49. As noted, the stated objective of IMF TA is to (a) assist countries in the design of appropriate macroeconomic and structural reforms, and (b) to strengthen members’ capacity to formulate and implement growth-oriented and poverty reducing policies. These are broad objectives, which in turn have been used to define a wide set of TA activities, including fact finding and analytical work to enhance the IMF’s knowledge in its policy dialogue and program design. These activities may include short-term TA to advise on particular policy issues or crisis situations. At the other end of the spectrum—and depending on the degree of institutional development—they may comprise longer-term TA using resident experts to improve a country’s capacity to diagnose problems and design and implement policies.

A. What Is Classified as TA?

49. As noted, the stated objective of IMF TA is to (a) assist countries in the design of appropriate macroeconomic and structural reforms, and (b) to strengthen members’ capacity to formulate and implement growth-oriented and poverty reducing policies. These are broad objectives, which in turn have been used to define a wide set of TA activities, including fact finding and analytical work to enhance the IMF’s knowledge in its policy dialogue and program design. These activities may include short-term TA to advise on particular policy issues or crisis situations. At the other end of the spectrum—and depending on the degree of institutional development—they may comprise longer-term TA using resident experts to improve a country’s capacity to diagnose problems and design and implement policies.

50. Present budgetary classifications tend to mask these distinctions, limiting the ability of management and the Board to assess the true effort of the IMF in medium-term institution building. This becomes important because—as discussed later in the report—there may be some factors that tend to bias the identification of TA needs toward short-term TA.

51. We recognize that these short-term activities are critical for the functioning of the IMF and for its capacity to advise and to design programs. In fact, the World Bank explicitly recognizes this through a special budget item called “Country Economic and Sector Work.” However, it is important to recognize that many of these activities are different from those aimed at improving countries’ own institutional capacities to diagnose problems and design, and implement policies. It is not easy to derive the total amount of IMF TA devoted to these short-term diagnostic activities relative to institutional capacity building. Our best estimate is that about one-third of TA may consist of short-term diagnostic or policy advice activities.28

B. Identifying TA Needs at the Country Level—Attempts to Make the Process More Strategic

52. The fact that TA is called to address such a broad set of objectives makes the process of identifying priorities at the country level particularly complex—especially under tight budgets. It requires identifying the appropriate balance and sequencing between short-term and longer-term activities, and how these interact with each other and with other IMF instruments. It requires a shared vision with the authorities on a medium-term policy framework that would allow prioritization among short- and medium-term TA needs as well as across sectors, departments, and IMF institutional initiatives. It will also require decisions on how IMF TA fits within a broader framework of capacity-building efforts by other donors.

53. Whether present practices of TA prioritization provide such a framework remains a question. In fact, one of the major findings of the 1999 IMF TA evaluation was the need for a more strategic orientation for TA and the need to make TA more proactive and preventive rather than reactive. As a response to the 1999 evaluation, Executive Directors supported recommendations to integrate TA better into a country’s overall policy framework as set out in Article IV consultations. Several Fund-wide and departmental specific initiatives were taken, including:

  • (a) To integrate TA better with Article IV consultations, staff piloted a series of TCs incorporating explicit TA discussions with the authorities—reviewing past experience and future needs. As mentioned earlier, neither area nor functional departments found the initiative cost effective as implemented. It has since been discontinued.

  • (b) To assure that TA priorities were simultaneously aligned with core areas of IMF expertise and new institutional mandates, a set of filters was introduced to provide criteria for the prioritization of TA activities.

  • (c) Functional departments have tried to sharpen their own strategic country focus (see Appendix 1, Tables A1.1A1.4). Diagnostic instruments, such as ROSCs and FSAPs, are now being used to identify country TA priorities and these instruments are now reaching a larger share of the membership. FAD has developed fiscal strategy briefs that review progress on structural reforms in fiscal systems and identify the remaining agenda for the future.29 MFD has been experimenting with assessment missions in order to engage the authorities on specific TA needs in the financial and banking sector. STA has recently introduced a country-centered approach to TA allocation based on country strategy notes.

54. The purpose of this chapter is to examine how these initiatives have influenced the process of TA prioritization in the post-1999 period and what lessons can be drawn. Specifically, it addresses the following questions:

  • What is the degree of alignment of IMF TA provided to a country with the key issues emerging from Article IV consultations?

  • Does the PRS process help identify capacity-building priorities in low-income countries?

  • How is TA responding to the new surveillance and thematic initiatives (such as the GDDS, ROSCs, and FSAPs), and new institutional mandates (AML/CFT and OFC assessments).

  • Why did the experiment with TCs not succeed and are there lessons for the future?

  • Do the new prioritization filters introduced in 2001 contribute to a more strategic focus for TA?

  • How are countries’ TA needs translated into final resource allocations?

C. Links Between TA Provided by the IMF and Article IV Surveillance

55. As mentioned, one of the conclusions of the 1999 internal evaluation was the weak link between Article IV consultations and IMF TA. The expectation of a link is reflected in the 2001 TA policy statement which says that “it is often during discussions with country authorities during Article IV consultations…that the needs for capacity building and policy strengthening are mutually recognized and the concomitant requirements for technical assistance clearly identified.”

56. We, therefore, examined the extent of alignment for a recent period by comparing the main policy challenges identified in Article IV consultation reports with the subsequent amount and composition of IMF TA in a sample of 18 countries (Table 4.1).30 To assess the strength of the link (defined as the extent to which the TA provided by the IMF reflected the broad policy priorities identified in Article IV surveillance), we drew on at least three Article IV staff reports spread out over the period 1998 to 2003 to identify the major policy challenges facing each country. This was compared with the subsequent allocation of TA resources across five areas: PEM and budget management, revenue, monetary and exchange rate policy, banking supervision and restructuring, and statistics. Scores were assigned to reflect the degree of alignment.31

Table 4.1.

