Review of the Experience with the Policy Support Instrument

This review is focused mainly on the implementation of the Policy Support Instrument (PSI) against the goals and expectations set out by the Executive Board. Some possible modifications to the PSI that have been considered by various stakeholders are also touched upon in the concluding section. This review was conducted in parallel with ongoing work on a new architecture of lending facilities for low-income countries (LIC). Based on the results of this review, the reform of LIC facilities is not expected to alter the case for the PSI as a complement to financing instruments.


This review is focused mainly on the implementation of the Policy Support Instrument (PSI) against the goals and expectations set out by the Executive Board. Some possible modifications to the PSI that have been considered by various stakeholders are also touched upon in the concluding section. This review was conducted in parallel with ongoing work on a new architecture of lending facilities for low-income countries (LIC). Based on the results of this review, the reform of LIC facilities is not expected to alter the case for the PSI as a complement to financing instruments.

I. Introduction

1. The Policy Support Instrument (PSI), introduced in October 2005, enables the Fund to support low-income countries that do not need Fund financial assistance. The PSI is intended for low-income countries that have secured macroeconomic stability and completed basic structural reforms but wish to remain engaged with the Fund to consolidate their macroeconomic performance and pursue important second-generation reforms to secure higher and sustainable growth. The engagement with the Fund through a PSI can support the countries develop their policy programs, build reform commitment through the Fund endorsed program, identify priorities for technical assistance and, like other arrangements, signal to others the implementation of a reform program meeting upper credit tranche conditionality. The adoption of such a reform program under the PSI also facilitates a rapid move to Fund financial support should conditions change.

2. At the time the PSI was launched, the Fund’s Executive Board called for a review after three years.1 The experience under the PSI in the intervening three years is only sufficient to draw some initial broad conclusions about the operation of the PSI. Seven PSIs have been approved for six member countries, and each has concluded or had at least two reviews completed (i.e., covering at least one year of program experience). This is a small sample on which to base an assessment of the instrument’s effectiveness; therefore surveys and some anecdotal evidence are also used to shed light on implementation. Nevertheless, the preliminary conclusion of the review is that the economic performance of PSI users has been at least as strong as that under PRGF-supported programs. Member countries also deem the PSI to be a useful alternative to the Fund’s conventional instruments in circumstances where there is no immediate need for Fund financing.

3. This review is focused mainly on the implementation of the PSI against the goals and expectations set out by the Executive Board. Some possible modifications to the PSI that have been considered by various stakeholders are also touched upon in the concluding section. This review was conducted in parallel with ongoing work on a new architecture of lending facilities for low-income countries (LIC). Based on the results of this review, the reform of LIC facilities is not expected to alter the case for the PSI as a complement to financing instruments.

4. This paper considers four dimensions of the experience under the PSI to date, drawing on empirical evidence and stakeholder input.

  • The characteristics of PSI users: The PSI is designed to meet the needs of a subset of PRGF-eligible members that are “mature stabilizers,” defined mostly in terms of four macroeconomic criteria but also in terms of institutional and policy development. The paper attempts to assess the extent to which PSI users have indeed been drawn from among these mature stabilizers.

  • Program design and implementation: The individual PSI-supported programs are assessed in terms of the members’ success in meeting the specific goals set out in their programs. Baselines for comparison in this regard are drawn from the previous PRGF-supported programs for PSI users and from contemporaneous PRGF programs implemented by similar comparator countries.

  • Comparative macroeconomic performance: The overall macroeconomic performance of PSI users in terms of variables such as growth and inflation is compared with PSI users under their previous PRGF-supported programs and of contemporaneous performance of similar comparator countries. Given the changing global environment, the latter provides a means of controlling for transient factors.

  • Perceptions of the PSI’s role: Finally, the paper summarizes input from a broad array of interested members and other stakeholders. Written surveys were sent to members that are PSI users, all other PRGF-eligible members, donors, and private investors. In addition, outreach efforts to civil society organizations (CSOs) and other stakeholders have been pursued through teleconferences, meetings, and seminars, including in countries that are PSI users. Country teams have also provided useful input.

