The interrelated key features 1 and 2 set out several aspects of the expected relationship between the PRGF-supported program and the PRSP. Specifically. PRGF-supported programs should be consistent with, and drawn from, the PRSP in a manner that takes into account national priorities. Further, PRGF-supported programs should concentrate on those parts of the poverty reduction strategy that are within the IMF’s areas of expertise while remaining cognizant of and consistent with those parts of the strategy that are outside the IMF’s areas of expertise.

The interrelated key features 1 and 2 set out several aspects of the expected relationship between the PRGF-supported program and the PRSP. Specifically. PRGF-supported programs should be consistent with, and drawn from, the PRSP in a manner that takes into account national priorities. Further, PRGF-supported programs should concentrate on those parts of the poverty reduction strategy that are within the IMF’s areas of expertise while remaining cognizant of and consistent with those parts of the strategy that are outside the IMF’s areas of expertise.

PRGF-supported programs have achieved a high degree of consistency with the supporting I-PRSPs/PRSPs on broad macroeconomic goals, targets, and projections, as well as on structural measures. The first two of the seven key features of the PRGF have five sub-elements relevant to this review: (a) PRGF-supported programs draw their main elements from the country’s PRSPs; (b) where relevant, joint staff assessments (JSAs) and staff reports should highlight flexibility in accepting country choices; (c) PRGF-supported programs demonstrate how macroeconomic and other policies have been influenced by growth and poverty objectives; (d) aspects of the PRGF-supported program that promote private sector development should be highlighted; and (e) PRGF contribution to the poverty reduction strategy should be focused on areas within the IMF’s areas of expertise and responsibility. Because (a) and (c) cover the core of both key features but also cover much the same ground, they are treated together at the outset; (b), (d), and (e) are treated individually in the following subsections.4

Draw Main Elements from a Country’s PRSP

PRGF-supported programs show a high degree of consistency with the underlying I-PRSPs/PRSPs in three main aspects—the broad objectives of the programs, the macroeconomic projections and targets, and the structural measures. The broad objectives of the PRGF-supported programs as set out in the LOIs/MEFPs were compared with the goals of the underlying I-PRSPs/PRSPs.5 While the goals of the 1-PRSP/PRSP typically cover areas beyond those of the PRGF-supported program (e.g., poverty diagnostics), it is generally possible to identify the broad objectives set out in the LOIs/MEFPs in the I-PRSPs/PRSPs. Macroeconomic projections and targets from the staff reports/LOIs/MEFPs have been compared to those of the 1-PRSPs/PRSPs (where quantitative projections are available in the latter documents) and to the macroeconomic projections and targets of the most recent preceding annual arrangement under the ESAF. These show a very high degree of consistency between the PRGF-supported program documents and the I-PRSPs/PRSPs. as well as a broad continuity with similar projections in previous ESAF-supported annual arrangements. Finally, the links between the specific structural measures in the PRGF-supported program documents (LOIs/MEFPs and staff reports) and the I-PRSPs/PRSPs were examined; while it was generally possible to find measures in the I-PRSPs/PRSPs that were at least consistent with, and often essentially the same as, those in the PRGF-supported program documents, cross references to specific measures are uncommon, and readers are left to determine the consistency by their own comparison of the documents.

Broad Macroeconomic and Macro-Relevant Objectives

The broad macroeconomic goals of PRGF-supported programs show a very high degree of consistency with the I-PRSPs/PRSPs underpinning these programs. The broad macroeconomic and macro-relevant goals set out in PRSPs have been compared with the goals of PRGF-supported programs as stated in the LOIs/MEFPs for 59 PRGF-supported program requests and reviews.6 For 44 of these, the goals of the PRGF-supported program were judged to be essentially the same as those set out in the countries’ I-PRSPs/PRSPs. For another 13 requests or reviews, the broad goals were partially consistent with those set out in the countries’ poverty reduction strategies. Only two LOIs/ MEFPs have broad macroeconomic goals that cannot be seen as having been drawn from the poverty reduction strategy. The two exceptions relate to the request for a PRGF-supported program by Ethiopia and the first review under this arrangement. In this case, the I-PRSP is almost entirely silent on macroeconomic and macro-relevant goals. The JSA for the I-PRSP indirectly acknowledges this; it cites only goals derived from the PRGF-supported program and notes that these goals were discussed with the World Bank and the authorities and were regarded as part of the interim poverty reduction strategy even though these were not part of the I-PRSP itself.

