8 Summarizing the Views of the IMF Executive Board

Ruben Lamdany, and Leonardo Martinez-Diaz
Published Date:
July 2009
  • ShareShare
Show Summary Details
Jeff Chelsky

Clear and accurate summaries of the decisions and views of the Executive Board, which reflect a sufficiently broad range of views from across the membership, are a key element of sound IMF governance. Throughout the Fund’s history, various vehicles have been used to achieve these objectives. These have evolved over time, taking into account the nature of the issue under discussion, changes in Board practice, and the evolution in the Fund’s approach to transparency. This note describes the main instruments used, reviews their evolution, and assesses the adequacy of current practice from the standpoint of ensuring continuity, clarity, and accountability of Board deliberations. It concludes that, while the process seems to be working well on the whole, minority views are inconsistently reported, and consensus views and decisions are not clearly distinguished from those of groups of Directors.


The ability to summarize clearly and accurately the views and decisions of the Executive Board is a key element of sound IMF governance. With respect to formal decisions of the Board, the constituency and quota system can provide a precise basis for ascertaining support for the issue at hand, but decisions can also be imbedded in meeting summaries, diminishing somewhat their clarity. The expression of Board views, however, does not always involve a clear “yes or no” decision but instead communicates the degree of support on the Board for a particular view or the range of opinion on the issue in question.

Throughout the IMF’s history, various instruments have been used to summarize the content and outcome of Executive Board meetings and/or to convey information about the factors and debate that went into a decision. These have evolved over time, taking into account the nature of the issue under discussion, changes in Board practice, and the evolution in the Fund’s approach to transparency. Reflecting the significant element of judgment involved in any effort to condense the content of an Executive Board discussion, Board summaries have at times been controversial, with questions raised about whether they accurately reflect the “sense of the meeting” and the extent of support for various positions.

This paper describes the main instruments used to summarize Board discussions, reviews their evolution, and assesses the adequacy of current practice from the standpoint of ensuring the continuity, clarity, and accountability of Board deliberations. The next section offers a brief history of mechanisms for summarizing the views of executive directors. The third section describes current practices, and looks at related legal and semantic issues. The fourth section analyzes a sample of summings up (SUs), assessing their clarity and accuracy and attempting to pinpoint the stage in their preparation at which inaccuracies have most often arisen. The final section draws tentative recommendations.

A Brief History

According to the Fund’s Rules and Regulations, the Secretary, under the direction of the Managing Director, shall be responsible for preparing a “summary record of proceedings of the Executive Board.”1 This summary record initially took the form of relatively brief minutes, which noted only the issue discussed and any decisions or agreements reached. Early Board guidance to “refrain from additions to the minutes setting forth in detail the positions of particular members of the Executive Board”2 suggests that minutes were not intended as a comprehensive record of all points expressed during a Board discussion. Instead, provision was made for the preparation of “memoranda,” upon request of the Chairman, that would contain a more comprehensive summary of important discussions and reflect the various points of view expressed.3 These “memoranda,” which could be seen as a precursor to today’s SUs, were to be circulated to discussion participants for approval before being incorporated into the official records of the Executive Board. However, they do not appear to have been used widely, if at all.

Initially, meeting minutes were circulated to all executive directors shortly after Board meetings for review and correction, and approval at the subsequent meeting.4 The Secretary was required to read draft minutes “aloud and in full” before their adoption, a practice which proved cumbersome given increases in the number and length of meetings, and which was amended several times, most recently in April 1978.5 By the 1970s, Board minutes had become quite detailed and had begun to describe interventions by specific directors.

“Stand-alone” summaries of policy discussions began to be systematically produced in the early 1970s to make the content of lengthy and complex discussions more easily digestible by ministers, who were at that time being called upon to consider fundamental reforms in the Fund’s mandate and policies.6 These were initially referred to as “Summings Up,” and they described the range of views expressed by directors during the discussion and provided commentary and preliminary conclusions by the Chair. By 1980, they were being referred to as “Concluding Remarks by the Chairman” to differentiate them from the SUs that were by then being prepared to summarize the views of executive directors following Article IV consultation discussions (see the next paragraph). However, in November 2001, reflecting a desire to regain greater control of the content of summaries of Board policy discussions, directors indicated that “as a general rule,” policy discussions should conclude with a formal SU.

Stand-alone summaries of bilateral surveillance discussions did not appear until 1978, following the transformation of Article VIII and Article XIV consultations into Article IV consultations.7 The primary audiences for these documents were the authorities of the country under discussion, the general membership, and Fund staff.8 The first of these SUs were a single page in length and always began by asserting that directors were generally in accord with the views expressed in the staff appraisal. This was not intended to imply that all directors necessarily agreed with every aspect of the staff appraisal. No reference was made to the views of individual directors. Non-consensus views were frequently ascribed to “most,” “many,”“some,”and “several” directors.