II. Have PSI Users Been Mature Stabilizers?

5. The PSI is potentially available to all PRGF-eligible members, but its use is explicitly intended for the subset of mature stabilizers. When the PSI was created, mature stabilizers were defined according to four criteria: such countries were expected to have “achieved a reasonable growth performance, low-underlying inflation (comparable to that obtaining in their major trading partners), an adequate level of official international reserves, and begun to establish (external and net domestic) debt sustainability.”2

6. For this review, some quantitative measures were applied to assess whether PSI users have generally been drawn from the ranks of mature stabilizers. Specific variables directly related to the above criteria—in some cases more than one variable per criterion—are used to screen PRGF-eligible countries. For each variable, a threshold is set at the average or an above-average level among PRGF-eligible countries, or near or above some conventional policy “rule-of-thumb” (e.g., reserves equal to three months of imports). Countries are then selected as comparator countries if they meet or exceed all the thresholds for all of the criteria. Because the term “mature stabilizers” implies a sustained level of economic performance, variables are averaged over five years, where data permit. In the specific case of inflation, where high variability can be disruptive to economies, a maximum annual threshold is included along with a threshold for the five-year average.

7. The PSI policy paper also put emphasis on “second generation” structural reforms that suggest some threshold of policy and institutional development. For the purposes of this review, this element is captured by adding an additional criterion on structural policies. Those countries below the mid-point of the “medium institutional capacity” category in the World Bank’s Country Policy and Institutional Assessment (CPIA) index are excluded from the class of comparator countries.

8. The countries that are classified as comparator countries on the basis of this methodology are shown in Table 1. It must be stressed that this list of comparator countries is only intended to establish a group of relatively similar countries for the purposes of this review. This is not a list of countries that should be considered current PSI candidates, as it is entirely backward looking and does not take into account 2008–09 data. Moreover, it should not be seen as an endorsement of the appropriateness of the PSI or other modes of Fund engagement for individual countries, as it does not take into account the strength of current or prospective policies.

Table 1.

Comparator Countries, classified by type of Fund engagement (2003–07)

article image

Nigeria did not meet the institutional quality criterion.

9. The group of comparator countries, as defined above, accounts for one quarter of PRGF-eligible members and encompasses all but one of the PSI users (See Table 1). Even the exception among PSI users, Nigeria, met four of the five criteria. Countries in the comparator group also include six surveillance-only countries (“other comparator countries”) and eight countries with PRGF-supported arrangements that started in 2003 or afterwards (“PRGF comparator countries”). 3 The evidence thus suggests that the PSI has largely been used by its target group. Moreover, the fact that the PSI has only been used by countries meeting these criteria (or nearly so), suggests that the Fund is able to reserve the PSI largely for its intended audience.

10. In practice, however, views have sometimes differed as to whether all PSI users have been mature stabilizers. During Board discussions, questions arose as to whether countries prone to policy slippages can be considered mature stabilizers, despite their recent progress in achieving macroeconomic stability. Such a bias could undermine the instrument’s credibility. In the cases of both Nigeria and Senegal, these concerns were addressed through successful implementation of front-loaded and ambitious structural reforms.

11. While the PSI appears to have gained a foothold among mature stabilizers since its inception, it has by no means totally displaced other forms of program engagement within this group:

  • Four comparator countries had new PRGF arrangements approved after the PSI became available in October 2005.

  • Of the five countries that had low-access PRGF arrangements approved after October 2005, two were from among the comparator countries. This is noteworthy insofar as countries that choose low-access arrangements typically do so in order to signal Fund endorsement of policies, much as the PSI is intended to do. Hence, as intended, the PSI can therefore be viewed as having been a voluntary, demand-driven tool, which usefully expanded the range of instruments for useful Fund engagement with low-income countries.4

12. Conclusions about the PSI should be interpreted cautiously, given the small sample sizes. Only a handful of PSIs and PRGF arrangements have been approved since the inception of the instrument, and most of these PSIs are still ongoing. Moreover, with PSI and comparator countries constituting a minority of PRGF-eligible countries to begin with, the relevant population is small for statistical purposes, and the PSI- and PRGF-supported program subsets of this population are smaller still. In this context, all conclusions about the PSI in practice should be regarded as tentative.

