Structural Conditionality in IMF-Supported Programs
Chapter 4 The Impact of the Streamlining Initiative
- International Monetary Fund. Independent Evaluation Office
- Published Date:
- April 2008
37. This section reviews the effects of the streamlining initiative. It introduces the Interim Guidance Note (IGN) and the 2002 CG, the two main documents that have regulated the Fund’s approach to SC since 2000, and discusses their goals and main requirements. It then examines the factors that affected their implementation in regard to the volume and composition of conditionality.
A. The 2000 Interim Guidance Note and the 2002 Conditionality Guidelines
38. In 2000, largely in response to the proliferation of SC in the mid-1990s and the criticism that this drew, IMF management issued an IGN. The IGN aimed at strengthening country ownership of reforms and attempted to lay out principles for narrowing the scope of SC through the requirement that structural conditions be “macro relevant.” A review of conditionality was subsequently launched and in September 2002 the Executive Board approved the 2002 CG, which replaced both the 2000 IGN and the 1979 Guidelines on Conditionality.
39. The CG re-emphasized the importance of national ownership and of effective coordination with the World Bank, and called for “parsimony” and “criticality” in the use of conditions. It stated that conditions would be established only if they were “of critical importance for achieving the goals of the member’s program” and that they “will normally consist of macroeconomic variables and structural measures that are within the Fund’s core areas of responsibility.”
40. The CG was generally expected to lead to a reduction in the number of conditions. 12 From the start, however, there was some debate about what criticality implied for the sectoral composition of SC. Though the CG left some room for conditions to be established outside the Fund’s core areas of responsibility, it set the expectation that such instances would be exceptional and it required that in these cases staff provide a more detailed explanation of the conditions’ critical importance. The CG called for the application of a “lead agency framework” when dealing with reforms outside the Fund’s core areas of responsibility. In particular, it stressed the need for close collaboration with the World Bank in establishing and monitoring these reforms. The language of the CG, therefore, left some discretion to allow the Fund to deal with country-specific circumstances.
B. Volume and Composition of Conditionality: What Happened in Practice and Why?
41. There is no evidence of a reduction in the number of structural conditions following the introduction of the streamlining initiative (Figure 2). The average number of structural conditions per program/year remained at around 17, when comparing the four-year period following the start of the streamlining initiative (2001–04) with the four-year period that preceded it (1997–2000). In PRGF arrangements, this average declined slightly from around 16 to 15 conditions, while in SBAs it rose from 18 to 19. But none of these changes is statistically significant. Neither did significant changes take place in the shares of PAs, PCs, or SBs in the total number of conditions. These findings are somewhat surprising because the CG argued that “arrangements supporting longer-term programs will have more extensive SC […] because structural reforms are more likely to be critical for the achievement of longer-term goals,” and it also explained that excessive reliance on PAs could tax implementation capacity and hinder effectiveness.
Figure 2.Average Number of Structural Conditions Per Program Year, 1995–2004
Source: MONA database and IEO staff estimates.
42. The streamlining initiative did affect the thematic and sectoral composition of structural conditions. Both in PRGF arrangements and SBAs, SC shifted out of privatization of SOE and trade reform toward tax policy and administration, PEM, and financial sector reform. These findings are consistent with internal IMF studies (IMF, 2005a, b) that show that the number of conditions did not decline and that SC shifted toward Fund core areas: economic management and the financial sector. Some of these sectoral shifts reflected a change in approach rather than a response to the underlying economic situation. For example, the analysis in Background Document Chapter II indicates that the number of conditions in financial sector and privatization was correlated with the corresponding level of distortions in the sector. It appears that this relationship shifted at around the time the initiative was launched, i.e., for a given degree of distortion, the number of conditions increased in the financial sector and fell in privatization. In some sectors the CG reinforced shifts that had begun earlier. For example, in the financial sector the number of conditions began rising in 1998, linked to the growing attention being paid to financial sector surveillance after 1997.