Executive directors not intervening in a particular discussion were deemed to be supporting the views of staff.9 This convention derives from the requirement in Rule C-10 of the Funds’ Rules and Regulations that “the Chair shall ordinarily ascertain the sense of the meeting in lieu of a formal vote …” where “sense of the meeting” is defined as “a position supported by executive directors having sufficient votes to carry the question if a vote were taken.” However, when there is no explicit decision to be taken, and a range of views have been expressed on a particular issue, the Chair has significant discretion as to how to interpret the silence of an executive director. Partly to improve the efficiency of Executive Board meetings by avoiding a situation where every director feels the need to intervene to voice agreement with the staff, this has been interpreted by the Secretary’s Department (SEC) as a presumption that silence equals agreement.10 This contrasts with the situation where a formal vote is taken. In these cases, for decisions requiring a simple majority of the votes cast, the silence of an executive director is interpreted as an abstention.

More than 20 years later—in 1999—SUs began to be prepared for Board discussions of Fund-supported programs (“use of Fund resources” or “UFR”). The guidance for these was based on the Board’s long-standing practice of not voicing “reservations or divergent views [which might] reveal the absence of full consensus at least on some aspects of the program, and could potentially weaken the objective of building confidence in the member’s program.”11 These SUs therefore differed substantively from those for Article IV consultation discussions in which differing views and criticism could be reflected.

Prior to 1999, the Board had relied on the executive director representing the borrowing member to convey the views of the Board to the authorities following UFR discussions. But, by the late 1990s, some directors were seeking greater assurance that the specific concerns expressed during the discussion would be fully and accurately conveyed to the borrowing country authorities. In response, “internal” SUs began to be produced to reflect comments that were critical of programs or questioned their viability. For combined Article IV/UFR discussions, this resulted in the preparation of more than one SU In 2001, the Board agreed to discontinue this practice and instead decided to augment Article IV SUs with text that could describe critical views on programs but that was not included in the version of the SU that was published as part of the Public Information Notice (PIN).

Stand-Alone Summaries Today

This section describes the use of stand-alone summaries today, first describing the preparation and approval process and subsequently noting some legal and semantic aspects.

SUs are currently prepared for all Article IV and UFR discussions except stand-alone discussions on misreporting and repurchase expectations. For a combined Article IV/UFR discussion, the SU contains a section on key policy issues followed by a separate short section at the end on key program issues; only the first part can be published. SUs are also prepared at the conclusion of virtually all policy discussions.

Since 2005, and with few exceptions, Chairman’s concluding remarks have only been used to summarize Executive Board seminars (rather than regular meetings). These remarks include the views of the chair12 and cannot, therefore, convey a formal decision of the Board. However, they can be used to describe the status of policy discussions on issues under ongoing consideration, in order to identify areas of preliminary consensus and unresolved issues that will need to be addressed.13

Chairman’s statements (which are not the same as Chairman’s concluding remarks) are prepared after the Executive Board adopts a UFR decision or completes a discussion of a country’s participation in the Heavily Indebted Poor Countries (HIPC) Initiative or of a Poverty Reduction Strategy Paper (PRSP)-related document. According to SEC’s guidelines, “unlike the Summing Up, the Chairman’s statement should not attempt to cover the discussion as a whole or reflect divergent Directors’ views, but rather convey a few (three to four) points on which the Board placed emphasis. It is drafted as a statement from the Chairman or Acting Chair summarizing the views of the Board. . . . The Chairman’s Statement is much shorter than the Summing Up, hence necessitating judgment as to the areas for emphasis. It does not attribute statements to Directors.” It is this characteristic that largely differentiates a “Chairman’s statement” from a “Summing Up.”

The remainder of this paper focuses on SUs, reflecting their predominant role as a record of the views of the Fund’s sole decision maker, the Executive Board.

Preparation and Approval of Summings Up

To facilitate their preparation, SUs are drafted in advance of the Board meeting by the department (s) responsible for the staff report that is to be discussed. The draft SU is prepared under the assumption that directors will generally agree with the staff analysis and recommendations contained in the staff report. It is sent to SEC three working days before the Board meeting to be amended, if necessary, to ensure consistency with usual practice.14 In the case of Article IV consultation staff reports, the area department also transmits the draft SU to management. Directors do not see the draft SU at any point.

There are several opportunities for the views of directors to be reflected in the SU. The draft SU can be amended to incorporate views expressed in written statements circulated before the Board meeting (“grays”) (see Box 1). During the Board meeting, the Secretary, under the direction of the Chair, makes changes to the draft SU to reflect oral interventions by directors. At the close of the discussion, the revised SU is read out by the Chair and directors have the opportunity to intervene if they disagree with the characterization of the discussion. When disagreements arise, particularly with the description of the extent of support for a particular view, the Chair generally indicates that the Secretary will “consult the record” and review the language after the meeting.

