Lessons from the Real-Time Assessments of Structural Conditionality

Lessons from the Real-Time Assessments of Structural Conditionality


Lessons from the Real-Time Assessments of Structural Conditionality

I. Introduction

1. The IMF is moving to strengthen the effectiveness of the programs it supports by streamlining and focusing conditionality and fostering greater national ownership of these programs. The structural conditions attached to IMF financing are to be focused on those measures that are critical to the scope and nature of conditionality. The overriding principle is to concentrate conditionality on those measures that are critical to achieving a program’s macroeconomic objectives and on measures that are macro-relevant and are within the Fund’s core area of responsibility. These basic principles were presented in the Interim Guidance Note on Streamlining Structural Conditionality issued to Fund staff by the Managing Director in September 2000.

2. At the Executive Board discussions of Conditionality in March and July 2001, Directors endorsed the principles laid out in the Interim Guidance Note, and agreed that experience in their application to specific country cases would help draw the line between measures that are critical to individual program objectives and those that are relevant but not critical. They recognized that striking the right balance would not be an easy task and would typically be a matter of judgment. If the criteria for conditionality were defined too narrowly, there was a risk that important areas of reform—that might, indeed, prove critical ex post—would not be properly covered. On the other hand, if defined too widely, the objective of the streamlining initiative itself could be undermined. Directors therefore encouraged staff to build up experience with the application of the Interim Guidance Note in various country circumstances, while learning from past cases and taking into account further guidance from the Board. The feedback provided by the Board during discussions of new Fund-supported programs (which came to be known as “real-time assessments of conditionality”) should be an integral part of this guidance. This paper draws on these real-time assessments to discuss how the balance in the scope and detail of Fund conditionality may be refined, and, in particular, how to assess which policy measures are critical to program objectives.

II. The Real-Time Assessments

3. The experience with streamlining and focusing structural conditionality in individual countries is assessed on the basis of 19 representative case studies discussed by the Executive Board since the issuance of the guidance note and through November 2001. Two categories of programs were chosen for this task. First, the sample included the 13 new programs (including second- and third-year arrangements under the PRGF) brought to the Board since mid-May 2001, when text boxes describing the extent and justification for the choice of conditionality were included for the first time in Fund documents.1 Second, the sample was supplemented with 6 other cases that were notable either for their extensive conditionality (Turkey, negotiation of SRF); their significant streamlining (Latvia); or particularly lively Board discussions on the subject (new programs for Georgia and Vietnam, and reviews for Gabon and Mauritania). The presentation of the case studies is selective, in the sense that not every country is discussed under each aspect. However, to provide a “complete” picture of all 43 new programs discussed by the Board during the one-year period following the issuance of the interim guidelines, the discussion of case studies is complemented by an update through November 2001 of the numerical analysis included in the earlier Review of Initial Experience, broadly confirming its main conclusions (see Annex).

4. Staffs implementation of the guidance note resulted in considerable variation in the extent of conditionality, which was largely, but not exclusively, a reflection of different country-specific circumstances. In financial crisis situations, where structural weaknesses (e.g., in the banking sector) were a major vulnerability factor and the financing need was sizeable (Turkey), structural conditionality was more extensive and detailed. On the other hand, in less acute circumstances (Latvia), the staff was more willing to streamline. Beyond macroeconomic factors, staffs judgment in individual cases was also influenced by (i) a country’s track record, the lack or weakness of which was typically associated with more extensive conditionality (Gabon, Mongolia, Yugoslavia); (ii) considerations of political feasibility, shifting the balance toward stronger streamlining (Bolivia, Vietnam); and (iii) limited administrative capacity, implying a similar effect (Kyrgyz Republic, Mauritania). Remaining variations beyond these factors were largely a reflection of the considerable room for judgment provided in the Interim Guidelines.