Strength of the Link Between Article IV Consultations and IMF TA

article image
Source: IMF documents.1 = weak or no apparent link; 2 = partial link; 3 = close link between surveillance and TA from IMF; n.a = not applicable (for these countries, responsibility for subject area rests with the regional central bank (Banque Centrale des Etats d’Afrique de l’Oeust)). Countries in bold are those six chosen for in-depth field work.

57. Of course, there may be good reasons why the link between TA and Article IV surveillance is not strong. Other factors, some of which are embodied in the allocation filters contained in the 2001 Policy Statement (discussed below), may dominate resource allocation decisions. For example, a country’s track record of implementing reforms identified in previous surveillance or in making use of previous TA may be poor suggesting that future TA may not be used effectively (e.g., Yemen). Other TA providers may already be active in support of TA needs identified by IMF surveillance (e.g., statistics in Yemen, tax and customs administration in Zambia, and banking in Ukraine).

58. Not surprisingly, the results vary considerably across countries and subject areas. However, on average they imply only modest progress since the conclusion of the 1999 IMF evaluation that the link between surveillance and TA was weak. There are some interesting patterns among the group of countries, a number of which are consistent with information gathered during interviews with staff and our case study evidence:32

  • The strongest link was seen for Cambodia, Ukraine, Yemen, and Zambia, all of which were among the heaviest users of IMF TA in our sample. It is noteworthy that Cambodia and Yemen were the only two countries in the sample to have Technical Cooperation Action Plans (TCAPs), which were to serve as medium-term frameworks for delivering TA. The weakest links were seen for Bangladesh, Belarus, Burkina Faso, and Vietnam—a diverse group, with no apparent commonality, except perhaps the absence of an IMF-supported program for all or much of the period under consideration (FY1999–2003).

  • On average, there was no major difference in the strength of the link between low-income and medium-income countries, or between those countries defined as Heavily Indebted Poor Countries (HIPCs) and non-HIPCs. However, HIPCs showed a somewhat stronger link than non-HIPCs with respect to TA in PEM and budget management.33 This is not surprising given the emphasis placed on PEM and budget management for countries participating in the HIPC Initiative stemming from the desire to ensure that resources released from debt service were used for productive purposes.

  • The link between TA and Article IV surveillance appears—on average—strongest in the area of banking supervision and restructuring, and lowest in statistics. One possible explanation is that some banking problems flagged during Article IV consultations may require more timely action to address situations with the potential to spill over to other parts of the economy (although much of the TA in banking has been of a preventive nature and might therefore not be so time sensitive). On the other hand, problems related to quality of data rarely imply such urgency. That said, one would still expect a relatively timely response to serious shortcomings in a member’s statistical capacity.34

  • The weakest alignment with TA on statistics occurred for the three French-speaking African countries in the sample, while TA on statistics to the Anglophone African countries was closely aligned. This is consistent with views expressed by some staff that—at least until recently—STA did not have sufficient French-speaking experts, which undermined their ability to respond to needs in West Africa. Among our case studies, Niger provides a good example of this phenomenon.

  • This exercise was only able to look at average patterns within a broad period—particularly because some of the links between the Article IV consultations and TA occur with a lag. But there are also important yearly fluctuations in the volume of TA that cast doubts whether (at least in the case of non-PRGF countries) Article IV consultations either develop a long-term view, or if they do, whether they effectively guide TA allocations. A good case is Ukraine which received a yearly volume of TA in the range of 1000–1700 person-days in the period FY1995–2002. That volume collapsed to about 100 person-days in the FY2003–04 period, when most IMF resident experts were phased out, owing to a better economic environment that improved the fiscal and banking sector situation, no new financial arrangements, and increased reliance on other TA providers. As we have seen elsewhere, the existence of a program seems to have been a major determinant of the large volume of TA. It is unlikely that institution-building needs of Ukraine or changes in the track record of implementation alone can explain such a wide fluctuation.

D. How Well Are TA Priorities Linked to the PRS Process?

59. Do PRSPs, which were introduced in December 1999, provide a suitable diagnosis of the capacity constraints faced by low-income countries in designing, implementing, and monitoring poverty reduction and growth strategies? One would expect the PRSP and subsequent annual progress reports (APRs) to identify key bottlenecks in technical and institutional capacity and indicate how TA could contribute to their amelioration. To the extent that this is the case, one would also expect significant alignment between capacity-building priorities set out in PRSPs and the composition of TA delivered by the IMF in its areas of expertise. To address this question, we conducted a panel data analysis of 48 PRSP countries, and then reviewed PRSPs and APRs for a sample of 10 countries drawing on evidence presented in the recent IEO evaluation of PRSPs and the PRGF as well as on the case studies included in this evaluation.

Are PRSPs good diagnostic instruments for TA needs?

60. The overall documentation prepared as part of the PRS process was of only limited value in diagnosing capacity-building needs in a way that would constitute a sound basis for planning TA activities (Table 4.2). Many PRSPs provide general indications of areas where there is a need for capacity building—especially tax administration, public expenditure, and the payments system. Some do not go this far and only outline broad directions for policy change, such as “improving macroeconomic management.” About half of the PRSPs we examined identified capacity-building needs in a way that is specific enough to give some indication of the issues that need to be addressed and thereby offer a partial basis on which TA providers and recipients could build a detailed assessment of TA needs. This general assessment hides significant variation across countries: in a few “good practice” cases (Ethiopia, Niger, and Mozambique), there is an explicit focus on capacity building in the PRSP, and, therefore, broad TA needs are clearly identified. The diagnosis offered in APRs is typically less extensive than in PRSPs, although one might expect experience gained in implementing the PRSP to help countries identify capacity bottlenecks and prioritize their needs.

Table 4.2

Extensiveness of Diagnostic of Capacity-Building Needs in PRSPs 1/

article image
Source: IMF documents.