In the case of Article IV discussions, the SU is submitted for review following the Board meeting, first to the relevant department and then to the executive director representing the country under discussion. Post-meeting changes are supposed to be limited to those needed to ensure the accuracy of the statements and to delete “market-sensitive” material in accordance with the Board-approved policy on deletions. Any amendments that change the policy message of the SU are supposed to be taken back to the Board for consideration.15 Once finalized, the SU is entered into the record of the meeting and issued as a formal Fund document. If the Board and relevant authorities agree, it is released publicly as part of a PIN.

Summings up pertaining to policy discussions are treated somewhat differently from those for Article IV discussions. This stems from change in practice introduced in 1999 resulting from the desire of a number of directors to be better able to ascertain the conformity of the SU with the content of Board discussions of policy issues. At that time, executive directors also requested that “to the extent possible, discussions on policy items should be concluded with a formal decision, rather than a summing up, and the draft decision should be clearly set forth in the relevant staff paper.”16 And while the SU for a policy discussion was still read out at the conclusion of the discussion for comments, it was subsequently circulated to directors by e-mail. Initially, executive directors were given two hours to submit to the Secretary (by e-mail, copying all other directors) written comments of a “material nature.”17 In June 2007, following a recommendation by the APC, the period for comment was extended to one business day. The Secretary attempts to incorporate comments received while strictly adhering to the record of the discussion. Once an amendment has been authorized by the Chair, the text is re-circulated, with no further changes allowed. If the Chair feels that some points require directors’ further consideration, the SU is brought back to the Board.

Box 1.The Sharp Increase in the Use of “Grays”

Executive Directors often circulate written statements in advance of Board discussions (so-called “grays”). The use of grays has increased significantly over the past decade, particularly since 2001 (see figure below). The number of grays per meeting has increased more than ten-fold since 1997, resulting in a more than eight-fold increase in the number of pages per discussion. On average per Board item, almost two-thirds of directors now issue grays, compared with fewer than one in twelve in 1997. The sharp increase in volume is the result of an effort on the part of executive directors to reduce the amount of time spent in Board meetings. At its January 2002 meeting, the Agenda and Procedures Committee (APC) noted that “the main role of grays had evolved since their introduction in 1987, from a vehicle for advance consideration of strongly held views to a means of saving Board time and increasing efficiency.” A similar motivation appears to have been behind the APC’s discussion in June 2004 of “Voluntary Best Practices for Choice Between Grays and Oral Statements and on the Character of Grays and Oral Statements.”

While the advance circulation of grays provides a basis on which staff can refine a draft SU, SEC needs adequate time and capacity to digest and synthesize what is often a large volume of text. Directors are encouraged to submit grays by noon the day before the relevant board meeting. However, this deadline was set before the sharp increase in the number and volume of grays, and has not changed.

According to staff, a majority of grays often arrive after the noon deadline, with a few arriving even on the day of the Board discussion. The responsibility for this does not rest solely with the authors of grays. Executive directors usually await the circulation of the “Buff” statement from the director who represents the country under discussion before sending out their grays. If the Buff is distributed late, the circulation of grays is delayed. Not only does this complicate the preparation of the draft SU, it makes it less likely that directors themselves will have read and absorbed one another’s views prior to the Board meeting.

Source: Secretary’s Department, IMF.

Legal Status of Stand-Alone Summaries

Formal votes are rare at the Fund, as suggested by Rule C-10 of the Fund’s Rules and Regulations which indicates that the Chairman ordinarily ascertains the “sense of the meeting”18 in lieu of taking a formal vote. Therefore, while SUs are intended to provide a clear and concise vehicle to communicate the main thrust of a Board discussion, SUs for policy discussions can also have legal weight. Specifically, “in cases where a decision was reached but no formal text was proposed for approval, the SU will, in its relevant part, be regarded as a record of the decision taken.”19 The “sense of the meeting” is deemed to have been accurately described when “the required voting majority would be very comfortably satisfied if there were to be a vote taken and all, or almost all, directors can go along with the majority view in the sense that they would not vote against it.”20 This formulation is intended to encourage the development of a view that as large a majority as possible can share.

Language in Summings Up: The Use of “Code Words”

For much of the Fund’s history the prevailing view was that there was a “great virtue in being deliberately vague in reporting on executive directors’ various positions…as long as the debate on an issue has not yet moved forward enough to allow the Chairman to take the sense of the meeting.”21 This was particularly the case for Board reports to the Interim Committee,22 where there was a desire to avoid limiting ministers’ room for maneuver in formulating their guidance to the Executive Board. “Constructive ambiguity” was part of the justification provided by the Fund Secretary for the use of “code words” (e.g., “some,” “many,” and “a number”) in lieu of precise numbers to summarize the views expressed during Board discussions.