5. Directors supported the notion that country-specific circumstances warranted differences in the coverage of conditionality, but leaned toward narrowing differences in the extent of conditionality across countries.2 They broadly endorsed more extensive conditionality in crisis situations (Turkey)—and, more generally, when access levels were high. Also, somewhat tighter conditionality (relative to the average program) was welcomed in countries where the track record under past programs was weak (Gabon, Mongolia) or needed to be newly established (Yugoslavia). The Board was further willing to take account of political factors and administrative capacity in determining the right balance of structural conditionality, in recognition of the importance of country ownership for the success of programs.3 Nevertheless, in cases in which conditionality remained extensive (Gabon, Yugoslavia), Directors predominantly questioned the inclusion of certain measures under Fund conditionality while they demanded explanations for the exclusion of measures in highly streamlined cases (Bolivia, Latvia, Vietnam). Thus, the Board broadly supported a differentiated approach to streamlining based on different country circumstances, but not to the extent exercised by the staff. Box 1 illustrates this point on the basis of three cases, two (Bolivia and Mauritania) representing fairly vigorous and one (Gabon) very little (if any) streamlining of structural conditionality.4 While a number of Directors commented favorably on the streamlining approach in Bolivia’s and Mauritania’s programs—especially on the attention paid to implementation capacity and political constraints—several concerns were voiced on whether the staff had gone too far. On the other hand, in Gabon’s program (which was negotiated prior to the issuance of the guidance note), most Directors felt that the extent of conditionality, particularly in non-core areas, was excessive and hampered ownership.

Programs With Extensive and Limited Streamlining: The Cases of Bolivia, Mauritania, and Gabon

Mauritania and Bolivia were the two countries that streamlined the most over the sample period. The number of conditions in the Mauritanian program declined from 35 in the first year of the program to 6 in the second year, while the number of conditions in the Bolivia program declined from 25 in the second year of the program to 7 in the third year. Gabon’s program, on the other hand, which was negotiated before the issuance of the Interim Guidance Note, displayed very little (if any) streamlining of structural conditionality. It included 60 conditions (9 PAs, 11 PCs, and 40 SBs).

During the third review under the Mauritanian program, staff justified the decline in conditions on the basis of streamlining conditionality, previous track record, and level of ownership, while acknowledging the limited administrative capacity of the government to implement all the conditions attached to the initial program. The new conditions fell within the Fund’s core areas and were critical for the achievement of the program’s macroeconomic objectives. While the decline in conditionality was substantial, a number of Directors did not feel that too much streamlining had taken place between the first and second years of the program. Rather, the large decline was an indication that the number of conditions in the initial program were excessive, especially given the limited implementation capacity of the country. Nevertheless, other Directors had doubts about the sharp reduction in the number of measures. There was general recognition that the missed PCs on pricing policy, private sector development, public enterprise reform, and privatization could be waived, but the exclusion of conditionality in the areas of customs receipts, import surcharges, municipal tax reform, and public procurement was questioned. Directors requested, in general, a more explicit rationale for the inclusion and exclusion of certain measures which prompted the insertion of conditionality boxes in future country reports. Moreover, they advised the staff to take more account of the implementation capacity in program countries in the future.

In the Bolivian program, staff explained the sharp decline in the number of conditions with the need to focus on critical measures and to facilitate the World Bank’s leadership in conditions that were not in the Fund’s core area. Directors provided a mixture of views on this experience. While a number of them welcomed staffs efforts in streamlining conditionality, indicating the sea change in policy that had taken place, others felt that the reduction in conditionality was heavy handed, especially since 18 of the original 25 conditions were either partially or fully unmet. Various views were presented on the decision to exclude tax and labor market reforms from conditionality. According to some Directors, since these policies were not covered by the World Bank, it would have been appropriate to include conditionality in the Fund-supported program. Others, however, viewed the exclusion of these measures as a candid recognition of the immediate political realities in Bolivia, since the needed political consensus on these issues had not yet been achieved. Some Directors noted that the current exclusion of these conditions from the program should not be viewed as permanently streamlining them out of conditionality. There would be an opportunity for the staff to take up these issues at a later date.

