Chapter 7: The Implications of Prolonged Use for the Member Country and the IMF
- International Monetary Fund. Independent Evaluation Office
- Published Date:
- April 2003
1. In this chapter, we examine the effects of prolonged use on institutional development within the borrowing countries and on the IMF’s financial and human resources.
Implications for the Borrower: Impact on Institutional Development1
2. There are two possible channels through which prolonged use of IMF resources could have an impact on a country’s institutional development: (i) it may help, or hinder, the development of technical skills within the departments involved in the negotiation of IMF programs; and (ii) it may foster or undermine the country’s policy formulation process; including the process of policy closure, that is, the way in which the political and bureaucratic system reaches consensus on a final strategy. The second channel is closely linked to the issue of ownership of IMF-supported programs.
Impact of prolonged UFR on the buildup of economic management skills
3. Government officials in all three country case studies and in most other prolonged user countries generally agreed that the prolonged involvement of the IMF had resulted in some positive transfer of technical economic management skills. However, views differed as to the extent of that transfer.2 In a few cases, including Morocco and the Philippines, the prolonged time spent under IMF-supported programs was deemed to have made a major contribution to the development of technical and analytical skills. However, in several others, in particular Pakistan and Senegal, it was felt that the knowledge being transferred by prolonged exposure to IMF missions was too narrow and specialized to be of much use outside the context of IMF programs and that the ready availability of IMF models and methodologies weakened the incentives for local technicians to develop their own tools. The potential buildup of technical skills was also limited by the relatively low retention rate of technocrats having benefited from the IMF’s training.
4. Notwithstanding the perception that there was some transfer of technical skills, officials in all three country case studies, as well as a number of respondents to the questionnaire, commented that programs had paid insufficient attention to institutional reform and to the development of adequate implementation capacity. Perhaps this is not surprising given the relatively short time span of individual IMF-supported programs. Even in cases where there has been a succession of programs following on one from another, the short-term horizon of each individual program naturally limits the extent of attention paid to capacity building. Respondents acknowledged the potential role of technical assistance (TA) and also recognized that IMF-supported programs made access to IMF TA easier, but they felt that in general IMF-provided TA was too transitory to have a sustained capacity building impact, in particular due to the lack of implementation followup. A general recommendation often made was that IMF TA should have a more direct focus on the ways and means of policy implementation.3
5. Finally, in several countries, including the three case studies, many commentators within and outside government were of the view that the nature of the IMF’s involvement and its leverage could have given it scope for dealing with governance issues more than it did (for instance, in relation to connected lending practices in Pakistan’s banking sector, or with tax evasion in all three countries).
Impact of prolonged UFR on the policy formulation and closure process
6. With a few exceptions, comments received from country authorities regarding the impact of prolonged UFR on the policy formulation process were generally quite negative, primarily because prolonged use was not conducive to the development of a capacity to generate “homegrown” policies. Instead, it tended to create a situation where the policy formulation process revolved around negotiations with the IMF, locking domestic policymakers into a reactive rather than a proactive role. A typical comment was that IMF missions’ briefing paper instructions become the framework of reference for negotiations, thereby giving too little scope for the discussion of alternative policy options and tradeoffs. However, many country authorities acknowledged that the transformation of the ESAF into the PRGF and the creation of the PRSP process had begun to change the mode of relations with the IMF in a positive way, though it remained to be seen how this would evolve over time.
7. In several countries, including Pakistan and Senegal, it was also pointed out that the monitoring of program implementation—required not only by the IMF’s own procedures but also by those of other donors whose financing was linked to the IMF program in some way—was a significant strain on limited administrative resources. As a result, local civil servants often simply did not have enough time left to focus on developing policies on aspects not covered by the program itself, or even to think through the implementation of the policies adopted under the program.
8. Some country authorities noted that the perceived need to consult the IMF on every single initiative in areas falling under the program’s purview led to an inhibition of local policy formulation and into the attrition of policy development capabilities. This is a problem that could potentially arise in all IMF-supported programs, but it obviously becomes much more serious in a prolonged use situation, where the effect of attrition over time and the development of a culture of deferring to the IMF can be substantial. However, there were exceptions where the domestic policy formulation process was more centerstage, usually with beneficial results. For example, Pakistan was able to develop, largely on its own and outside the context of a program, a very far reaching tax administration plan in 2000. This example illustrates that where there is sufficient political will, the inhibition of local initiative can be rapidly overcome, even after more than a decade of use of IMF resources. Likewise, authorities in a couple of other countries noted that the policy formulation process had been largely country-driven, with policies implemented under IMF-supported programs being only part of their own, “homegrown” agenda. This suggests that even in situations where prolonged use is unavoidable, it is possible to ensure that policies are more homegrown and therefore more likely to be owned.