Twenty-five years ago, after executive directors pressed for greater clarity on the methods used to summarize their views, the Secretary clarified the specific definitions of individual code words in a 1983 statement to the Board.23 Despite this, the meaning of the code words remains clouded by a number of factors, providing scope for discretion in characterizing the degree of support for a particular position. First, as confirmed during several Board discussions, these words are intended to reflect both the number and voting share of directors supporting a particular view, although there was little if any guidance on which factor should dominate. Second, the meaning of code words could change depending on the required majority for a particular decision, implying that the same code word could have different meanings within a single SU.24 Third, not all speakers take unequivocal or unconditional positions, and indications of flexibility (or reference to a diversity of views within a single constituency), could conceivably place a speaker in more than one group. Fourth, the list of code words is not exhaustive. For example, reference to “several” directors frequently appears in SUs but has not been defined. Finally, there is some overlap between the definitions with “a number” and “some” both referring to support by six directors.

Since the creation of this system there has been a pronounced shift at the IMF toward greater transparency, including a presumption in favor of publication for SUs (except for UFR discussions). As a result, current practice is outdated and undermines efforts at transparency, particularly with respect to external audiences for whom the code words have never been clearly defined.

Accuracy of Executive Board Summings Up

There are frequent debates within the Board on the extent to which a particular SU adequately reflects the views expressed by directors during a discussion.25 A 1998 ethnographic study of IMF documentation (Harper, 1998) concluded that, over time, staff preparing draft SUs had become adept at anticipating what would be acceptable to directors, and hence that significant changes to draft SUs during meetings were unnecessary. This conclusion is consistent with the fact that, throughout much of the 1990s, following the reading of the SU, directors rarely raised significant objections at the close of Board discussions. But since the late 1990s, perhaps reflecting the increased likelihood that SUs and concluding remarks would be published as part of a PIN,26 questions about the content or wording of an SU have become more frequent, particularly following policy discussions. That policy SUs tend to be more controversial than those prepared for country discussions is consistent with the results of a recent IEO survey of current and former members of the Executive Board in which respondents expressed a lower level of satisfaction with SUs from policy discussions than country discussions.

Accuracy of Summings Up for Article IV Consultations

The starting point for the preparation of an SU is the staff report on which the Board discussion will be based. In the case of an Article IV consultation report, the staff report concludes with a staff appraisal that summarizes the views and recommendations of staff. In principle, if directors fully agree with the views expressed in the staff report, there should be little substantive difference between the conclusions of the staff report and those in the SU. In practice, the views of directors often differ from those of staff and it is the former that should be reflected in the SU.

Do SUs accurately reflect the views of executive directors? This section attempts to answer this question for SUs of Article IV consultation discussions. A similar exercise for SUs of policy discussions follows.


A recent Article IV consultation for each of six countries was chosen for analysis: Albania, China, Liberia, Italy, Pakistan, and the United States (see Table 1). The countries were randomly selected from each region. The Article IV staff reports chosen were prepared over a relatively short period (November 2005 to February 2007) to minimize the likelihood that the sample would reflect a significant change in guidance to staff or work practices within SEC. Also noteworthy is the large number of interventions for these discussions (more than 90 percent of chairs intervened, on average).27 This is relevant because of the convention for Executive Board discussions that an executive director’s silence is “taken as support for the thrust of the staff appraisal and/or proposal.”28

Table 1.Summary of Article IV Case Study Characteristics
Number of Directors intervening (Excluding ED Representing Country)Number of GraysPages of GraysNumber of Late Grays1Voting Share Represented by Late Grays (in percent)Number of References to Minority Views in Draft SUNumber of References to Minority Views in Final SU
Albania—2006 Article IV Consultation and Reviews Under the PRGF and EFF1715312400
China, P.R.—2006 Article IV Consultation23237741407
Italy—2007 Article IV Consultation2323540010
Liberia—2006 Article IV Consultation and Staff Monitored Program1814282600
Pakistan—2005 Article IV Consultation, Discussion of EPA23225732824
United States—2006 Article IV Consultation212165310916

Defined as grays issued on the day of the discussion. This figure understates the number of late grays since, according to SEC guidelines, a gray is considered ate if it is issued after noon of the working day before the Board meeting at which the discussion occurs. However, data on time of issuance was not available and the analysis had to, instead, rely on the date of issue.

Defined as grays issued on the day of the discussion. This figure understates the number of late grays since, according to SEC guidelines, a gray is considered ate if it is issued after noon of the working day before the Board meeting at which the discussion occurs. However, data on time of issuance was not available and the analysis had to, instead, rely on the date of issue.

The draft SU for each Article IV consultation was compared first with the staff report and then with the final SU.29 Any substantive differences between the draft SU and final SU were compared with the minutes of the meeting at which the staff report was discussed (and with all gray statements submitted in advance) to determine the extent to which the any changes could be ascribed to the statements of directors.

Given the small number of cases chosen, patterns that emerge from this analysis cannot be considered conclusive. Nevertheless, the exercise provides several interesting insights into the SU process.