During the first and second reviews under Gabon’s program, staff acknowledged that the extensive conditionality might adversely affect program ownership, but explained the number of conditions with the authorities’ wish to restore credibility in light of previous policy slippages and with the large budgetary impact of public enterprises. The majority of Directors saw a need to reduce conditionality, calling for more careful prioritization and sequencing of measures—the lack of which was partly to blame for six unobserved performance criteria (four of which structural). One Director stressed that the desire to impress creditors and donors with overly ambitious conditionality was the wrong approach to designing a properly-owned program. Conditionality should be matched with the country’s administrative capacity. Others, supporting this view, noted that the Fund should be more patient for the sake of ownership. One Director was concerned, however, that too much patience would lead to missed opportunities, and did not support a reduction in conditions in light of Gabon’s poor track record.

6. Regarding the overall extent of streamlining, Directors broadly supported staffs choice in SBAs, but favored, on balance, less streamlining in the PRGFs included in the sample. In particular, a number of Directors requested more performance criteria in a number of PRGF programs where conditionality was sparse (Benin, Bolivia, Kyrgyz Republic, Mauritania, and Vietnam), preferring to maintain some conditionality for measures that they believed could potentially prove critical and were not adequately covered by the World Bank or other institutions. To some extent, the Board’s reactions reflected the specific sample: while the number of conditions in new programs for the sample were in line with those in the entire set of new SBAs and PRGFs, the extent of streamlining in the selected PRGFs was particularly strong, reflecting the relatively high number of conditions in their previous programs (see Annex). However, it could also suggest the preference of some Directors to be cautious in applying the general principle of parsimony to specific country cases. Irrespective of the expressed preference for less streamlining in the selected PRGFs, a few Directors were concerned that from the country’s perspective, the initiative of streamlining conditionality had not resulted in fewer conditions when measures from all IFIs were summed together.

7. In terms of implementing conditionality in the core areas (defined to include fiscal, financial, and exchange rate policies, in some of which areas responsibility is, however, shared with the World Bank), staff focused on fiscal measures, typically centered on improving expenditure management and/or enhancing revenue performance. Measures to improve public resource management were dominant in PRGF programs (Azerbaijan, Benin, Bolivia, Mali, Mauritania, and Sierra Leone—all HIPCs with the exception of Azerbaijan). This is consistent with the emphasis in recent Fund documents on the need to maintain control over expenditures and with the requirement for HIPC countries reaching their decision points to monitor the use of additional budgetary resources flowing from debt relief. Many PRGF countries also had conditions on increasing the efficiency and yield of the tax system (Bolivia, Mongolia, Sierra Leone) or containing wage increases (Benin and Honduras). Fiscal measures in SBAs were more varied, ranging from improving municipal finances and reforming the tax system in Lithuania to implementing an integrated information system for public finances in Gabon. In general, since fiscal weaknesses are often at the root of countries’ macroeconomic problems, the macro-relevance of these conditions was typically taken for granted by the staff and not explicitly discussed in the reports.

8. Measures in the financial sector centered on bank liquidation/privatization and strengthened bank supervision, introduced with the explicit aim of improving the soundness of the banking system. In the Kyrgyz program, banks would be liquidated if they persistently failed to meet the minimum capital requirements or became insolvent, in Benin and Lithuania an action plan/tender needed to be formulated to privatize the remaining state bank, and in Bolivia, the approval of financial sector legislation was required to strengthen bank resolution procedures and facilitate prompt corrective action for problem banks.

9. Directors commentedfavorably on the design of conditionality in core areas, stressing the importance of conditions in the fiscal domain. Indeed, in the entire sample of Board discussions, no request was made for further streamlining in the fiscal area. Specifically, Directors highlighted the appropriateness of conditionality on local authority finances (Lithuania, Mongolia) and public sector wage increases (Honduras, Romania). In the financial sector, Directors were also generally content with the degree of conditionality, stressing, however, the need to ensure that measures are internally consistent (e.g., in Bolivia, Directors emphasized that measures for encouraging bank lending may be difficult to reconcile with initiatives for strengthening prudential supervision).5