The higher the “score,” the more the PRSP helped set out priorities for TA needs. The detailed “coding” that was used to attribute scores is as follows: 1=PRS document does not go beyond acknowledging TA received or promised by the Bretton Woods institutions or other partners; 2=PRS document identifies capacity-building needs only in very general terms, without clear linkage to TA needs; 3=PRS document identifies clearly specific capacity-building needs, wherefrom TA needs can easily be derived; and 4=PRS document contains or builds upon full-fledged capacity-building strategy with clearly identified TA needs.

61. These results are consistent with those of other evaluations. The World Bank OED’s evaluation of the PRS process finds that, although capacity constraints were frequently mentioned, there was little effort to address them as part of the ongoing dialogue with the authorities. As a result, the PRSP has had a limited impact in identifying initiatives supporting capacity enhancement.35 In a similar vein, a recent FAD assessment of TA in PEM in Anglophone African countries finds that for the most part PEM issues were not addressed in countries’ PRSPs.36 The study suggests that PEM issues could be integrated into national Poverty Reduction Strategies as a way to give more consideration to the institutional and governance environment, and promote ownership.

Is the volume of TA provided by the IMF influenced by having a PRSP?

62. To examine whether there is a relationship between the PRSPs and the volume of TA provided by the IMF, we use a panel of 48 PRSP countries between FY1998 and FY2004, where the variable to be explained is the annual volume of TA received. As shown in Table 4.3, equations 1 and 2 analyze the impact of the Interim Poverty Reduction Strategy Papers (I-PRSPs) and PRSPs on the total volume of TA, and equations 3 and 4 show the impact of the I-PRSPs and PRSPs on the volume of TA in the fiscal area (FAD).37 These four equations are consistent in showing that there is no evidence of a significant impact of the I-PRSPs or PRSPs on the volume of TA, once we control for other factors such as the presence of a PRGF-supported or an EFF-supported program. In other words, the average volume of TA provided by the IMF after the PRSP was put in place (controlling for the income level and the presence of IMF-supported programs) was not significantly different from the average volume of TA provided by the IMF prior to the PRSP exercise. These equations also suggest that IMF-supported programs, not the PRSP, are the major driving force behind TA provision, confirming the results of Chapter 3.

Table 4.3.

Panel Data Regression Results 1/

article image
Source: IMF.

Robust fixed effects.

Significant at a 95 percent level of confidence.

Significant at a 99 percent level of confidence.

63. The evidence found in our country case studies is for the most part consistent with this regression analysis. For example, in both Niger and Yemen, total TA fell off dramatically in the period following the adoption of the PRSP. A longer-term view of institution-building needs and absorptive constraints might have prevented the very large swings in the volume of TA provided to Yemen. In Niger, however, the decline appears to have been motivated by a shift in the burden of TA delivery from FAD to the newly established regional TA center (West AFRITAC), which had competing regional priorities and fewer resources to allocate. In Zambia, neither the level nor the composition of TA appeared to respond to the PRSP but rather was motivated by the need to address implementation problems in the PRGF arrangement. A similar phenomenon was evident in Ethiopia, although, in this case, TA allocation appeared to be driven by the country’s medium-term development strategy that preceded the PRSP. In the case of Honduras the PRGF—which was not based on the framework of the PRSP38—was the umbrella under which most TA activities were aligned. In Cambodia the TCAP, and not the PRSP, has been the driving force behind TA provision.39

64. The inability of most PRSPs to influence the level and composition of IMF TA was also confirmed by discussions with a number of IMF staff, including departmental TA coordinators. They argued that PRSPs did not provide an adequate basis for articulating TA needs or setting TA priorities for individual countries. In fact, interviews with TA coordinators in area departments revealed that only in the African Department (AFR) was the PRSP looked to as a potential significant driver of decisions on the allocation of TA resources. The disconnect between TA resource allocation decisions and PRSPs was further reflected in the fact that the PRS process is barely mentioned in the IMF’s Policy Statement on TA adopted in April 2001.40

65. An additional factor hindering the link between the PRSP and TA allocation is that commitments to TA activities tend to be made for a relatively short time horizon—usually one fiscal year—which may not mesh easily with the longer-term perspective of PRSPs. However, the significance of this factor should not be overstated because it appears to have been less of a problem for IMF-supported programs that extend over a similar period (PRGF and EFF programs) and for which links with TA appear to have been considerably stronger. Thus, while the short-term budgetary framework presents some challenges to using the PRSP to guide TA allocation, there is still scope for TA allocation to better reflect PRSP priorities.

66. While PRSPs seem to have had a weak direct impact on TA, there is potential for an indirect impact by way of PRGF-supported programs that are themselves intended to be aligned on PRSP priorities. However, as discussed in the IEO’s evaluation of PRSPs and the PRGF, the degree of alignment between PRSPs and PRGF-supported programs is very uneven in large part because most PRSPs do not yet provide an effective strategic roadmap (IEO, 2004). The result is that any observed alignment between IMF TA and PRGF-supported programs does not (at least at this point in time) necessary translate into a link between IMF TA and any capacity-building priorities articulated in PRSPs.

67. The creation of regional technical assistance centers, for example, in the Africa region (AFRITACs) may present an opportunity for change. These centers’ explicit aim is to assist countries to address the capacity-building needs identified by the PRS process. However, as noted, their ability to do this is influenced by the fact that PRSPs typically did not offer a suitable diagnosis of TA needs or capacity constraints. At this stage, it is too early to judge whether the regional centers are having a substantial impact on alignment of IMF TA with PRS priorities. In our case studies, only Niger was served by an IMF regional TA center, and there is insufficient evidence to assess the extent to which the PRSP has had an impact on TA delivery from West AFRITAC.

E. Inputs of Other Surveillance-Related Activities on TA Allocation

68. As part of a wider strategy for strengthening the international financial system, the IMF has embarked in recent years on a series of new surveillance-related initiatives such as FSAPs and ROSCs, and assessments of AML/CFTs and OFCs.