The degree of consistency in macroeconomic goals is greatest for PRGF-supported programs underpinned by full PRSPs, only slightly less for PRGF-supported programs presented to the Board together with I-PRSPs, and still substantial for PRGF-supported programs presented to the Board subsequent to the endorsement of the I-PRSPs, but before the presentation of the full PRSP (Table 2). For program reviews underpinned by full PRSPs (either completed by the Board together with the endorsement of the full PRSP or subsequent to it), the objectives of the PRGF-supported program could be drawn fully from the PRSP in eight of nine cases and partially in the ninth. Requests or reviews approved by the Board at the same time the I-PRSPs were presented for endorsement also displayed a very high degree of consistency in macroeconomic and macro-relevant goals between the I-PRSPs and LOIs/MEFPs with goals in 23 of 29 LOIs/MEFPs that are fully consistent with the I-PRSPs, 5 where PRGF-supported program goals were partially consistent with those in the I-PRSPs, and only 1 where the goals of the program were not consistent with goals set out in the I-PRSP. Finally, a majority of LOIs/MEFPs coming to the Board after the I-PRSP, but before the PRSP, were judged to be fully consistent, and all but one of the remainder were judged partially consistent with the preceding I-PRSP. In this last case, changes in economic conditions or the emergence of new issues in the intervening period explained the discrepancy.

Table 2.

Are the Broad Macroeconomic and Macro-Relevant Goals of PRGF-Supported Programs Drawn from l-PRSPs/PRSPs?

article image
Sources: IMF country documents; national authorities, country policy intentions documents.

The cases in which the goals could be only partially derived from the I-PRSPs/PRSPs are mostly attributable to very general or partial specification of macroeconomic goals in these documents rather than to any inconsistency or conflict between the two. For example, Armenia stands out as having a relatively substantial specification of macroeconomic goals in the I-PRSP. In this case, however, the need to raise tax revenue is no more than implicit, while the PRGF-supported program documents and the JSA present raising tax revenue as a central goal of the PRSP strategy and the PRGF-supported program. The closest case among these where an actual conflict may be evident between the I-PRSP (issued June 1, 2000) and the PRGF-supported program is the case of Senegal, where the JSA notes differences in macroeconomic projections between the two and the omission of major structural issues in the I-PRSP. Even here, though, the differences in outlook are modest, and disagreements are confined largely to a looseness of the relationship between the I-PRSP and the PRGF-supported program rather than to any inconsistency between the two.

Macroeconomic Targets and Projections

The macroeconomic targets and projections underpinning PRGF-supported programs are very similar, though not identical, to those of the contemporaneous I-PRSPs/PRSPs in almost all cases. Projections for the current and three projected years were collected from both I-PRSPs/PRSPs and PRGF-supported program staff reports on inflation, growth, the current account balance, official transfers, and several balance-of-payments financing items as another indication of consistency in broad macroeconomic projections.7 Projections were identical in 18 of the 33 cases where a PRGF-supported program request or review was presented to the Board together with an I-PRSP/PRSP, and differed by only very small amounts in seven additional cases. The one case in which there are significant differences in projections that appear to be intentional is that of Cambodia, where the JSA notes that the projections in the I-PRSP are those discussed with IMF staff, but economic disruption caused by severe floods after the I-PRSP projections were made final had resulted in revisions to the PRGF-supported program projections. In the remaining seven cases, all from 2000, no table of macroeconomic projections was included in the I-PRSPs/PRSPs, and the text contained too little detail for a meaningful comparison to be made. While these results indicate that consistency of the PRGF-supported program with the I-PRSPs/PRSPs is taken seriously, they do not rule out the possibility (reported in several instances by some donors, NGOs, and authorities) that the PRGF-supported programs influence the I-PRSPs/PRSPs rather than the other way around. The close correspondence in form and content between several I-PRSP/PRSP projection tables and the standard Selected Economic Indicators tables for the same country in IMF staff reports is also suggestive in this regard.