10. Conditions in non-core areas varied considerably across countries and were mostly justified by the staff on the basis of their critical impact on the public finances. Areas covered by the Fund’s conditionality included privatization (Gabon, Kyrgyz Republic, Lithuania, Mauritania, Yugoslavia), public enterprise reform (Azerbaijan, Gabon, Kyrgyz Republic, Mali, Romania, Yugoslavia), trade policy (Yugoslavia), pricing policy (Mali, Yugoslavia), civil service reform (Mali), labor market reform (Lithuania), and governance (Azerbaijan, Kyrgyz Republic, and Honduras). In many (but not all) cases, staff reports provided some explicit justification for the inclusion of measures in these non-core areas. The dominant explanation was that the measures were critical to the country’s fiscal performance, e.g., limiting budgetary subsidies, safeguarding the public finances, reducing the burden of external debt, or reducing quasi-fiscal deficits. For measures not motivated directly by their budgetary impact, explanations tended to be related to the need to promote investment and growth (Kyrgyz Republic, Yugoslavia).

11. The exclusion of measures that were relevant to macroeconomic objectives, and had typically been subject to conditionality in earlier programs, was explained in most cases by reforms’ being outside the Fund’s core area. Often this went hand-in-hand with coverage under World Bank conditionality (e.g., on privatization in Azerbaijan, Latvia, and Romania, energy sector reform in Azerbaijan and Georgia, or civil service reform in Sierra Leone).6 Other justifications for the exclusion of individual measures were that (i) conditionality was no longer needed, reflecting progress already achieved (energy sector reform in Lithuania), (ii) improvements would be achieved indirectly through conditionality in other areas (governance in Romania); and (iii) the prospective benefits of the reforms were slow to materialize and intermediate targets were difficult to establish (labor market reform in Brazil).

12. The coverage of reforms outside the Fund’s core areas of responsibility turned out to be the aspect of conditionality on which views diverged most widely among Directors. Opinions differed both across and within countries on how specific areas (in particular, privatization, public enterprise and civil service reform) should be treated under Fund conditionality. This diversity of views appears to reflect a variety of factors. First, individual Directors expressed different preferences about the overall extent of streamlining in general, abstracting from individual circumstances. Second, the determination of whether reforms are critical—and thus, according to the Interim Guidance Note, should be included under Fund conditionality—or relevant but not critical, is difficult and often a matter of judgment. Third, Directors found it difficult to assess whether a measure was critical when staff reports failed to provide explicit explanations on how their decision was made.

13. Board comments on specific conditions in non-core areas again revealed a tension between the desire to tailor conditions to individual country circumstances and concerns about uniformity of treatment. Overall, there were few areas where Directors uniformly requested more or less conditionality: individual demands for less conditionality in a certain area in one country (e.g. labor market reforms in Lithuania) were typically matched by requests for more conditionality in the same area in others (Bolivia, Yugoslavia). The only exceptions to this rule were conditions on governance (Box 2), where no Board member requested further streamlining but several either applauded the inclusion of conditions or were concerned about too much streamlining. When commenting on the appropriateness of specific conditions, Directors took note of individual country circumstances. For example, they welcomed the inclusion of Fund conditionality on privatization in some programs because they were a critical component in reducing fiscal pressures and would enhance the investment climate (Yugoslavia). On the other hand, they were content with its absence from the Honduran program because the telecommunications company under consideration was in surplus (Box 3). However, concerns about equal treatment were voiced whenever certain conditions were included in some countries and not in others in seemingly similar circumstances (e.g. civil service reform in Mali but not in Sierra Leone).

Conditionality on Governance

In recent years there has been a strong impetus to improve the quality of governance in countries with Fund-supported programs (associated with the fostering of efficiency and private sector confidence) and thereby stimulate growth. The Fund presented its views on the importance of governance in a public document issued in 1997 (Good Governance: The IMF’s Role) and later in a Board document issued in February 2001 (SM/01/30). As a result, governance-related measures have played an increasingly important role in Fund conditionality. Measures to strengthen governance typically include improvements in the management of public resources (e.g. the treasury, central bank, public enterprises, civil service, official statistics) and aim, more generally, at the development and maintenance of a transparent and stable economic environment.