Results and Observations

All but the SU for Pakistan began with the statement that “directors supported the thrust of the staff appraisal.” There was a tendency, even when there was broad agreement between staff and directors on the appraisal in the staff report, for the diagnosis of problems and associated recommendations in the SU to be less candid than those in the staff report. The extent to which this was the case varied considerably from country to country, being most pronounced for China and Italy and least so for Liberia. To get a sense of the stage at which these changes were incorporated, the analysis compared the staff appraisal in the staff report with both the draft SU and the final SU.

From Staff Report to Draft Summing Up

In the draft SUs (DSUs) reviewed, there was a tendency for the diagnoses and advice to be less candid than those in the staff reports. There were multiple examples of this in the majority of the case studies, albeit with considerable variation in frequency and significance. This dilution of messages before receiving input from executive directors was particularly pronounced in the case of China. The DSU was most similar to the staff report in tone and substance for Italy and Liberia.

There were some noticeable patterns in the way messages were softened. For example, clear warnings about vulnerabilities were sometimes replaced with relatively bland statements. Precise language was often replaced with vague wording. Negative modifiers (e.g., “weak,” “uninviting,” “underdeveloped,” “vulnerable”) were frequently replaced with calls to “improve” or “strengthen” some aspect of policy, albeit from an undefined starting point or level. In many cases, important adjectives found in the staff report did not appear in the DSU.

It would be difficult for directors listening to an SU as read out at the end of an Article IV discussion, particularly those directors for whom English is not the first language, to undertake a thorough comparison between the language in the staff report and the SU and to identify small differences in wording. The consequences of this for institutional accountability and credibility are not negligible, particularly for those countries for which the Fund publishes the PIN but not the associated staff reports. This suggests a need to justify any substantive deviation from the language in the staff report, on the basis either of views that directors have expressed in the Board discussion itself or of exogenous circumstances (such as the severe earthquake that hit Pakistan after the staff report had been finalized).

From Draft to Final Summing Up—The Impact of Directors’ Comments

For all six discussions, most directors issued gray statements in advance.30 Almost 90 percent of these statements were issued at least the day before the Board discussion, giving staff at least the evening before the discussion to read them. Late grays (defined as those arriving on the day of the discussion) corresponded, on average, to just over 10 percent of the Fund’s voting shares, suggesting no obvious link between the voting weight of a chair and its tendency to issue statements late. As noted, the benchmark used to assess the impact of the views expressed by directors was the DSU prepared by the relevant area department and transmitted to SEC prior to the Board meeting. In comparing the DSU with the final SU, changes attributable to directors’ comments can at least be partially isolated.

The record for accurately adapting the DSU to directors’ comments was mixed, but on balance, directors’ comments did have a noticeable impact on the final SU, causing significant changes to the DSU for all but one of the countries surveyed (Liberia). The treatment of China’s exchange rate regime was among the clearest examples of directors influencing the final SU. In several cases, issues that were treated in the staff report but not in the DSU were inserted into the final SU (e.g., a reference to a “soft landing” in the U.S. Article IV; and a reference to the Extractive Industries Transparency Initiative for Liberia). In terms of direction, directors tended to be more positive than either the staff report or DSU.

Reporting of Minority Views in the Final Summing Up

The use of code words in the sample conformed to the 1983 Guidelines. In several instances, however, a view was ascribed to “directors” where it was supported only by a minority. The extent to which minority views were referenced at all differed markedly across countries. While the SU for the U.S. Article IV consultation made extensive reference to a diversity of views, the SUs for Albania, Italy, and Liberia made few if any such references.

The decision on whether or not to report minority views is subject to a degree of judgment and a number of important considerations. For example, in capturing the “sense of the meeting,” the Chair may frequently seek to find (or even encourage) a consensus view among executive directors.31 There is also a need to avoid overly long and convoluted SUs that describe a multiplicity of views on a single subject (particularly, given the recent adoption of word limits on SUs). However, on balance, there was scope to give greater attention to differences of opinion.32 This was particularly true for Italy’s Article IV discussion, where there was significant disagreement among directors and between directors and staff, particularly with respect to the treatment of competitiveness, but to which the SU did not refer, leaving the inaccurate impression of an uncontroversial assessment.

In reviewing directors’ statements for the discussions in the sample, there was often a lack of precision which made it difficult to understand their position on, or the extent of their support for, a particular point. Several statements contained instances of directors agreeing with a staff assessment or recommendation but following their agreement with major caveats, making it difficult to interpret the view expressed. Another practice that undermined the clarity of directors’ positions was the use of the phrase, “I note staff’s view that. . .,” which if taken at face value, implies neither support nor disagreement for staff’s view. It is possible that a neutral position was not intended.