9. Similarly, it was mentioned, most notably in Pakistan, that once several successive governments had made a habit of justifying their policy decisions to the public by invoking the exigencies of IMF-supported programs, it was very difficult to reestablish a sense of ownership even for those policies that were genuinely “homegrown”: as soon as they were incorporated into an IMF-program letter of intent (LOI), they became perceived as IMF-imposed. The practice of trying to include a country’s entire reform agenda in the IMF’s LOI, rather than relying more on the country’s own policy documents, exacerbates these perceptions.
10. More generally, the absence of a broad public debate on core policy “planks” resulted in giving policy decisions a flavor of “fait-accompli” to many stakeholders, even when there was, in fact, flexibility in the negotiations. Once again, this perception tended to undermine ownership. For example, in the Philippines under the Ramos administration in the early 1990s, considerable progress had been made on a range of structural reforms that emerged from the domestic political process and may have been more politically acceptable since they did not feature prominently in IMF conditionality. These included an extensive privatization program; the opening of certain sectors to foreign participation; and improvements in customs administration. The long-delayed decision to restructure and recapitalize the Central Bank was also implemented in 1993, in between programs. In contrast, some of the structural measures central to IMF conditionality in the period, including reform of the oil pricing system and tax re-form, encountered serious political difficulties, probably in part due to a lack of ownership.
11. Many officials interviewed in Senegal and Pakistan pointed to the overly exclusive nature of the dialogue established by the IMF with the Finance Ministry and the Central Bank as a negative feature. In their view, this process de facto gave these agencies excessive power to make decisions and undertake commitments that have implications in areas which are the responsibility of other ministries and departments, without sufficient consultation, which led to lack of ownership of some parts of the government and possible problems in implementation. While this problem arises in all IMF-supported programs, it becomes especially acute in situations of prolonged use because the process is seen to marginalize a number of agencies over an extended period of time.
12. It is interesting to note that IMF staff members involved in the negotiations in the three country cases have commented that they made every effort to associate line ministries in matters related to decisions affecting their field of competence. They also noted that initiatives to keep the negotiating circle small often came from the authorities’ side. However, associating line ministries is not enough if these ministries are effectively excluded at critical decision making points. We recognize that the perception of powerlessness of the line ministries and the resulting lack of broad ownership may have stemmed primarily from shortcomings in the domestic policy closure process, which may be unfairly blamed on the IMF and its procedures. However, the end result is clearly negative and the problem is not just one of perception, since it can compromise implementation.
13. In fairness, it should be noted that perceptions were not universally negative. In Morocco and a few other countries, the authorities noted that the formulation of IMF-supported programs had generally involved a relatively large group of stakeholders, even though negotiations per se were held strictly with the Finance Ministry and the Central Bank. This had contributed to educating managers across economic institutions about the benefits of prudent macroeconomic management and openness to the world economy. The Philippine authorities likewise credited the IMF’s prolonged involvement with the strengthening of the cohesion of the economic team and said the IMF had played a useful role in interacting with a broader set of policymakers, including in Congress on the rationale for the IMF’s prescriptions, even though they felt that this function had been underused. This suggests that efforts to broaden consultations, especially with line ministries but also with other agencies, should receive more attention and the IMF should work with the authorities to evolve country-specific modalities toward this end.
14. Yet another mechanism whereby IMF programs have an impact on policy closure is when they are used by some parts of the political establishment in an attempt to “tie the hands” of the incoming government with a program negotiated by the previous administration. Examples include Pakistan in 1988—in the transition from a military regime to an elected government—and also the Philippines in 1998—in the transition to the Estrada administration. While some argue that, in those cases, the presence of IMF-supported programs had a stabilizing influence, the overall result seems to have been that the resulting program was not strongly “owned” by the incoming government, which led to weak implementation and ultimately a failure to meet program objectives.