Among the sample under study, governance-related measures were included in 6 of the 19 programs with conditions focused on proper audits (Azerbaijan, Gabon, Lithuania), and accountability in government operations and simplifying regulations on tax and customs administration (Kyrgyz Republic). Interestingly, no Director felt that too many conditions were included in this area. Instead, there were calls for additional conditionality for the Kyrgyz Republic and Georgia, to stem the continuation of widespread corruption in these countries and reverse their previous weak track records. Directors felt that in these countries strengthened governance—and the associated improvement in the business environment—was crucial for the successful implementation of the program.

Conditionality on Privatization

There was general agreement that privatization was a relevant issue in many countries, but differences emerged on the appropriate use of conditionality on privatization in individual cases. In its justification of conditionality, the staff distinguished between privatizing public enterprises in the financial sector and those in the non-financial sector that were burdening the budget.

Directors were in general agreement on this distinction, arguing that the privatization of state banks was critical in strengthening the soundness of the financial system. They broadly supported requests for action plans/tenders to privatize the remaining state banks in Benin and Lithuania and for the performance criterion on privatizing the largest state-owned bank in Romania. There was more disagreement with the staff on conditionality placed on privatizing public enterprises outside the banking sector. Although, Directors supported the decision for the World Bank to take the lead role in monitoring conditionality on privatizing the telecommunications company in Honduras, because it was running a surplus, some were surprised about the absence of measures on privatizing infrastructure in Latvia, arguing that privatization measures in this sector had experienced noticeable delays in the past. Moreover, a concern was voiced that the privatization of SOEs in Romania was not subject to conditionality.

In designing conditionality on privatization, Directors acknowledged that the timing was fraught with uncertainty, as its implementation is generally not under the government’s full control. Thus, Directors were not surprised about a request to waive the condition on privatizing AgroGabon because of difficulties in finding a buyer within the specified time frame. To obviate such concerns, it was suggested to reclassify conditions on privatization as structural benchmarks rather than performance criteria and to specify them less precisely. In this light, Directors supported the use of structural benchmarks related to privatization in Yugoslavia and Lithuania. In addition, setting conditionality within the time frame of a Fund-supported program may run counter to the advice of the World Bank which in many cases would be the lead agency in this area. For example, many Directors accepted that cash collection targets were not useful for dealing with problems in Georgia’s energy sector, agreeing with the Bank’s preferred approach of a longer-term privatization strategy. This underscores the tension between choosing conditionality based on short-term revenue gains versus longer-term efficiency considerations.

14. In striking the right balance, Directors agreed with the staff that, in most cases, the fiscal impact was central to justifying conditionality in non-core areas. The budgetary implications linked to the removal of external imbalances were a recurrent theme in comments about the inclusion or exclusion of measures under Fund conditionality. For example, the Board supported measures to correct the primary deficit in the energy sector in Armenia because it had been an ongoing drain on public finances, and it suggested conditionality on the energy sector in Mongolia in the event that developments in the sector appeared likely to undermine fiscal performance. On the other hand, Directors questioned the application of conditionality to electricity pricing in Yugoslavia on the basis that its budgetary impact had not been established. As a general tendency, the weaker and less direct a measure’s impact on the fiscal accounts, the more likely Directors would ask for an explicit justification of its inclusion under Fund conditionality.

15. The Board also supported the use of growth as a motivation for conditionality, but mainly in PRGFprograms, and endorsed a tailored approach to conditionality in capital account crises. If weak growth was seen as hampering longer-term fiscal and external sustainability, Directors were supportive of reforms identified as priorities for removing growth impediments (e.g., governance and transparency in Azerbaijan), and in some cases voiced their consent even if the measures were associated with some short-term fiscal revenue losses (trade reform in Yugoslavia). However, they questioned conditionality on these grounds, when the association between the condition and improved growth prospects was weak (such as the reorganization of petroleum product distribution in Mauritania), and generally emphasized the need for staff reports to prioritize and explain conditionality under this justification. In general, the Board was less supportive of conditionality based on growth objectives in stand-by programs (e.g., labor market reforms in Lithuania), but endorsed a comprehensive structural agenda in Turkey, motivated by the need to restore market confidence.