43. The streamlining initiative did not lead to better explanations of why conditions, either core or non-core, were critical, nor of how these conditions would contribute to program objectives. Well-specified medium-term roadmaps were present in PRGFs only in countries whose Poverty Reduction Strategy Paper (PRSP) contained a well-developed medium-term policy assessment. To examine how program documents have justified the conditions being imposed, the evaluation compared documents from a sample of IMF arrangements approved during 2004–05 with those for the 13 in-depth case study programs (approved in 1999–2003). It found that the use of specific structural conditions was no better explained in the recent programs than in the earlier ones. 13 In most cases, program goals were very broad, the specific measures subject to conditionality were far removed from the ultimate objective they were supposed to serve, and the intermediate goals or mechanisms through which the measures would affect the goals were not specified. The more recent PRGFs continued to fall short of providing clear policy roadmaps. Only for Mozambique, whose PRSP has a well developed medium-term policy assessment framework, was it found that the PRGF documentation conveys an adequate multi-year perspective.
44. The World Bank’s sectoral distribution of conditionality changed in the same direction as the Fund’s over the same period, and indeed the Bank’s shift away from conditionality on privatization and SOE restructuring was more pronounced than that of the IMF (World Bank, 2005). Fund arrangements often included detailed PC and PA in areas of core World Bank competency while the Bank imposed only vague or no conditions at all. At the same time, both organizations increased their involvement in the financial sector and in basic fiduciary issues. Given the above-mentioned lack of roadmaps in program documentation, it is difficult to judge whether these developments reflect serious problems in Bank-Fund cooperation or better prioritization of structural reforms by both institutions. In any case, these developments run contrary to the expectation that the Bank would lead the work in non-macro core areas, while the Fund would reduce its involvement. This finding raises questions on how the Fund should address critical reforms in IMF non-core areas, particularly when the World Bank is not addressing them.
45. The evaluation findings are not fully consistent with the aims of the CG, since (i) the number of structural conditions did not diminish; and (ii) arrangements continued to include conditions that do not appear to have been “critical to the program objectives,” while some critical issues, particularly in non-core areas, may have been ignored.
46. Several factors contributed to the lack of progress in reducing the number of conditions. A review of directors’ discussions on arrangements submitted to the IMF Board during 2003–04, as well as the discussions during country programs and the bi-annual reviews of conditionality, shows that the lack of progress was not of great concern to the Board or to management. 14 In specific country programs, objections to excessive conditionality by some executive directors were offset by the interest of other directors, usually from donor countries, in the role that these conditions could play in monitoring their own initiatives. Generally, directors calling for more conditions had a much stronger interest in the programs in question. 15 A similar example relates to the support by many EU members for multiple and wide-ranging conditions in arrangements with countries that are in the EU accession process. In a few cases, the authorities themselves asked to add conditions that they expected would help them signal to donors a change in their commitments to specific reforms. Moreover, during the 2005 review of the CG, directors noted that the number of conditions had not been reduced, but they agreed that numbers were “at best a crude metric.”
47. The Executive Board, management, and staff have argued that the shift of conditionality toward the core areas is consistent with the implementation of the CG. But views differ on whether programs have sufficiently addressed criticality. In particular, Board signals have been mixed, with some directors arguing that there is scope for further streamlining, and others expressing regret that streamlining conditionality might have gone too far and calling for greater coverage of growth- and efficiency-related reforms.
48. The staff survey conducted for this evaluation found widely differing views among staff on what the CG require in terms of focus (and numbers) of structural conditions, and on how the criticality test is being interpreted. About two-thirds of the staff surveyed believed that being within a core area of Fund expertise was a necessary condition for a measure to be subject to conditionality; while about 20 percent thought that the measure needed to be critical for the achievement of the program’s stated objectives independently of whether it was in the core. Notably, a majority of survey respondents felt that the streamlining initiative had weakened some programs, and about half believed that critical non-core conditions were being left out either “often” or “very often.” During focus group discussions, staff noted that in designing conditionality they paid close attention to the CG, as well as to executive directors’ past reactions to conditionality and program design—but that these had not provided clear guidance on the issue of criticality.