69. These initiatives are intended to encourage good practices, identify potential weaknesses in domestic institutions and policies, improve transparency, and broaden IMF surveillance efforts through more in-depth analysis in those areas within the IMF’s expertise. The number of countries taking part in these initiatives has increased sharply since their inception in FY2000, and as of the end of FY2004, more than 100 economies were covered by over 500 ROSC modules and 58 FSAPs. The majority of the ROSC modules—64 percent—covered financial sector standards and were derived from FSAPs. Fiscal transparency and data ROSC modules accounted for 13 percent and 10 percent of the total, respectively.41

70. The evolution of TA has mirrored to a large extent the emphasis the IMF has placed on these initiatives.42 Follow-up activities—beyond the original diagnostic mission—frequently require an expanded set of TA activities from functional departments. Table 4.4 shows the composition of TA between that associated with regular policy reform and capacity-building activities and those specifically associated with the follow-up of these new surveillance activities.43 It shows a sharp increase in the TA associated with these new initiatives at the expense of the regular TA supporting policy reform and capacity building.44 The growing significance of these follow-up TA activities is even more evident when we look at each functional department. For example, follow-up TA in MFD went from around 1 percent of its total TA delivery in FY2001 to 27 percent in FY2004 (Figure 4.1). Similarly, follow-up TA in LEG went from less than 1 percent of its total TA delivery to around 20 percent over the same period, and in STA, the share doubled to 61 percent.45

Table 4.4.

TA Delivery by Key Policy Initiatives and Concerns, Excluding INS 1/

(In person-years)

article image
Source: OTM.

TA delivered in the field.

Figure 4.1.
Figure 4.1.

Follow-Up TA of ROSCs, FSAPs, AML/CFTs, and OFCs by Functional Departments

(In percent of each department’s total TA) 1/

Source: OTM.1/ TA delivered in the field only.

71. At the country level, our case studies provide additional evidence on the impact of these initiatives. In Zambia, TA priorities and the design of interventions have started to reflect the findings of the FSAPs, and the HIPC Initiative’s expenditure-tracking exercise. For instance, TA was provided to help the country develop a comprehensive development plan for the financial sector as recommended by the FSAP. TA was also provided to improve Zambia’s own assessment of current statistical practices under the GDDS framework. In the case of Ukraine, the 2003 FSAP identified a number of vulnerabilities in the banking sector which motivated further TA activity in banking supervision in late FY2004.

72. In Honduras, based on the FSAP, the government has embarked on a financial-system reform that aims at strengthening the regulatory framework and ongoing supervision, in addition to improving mechanisms for AML/CFT.46 In Yemen, substantial TA resources were devoted to banking supervision and reserve management following the FSAP.

73. In sum, the aggregate numbers along with the case studies suggest that the IMF’s new surveillance initiatives and institutional mandates are increasingly influencing the direction of TA. This trend was also confirmed by our discussions with IMF staff, including TA coordinators. In their view, these new initiatives have now become more important drivers of TA needs than Article IV surveillance.

74. This is not necessarily an undesirable development—if it were to reflect a strategic decision by the IMF that provision of TA in these areas is an overarching priority (e.g., because of the global public goods they involve and the comparative advantage of the IMF in providing them). However, it does raise the question of how priorities suggested by the more “global” Fund-wide initiatives are meshed with TA priorities derived from a policy framework driven by the country itself. Without such a framework, there is a danger that these new surveillance instruments and departmental initiatives may become ad hoc individual drivers of TA, without a coherent process to weigh priorities that may emerge from discussions with the authorities.

F. The Experience with the Technical Consultations Initiative

75. Technical assistance consultations (TCs) were supposed to have encouraged a more strategic outlook for TA by creating a country-centered approach to IMF TA delivery, providing better information to help review TA effectiveness and country’s commitment to making effective use of such assistance. The TC would comprise an assessment of past IMF TA, particularly implementation, and a forward-looking discussion of TA needs. The TC initiative was meant to lead to more consistent planning, improved follow-up, better coordination with other TA providers, and stronger country ownership. We have, therefore, examined this initiative in some depth.

76. TCs were to be implemented gradually through a series of pilot studies and then reviewed after a period of time. A procedure guidance note was issued in October 1999 to guide the pilot studies.

77. During 2002, OTM conducted a comprehensive review of 29 pilot cases and concluded that TCs have not proven as useful as expected and suggested that they be discontinued.47 It concluded that “TC notes … have tended to focus on a factual description of past TA activities, without much added insight on key issues, underlying policy factors, and future direction,” resulting in less informative discussions with country authorities. Some functional departments also noted the duplication of efforts with other staff reports. These disappointing results were attributed to the multiple objectives envisioned for the TCs and the choice of an already overloaded vehicle, the Article IV consultation, to implement them. Executive Directors, while acknowledging that TCs have not been entirely effective, added that, “continuing efforts to enhance the effectiveness of TCs as originally envisaged also should not be ruled out.”48

78. Although TCs have been discontinued, we took a second look at the pilot cases and made an assessment of what happened and what could be learned for the future. For this purpose, we examined in detail a diverse subgroup of 12 TCs among the 29 pilot cases, about half of which are PRGF-eligible countries. We divided the analysis between (a) an evaluation of the TC note contained in the Article IV staff report appendix, and (b) how well the main text of the Article IV report summarized the main thrust of TA discussions and the identification of priority needs. The following questions guided the analysis:

1. Evaluation of the TC in the Article IV staff report appendix

Q1: How clearly were past TA activities described?

Q2: How well was past TA effectiveness being assessed? How clearly was effectiveness defined (e.g., implementation record of recommendations, impact on ultimate objectives). Were these views shared by the authorities?

Q3: How well are the key factors that may have affected effectiveness discussed?

Q4: How clear was the identification of future TA needs, irrespective of who may be able to deliver the TA (either the IMF or other donors)?

Q5: Was the link between TA needs and key policy and institutional issues raised in the context of program and surveillance work clearly explained, e.g., the rationale behind these needs?