16. Directors welcomed the initiative of including a box on conditionality in staff reports but felt that the information on the choice of conditionality could be expanded, especially on the division of labor between the Fund and the Bank. Requests for additional information were numerous (Georgia, Latvia, Lithuania, Mauritania, Romania, Vietnam, Yugoslavia), and were, in most instances, linked to concerns that potentially critical measures (of fiscal importance) were not covered under the Fund’s conditionality (e.g., privatization in Romania and energy sector reform in Georgia). In cases where Fund conditionality on relevant reforms was discontinued on the basis that it was the World Bank that was taking the lead, Directors requested that staff reports include detailed information on the structure and timing of Bank conditionality, the progress achieved, and the implications for the fiscal situation and the program in general.7 Moreover, they requested that the appropriateness of structural conditionality and the division of labor be kept under review throughout the program period.

17. In terms of modalities, Directors in some cases took issue with staffs choices between structural benchmarks and performance criteria, while they were generally satisfied with the selection of prior actions. In several cases, the inclusion of large numbers of prior actions was welcomed as a signal of the authorities’ commitment (Romania, Vietnam). Moreover, in countries that were radically transforming their economies, numerous prior actions were regarded as necessary to jump start the reform process (Azerbaijan, Mongolia, Yugoslavia). With respect to structural benchmarks, there were few requests from Directors for further streamlining despite the fact that the proliferation of benchmarks had accounted for most of the expansion of conditionality in recent years.8 This may suggest that Directors were now satisfied with the scope and number of structural benchmarks in programs. Indeed, in some areas, notably privatization, Directors expressed a preference for using structural benchmarks over performance criterion.

III. Lessons

18. A number of broad lessons emerge from the real-time assessments of the experience with streamlining and refocusing of conditionality, some of which strengthen previous conclusions on the proper approach to conditionality.

19. First, to ensure focus and consistency in the application of structural conditionality and support program ownership, measures covered by conditionality must have a clearly identifiable link to the program’s macroeconomic objectives, grounded in an analysis of the country’s balance of payments problems. Achieving a sustainable balance of payments is the central objective of all Fund-supported programs, and Directors supported the staff’s choice of conditionality when it pertained directly to this objective, especially with regard to measures that underpin fiscal adjustment. Moreover, any extension of conditionality beyond the Fund’s core areas of responsibility must be strongly substantiated. Beyond these common goals, Directors acknowledged that other considerations may come into play in determining program objectives and the appropriate conditionality, while emphasizing two points:

  • In countries suffering sudden and massive outflows of private capital, a critical mass of front-loaded reforms may be required to restore market confidence. Conditionality should nonetheless be focused on reforms that tackle the roots of the crisis and limit resulting vulnerabilities.

  • In countries with PRGF programs, where the program objectives include long-term growth and poverty reduction, the choice of conditionality is more wide-ranging, and would likely include measures to enhance a country’s export and growth potential. More generally, in cases where structural weaknesses are at the root of protracted macroeconomic imbalances, conditionality may need to address longer-term fiscal and external sustainability but, in these cases, the measures would be expected to have a demonstrable impact on the objectives.

Further to these general principles, Directors emphasized that the scope and modalities of conditionality should be assessed in the context of the overall program design, the role being played by other institutions, and the country’s specific circumstances, including its track record and implementation capacity.

20. There is a need to reaffirm that the presumption of coverage should be one of parsimony rather than one of comprehensiveness—thus requiring a stronger burden of proof for the inclusion of specific structural measures as conditions. This presumption was agreed by the Board as a general policy in March 2001.9 The examination of the country cases has brought up the tension between including measures on the grounds that they might arguably turn out to be critical and the objective of parsimony. This points to a need to establish mechanisms to ensure that the case-by-case application of conditionality does not result in a recurrence of the proliferation of structural conditionality that occurred in the past.