III. Have PSI-supported Programs Been Designed in Line with Expectations and Implemented Successfully?

A. Expectations for the PSI on Program Design5

13. The Executive Board took the view that the content of PSI-supported programs should closely follow that for the PRGF. Most importantly, it was stipulated that PSI-supported programs should meet the standard of upper credit tranche conditionality (the same standard as for programs supported by Fund facilities such as the PRGF). 6 This was in part to facilitate transition to a Fund financial arrangement should a balance of payments need arise, but also to anchor expectations in a well-established standard and avoid any suggestion that the PSI represents a weaker or otherwise different standard than that associated with the PRGF. The Executive Board also determined that the standard Guidelines on Conditionality should apply to PSIs. Finally, while the PSI was expected to apply the same standard for program design, there was an expectation that documentation for PSI-supported programs could be more concise than for financial arrangements.

B. Comparisons of PSI Program Design and Implementation with the PRGF


14. To ensure an even-handed basis for assessment, the content and implementation of PSI-supported programs is compared with that of PRGF-supported programs with similar comparator countries and, separately, with other LICs. A comparison to all PRGF-supported programs alone would be potentially misleading, as PSI users are effectively pre-screened for higher levels of macroeconomic performance, while the wider group would include countries emerging from conflict or struggling to establish or maintain macroeconomic stability. PSI-supported programs are, therefore, compared primarily with the “PRGF comparator countries” (see Table 1) and with the previous PRGF-supported programs for PSI users that were identified in Section II. Using previous PRGF-supported programs of PSI users as comparators helps control for country-specific factors that may influence program design and implementation.7


15. The number of conditions is somewhat lower in PSI-supported programs than in contemporaneous PRGF-supported programs. Specifically, PRGF-supported programs with both mature stabilizers and other LICs had roughly one quarter more conditions per review than PSI-supported programs (Figure 1). The lower number of conditions in PSI-supported programs extended to both quantitative and structural conditions. This is true despite the fact that the Nigerian authorities chose to include extra conditions in their PSI-supported program as a commitment mechanism, resisting attempts by Fund staff to streamline. The number of conditions in PSI-supported programs was roughly similar to that in the previous PRGF-supported programs for the same PSI countries, suggesting that country-specific factors in conditionality may partly explain the lower number of conditions.

Figure 1.
Figure 1.

Extent of Conditionality

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Source: Fund staff estimates.

16. PSI-supported programs have a somewhat higher level of adherence to structural conditionality than contemporaneous PRGF-supported programs (Figure 2). As with patterns for the number of conditions, the adherence to conditionality in PSI-supported programs most closely resembled that of the previous PRGF-supported programs in PSI countries, suggesting again that country-specific factors are important.

Figure 2.
Figure 2.

Implementation of Structural Conditionality

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Source: Fund staff estimates.

Areas of Structural Conditionality

There are broad similarities between the subject areas of structural conditionality in PSI-supported programs and the various groups of PRGF-supported programs. While the pattern of fewer conditions per review in PSI-supported programs extends to the various categories of structural conditions, there is little clear pattern to the differences across subject areas (Figure 2). Revenue mobilization, debt and expenditure management, and banking and financial sector conditions were among the most common in all groups of program countries. Among these areas, debt and expenditure management measures were relatively more common for PSI-supported programs, while revenue mobilization was relatively less important in PSI-supported programs than banking and financial sector conditions. Most other areas had few conditions for any of the groups of programs. Within each group of conditions, there are varying levels of sophistication in needed reforms, and the lack of clear differentiation between PSI and PRGF-supported programs should be ascribed to the variety of individual country circumstances.


Areas of Structural Conditionality 1/2/ Shares of Conditions

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Source: Fund staff estimates.1/ Structural conditionality includes prior actions, assessment criterion, and benchmarks.1/ These figures represent the average number of conditions per program review.3/ Institutional reform includes legal and regulatory, labor market, and corporate reform.