2. Evaluation of the main text of the Article IV staff report

Q6: How well does the staff report summarize past TA effectiveness?

Q7: How clearly are future TA needs presented?

Q8: How well does it summarize the rationale for the needs, e.g., their link to the future policy agenda?

79. Answers to these questions were rated between 0 and 3, where 0 refers to a case where the questions are not addressed; 1 = when these were addressed in a perfunctory way; 2 = they were addressed, but not in any significant depth; and 3 = they were addressed with good efforts at deriving strategic conclusions. Table 4.5 summarizes the results.

Table 4.5.

Experience with the TC Initiative 1/

article image
Source: IEO staff.

Scores represent average across all TA areas.

Although there is no separate TCs annex, the evaluation was based on the content of the main report’s section on TA matters.

80. The results are not encouraging. The pattern that emerges is that at best TCs did a good job in listing past TA activities, a rather mechanical exercise (Q1). However, they did not contribute much in assessing past effectiveness and its determinants, and the rationale for future priorities—which would have called for links to the overall policy framework. It is not surprising that the documents did not elicit much interest from the Board—the depth of analysis that would have stimulated a discussion was in general absent.

81. An in-depth examination of the 2001 Honduras TC, one of our country case studies, sheds further light on why this exercise was not particularly useful. It reads as a mechanical and proforma exercise. It describes past TA activities and lists current requests from the authorities, sector by sector. This information could have been derived directly from functional departments’ documents. It does not reflect a process of analysis and substantive discussion with the authorities on longer-term TA needs and priorities. The TC does not convey a sense that area departments have an overall TA strategy emanating from a medium-term policy framework that would allow a comparison of TA priorities across functional department lines. Consequently, it is not surprising that the exercise was not found useful.

82. Part of the problem may have been related to the specific implementation guidelines for the TCs, which called for staff to reach a broad understanding with the authorities on future TA priorities but not to “precommit the IMF to any particular course of TA action.” This qualification appears to have reduced the incentives for both staff and the authorities to devote sufficient effort to TA discussions, particularly in the face of other demands on their time during Article IV consultations.

83. We believe the basic idea and objectives behind the TC remain valid, in spite of their potential being far from realized in the pilot cases. Paraphrasing the findings in one of the TCs reviewed, there is a need to move from ad hoc TA requests to TA provided strategically to support economic policies and economic reforms. However, for the TC to play this role it requires more careful preparation, better supporting incentives, and more intense involvement of both the authorities and the area departments. If implemented in the PRGF-eligible countries, such an exercise should be clearly linked to the PRS process.

G. The Experience with Prioritization “Filters”

84. As part of the “Policy Statement on IMF Technical Assistance” adopted in April 2001, the Executive Board endorsed a set of nine “filters” intended to guide priorities for the allocation of TA resources (Box 4.1). To some extent, this action was a remedial response to the inability of Article IV surveillance to guide meaningfully the allocation of TA resources.

85. In the view of this evaluation, the system of “filters” as currently formulated is not helpful as a means to strategically prioritize TA. A number of filters were found to be too general and all-encompassing to be operationally meaningful, while others could result in unintended biases if followed mechanically. There are also some inconsistencies among the filters and other objectives being pursued within the IMF. Taken together, we found the system of filters to be “overdetermined,” implying little de facto guidance for priority setting. As is, they cannot compensate for the failure of other initiatives such as the TCs or the lack of a country-centered process to set TA priorities.

86. Filter 1 calls for TA to fall within the IMF’s core areas of specialization. These areas are listed in an annex to the TA Policy Statement and cover a wide range of topics related to policy advice and institution building. For the most part, the TA activities reviewed for this evaluation have fallen clearly within these areas of specialization. There have been a few exceptions (such as micro-financing and fiscal decentralization) for which the extent of the IMF’s mandate and expertise vis-à-vis the World Bank is somewhat ambiguous. On balance, however, IMF TA has been generally consistent with this filter.

Prioritization Criteria for Technical Assistance 1/

In light of demand pressures and competing needs for limited IMF TA resources, a new process of prioritization was adopted in 2001. This process consisting of nine criteria was created to assist staff in allocating resources, and link TA more explicitly to the IMF’s key policy initiatives. The weight given to each of these factors would shift depending on management’s perception of the IMF’s changing priorities.

Filter 1: Core specialization of the IMF. Assistance provided must fall within areas recognized as being within the IMF’s core areas of specialization.

Filter 2: Main program areas. TA must be clearly directed toward (1) preventing or containing crises and their contagion effects in nonprogram countries; (2) implementing sustainable debt relief and poverty reduction programs for low income countries; (3) fostering and maintaining macroeconomic and financial sector stability in countries not currently using IMF resources; (4) promoting regional capacity building and integration initiatives; and (5) rehabilitating basic economic and financial institutions in post-conflict situations.

Filter 3: Key policy initiatives. TA should be focused on supporting the IMF’s key policy priorities and initiatives. While these will vary over time, at present they are (1) standards and codes; (2) FSAP; (3) HIPC programs; (4) safeguarding IMF resources; (5) OFC assessments; and (6) policy reforms/institutional capacity building in support of achieving macroeconomic viability.

Filter 4: Impact and commitment. A country’s track record and degree of ownership and commitment should receive a large weight in appraising requests for TA. These factors could be offset by other considerations such as dealing with systemic countries, crisis situations, or when there has been a significant change in country leadership.

Filter 5: Regional diversity. The principle of uniformity of treatment of members requires some diversity in TA resource allocation and small countries’ needs—where TA can often have high impact and where other sources of assistance may be unavailable—are not ignored.

Filter 6: Availability of external financing. Although availability of external financing should normally not be a decisive factor, cooperation with other donors providing funding for IMF TA and/or to support other elements of assistance which the IMF cannot easily provide should be positively taken into account when appraising requests for TA.

Filter 7: Nature of the request. Requests for a policy or diagnostic mission, or short-term expert, are less costly than requests for a resident expert and can often be met without placing a significant burden on TA resources.

