not least because structural measures tend to be difficult to quantify. Nonetheless, there is a statistically significant positive association between achieving fiscal adjustment and the number of structural fiscal measures in the program, and between higher output growth and the number of efficiency-enhancing reforms.

Overall, the results suggest broad alignment between program goals and various macroeconomic policies and structural reforms. By the same token, this also means that policy slippages were reflected in program targets being missed.

More work to be done

These papers are spurring further analytical work. Among the topics are how sound macroeconomic policies and sustained growth can be fostered while ensuring debt sustainability. For programs in middle-income countries, analysis is focusing on how to enhance the catalytic response of financial markets—especially to help prevent crises by coupling a sufficiently vigorous policy response with IMF financing at times of heightened vulnerability.

The present study also suggests ways of improving programs, including a more clearly defined role for medium-term debt dynamics in program design; better modeling of medium-term output growth; and, for emerging market countries, a better understanding of the nexus of forces in the financial, public, and external sectors in driving capital flows and crisis dynamics. Beyond these general lessons, there are issues in the use of specific policy instruments, including greater scrutiny of the consistency of the exchange rate regime with program objectives and other macroeconomic policies; the need to sterilize large donor or capital inflows in the monetary program; greater emphasis on sustaining fiscal adjustment efforts and the need to design the fiscal program accordingly; and closer alignment of structural measures with program goals.

These would not be revolutionary changes, but they do represent shifts in emphasis that, if undertaken, would contribute to better designed and better implemented IMF-supported programs, ultimately leading to better economic outcomes.

The Design of IMF-Supported Programs was prepared by a staff team headed by Atish Ghosh and comprising Charis Christofides, Jun Kim, Laura Papi, Uma Ramakrishnan, Alun Thomas, and Juan Zalduendo, assisted by Barbara Dabrowska, Siba Das, Olivia Carolin, and Neri Gomes, under the supervision of G. Russell Kincaid and Mark Allen. The papers are available on the IMF website (

Other Resources Citing This Publication