17 The comparisons on program design also suggest that the PSI’s fixed review schedule does not weaken program implementation or design (see also Box 2). Adherence to conditionality is similar in PSI and PRGF-supported programs, and PSI users have not needed more waivers. Nor does conditionality appear to be lighter in PSI-supported programs; PSI users have similar numbers of conditions under both the PSI and their previous PRGF arrangements. Also, Section IV will show that PSI users do not have worse macroeconomic outcomes than other comparator countries, as might be the case if reviews were allowed to proceed despite slippages in implementation. Taking these findings together with the fact that all reviews under PSI-supported programs have been completed on schedule, it seems reasonable to conclude that the fixed review schedule has been effective in encouraging policy discipline. A related implication of the relatively strong performance of PSI users is that the presence of Fund financial resources is not essential to good policy implementation and outcomes, given that the PSI users do at least as well on these measures as members supported by PRGF arrangements.

Senegal: Promoting Structural Reforms under the PSI

An assessment of Senegal’s performance under its three PRGFs since 1994 was critical of Senegal’s performance on structural reforms under its three PRGFs since 1994. It concluded that only about three-fifths of structural performance criteria and about half of the structural benchmarks were met on time. Instances of delays or noncompliance exceeded the observance of structural conditionality continuously since mid-2000, and some reforms succeeded only after several attempts.

This sharply contrasts with Senegal’s track record of structural reform under the PSI. Since the start of the PSI program, Senegal has met all prior actions and all structural assessment criteria. It has also implemented two-thirds of the structural benchmarks on time, and the remainder either partially or with a slight delay. Some of the reforms, like those related to transparency in the financing of the new airport, had been a major obstacle and delayed completion of past reviews under the PRGF.

Both the authorities and the staff team considered that the PSI’s predetermined review schedule was an important factor behind the improved implementation of structural reforms. The Minister of Finance reportedly expressed the view to management and staff that the schedule imposed significant discipline on decision-making within the government and led to more timely decisions. One instance of this was the major effort made to keep the program on track when budgetary slippages were uncovered in 2008; strong corrective actions were quickly taken, laying the foundation for a positive staff recommendation to complete the second PSI review. Prompt action in this case helped put Senegal in a stronger position to face the impact of the global economic crisis that was breaking at that time.

C. Technical Assistance

18. There do seem to be differences in the focus and scope of technical assistance between PSI users and other country groups. The average number of completed TA projects per country during 2003–08 was higher for PSI users (6.4) than for other comparator countries (5), although the difference was small in comparison with PRGF-supported program countries (6.25). There were large differences in the areas of concentration – expenditure management and related topics comprised almost 50 percent of TA projects for PSI users, but only a quarter for PRGF countries, and less than 5 percent for other comparator countries.

Figure 3.
Figure 3.

Types of Technical Assistance

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Source: Fund staff estimates.

D. Is PSI Documentation More Concise?

19. Documentation for PSI-supported programs has been somewhat more concise than for PRGF-supported programs, in line with expectations. The length of letters of intent (including MEFP/TMUs) in PSI requests for approval documents averaged 22 pages, versus 27 and 28 pages in contemporaneous PRGF cases and in previous PRGFs in PSI countries, respectively. A comparison of the length of main staff report texts in requests for approval (excluding those combined with Article IV reports) also shows some evidence of streamlining: 35 pages in PSI cases versus 42 pages in contemporaneous PRGF cases and previous PRGFs in PSI countries (though the latter could also reflect broader efforts to limit word counts and the length of reports to the Board).

IV. Has the PSI Led to Consolidation of Macroeconomic Performance?

20. The goals of PSI-supported programs include consolidation of macroeconomic stability and improved growth. Data limitations hamper comparisons of macroeconomic performance under the PSI. The available information allows only very tentative comparisons of macroeconomic developments in PSI-supported program countries versus other countries. PSI-supported programs have been in place for at most three years and most started only in 2006 or 2007. In contrast to the program performance information discussed in Section III, most macroeconomic data are available only with long lags and at annual frequencies for LICs, which rules out using 2008 macroeconomic data for these comparisons. Moreover, as the periods covered by PSIs or PRGF arrangements do not generally match calendar years, it is rarely possible to match program periods to the periods covered by macroeconomic data.8

21. Subject to these caveats, it would appear that countries with PSI-supported programs have performed as well as, or better than, other groups of LICs by most measures (Figure 4 and 5). GDP growth, international reserve cover, and fiscal balances were generally stronger for all three groups of comparators than for other LICs or the previous PRGF-supported programs of PSI countries, although there was little differentiation among the groups of comparator countries. Inflation showed little differentiation either among comparator countries or between comparator countries and other LICs. The clearest differentiation between PSI countries and others was on foreign direct investment, where PSI countries were associated with higher levels of FDI inflows.