Filter 8: Regional approach. The IMF will actively seek opportunities for using regionally based approaches to providing TA.

Filter 9: Presence of other TA providers. In deciding whether to provide assistance, the IMF will consider whether other providers are actively engaged in similar efforts. Duplication should be avoided and all efforts at coordination will be made.

1/IMF (2000b).

87. Filter 2 requires that TA must be clearly directed toward supporting one of five main program areas. These are defined to be (1) crisis prevention or containment in nonprogram countries (principally systemically important nonindustrial and emerging market economies); (2) implementing sustainable debt relief and poverty reduction programs for low-income countries; (3) fostering and maintaining macroeconomic and financial sector stability in countries not currently using IMF resources (largely middle-income and transition economies); (4) promoting regional capacity-building initiatives, including training; and (5) rehabilitating basic economic and financial institutions in post-conflict situations.

88. With respect to low-income countries, program areas (2) and (5) are sufficiently broad as to encompass virtually all aspects of Filter 1. Consistent with the criticism of the 1999 TA evaluation, the emphasis implied by (1) and (3) is on preventive (rather than reactive) TA, as suggested by their applicability to nonprogram countries. What is less clear, however, is the role of IMF TA in middle-income countries making use of IMF resources (and not emerging from conflict). This group seems to be excluded from all but program area (4), which deals only with TA at the regional level. This is clearly neither the case in practice, nor likely in intent. Finally, program area (4) is essentially not a substantive program area but focuses on promoting efficiency through seeking economies of scale in TA delivery or in promoting regional integration. Were these to be seen as worthwhile objectives, they might best be contained elsewhere among the filters. On balance, there are shortcomings in the coherence and value added of this filter that should be addressed.

89. Filter 3, which calls for TA to be focused on supporting the IMF’s “key policy priorities and initiatives,” includes an extensive list of policy initiatives including the residual category of “policy reform/institutional capacity building in support of achieving macroeconomic viability.” This category alone is so broad that almost any activity can be classified under it. Thus, although the intention may have been to give priority to TA that supported several global initiatives (see preceding section) as formulated, the filter is too broad to serve such a purpose.

90. Filter 4 calls for a country’s past track record and its degree of “ownership and commitment” to be given more weight in setting priorities for future TA. Track record should be taken into account when functional departments discuss individual requests.

91. The challenge is how to implement these criteria in a consistent and transparent fashion. We find a variety of experience in our case studies. A weak track record in the implementation of TA in the period examined has constrained TA allocations in the areas of tax administration in Ukraine and statistics in the case of Yemen. On the other hand, in spite of the weak track record in the area of PEM in Zambia and customs administration in Cambodia, these countries kept receiving significant amounts of TA.

92. Linking past track records in the implementation of TA to future TA requires some analysis of the reasons for weak implementation as well as how to generalize judgments across TA activities within the country. (This issue is addressed in Chapter 6.) First, there are different types of track records that are relevant for different decisions regarding future TA. For example, the reluctance of the authorities to follow up on a specific piece of policy advice may not be relevant to guide future TA decisions in the area of long-term institutional support. Second, there may have been good reasons why implementation was slow. The case studies show that some implementation problems may have resulted because the recommendations were not convincing, were premature, or were not well adapted to the characteristics of the country, rather than a clear lack of commitment by the authorities. The implementation of a single treasury account in Ukraine, and TA in the area of monetary policy and government securities in Zambia are some examples.

93. Likewise, in order for a track record to influence TA allocations across countries, the process through which these decisions are made must be coherently analyzed and recorded across the institution. Only then can a filter of this type be applied more objectively and consistently. The implementation of a monitoring module of the TAIMS system could eventually allow such progress. (This is discussed further in Chapter 6.)

94. Filter 5 calls for diversity in the allocation of TA resources, including ensuring that the needs of small countries are not ignored. We found in Chapter 3 that the allocation of TA does conform to this criterion and that consumers of IMF TA have very diverse characteristics. As a matter of fact, many small countries receive large amounts of TA relative to the size of their economy.

95. Filter 6 indicates that—while not normally a decisive factor—the availability of external financing should be positively taken into account when apprising requests for TA. The findings of Chapter 3 (e.g., Figure 3.7) are consistent with this filter: countries receiving larger volumes of IMF TA relative to the regression norm often are those also receiving higher volumes of external financing. This is to be expected—it is part of the agreement with donors.

96. A more relevant question is what happens to the allocation of IMF-financed TA in response to external financing allocated to specific countries or subregions. Our estimate—using field data on TA—is that an increase of 10 person-days of externally financed TA is associated with an increase in IMF-financed TA of 1 person-day.49 These results suggest that some IMF-financed TA does get diverted from countries that are not receiving externally financed TA to those that do. Whether this is a concern depends critically on what factors drive this reallocation. If it reflects a situation where the IMF increases its own financed TA to particular countries when TA becomes highly productive—and as a result it also mobilizes additional external TA resources—that diversion could be welcomed. It would have been a part of the IMF reassessing priorities across countries, independently of having received extra external financing. On the other hand, if the above association reflects simply a reactive role of the IMF—more IMF-financed TA because donors increase their financial support—the implications are worrisome. It will signal a “procyclical behavior” under which IMF-financed TA responds to the priorities as seen by donors and may generate a misallocation of the IMF’s own TA resources.

97. In any event, this filter cannot substitute for the requirement that the IMF needs to develop a coherent country-centered policy framework from which to derive TA priorities—and in that context interact with donors as far as financing is concerned.

98. Filter 7, which has its origins in the 1999 evaluation of IMF TA, calls for priority to be given to short-term (or less resource intensive) requests for TA relative to requests for longer-term (resident) experts. Consistent with this emphasis, we see a shift beginning in FY2000 in the composition of TA away from long-term TA toward short-term missions (Figure 4.2). However, as discussed later in Chapter 5, a blanket assertion of the greater effectiveness of short-term missions is not consistent with our findings. Rather, we conclude that there is no a priori superior modality, and the most effective modality for TA delivery will depend heavily on individual country circumstances, particularly the level of institutional development.50 We, therefore, question the appropriateness of Filter 7, which imparts an a priori bias that may not be based on strong evidence.