Accuracy of Summings Up for Policy Discussions

SUs for policy discussions can be particularly controversial, particularly when they contain Board guidance or decisions on IMF policies. However, given the consensus-based culture of the Fund’s Executive Board, the Chair frequently seeks to avoid narrowly circumscribed positions that have the support of slim majorities. This could potentially require using somewhat imprecise language. But if this comes at the expense of clarity, it undermines the Board’s ability to provide effective guidance to staff and management. The need for clarity on decisions, and with respect to Board guidance to Fund staff, is evident in the results of the survey referred to above (Figure 1).

Figure 1.How Are Minority Views in Board Discussions Treated in Summings Up?

(In percent)

Source: IEO Survey.


The minutes from two policy discussions—“Review of the Ex Post Assessment of Issues Related to the Policy on Longer-Term Program Engagement” (LTPE)33 and “Precautionary Arrangements—Purposes and Performance”34—were reviewed to ascertain the extent to which the views of directors were taken on board in producing the SU, and the clarity of any decisions that were taken during those discussions. These policy papers were chosen because they were not on issues that required several meetings to resolve and because they involved issues that are not currently under active consideration by the Executive Board. Also, for consistency, both discussions were chaired by the same member of senior management. Both SUs were subsequently published. Given that only two SUs were reviewed, observations are only illustrative.

Results and Observations

In the case of the Board discussion of Ex Post Assessments (EPAs), the SU gave a more positive impression than either the staff report or directors’ comments. Criticism from directors was muted and generalized in the SU, while positive achievements described in the staff report were presented in greater detail. The SU failed to describe the diagnosis for several of the shortcomings in the Fund’s experience with EPAs cited by directors, and instead presented a “directional” policy remedy (a similar pattern to that observed for Article IV SUs). For example, whereas the staff report noted that “in most cases, critical aspects of program design were covered to a limited degree,” the SU stated that “the value of EPA reports could be enhanced by better selectivity and focus on a few critical issues.”

Particularly noticeable was the failure of the SU to reflect the widely held and extensively discussed concern among directors with the lack of independence of EPA mission chiefs. Only the resulting recommendation was cited in the SU (i.e., “most directors” considered that EPAs “should be led by a mission chief from a department other than the home area departments”). Also, the SU contained no mention of the fact—contained in the staff report and cited by a number of directors—that a significant share of EPA mission leaders had indicated that they felt pressured to change key messages in their reports.

The SU made very few references to dissenting opinions. The most obvious exception (cited in the previous paragraph) was in response to the suggestion of staff that there be flexibility to combine EPAs with Article IV consultations (i.e., “most” directors argued that the EPA team “should be led by a mission chief from a department other than the home area departments”). However, in contrast to the rest of the SU, only “a number of directors” wanted to give discretion to area departments and management to decide whether to prepare stand-alone EPA reports. The only other “minority view” reported was in opposition to the proposal not to include undisbursed precautionary arrangements in the definition of longer-term program engagement. Consistent with established understandings, the phrase “the view was expressed” was used to indicate that this view was held by a single large chair alone.

In this SU, the guidance given was clear (e.g., “directors agreed that . . . the definition of LTPE should be unified for PRGF and GRA users),” with a minor exception. On how to treat time spent under precautionary arrangements that remained undrawn, the SU indicated that “directors broadly agreed. . . .” The insertion of the word “broadly” could reflect the opposition of a single director but, in conveying the result of the decision (which was agreement by the Board), it was unnecessary and undermines the clarity of the guidance.

Unlike the EPA discussion, the SU from the discussion of Precautionary Arrangements (PA) extensively reported minority views (making 16 separate references in three pages of text). To some extent, this may reflect the fact that the “Issues for Discussion” section of the PA staff report posed relatively few direct questions to directors, making the achievement of consensus less critical.35 However, the abundance of minority views on a range of issues reported in this SU created a somewhat disjointed picture, making it difficult to derive a sense of where the debate stood on the impact and usefulness of precautionary arrangements. The SU reads like a “grocery list” of opinions on individual issues, without an easily extractable perspective on the more controversial and important considerations. Also, some of the views ascribed to only sub-sets of directors were in reference to statements of fact, which implies a lack of confidence in the information provided in the SR. This does not appear to have been intended.

Improving the Summing Up Process

In June 2007, the Board’s Agenda and Procedures Committee (APC) held a discussion on improving the process for formulating and finalizing SUs to ensure they accurately reflect directors’ comments. While this discussion concerned only the SUs for policy items, it yielded a number of useful suggestions, including to extend the period for executive directors to review policy SUs. The discussion focused on what staff could do to improve the accuracy of SUs, but given the occasional lack of clarity in many of the directors’ statements, there is much that executive directors could do to facilitate the production of clear and accurate SUs.

For example, interviews with staff indicated that there is often a serious time constraint in sifting through dozens of pages of text the night before a Board meeting in an effort to reflect the views of directors in a draft SU. This task has become more difficult since 2000 with the very large increase in the volume of grays, despite calls for shorter statements that highlight key points. Much of the volume comes from a repetition of arguments made in the staff report or from the description of known facts and statistics. Since the convention at the Board is that an executive director’s silence implies support for the position of staff, much of this text is unnecessary for the SU.