15. In a few countries, including Senegal and the Philippines, it was pointed out that the IMF was not always sufficiently sensitive to the role of parliaments in the decision-making process. For instance, in Senegal, the ad hoc measures adopted in the course of a program under the IMF’s pressure, particularly in the taxation area, often rendered meaningless the role of parliament in the budgetary process. This was a source of resentment from parliamentarians, potentially leading to difficulties in implementing other parts of the program at a subsequent stage. In other cases, such as the Philippines, Congress generally had the ultimate decision-making power in the approval of key reforms prescribed by the IMF, making it difficult to comply with program conditionality within the time constraints set by the program. As pointed out in the Philippines country study, the IMF team did interact informally with Congress on the rationale for the IMF’s recommendations, but the problem was not necessarily overcome. Ultimately, it is the responsibility of country authorities to determine how parliaments should be associated with the government’s negotiations with the IMF, but a greater mutual awareness by Fund missions and parliamentarians of their respective roles and objectives would help smooth program implementation.
16. Many of the problems mentioned above would arise in any IMF-supported program, but they become more serious in the case of prolonged use because the process of policy formulation and closure is seen to be affected over a much longer period. In principle, these problems should be reduced for PRGF countries with the introduction of the PRSP process, though it was too soon to judge in the cases of Pakistan and Senegal, or in the responses from other prolonged users, how much difference this was making in practice.
Implications of Prolonged Use for the IMF
Impact on IMF finances
In the General Resources Account (GRA)
17. As noted in Chapter 2, over the last 30 years, prolonged users have attracted a substantial share of IMF’s GRA resources: over 20 percent of commitments per year on average (and close to one-third in the last 10 years). These substantial commitments to prolonged users do not appear to have significantly constrained the ability of the IMF to extend loans to its members in need, as total commitments only rarely came close to the ceiling of available resources, and even then only for brief periods of time. Of course, since the IMF manages its liquidity position in part by decisions on the level of financing provided to particular countries, as well as through periodic reviews of its resource base, the fact that commitments only rarely reached the ceiling does not mean that the extent of financing to prolonged users did not affect the access level of other users or decisions on the magnitude of quota increases. Nevertheless, periods of resource scarcity were not associated with a sharp increase in commitments to prolonged users (see Figure 7.1).4
Figure 7.1.Available IMF General Resources and Commitments
Sources: IMF Treasurer’s Department and IEO calculations.
18. Although prolonged use did not preempt resources on a scale that constrained total lending, it did have a significant effect on the revolving character of IMF resources. This impact can be measured indirectly5 by comparing the average length of lending cycles between prolonged and “temporary” users, taken as fixed groups, the latter being used as a proxy for standard lending cycle length (hence for standard revolution speed of IMF general resources): resources lent to prolonged users took, on average, almost twice as long as resources lent to “temporary” users (15 years against 8 years, respectively) to become fully available again to other members.
In concessional Trusts
19. The resources of the ESAF/PRGF Trust are also intended to have a revolving character, albeit over a longer time frame than the GRA. For the reasons already discussed, that time frame could not be defined authoritatively, and instead was approximated by the average lending cycle length of “temporary” PRGF-eligible users. The comparison with prolonged users shows, once again, that the latter take much longer to make the IMF resources they use available again for other members: 21 years on average, against 14 years for “temporary” users (over 1971–2000).6
20. As in the GRA, the scale of commitments to prolonged users in relation to total resources available for lending rarely came close to representing a significant constraint on overall lending. Nonetheless, commitments to prolonged users averaged just under half of total commitments of ESAF/PRGF resources since 1989 (Figure 7.2).7
Costs imposed on the membership through the accumulation of protracted arrears8
21. The problem of arrears to the IMF is much broader than prolonged use, and there is no obvious direct relationship between the two issues. However, close to 70 percent of countries which, at some point over the last 30 years, incurred protracted arrears to the IMF are prolonged users, and about one-third of the countries identified as prolonged users in this study incurred protracted arrears to the IMF at some time (Figure 7.3). In most cases, arrears were accumulated after a period of prolonged use. However, in a number of cases, prolonged use began after the clearance of arrears; for these cases, prolonged use might have been part of a “defensive lending” strategy, aimed at preventing the appearance of new arrears.
Figure 7.3.Evolution of Arrears to the IMF
Sources: IMF Treasurer’s Department and IEO calculations.
Notes: In these two figures, prolonged users are treated as a fixed group. Otherwise, by definition, few countries in arrears would qualify as prolonged users except in the year in which the arrears are first incurred, since their arrears make them ineligible to make further use of IMF resources. Arrears include both overdue repurchases and unpaid charges.