99. Filter 8, which encourages the IMF to respond to cost pressures by using regionally based approaches to TA delivery (including regional TA centers like PFTAC, CARTAC, and the AFRITACs), is consistent with the trends shown in Figure 4.2. There does appear to have been a shift toward delivering TA through these regional centers (see Table 3.3).

100. Filter 9 calls for the IMF to take into account whether other providers are actively engaged in TA activities when it decides to provide its own TA. The intention is to avoid duplication given the limited resources at the IMF’s disposal. The filter recommends that, in such situations, all efforts at coordination will be made. However, as discussed in the next chapter, for such coordination and the associated division of labor to be effective, it should be considered at an early stage. And this requires that TA priorities be selected from an overall policy framework with participation of the authorities and interested donors.

101. To summarize, we have not found that the system of filters provides a meaningful or coherent basis from which to guide TA resource allocation. We, therefore, recommend either discontinuing these filters or revising them significantly to remove inconsistencies and instead encourage more flexible approaches adapted to individual country needs and circumstances.

H. How Are Countries’ TA Needs Translated into Final Resource Allocations?

How does the process work today?

102. The process of allocating resources to TA has many levels and participants and differs across departments. The global allocation of resources to the various functional departments emanates from the budget approved by the Executive Board. Each functional department allocates its resources to its different mandates, such as diagnostic programs (FSAPs or ROSCs), institutional initiatives (AML/CFT or OFC assessments), program support, Article IV consultations, and TA. Based on the projected resources allocated to TA, an iterative process between the functional and area departments, with occasional intervention from OTM, takes place to assign resources to regions, countries, and specific TA activities. The process culminates in the preparation of a yearly Regional Allocation Plan (RAP).

103. It should be noted that there is no global TA budget allocated to each functional department. There is also no multiyear budgetary process—a fact that impairs considering TA in a medium-term framework and the possibility of reallocating TA resources across functional departments. Rather, a de facto TA budget emanates from a decentralized process. At the same time, the increasing institutional demands on the functional departments—as new initiatives are launched without corresponding funding—may crowd out TA activities.

The RAP process

104. In December and January of each year, the discussion of the next financial year’s RAP begins with each functional department eliciting views about TA priorities from both their own department and from area departments through a series of bilateral meetings and informal consultations. These TA demands are driven by the needs of institutional mandates, program and surveillance activities of the area departments, and direct requests from country authorities.51

105. The preparation of the first draft of a RAP starts around late January, when functional departments send area departments their proposals for the allocation of TA resources to countries in their regions, specifying the topics to be covered. For a request to be included in the RAP it is expected to have passed through the filters described earlier. The area departments provide feedback on the draft RAP and suggest new missions or changes to the proposed timing. Typically, these responses combine feedback from heads of mission for each country in the department, with little explicit or systematic effort to sort out departmental priorities (either within or across countries) or contain the aggregate request within a preestablished resource limit. The staff of some area departments have argued that they are in no position to establish such priorities or to assess the relative urgency of need at this stage of the process. As a result, it falls to the functional departments to sort out and prioritize which requests to accept, which to put on reserve, and which to drop. The fact that area departments do not provide enough guidance on TA priorities across their own countries is a weakness in the present system.

106. The second round of consultations between area and functional departments is concluded by end-March and is then sent for final approval to management in April. In preparing a revised RAP, the functional departments try to ensure some flexibility to deal with unforeseen requests after the RAP has been finalized. During the year, area and functional departments have regular reviews of the implementation of the RAP to discuss unforeseen requests, timing of missions, or possible reallocations (Box 4.2).

FAD’s Regional Committee Information System

A good example of making the process of allocating TA resources more transparent is the current practice in FAD. Since mid-2002, FAD has made available on the IMF’s internal website the deliberations of its regional committee. This committee usually meets once a month under the chairmanship of the Deputy Director or a Senior Advisor, and is in charge of assessing the merits of all TA requests submitted to FAD in order to make a recommendation to the Director to approve or reject these requests. The summaries of these deliberations include a list of all requests, divisional assessments of each request, a summary of the committee’s discussions, and the justification for the final decision. As a result, the logic for FAD’s decisions has become clearer and more transparent. For example, a poor track record, a lack of clear commitment from the authorities, or the request not being within the area of FAD expertise have often been mentioned as key factors behind TA rejections. Similarly, the systemic importance of a country, the good performance of a resident advisor, or the support of the area departments for the request, are mentioned as important factors influencing the approval of some requests. Overall, FAD is enhancing accountability by making more transparent the process of allocating TA resources.

107. In the view of TA coordinators, the RAP process is working well as a clearing house for TA requests. It has allowed systematic interaction and feedback among departments and enough flexibility after resources have been allocated to deal with unexpected requests and shocks. However, in our view, the key limitation is not the RAP as a clearinghouse, but the lack of a process to identify TA country priorities from a coherent medium-term framework that meshes Fund-wide and country-driven priorities (a function that was to have been served by the TCs) and the weak role of area departments in identifying priorities across countries.

The impact of budgetary constraints on TA priority setting over the medium term

108. A critical constraint on TA priority setting and resource allocation in the context of a coherent medium-term strategy is the absence of a multiyear budgetary process within the IMF.52 This, in effect, makes it difficult for functional departments to make TA commitments to member countries beyond a single financial year, potentially generating a bias in favor of short-time TA activities. Moreover, TA is financed—not by a separate budget item—but out of departmental budgets, decisions on which are not usually made until just prior to the new financial year. While, to a large extent, the size of overall departmental budgets is predictable, there are uncertainties, particularly because of the budget constraint under which the IMF has been operating in recent years. Parallel to decisions on the overall administrative resource envelope are departmental decisions on resource allocation between activities (of which TA is only one) and among countries served by each department. The latter is further complicated by unexpected TA demands or changing priorities during the fiscal year.