Figure 4.
Figure 4.

Macroeconomic Performance in PSI-Supported Programs and Comparator Countries

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Sources: Staff Reports; WEO for surveillance countries and non-mature stabilizer LIOs1/ The pre-program year is the calendar year preceding the year of approval for PRGF arrangements and PSIs approved before June 30, the year of approval for PRGF arrangements and PSIs approved after mid-year, and calendar year 2005 for surveillance countries.2/ Due to recent approval, only pre-program data are available for Senegal and is excluded from the PSI data.3/ Excludes Somalia for which data is not available.4/ Excludes Zimbabwe.5/ Excludes debt forgiveness grants (HIPC/MDRI) and Timor-Leste.
Figure 5.
Figure 5.

Macroeconomic Performance in PSI-Supported Programs

Citation: Policy Papers 2009, 028; 10.5089/9781498335966.007.A001

Sources: Staff Reports.Note: Figures in parenthesis next to the country labels in the legend indicate month and year the PSI program was approved.

V. Has the PSI Met the Expectations of Low-Income Members and Donors?

A. External Input for the PSI Review

22. Some aspects of the PSI’s goals need to be assessed primarily by members and other stakeholders. The three preceding sections focused on goals and expectations for the PSI that can be assessed through macroeconomic and program performance data that are available to staff. However, there are other goals and expectations for the PSI that are best assessed by low-income members making use of the PSI, other low-income members, donors, and other stakeholders. These include, specifically, the extent of country ownership of programs, and the PSI’s value in delivering clear and timely signals to donors and other stakeholders.

23. A variety of approaches were used in soliciting input from Fund members and external stakeholders.

  • Questionnaires were sent to four groups: (i) PSI users; (ii) other PRGF-eligible members; (iii) bilateral and multilateral donors and creditors; and (iv) foreign private investors in Africa (the only region with PSI users to date).

  • CSOs (LIC-based and others) were invited to participate in informal meetings in London in October 2008, an international teleconference in November 2008, and at a CSO seminar during the April 2009 Spring Meetings. CSO input was also solicited through the Fund’s external website. In addition, two CSOs have provided their own assessments of the PSI.

This section and the following one draw on these assessments.

B. External Views on the Implementation of the PSI Relative to its Goals9

24. PSI users and donors were generally positive on their experience with the instrument. The responses to almost all questions soliciting a positive or negative assessment were heavily skewed toward the positive.

  • Among the PSI users, most agreed strongly that the PSI was more appropriate to their needs than a PRGF-supported arrangement would have been (the others were neutral).

  • Donor summary views were similarly positive; most of those responding to a question on the usefulness of the PSI versus PRGF arrangements concluded that it was useful in making aid decisions; two-fifths of donor responses indicated that the PSI was more useful than a surveillance-only relationship in making aid decisions, and none considered it less so.

25. The views of other low-income members and private investors were also positive. Three-quarters of LICs that have not used the PSI expressed the view that, in the absence of a need to borrow from the Fund, the PSI would be an attractive alternative vis-à-vis a surveillance-only relationship with the Fund. Half of those responding considered that the PSI would be an attractive alternative to a low-access PRGF arrangement. Private investors were also positive, reporting that the PSI was more useful for making business decisions than a surveillance-only relationship, with two of the three taking the view that a PSI is more useful than a PRGF arrangement for that purpose.

The policy support role

26. PSI users and donors were most positive on the policy support aspects. All PSI users responding to the survey agreed their PSI-supported programs had promoted close policy dialogue with the IMF and that country ownership had been sufficiently emphasized. Donors were similarly positive on both aspects of the experience with the PSI. Some PSI country authorities have expanded upon the survey responses in this regard. One PSI user took the view that “the PSI in its current form is a strong system that has supported us in maintaining a firm grip on our macroeconomic objectives and targets. The program targets and structural reforms embedded in the PSI have enabled us to closely monitor and assess our own performance, and make adjustments where necessary to ensure that we remain on course.” Similarly, another reported that “policy advice through PSI has improved macroeconomic management.” Finally, yet another PSI user noted in a separate letter that “the PSI has helped to improve the [country’s] economic performance and leverage the government reform program.”