While the increase in the use of “grays” has shortened the amount of time spent in the Board, this may have come at a cost. At the APC’s June 2004 discussion of “Voluntary Best Practices for Choice Between Grays and Oral Statements and on the Character of Grays and Oral Statements,” several directors expressed concern with the proliferation of grays, arguing that the purpose of grays “should be to improve the efficiency and quality of the Board’s work, and not necessarily to reduce time spent in Board sessions. The Board’s goal should always be to reach good decisions through an inclusive and fair process, and to allow minority views to be expressed” [emphasis added].36 These concerns appear to have had some validity as the large volume of grays received in the 24 hours prior to the Board meeting has likely had a negative impact on the quality of SUs, which are a critical part of the Board’s decision-making process.

Directors and their staff can facilitate the production of SUs that better reflect their views by making their statements clearer and more focused. Closer adherence to existing Board guidelines on Article IV interventions and grays—including greater reliance on association with the views expressed by other directors—would also help. The deadlines for submitting Buffs and grays should be brought forward to 72 and 48 hours in advance of the Board discussion (respectively).

The definition of individual code words should be made public given that a large percentage of SUs are now published. The frequent use of “several,” which is not defined, should be avoided, or it should be defined, and overlap between “some” and “a few” should be eliminated. The code words could also be recalibrated to reflect the increase in the size of the Board since their adoption.

SUs should be clearer when a position is supported by a sufficient number of directors to represent a “Board” view. When this is the case, the position should be attributed—without qualification—to “the Board,” rather than a sub-group of Directors, while not ignoring the expression of minority views.

Vague language should be avoided unless explicitly warranted by directors’ comments. The starting point for drafting a SU should be the language in the staff report, since this is the document on which directors base their comments at Board meetings. While it is reasonable for staff to attempt to anticipate the range of views that directors might express, this should not manifest itself as a watering down or clouding of language in an attempt to avoid controversy or disagreements among directors.


    GianvitiF.1999Decision Making in the International Monetary Fund,” inCurrent Developments in Monetary and Financial Law Vol. 1 (Washington: International Monetary Fund).

    HarperRichard H. R.1998Inside the IMF: An Ethnography of Documents Technology and Organizational Action (Cambridge, United Kingdom: Cambridge Academic Press) pp. 25558.

    International Monetary Fund1993“Executive Board Work Procedures,” SM/93/18,January25 (Washington).

    International Monetary Fund2001“Guidelines for Efficient and Effective Interventions in Article IV Discussions,”January8 (Washington).

    International Monetary Fund2005“Compendium of Executive Board Work Procedures,”December (Washington).

    Van HoutvenLeo2002“Consensus Decision Making in a Cooperative Institution,”in Governance of the IMF: Decision Making Institutional Oversight Transparency and Accountability IMF Pamphlet Series No. 53 (Washington: International Monetary Fund).

    World Bank2002“Board Procedures and Practices,” Part A-III of Executive Directors’ Handbook (Washington).

    World Bank2005“Board Effectiveness: Making Board Decision-Making Processes More Efficient”October27 (Washington).

The author would like to thank Thomas Bernes, Leonardo Martinez-Diaz, Alexander Shakow, Alexander Mountford, Nils Bjorksten, Joanne Salop, and Iqbal Zaidi for helpful comments and suggestions. The author is also grateful to Roxana Pedraglio, Alisa Abrams, and Borislava Mircheva for research assistance, Annette Canizares, Arun Bhatnagar and Jeanette Abellera for administrative assistance and Rachel Weaving for insightful editorial suggestions. Remaining errors are solely the responsibility of the author.

IMF Rules and Regulations, Rule C-14, adopted September 25, 1946.

Minutes of Meeting No. 177 of the Executive Board, June 10, 1947.

Minutes of the Thirteenth Meeting of the Executive Board, May 29, 1946, Item 3.

Rules of Procedure for Meetings of Executive directors,” adopted at the fifth meeting of the IMF Executive Board, May 13, 1946.

Draft minutes are now only required to be submitted for approval “within a reasonable period of time.” In practice, directors generally receive minutes for review between 10 days and two weeks after a meeting.

See, for example, “Valuation and Yield of the SDR, Summing Up and Comments by the Managing Director,” Executive Board Meeting 73/120, December 19, 1973.

The first stand-alone summary for a bilateral surveillance discussion was for the 1978 Article IV Consultation with the United Kingdom.

Publication was not envisaged at this time; it was only in the late 1990s that the Board began to permit the issuance of PINs based on the SU.

IMF Legal Department’s memorandum on the meaning of silence, April 16, 1987

IMF Secretary’s Department, “Compendium of Executive Board Work Procedures,” Section 6(e).

“Summings up in the Context of the Use of Fund Resources,” SM/99/48, February 23, 1999.