109. The net effect of all this has made IMF TA less predictable to country authorities, including with respect to the IMF’s longer-term strategy to wean countries off some of the more intensive forms of TA as administrative and institutional capacity is developed. (This problem is identified in some of our case studies, such as Niger.) Effective coordination with other TA providers—who might be able to address gaps in IMF TA—is also undermined, particularly given the lags in their own resource allocation processes. These drawbacks to the present system are of greater consequence for the “capacity-building” aspects of TA, where coordination with other stakeholders within a medium-term framework is more important than for short-term policy advice and “firefighting” types of TA activities. This suggests the need to consider administrative mechanisms that would permit staff to make commitments at the country level within a medium-term framework.

110. The absence of a multiyear budgetary framework—and an associated RAP process that identifies TA priorities over a longer horizon—also reduces the possibility of identifying early on the need to gradually reallocate TA resources across areas of expertise and across functional departments.

I. Conclusions and Recommendations

111. The major conclusions from the earlier analysis are

  • The follow-up to surveillance-related institutional initiatives, such as ROSCs, FSAPs, and AML/CFT assessments has become an important driver of TA. These initiatives appear to have become more significant in identifying TA priorities than those emanating from Article IV surveillance or the PRS process.

  • Functional departments have taken a number of important initiatives to be more strategic in making intra-sector choices of TA priorities and looking at track records in implementation. The preparation of country strategy briefs in each sector, assessment missions, and a more country-centered system for evaluating TA requests have the potential to improve the resource allocation within departmental lines. However, these are not a substitute for a broad country policy framework that would be able to compare countries’ TA needs across departmental lines.

  • The PRS process was expected to be a key vehicle to identify capacity-building needs in low-income countries. This has not happened yet, not least because of shortcomings in PRSPs themselves.

  • According to staff, the RAP process is working well as a clearinghouse of TA requests. What is lacking is a medium-term priority-setting mechanism through which country TA priorities as well as priorities across countries can be established. To do this area departments need to play a more active role.

  • The present system of TA filters is not a substitute for the lack of such framework. The filters are at best ineffective since they are overdetermined. They should be either revised or discontinued.

  • To an important degree, the short-term focus of the current budgeting and RAP processes within the IMF works against medium-term planning for TA. Without a mechanism to enhance predictability in the provision of IMF TA over the medium term, it will be difficult for the IMF to coordinate and leverage its TA efforts with those of donors and other providers of TA as well as to identify the need to reallocate TA resources over time. This is likely to be more of a problem for TA focused on capacity building than for TA oriented toward short-term policy advice.

  • There is no mechanism through which medium-term TA resource allocation among countries and topics can be critically assessed by the IMF. While OTM prepares regular reviews of TA, these are essentially backward-looking documents. What is missing is a forward-looking and strategic discussion of whether or not the allocation resulting from decentralized decision making adequately reflects the medium-term capacity needs of countries as seen by both the IMF and the authorities.

112. Together with the authorities, the IMF should derive TA priorities out of a medium-term policy framework that takes into account country-specific strategic directions. A number of managerial approaches to establishing such a framework are possible, but the following suggestions could be considered.

  • In low-income member countries, the PRS process provides the natural vehicle to identify TA capacity-building priorities in the core areas of expertise of the IMF, although it has typically not yet been used effectively for this purpose. The PRSP should form the basis for a dialogue between the IMF and the authorities on TA priorities for the medium term, with as clear an indication as possible of what is expected from each side in order to improve accountability. It should also provide the vehicle for collaboration and division of labor with other donors and hence help the overall coordination of TA. The IMF needs to engage countries to help them articulate their medium-term capacity-building needs in its areas of responsibility, in accordance with the PRSP.

  • For other member countries, the institutional framework for deriving medium-term strategic priorities is less clear-cut, and a variety of approaches may work best, depending on country circumstances and the intensity of the members’ demand for IMF TA. In cases where there is a relatively high provision of IMF TA, the framework may require periodic in-depth TA consultations with the authorities (possibly every three years) comprising a retrospective analysis of past progress and a forward-looking exercise to identify priorities. To avoid the lack of incentives of the past TC initiative, the resulting TA priorities need to feed more directly into the IMF’s own mechanisms for committing its TA resources. For other countries, a less comprehensive approach, possibly built around Article IV consultations, may suffice.

  • Resident representatives could play a greater role in developing these frameworks and this role should be explicitly acknowledged in the TORs for these positions.

  • The RAP process should continue as the main annual interaction between the area and functional departments for determining TA activities. In addition to incorporating TA activities stemming from the medium-term framework, it will also incorporate new TA activities that may emerge from the needs of programs or from unexpected new demands from the authorities. This will prevent the TA program from being preempted by short-term needs while at the same time it will recognize the need to accommodate these types of demands.

  • As the IMF moves into a multiyear budget framework, the RAP could also evolve toward a multiyear RAP helping reallocate TA resources within the institution over a longer horizon. For this process to work, each area department would need to spell out their proposed TA country priorities (based on each country policy framework and with the authorities, as discussed earlier), including ranking of priorities across countries in the departments, the list being more notional for the outer years. This approach would allow a comparison between demands from the area departments in the outer years and the existing supply and composition of skills in the functional departments; it would help identify pressure points in the future that would call for a gradual reallocation of resources within and across functional departments. This multiyear budget process and the associated multiyear RAP would make the allocation of TA resources more responsive to demand as time passes.

  • This strengthened role of area departments in preparing a multiyear plan of TA priorities (including prioritization across countries) will require both incentives and resources. Senior management in the area departments should ensure that staff considers TA as an integral part of the assistance strategy to countries, with due recognition in staff performance assessments.