Growth, inflation, and current account objectives in PRGF-supported programs show a high degree of continuity with the preceding ESAF-supported programs (Table 3). Average real GDP growth rates are very similar in both cases. The median growth rate is projected at slightly higher levels in PRGF-supported arrangements, with 17 countries showing a higher average growth under the PRGF-supported program projections and 17 showing a higher average growth under the ESAF-supported programs. PRGF-supported programs also target larger increases in growth (1.5 percent against 0.5 percent for programs under the ESAF).8 Median inflation rates for the PRGF-supported programs are projected to be roughly one percentage point lower than for ESAF-supported programs. Eighteen countries in the sample show a somewhat higher average inflation rate under the PRGF-supported program, and 15 show a somewhat higher average inflation rate under the ESAF-supported program. Reductions in inflation were slightly smaller under PRGF-supported programs. The median current account balance-to-GDP ratio for programs supported by arrangements under the PRGF is slightly worse than under the ESAF-supported programs; the median improvement in the current account balance over the three-year projection period is 0.6 percent of GDP for both the ESAF and PRGF subsamples. Nineteen countries had a better current account balance (smaller deficit or higher surplus) under the ESAF-supported program than under the PRGF-supported program, and 11 had the reverse pattern (comparable data are unavailable for the remaining five). The differences between these comparisons may owe as much to the context of the global economic environment as to countries’ histories; the lower level of inflation under PRGF-supported programs (mostly 2001 projections) may be attributable in part to the global slowdown in growth, and in part to some success in reining in inflation under the previous ESAF-supported programs (predominant projections from 1998–99). In any case, they do not seem to show a substantial departure from targets and projections for ESAF-supported programs in the sample.

Table 3.

Median Economic Assumptions in PRGF- and ESAF-Supported Programs:

article image
Source: IMF staff estimates.

Third projected year less current year.

GDP deflator where available, CPI otherwise.

In percent of GDP.

Excluding transfers.

Official transfers are discernibly higher in PRGF-supported program projections (4.4 percent of GDP) than in those for ESAF-supported programs (3.8 percent of GDP). The higher levels of projected grants under PRGF-supported programs are not purely a consequence of assistance under the heavily indebted poor country (HIPC) initiatives. Nine HIPCs had higher average official transfers under PRGF-supported projections and eight had higher official transfers for their last ESAF-supported annual arrangement. Among non-HIPC countries, six had higher levels of official transfers in PRGF-supported program projections versus only three in the ESAF-supported program projections. Results for nine countries are not available or are not comparable. The median level of official transfers under PRGF-supported programs declines by 0.6 percent of GDP between the current year and the third projection year—a somewhat smaller decline than under ESAF-supported programs, which showed declines over the projection period in a large majority of individual cases.

Structural Measures

Structural policies of PRGF-supported programs are generally consistent with the overall poverty reduction strategy. About 60 percent of all structural conditions could be derived directly from or were at least consistent with the I-PRSPs/PRSPs, and 24 percent were explicitly derived from these documents. Moreover, consistency with the I-PRSPs/PRSPs has been significantly better in programs supported by new arrangements under the PRGF approved since July 2000. About 80 percent of structural conditionality in these programs was consistent with the I-PRSPs/PRSPs, versus 45 percent in transformed PRGF-supported programs. In the cases of Chad, Georgia, Guinea-Bissau, Kenya, Lesotho, Macedonia, and Rwanda, for example, the link was particularly strong and explicit. These examples suggest that consistency between the I-PRSPs/PRSPs and PRGF-supported program documents will improve over time as transformed PRGF-supported arrangements are replaced by new PRGF-supported arrangements.