27. PSI users and other LICs generally agreed that the October 2008 modifications to the Exogenous Shocks Facility made the PSI more attractive. The PSI was designed in tandem with the Exogenous Shocks Facility (ESF), and the PSI is intended to facilitate the transition to financing under the ESF. However, some features of the ESF as originally designed, notably the required termination of a PSI-supported program upon approval of ESF support, made it less attractive to potential users. Most PSI users and other LICs agreed that the subsequent modifications to the ESF—which can now be used in conjunction with the PSI—have had the intended effect of making it more attractive. Since these modifications were adopted, two PSI users (Senegal and Tanzania) have made use of the ESF during their PSI-supported programs (see Box 3 on Tanzania’s experience).

Tanzania: The Role of the PSI in Facilitating a Request for Support Under the ESF

During the IMF-World Bank Spring Meetings in late April 2009, the Tanzanian authorities requested balance of payments support from the Fund under the high-access component of the ESF (110 percent of quota). Having a program in place under the PSI expedited the request, which was approved by the Board only a month later, on May 29. Among the key factors that facilitated the process were the fact that:

  • The PSI program had been broadly on track since it was approved in February 2007;

  • The policy content of the PSI-supported program met the standard of upper credit tranche conditionality, which is required for a high-access ESF; no new conditionality was needed; and

  • Regular monitoring and program reviews under the PSI served as an early warning system, enabling a rapid assessment of the impact on Tanzania of the current global economic crisis and timely support by the Fund.

In the country team’s view, the close relationship and mutual trust between the authorities and Fund staff built up through frequent policy discussions allowed for considerable candor and a constructive problem-solving approach that mitigated a potentially more acute crisis situation.

The signaling role

28 Both PSI users and donors had positive assessments of the signaling role of the PSI, although they were less positive on this than on the policy support aspects. All PSI users and a large majority of donors agreed that the Fund’s assessments of performance under PSI-supported programs had delivered clear and timely signals of the strength of policies. In particular, both donors and PSI users agreed that the fixed review schedule has improved the timeliness of signals. However, PSI users were less positive on the usefulness of the PSI in attracting private sector interest, and two of the respondents believed that foreign investors regarded signals from other programs as stronger and clearer. A majority of PSI users and donors considered that the fixed schedule of reviews had helped coordination among donors and country officials, with only one PSI user and one donor expressing disagreement with this view.

Program design

29. PSI users and donors were also generally positive on program design aspects of the PSI, but some expressed reservations regarding structural conditionality.

  • There was a high degree of satisfaction with the design and implementation of macroeconomic policies.

  • Assessments of the structural content in program design and implementation were generally positive, and all PSI users and most donors agreed that structural measures were focused in areas critical to macroeconomic stability.

  • However, some other LICs indicated that they might have considered a PSI had structural content been more limited; more than half of these non-PSI users ranked having too many structural conditions either first or second among potential drawbacks to a PSI. With the recent Executive Board decision on the reform of structural conditionality, this should be less of a drawback in the future.

CSO views

30. CSO representatives in PSI countries had a fairly positive view of the PSI. Resident representatives in PSI countries held seminars and solicited input in other ways from local CSO representatives. The views of these representatives varied, but were positive on balance. However, one CSO representative, referring to a specific country case, suggested that the PSI could have been more demanding on transparency.

31. Two CSOs based in advanced economies took a more negative view of the PSI than other observers did. These CSOs provided studies that judged the PSI against standards and goals other than those set out by the Executive Board in 2005. In particular, both implicitly rejected the idea that the PSI should use upper credit tranche conditionality as the standard for program design. One study called for higher inflation targets and more relaxed fiscal policies than under PRGF-supported programs, and urged donors and investors to reject a signaling role for the Fund and eschew “orthodox” policies.10 Another study has not yet been finalized, but was shared in draft with Fund staff. It concludes, in part, that there is no significant difference between PSI- and PRGF-supported programs, in particular in terms of conditionality; that there is no risk to donor or private investor flows from the absence of a Fund-supported program; and that there is little or no risk that policies would become less “responsible” without the anchor of an IMF-supported program. On the basis of these conclusions, it urges that the PSI be discontinued.