49. The Fund’s internal review process 16 emphasized the need to strengthen critical reform policies in IMF core and non-core areas alike. Sometimes reviewers called for additional conditionality, but they very seldom suggested dropping other specific conditions. The review process pointed out weaknesses in program documentation, sometimes requesting greater clarity of program objectives and of the strategies to achieve them, and calling for a longer-term outlook in the design of PRGF. But the review process had little impact on the quality of program documentation; for the ten PRGFs approved in 2004–05 that were examined for this part of the evaluation, most of the documents had objectives that were too broadly stated, weak explanations of the intermediate strategies to reach them, and no thorough explanation of why certain conditions were included. One important cause for these shortcomings was raised by the reviews themselves—programs appeared to be rushed and premature. Sometimes the rush may have only affected the quality of the documentation, but in other instances it may have also hindered the program design and its effectiveness, particularly in countries with a weak track record in reform. The rushed timetable was often set by exogenous processes, e.g., meeting deadlines for the HIPC process, while in other instances it was determined by an attempt to support a reformist group within the government.
For example, the “Operational Guidance Note on the Conditionality Guidelines”—issued in 2003 and amended in 2006—stated that the 2002 CG would “replace the somewhat weaker standard of macro-relevance in the earlier [IGN]” by a test requiring that any variable selected for formal conditionality must be of “critical importance” in that “if it were not implemented, it is expected that the goals [of the program] would not be achieved.” Also, during the most recent review of conditionality, staff noted that “The guidelines were also expected, broadly speaking, to lead to reduced numbers of structural conditions” (IMF, 2005a, p. 14). As for expectations outside the IMF, Abdilina and Jaramillo-Vallejo (2005) note that “The streamlining effort entailed reducing the number of conditions in programs while sharpening their focus” (p. 85).
Good justifications were provided in three of the ten programs (Ukraine, Bulgaria, and Mozambique), but only moderately good explanations were given in another two (Burundi and Georgia), and fairly poor ones in the rest (Republic of Congo, Gabon, Honduras, Peru, and Zambia).
The sample includes all programs requested in 2004 and some requested in 2003—split in half between SBAs and PRGFs. The overall group includes Argentina (SBA, 2003), Bangladesh (PRGF, 2003), Bolivia (SBA, 2003), Ghana (PRGF, 2003), Kenya (PRGF, 2003), Nepal (PRGF, 2003), Paraguay (SBA, 2003), Macedonia (FYR) (SBA, 2003), Gabon (SBA, 2004), Bulgaria (SBA, 2004), Croatia (SBA, 2004), Peru (SBA, 2004), Romania (SBA, 2004), Ukraine (SBA, 2004), Burundi (PRGF, 2004), Georgia (PRGF, 2004), Honduras (PRGF, 2004), Mozambique (PRGF, 2004), Republic of Congo (PRGF, 2004), Zambia (PRGF, 2004), and Mali (PRGF, 2004).
About half of directors’ interventions did not call for either more or less structural conditionality, and several indicated satisfaction with the proposed volume of structural conditionality. Of the remainder, those calling for more structural conditionality outnumbered those calling for less by a wide margin. In some cases, directors acknowledged that the number of conditions was very large but accepted it nonetheless, due to country-specific circumstances—particularly in EU accession countries and countries with poor past records of implementation. This finding is in line with staff responses to a survey, where one-third noted that directors had raised concerns about insufficient structural conditionality. In contrast, only 6 percent indicated that directors had expressed reservations about excessive conditionality.
The internal review is conducted by staff in departments other than the corresponding area department. The analysis of the internal review was based on documents for ten arrangements approved in 2004.