Nevertheless, there is considerable scope for improving the consistency between the PRGF-supported program and the authorities’ poverty reduction strategy. About one-third of the structural conditionality measures included in PRGF-supported programs still seem to have little relation to the I-PRSP/PRSP. To some extent, this may occur because program negotiations take place some months after the I-PRSP/PRSP is made final; it may also be attributable to new developments. If so, it would be useful to report in PRGF program documents on discussions with the authorities about their views on the continuation of structural reforms and how these views will be reflected in the I-PRSP/PRSP or subsequent progress reports.9 Staff reports and LOIs/MEFPs generally failed to cross-reference or note the consistency with the I-PRSPs/PRSPs, even in circumstances where there is a high degree of consistency. The same conclusion was noted in the submission from the United Kingdom for the PRGF review.10

Coherence of PRSPs and PRGF-Supported Programs and the Emphasis on Growth

The foregoing analysis points to a high level of consistency of PRGF-supported programs with the I-PRSP or PRSP. But this should not be read as suggesting that the PRGF-supported programs can be fully derived from I-PRSPs and PRSPs. First, while PRGF-supported programs are consistent with their I-PRSPs and PRSPs, the latter often set out macroeconomic policies solely in broad terms, and quantitative frameworks are typically skeletal, perhaps reflecting capacity constraints in some countries. This leaves a substantial amount of detail to be filled in by the PRGF-supported program. In addition, the consistency between the PRGF documents and the I-PRSPs and PRSPs does not rule out cases in which the macroeconomic goals and policies for the I-PRSP or PRSP are derived from previous PRGF-supported programs or ongoing program discussions. Such “reverse causality” has been reported on many occasions by donors and NGOs in instances where either authorities felt constrained to remain within the framework set out in earlier PRGF or ESAF requests or reviews, or they adopted the PRGF framework in the absence of any independent contribution on the macroeconomic framework from the PRSP process.

Nonetheless, despite some “chicken-and-egg” controversy, national authorities in countries with PRGF-supported programs generally concurred with the assessment that PRGF-supported programs are consistent with I-PRSPs or PRSPs. In response to a survey (Appendix II), some 87 percent agreed (or agreed strongly) that their PRGF-supported program is consistent with the government’s underlying poverty reduction strategy as set out in the I-PRSP or PRSP. Similarly, 82 percent agreed or strongly agreed that the goals of the PRGF-supported program are “consistent with the economic goals of the government and other national institutions.” However, some queried the adequacy of consultation about the macro framework in the context of the I-PRSP or PRSP. Some 59 percent agreed that the PRGF gave the authorities more opportunity to influence program design, but 35 percent were neutral and 6 percent disagreed.

As well as indications of an increase in country ownership of PRGF-supported programs, there is broad evidence that the early PRGF-supported programs are placing additional emphasis on growth as a means of poverty reduction in program design, and doing so to a greater extent than under previous IMF supported programs, including on private sector development. Based on the reform agenda set out in PRGF-supported programs, macroeconomic projections and assumptions also show an increased emphasis on growth and expectations of increased external assistance. National authorities have expressed the view that PRGF-supported programs place sufficient emphasis on growth as a means of poverty reduction (more than four-fifths of the respondents agreed or strongly agreed, Appendix II) and also that the emphasis on poverty reduction is higher than under previous ESAF-supported programs (more than four-fifths of the respondents agreed or strongly agreed, with nearly half agreeing strongly). Growth is critical for achieving poverty reduction, and attention to the sources of growth is essential in developing appropriate policies and projections. Notwithstanding the preceding views, it is necessary to remember that these are expectations at the program design stage, and the outcome will depend on both the effectiveness of program implementation and the international economic environment in which countries undertake these efforts.

Highlight Flexibility in Accepting Country Choices

Relatively few staff reports for PRGF-supported program requests or reviews explicitly indicate areas in which alternative policy options were considered or in which the final choices were different from those initially considered by the authorities or by IMF staff. About one-fifth of the reports (12 of 63) identified such circumstances covering both macroeconomic and structural issues. The staff report for Lesotho (March 2001) provides a good example. The document describes the process of moving from the broad policy objectives in the I-PRSP to the specific fiscal commitments in the program, noting how a compromise was reached between positions of the staff and authorities that was compatible with the authorities’ macroeconomic goals. The staff report also highlights continuing differences between the staff and the authorities on the promotion of specific economic activities and on import liberalization. The staff report for the third review of the PRGF-supported program for Ghana (June 2001) provides another example of staff flexibility; it cites an accommodation of the authorities’ preferences with respect to the pace of interest rate adjustments and foreign exchange liberalization, where in both cases the staff had advised faster actions.