VI. Conclusions and Issues for Discussion

32. Several conclusions emerge from the experience with the PSI and the views of PSI users, donors, and other stakeholders.

  • The PSI has been effectively targeted to mature stabilizers, and it has been sufficiently attractive as an alternative to PRGF arrangements or a surveillance-only relationship to attract a significant subset of these members as users.

  • The extent of conditionality and its implementation have been broadly similar between PSI- and PRGF-supported programs with mature stabilizers, but with somewhat fewer conditions and somewhat higher levels of adherence to conditionality in the case of the PSI-supported programs.

  • PSI countries have generally had better macroeconomic performance during their PSI-supported programs than under their previous PRGF-supported programs, and macroeconomic performance is roughly comparable to the contemporaneous performance of other comparator countries.

  • The views expressed by members with PSIs and donors are generally quite positive on the experience with the PSI thus far. Other low-income members indicated a broad willingness to consider PSI-supported programs in the event that they had little or no need for Fund financing.

  • Views of members using the PSI and donors are most favorable on policy support aspects of the PSI, but they are also generally positive on signaling and program design aspects of the instrument.

  • Critiques of the PSI by a few advanced economies’ CSOs center on Fund engagement in general, rather than on a comparison with PRGF arrangements or surveillance-only relationships.

Issues for discussion

  • Do Directors agree that, so far, the PSI has generally succeeded in meeting the goals and expectations set out for it when it was launched?

  • Do Directors agree that the limited experience with the PSI (seven cases in the last three years) does not suggest a pressing need to modify its features at this time?

Annex —Review of the Policy Support Instrument: Survey Results

Questionnaires were sent to four groups: (i) PSI users; (ii) other PRGF-eligible members; (iii) bilateral and multilateral donors and creditors; and (iv) foreign private investors in Africa (as there were no PSI users from other regions). All six PSI users responded to the surveys, 18 of 69 other PRGF-eligible members, 19 bilateral donors and multilateral donors, and four private investors responded to the qualitative questions in the survey. The full quantitative results of the survey responses are set out below.

Table A1:

PSI-using Countries’ Responses to Questionnaire

(Responses are shown in percent of total responses received)

article image
article image
Table A2:

Donors’ Responses to Questionnaire on PSI-supported Programs

(Responses are shown in percent of total responses received)

article image
Table A3:

Other Low-income Countries’ Responses to Questionnaire on PSI-supported Program

(Responses are shown in percent of total responses received)

article image
Table A4:

Private Sector Responses to Questionnaire on PSI-supported Program

(Responses are shown in percent of total responses received)

article image

The deadline for the interim review was originally December 31, 2008, but was subsequently extended to June 30, 2009.


Policy Support and Signaling in Low-Income Countries


Countries with arrangements that started before 2003 were excluded, as those programs would have ended before the PSI was initiated.


See Policy Support and Signaling in Low-Income Countries and PIN/05/144


See the Secretary’s understanding of the PSI decisions in Implementation of the Policy Support Instrument. There exists no formal definition of “upper credit tranche conditionality,” but it is generally taken to mean a set of policies sufficient to correct balance of payments imbalances and enable repayment to the Fund. This paper uses the term conditionality and its variants as a shorthand to refer to both conditionality under use of Fund resources and the parallel monitoring modalities under the PSI. However, strictly speaking, the term conditionality only refers to policies established in the context of use of Fund resources and not to the monitoring modalities such as assessment criteria and benchmarks in PSI-supported programs.


All PSI users previously had at least one PRGF-supported program, with the exception of Nigeria.


For program countries data used are for the pre-program year and the first program year. For other comparator countries and other LICs, the data in the pre-program and program year columns are for 2006 and 2007, respectively.


The full quantitative results of the surveys are presented in the Annex.


ActionAid, The IMF’s Policy Support Instrument: Expanded Fiscal Space or Continued Belt Tightening? October, 2007. A transcript of a conference call between IMF and ActionAid staff is posted on the Fund’s external website at

Review of the Experience with the Policy Support Instrument
Author: International Monetary Fund