21. Strengthened cooperation with the World Bank will over time improve the coherence and focus of conditionality and will promote more effective support to countries. In policy areas where the World Bank is in the lead, Fund conditionality should be limited to measures that are critical for the success of the Fund-supported program and should draw on the Bank’s advice on the substantive content. The Executive Board should be kept informed about relevant aspects of Bank conditionality and about the country’s progress in implementation. In PRGF programs, a framework for coordination that guides the choice of Fund conditionality already exists and provides a good basis for enhanced collaboration. In middle-income countries, the staffs are currently working to strengthen cooperation. The specific interaction will depend on the circumstances of the country and the nature of the program, but the same principles of specialization, collaboration, overall streamlining, and focus will apply.

IV. Issues for Discussion

  • Early in the conditionality review, Directors agreed that the presumption for the coverage of conditionality would be one of parsimony rather than comprehensiveness—requiring a stronger burden of proof for the inclusion of specific measures as conditions. Are Directors satisfied with the application of this principle to date, or do they see a need to apply it more consistently in individual country cases?

  • In deciding whether conditionality should apply outside the Fund’s core areas, do Directors agree that the primary criterion should be the magnitude of the impact on the fiscal and external balances? Do they agree that in PRGF programs, structural measures oriented primarily toward the growth and poverty reduction objectives of these programs may also be considered macro-critical, but with the presumption that conditionality related to such measures will be based, to the extent possible, on advice from the World Bank? Do they agree that, more generally, in countries in which external imbalances reflect structural problems, conditionality may need to cover some measures designed to strengthen growth and export capacity, but only if it can be shown that the impact of such measures is likely to be material?

  • Do Directors agree that, in countries suffering sudden and massive outflows of private capital, conditionality should be focused on reforms that can be implemented forcefully and are clearly needed to tackle the roots of the crisis and limit resulting vulnerabilities?

  • While Directors will have the opportunity later this year for a full review of progress in Bank-Fund collaboration in program design and conditionality, they may wish to comment on the lessons for collaboration from these real-time assessments.

Streamlining Structural Conditionality

22. This annex provides an update of the numerical analysis included in the Review of Initial Experience, and is therefore subject to all the caveats noted in that paper.10 Between the issuance of the interim guidelines in September 2000 and end-November 2001, 43 new programs (including annual arrangements of three-year PRGF programs) were discussed by the Board.11 Of these, 13 were considered as case studies in the main body of the paper in addition to 6 other programs that do not meet the above definition but were of interest for the reasons mentioned in the main text.

23. The application of the Interim Guidance Note by staff generally resulted in fewer structural conditions, but there was substantial variation in the extent of the reduction, particularly among SBAs. In general, there has been a trend toward leaner conditionality in the 35 new PRGFs. The average number of conditions since the introduction of the interim guidelines was 11, a reduction of more than one-third with respect to previous programs (Table 1).12 Out of 35 arrangements, 25 had fewer structural conditions than in the corresponding earlier arrangement (Table 2). Among the 8 new SBAs, the average number of conditions since the introduction of the interim guidelines (normalized by program length) was 13.8, comparable to the average in previous programs. These findings are broadly similar to the results documented in the Review of Initial Experience. However, the average is heavily influenced by the diversity of SBAs. The median estimate of the number of conditions was 10.6—again close to one-third below the median number of conditions in comparable previous programs.13 This altered finding underscores the sensitivity of the results to changes in the sample of stand-by programs and should therefore not be taken too literally.

Table 1.

Structural Conditions in IMF Programs 1/

(Average number of measures per program year)

article image

As at November 30, 2001.

Excludes Sri Lanka, whose previous arrangements date back to the mid-1980s, and Yugoslavia, without a previous program.

Includes conditions for the completion of a review.

Table 2.

Comparison of Structural Conditions in New and Old Arrangements

article image

Ratio of total number of structural conditions (structural benchmarks, performance criteria, and prior actions) in new programs to total conditions in the respective benchmark program, in percent.

Number of conditions is normalized by program length.

24. The streamlining was mainly achieved by a reduction in the number of structural benchmarks. Indeed, in SBAs, the number of performance criteria and prior actions increased, on average. In the case of prior actions, this increase was driven by Croatia, Pakistan, and Yugoslavia (the latter of which was not included in the list of countries with previous programs).14 Without these three countries, the average number of prior actions would have fallen from nearly 6 in previous programs to 2 in the new ones.