Most staff reports do not explicitly cite such evidence of flexibility, even though there is frequently considerable change in the design of the program during the extended process of program negotiations. Part of the explanation for this silence may relate to the drafting approach taken in many staff reports. While difficult to quantify, it is clear that many staff reports are drafted in an impersonal style that avoids attribution of views to staff, authorities, or anyone else (e.g., through use of passive voice). Other reports describe the economic program and cite only IMF staff views in support of the program’s content. While understandable as a reflection of staff’s desire to remain consistently supportive of presenting the authorities’ program to the Board, these drafting styles likely contribute to the view held by some outside commentators that IMF-supported programs are developed through the unwavering imposition of a single, made-in-Washington model of economic policy. These drafting styles also stand in sharp contrast to that expected for Article IV staff reports where an explicit presentation of both sides of the policy dialogue is required. In this context, it may not be entirely coincidental that the examples of both Ghana and Lesotho cited above are taken from reports that were also Article IV staff reports, whereas the staff reports for stand-alone PRGF-supported program reviews for Ghana and Lesotho in the sample do not exhibit such differences of view.

Many donor agencies and civil society organizations (CSOs) have emphasized the value of providing more information in IMF staff reports on the evolution of program design. Country authorities have been more ambivalent about this because it might appear to qualify or compromise the IMF’s support for the program as a means of achieving the program’s goals and consistent with the need to maintain frank and confidential program discussions. On balance, this appears to be an issue on which IMF staff reports could go further, starting with cases where the authorities themselves are keen to show how elements of the program have evolved during the design stage.

Staff could also encourage authorities in countries with PRGF-supported programs to take measures to promote greater participation and national ownership and support these measures through their own efforts. The need for additional efforts in this regard was highlighted repeatedly in external comments for this review and has also been discussed by the Board in the context of the recent paper on strengthening country ownership.11 Staff should encourage an active and open discussion of options by the authorities through whatever means the authorities choose (e.g., releasing summaries of options under discussion or meeting with press and civil society), and mission teams and resident representatives should assist authorities in doing so (e.g., participating in seminars or press conferences). Authorities could also be encouraged to establish or make use of websites for this purpose.

Existing policies already encourage IMF missions and resident representatives to be in contact with legislators, trade union representatives, and other civil society organizations.12 Such contacts are already quite common, as has been indicated through surveys of IMF mission teams. There is a clear sense from external inputs to the Review of the PRGF and the Review of the PRSP Process, though, that these contacts have not gone far enough. Additional efforts may be appropriate for both mission teams and resident representatives. Nevertheless, care must be taken to recognize that the focus of the participatory process and the main means of developing national ownership should rest with the PRSP process, not the PRGF discussions. There has also been increasing focus on resident representatives’ work with civil society organizations to promote greater understanding of the IMF, and IMF-supported programs, and the importance of efforts in this area could be reaffirmed. Finally, it is important that the staff efforts be in support of the authorities’ own dissemination strategy and respect the modalities and institutions the authorities believe most appropriate for their circumstances.

Publication of PRSP and PRGF-Related Documents

Recent changes in IMF publication policy and practice have brought about nearly universal publication of the authorities’ policy commitments and substantial disclosure of staff documents assessing these programs. Both staff and authorities’ documents relating to the PRSP process (the I-PRSP/PRSP, and the JSA) and PRGF LOIs/MEFPs are presumed to be published and have been published for almost all Executive Board discussions in 2001. Staff reports for the PRGF, for which publication is purely voluntary, have been published for about half of PRGF requests or reviews in 2001.

These publication results represent a sharp shift from even the very recent past. Before 1998, well below half of the authorities’ policy intentions documents were published for PRGF countries. With the shift to a presumption of publication in 1999, publication of these documents has become the norm with only a handful of exceptions. Staff reports could not be released at all before 1999, and they could be released only on a narrowly restricted basis before the approval of the transparency decision (see IMF PIN No. 01/3, January 12, 2001). The presumption of publication also applies to JSAs, starting with the approval of the transparency decision, and all JSA documents issued since the approval of that decision have been published.