22. The sharp reduction of the volume of arrears in recent years and of the weight of prolonged users within the total implies that the costs imposed on the membership as a whole through these channels are small.9 However, the experience of the early 1990s shows that this was not always the case. Thus, enhanced attention to prolonged users’ ability to service their obligations to the IMF is warranted and this has indeed been prescribed by operational guidelines since 1990. Evidence from the case studies suggests, however, that sections of reports addressing these issues were often perfunctory, giving little idea of the potential risks in the event of markedly worse, but still possible, outcomes.
Pressures on the IMF human resources
23. The number of lending arrangements in effect each year has increased markedly since the late 1980s, from 34 in 1987 to 58 in 2000. As was explained above, that increase was, to a significant extent, due to an increase in prolonged use by certain countries (along with an expansion of the IMF membership in the wake of the collapse of the Soviet Union). Moreover, the share of arrangements under concessional facilities—which are typically more resource intensive than GRA arrangements, and among which prolonged users are over-represented—also rose significantly: from 30 percent in 1987 to over 50 percent in 2000. However, staff resources did not increase commensurately (Figure 7.4). This contributed to a rising work overload, especially from the mid-1990s onward.10
Figure 7.4.Use of Fund Resources and Staffing Indicators, 1987–2000
Sources: IMF budget, IMF Annual Report, and IEO calculations.
This section draws on the relevant findings of country case studies and on the responses to a questionnaire sent to the authorities of all prolonged users.
In one case, the authorities commented that it was still too early to judge whether the prolonged use of IMF resources had had a positive institutional impact, suggesting that the capacity building effect of programs, if any, had been limited thus far. In another case, the authorities were of the view that the prolonged involvement of the IMF had had “a debilitating effect on economic institutions,” particularly due to a “collapse of the long-term economic planning function.” A similar view with respect to long-term planning was expressed by the Senegalese authorities.
The effectiveness of IMF-provided technical assistance is a larger issue that goes beyond the scope of this evaluation. It is one of the items on the agenda for future evaluation by the IEO.
The concept of ceiling of available resources referred to here is defined on the basis of a liquidity ratio of 25 percent (i.e., 25 percent coverage of liquid liabilities by adjusted uncommitted usable resources, excluding borrowed resources). While the IMF does not have a mandatory minimum liquidity ratio, 25 percent is the lower end of the range historically observed for the minimum. A definition of available resources based on a more conservative liquidity ratio would reduce the amount of available resources. When its liquidity position becomes tight, the IMF has the re-course to activate its borrowing arrangements from creditor countries, which can temporarily increase available resources by about SDR 34 billion.
Ideally, this impact would be measured directly by (i) determining a notional “standard” speed of turnover of IMF resources and (ii) comparing it with the actual speed of turnover observed over a given period. However, the first step involves a number of judgments and assumptions on “normal” frequency or length of UFR and the blend of facilities in use at any given time, which it was not found practical to make in the context of this study.
As before, these figures understate the actual average lending cycle length owing to cycles that remained uncompleted in 2000.
In each year, only commitments to countries that were prolonged users in that year were taken into account.
Protracted arrears are defined as overdue obligations for six months or more.
Arrears impose costs on the IMF membership through two distinct channels, which have been designed to ensure symmetric treatment of creditors and debtors to the Fund. First, since 1986, the rate of charge (remuneration) is adjusted upward (downward) to cover the loss of income due to overdue charges on outstanding liabilities to the IMF. This burden-sharing mechanism replaced an earlier asymmetric mechanism under which the entire burden of arrears fell on borrowing members through an upward pressure on the rate of charge. Between 1987 and 2000, the adjustment averaged a little under 80 basis points per annum both ways. However, it has been much smaller in recent years (around or below 20 basis points) and only a fraction of it is attributable to prolonged users. It reached a peak of over 160 basis points (83 points on the rate of charge and 79 points on the rate of remuneration) in 1991, at a time when prolonged users represented the bulk of overdue obligations. Second, the IMF’s precautionary balances, which according to general guidelines should fully cover credit outstanding to members in protracted arrears, must receive higher allocations when the latter increase, implying a higher net income target and/or larger allocations to the SCA-1, which are financed by symmetrical adjustment to the rates of charge and remuneration.
According to the internal Budget Reporting System (BRS), in 2000, the shortfall in effective staff years (compared to the level needed to match the increase in UFR activity) exceeded 700 effective staff years. Over the last five years, prolonged users absorbed on average close to 130 staff years just for UFR-related work.