To emphasize that concluding remarks do not represent a decision of the Executive Board, the following disclaimer is placed on the first page of all concluding remarks: “These concluding remarks do not reflect decisions of the Executive Board but rather preliminary views expressed by executive directors in a discussion conducted in seminar format.” See Compendium of Executive Board Work Procedures.

A good example of this is “Quotas and Voice—A Possible Package of Reforms: Chairman’s Concluding Remarks,” August 23, 2006.

The Policy Development and Review Department (PDR) used to review draft Article IV SUs but the practice was discontinued earlier this decade.

According to IMF staff, there have been very few cases of this.

Summings up for Policy Items—New Procedures,” Memorandum from the Acting Secretary to Members of the Executive Board, May 12, 1999.

Accommodation was made to extend the period to four hours when the two-hour period coincided with a Board meeting or when the Board agenda was particularly heavy.

Initially, the concept used was “consensus,” defined by the Board as “a position by a majority, but not by all directors” (Minutes of Meeting No. 123, January 21, 1947). The Board subsequently decided to use “sense of the meeting” instead of “consensus.” This was defined as the “position supported by Executive directors having sufficient votes to carry the question if a vote were taken” (Minutes of Meeting No. 173, May 28, 1947).

Gianviti (1999). The General Counsel later clarified that for an “understanding” in a SU to constitute a decision, there must be a willingness on the part of the Board for the understanding to have legal effect without further Board action (which implies that a statement of intent in an SU, including a commitment not to take a particular action, cannot constitute a formal decision). Second, the requisite majority must support the “understanding” which, of necessity, must be consistent with the Articles of Agreement. (BUFF/00/169, November 17, 2000).

Compendium of Executive Board Procedures, Section 6(c), March 2007.

“The Definition of Code Words: Statement by the Secretary,” Executive Board Meeting 83/11, extract from EBM 83/11, January 12, 1983 on the Eighth General Review of Quotas–Draft Report to the Interim Committee.

Precursor to the International Monetary and Financial Committee (IMFC).

“A few” = 2 to 4 directors; “some”= 5 to 6 directors; “a number”= 6 to 9 directors; “many”= 10 to 15 directors; “most”= 15 or more directors, “nearly all”= about 20 directors, and “the view is held that”= the view of the United States. (See “The Definition of Code Words: Statement by the Secretary,” Executive Board Meeting 83/11, extract from EBM 83/11, January 12, 1983 on the Eighth General Review of Quotas—Draft Report to the Interim Committee. When the definitions were assigned, there were 22 directors on the Board. The size of the Board has since increased to 24 directors.)

For example, according to the Secretary’s 1983 statement, “while ‘many’ may perhaps be appropriate to describe the sense of the meeting on a matter requiring only a simple majority of views expressed, it would not suffice for an issue requiring a majority of 85 percent of the voting power of the Fund.”

In a recent IEO survey, only 16 percent of the Executive Board and 25 percent of senior IMF staff indicated that SUs “provided clear direction to staff and management.” Of the remainder of respondents, 71 percent of the executive directors, and 48 percent of IMF staff considered SUs to “sometimes” be vague and/or contradictory; 11 percent of the executive directors and 23 percent of staff described them as “often vague and/or contradictory.”

The percentage of Article IV consultations for which PINs are issued has risen steadily since the Fund began issuing PINs in 1997, and reached 94 percent of the 125 Article IV consultation discussions held in 2006.

This excludes the executive director representing the country under discussion, whose views are not reflected in the SU.

As noted earlier, the executive director representing the country under discussion is given the opportunity to review the SU before its finalization and to recommend the deletion of “market-sensitive” material. Our analysis did not attempt to assess the extent to which changes were made at this stage of the process.

Twenty out of 23 Executive directors, on average, amounting to an average of 52 pages of grays per discussion. This compares with an average for all discussions in 2006 of 16.7 grays consisting of 37.5 pages in total. The director representing the country under discussion does not issue a “gray,” and his/her views are not taken into account in preparing the SU.

This is less important than in the context of Board discussions (e.g., over policy issues) when executive directors are asked to provide guidance to the institution.

This conclusion is consistent with the findings of a recent IEO survey of the Executive Board and senior IMF staff which found that almost half of Executive Board respondents believed that attention to minority views in SUs was insufficient compared with only 26 percent for IMF staff (Figure 1).

SM/06/115, discussed May 15, 2006.

SM/06/120, discussed May 17, 2006.

The staff paper for the EPA discussion explicitly requested the guidance of Executive directors on at least 11 separate issues; with respect to the discussion of PA, guidance was explicitly sought on only three.

“Summary Record of the Agenda and Procedures Committee,” Meeting 04/2, June 15, 2004. Nevertheless, the balance of view among directors at this meeting was that “Grays had become a valuable tool for improving the efficiency of Board meetings, especially when directors wanted their views on record but did not necessarily want to use Board time to express their opinions.”

    Other Resources Citing This Publication