PRGF and PRSP Documents Published by Year:

(Percentage published; absolute figures in parentheses)

article image

For 2001, includes papers discussed by the Executive Board from January through November 2001 only, and therefore cover more documents than used for other aspects of the review.

Includes I-PRSPs, PRSPs, progress reports on PRSPs, and Status Reports on Implementation of PRSPs.

JSAs considered by the Executive Board in 2000 could be published only after the transparency decision went into effect on January 4, 2001. The only unpublished JSA for 2001 Executive Board consideration was issued to the Executive Board in December 2000; it was discussed only in January 2001.

LOIs/MEFPs were not published for one review each for Chad, Guyana, Senegal, and Yemen.

Stand-alone PRGF staff reports became eligible for publication under the transparency decision. Before that, and in the context of the pilot project for the voluntary release of Article IV staff reports, PRGF requests for staff reports could be released only for Board discussions after August 29, 2000, and published only after the January 4, 2001, transparency decision, except for earlier reports participating in the pilot project for voluntary release of Article IV staff reports (and then only combined with Article IV consultations).

The increased transparency of the IMF has expanded the audience for IMF documents (Box 3) and increased the importance of clearly setting out staff’s and authorities’ views. In the past, the implicit or indirect approach to explaining or hinting at difference of views between the staff and the authorities was less prone to yield misunderstanding because the readership was limited to a comparatively small set of Executive Board staff and national authorities familiar with the IMF and skilled at reading between the lines. Notwithstanding the now wider audience, many staff reports continue to avoid explicit discussion of such differences. Greater frankness about disagreements on policies and more direct discussion of the policy dialogue is needed to make such views clearer in staff reports. But the need for a more explicit discussion of the policy dialogue needs to be implemented in a manner that is consistent with the staff’s support for the program as a means of achieving its goals and that does not compromise the potential for frank and confidential program discussions.

Focus on Areas of IMF Expertise and Responsibility

Structural conditionality in PRGF-supported programs has concentrated on areas of IMF expertise consistent with the objectives set out for the facility. Streamlining structural conditionality is itself one of the key features of PRGF-supported programs, and the IMF staff’s efforts to focus the PRGF contribution to the overall strategy in terms of structural conditionality is covered in Section VI below. Outside the context of streamlining structural conditionality, there is little treatment of the role of other institutions or implementation of other parts of the poverty reduction strategy in IMF documents.

Two specific aspects of the authorities’ overall poverty reduction strategies are singled out in the various Board papers for special emphasis: measures to promote growth led by the private sector and measures directly targeting, monitoring, and alleviating poverty. Consistent with the objective of the PRGF, measures to promote private sector development are discussed prominently in PRGF-supported program documents. At the same time, conditionality on these measures has been selective because they often fall in the policy domain of the World Bank. The total number of conditions included in programs to promote private sector growth drops by more than half (from 3.7 to 1.45) between the ESAF and PRGF parts of the sample, and four-fifths of these measures in both ESAF- and PRGF-supported programs cover privatization and public enterprise reform. Only about one-fifth of the ESAF (18 percent) and PRGF (22 percent) requests and reviews surveyed had prior actions, performance criteria, structural benchmarks, or review topics directly focused on reducing poverty; these mostly related to expenditure allocations.

There remains substantial room for improvement in the presentation of the role of the World Bank and other institutions in supporting poverty reduction strategies in IMF staff reports. Few staff reports routinely describe the activities of the Bank (or other institutions) in promoting alleviation of poverty or place the PRGF-supported program in a broader context. The boxes on streamlining structural conditionality (see Box 5 in Section VI) have been some-what helpful in this regard; but these would tend to pick up the role of other institutions only to the extent that they are active in areas that have been subject to conditionality in recent IMF-supported programs. Measures to promote reduction of poverty have not generally been a part of conditionality of previous IMF-supported arrangements, and therefore they are less likely to be picked up in the descriptions of World Bank activity in structural conditionality boxes (or elsewhere in IMF documents), including descriptions in the standard appendix on World Bank Relations.