25. In terms of number of conditions, the 13 programs considered in the case studies were fairly comparable to the total sample, although the degree of streamlining differed considerably. The average number of conditions in the PRGFs analyzed in the case studies was 11.3, very similar to the number of conditions in the whole sample. The average number of conditions in the stand-by programs analyzed in the case studies was 12.7, one condition less than in the complete sample. Small differences between both samples were also reflected in the analysis of medians: the median PRGF program among the case studies had more conditions whereas the median stand-by program had fewer. The major difference between the case studies and the entire set of programs relates to the number of conditions in previous programs. These were considerably higher in the PRGFs selected as case studies, implying a larger degree of streamlining in terms of the reduction in the number of conditions.

26. A majority of policy measures that continued to be covered by conditionality fell within the Fund’s core areas, with particular emphasis placed on fiscal and financial sector reforms (Table 3). The Fund’s core areas accounted for about two-thirds of conditions in both PRGFs and SBAs. While their relative importance has remained fairly constant in new versus previous SBAs, conditions in the core areas have increased their prominence considerably in PRGFs. This increase has come mainly at the expense of conditions on privatization and public enterprise reform—both core areas of the World Bank.

Table 3.

Distribution of Structural Conditions in IMF-Supported Programs 1/

(IMF core areas shaded)

article image

Distribution of the total number of structural measures across sectors.


The text boxes focus on structural conditionality and include a discussion of conditionality in the current program, conditionality which has been dropped from previous programs or annual arrangements, and conditionality covered by the World Bank. The implied sample includes stand-by arrangements for Brazil, Romania, Lithuania, and Yugoslavia, and PRGFs for Armenia, Azerbaijan, Benin, Bolivia, Honduras, the Kyrgyz Republic, Mali, Mongolia, and Sierra Leone.


Directors’ views on conditionality presented in this section may not unambiguously reflect the overall Board assessment, because it can only draw on the limited number of views that were expressed in individual country discussions.


Indeed, following recommendations of a Board paper on strengthening country ownership of Fund-supported programs, resident representatives are to play a more proactive role in helping staff teams analyze and respond to issues of political economy (see SM/01/70).


The discussion on Gabon was in the context of a review and the appropriateness of streamlining conditionality during reviews is not evident. See SM/01/219 for a more detailed discussion of this issue.


In particular, the government was trying to effect a maturity transformation between the banking system which wanted to lend short and borrowers who wanted to borrow long. Although this action would not expose the government to credit risk, it would expose it to maturity and interest rate risk.


Romania’s program did include benchmarks on the privatization of some state-owned enterprises.


Guidance on improved coordination in the design of conditionality is currently being prepared jointly by the Fund and the Bank.


See SM/01/60 (2/16/01) and SM/01/219 (7/17/01).


See BUFF/01/36.


One important caveat is that the appropriate comparison to gauge progress in streamlining (namely, between a “new” program, negotiated after the issuance of the Interim Guidance Note, and an “old” program, negotiated just prior to the issuance under identical country circumstances) is not possible. More fundamentally, a comparison of the number of conditions, without analyzing its nature, is always an imperfect measure of the true extent of conditionality, see (SM/01/219, 7/12/01). Indeed, the number of conditions can be adjusted without any substantive change in conditionality (e.g. the privatization of three electricity companies could be expressed as one condition or three).


Gabon’s SBA was excluded, although it was approved by the Board in October 2000, because program negotiations were largely completed before the Interim Guidance Note was issued.


The actual scope of conditionality faced by a country may be higher (and the streamlining less than suggested by the above numbers), to the extent that conditionality under programs supported by other institutions has increased.


Such a comparison was not made in the paper on the Review of Initial Experience.


In the case of Croatia, the increase in the number of prior actions is misleading, as the bulk of them refer to specific wage measures which could have been addressed by one action. However, the authorities preferred a detailed strategy to substantially reduce the wage bill through a range of individual measures.

Lessons from the Real-Time Assessments of Structural Conditionality
Author: International Monetary Fund