Information about Asia and the Pacific Asia y el Pacífico
IEO Evaluation of Exchange Rate Policy

Chapter 4: The IMF’s Policy Dialogue and the Impact of Its Advice

International Monetary Fund. Independent Evaluation Office
Published Date:
August 2007
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details

How Effective Was the IMF’s Policy Dialogue with Country Authorities?

42. Policy dialogue between staff and the authorities is a key part of the IMF’s surveillance process. It is an input into the formulation of staff views and a channel of transmission of the resulting policy advice to the authorities, as well as of the authorities’ views to the Executive Board.1 While the official IMF view on a country’s policies is formulated only after discussion by the Executive Board and then transmitted to the authorities through the summing up of that discussion, staff views as endorsed or altered by management are earlier communicated in-country during policy discussions with the authorities.2 Significant changes of these positions at the Board level, though possible, are infrequent, limiting the impact of Executive Board views.3 The effectiveness of the staff’s dialogue with the authorities is thus of central importance for—though not synonymous with—the impact of the IMF’s surveillance activities (Figure 4.1).

Figure 4.1.How IMF Views Connect to Outcomes: A Closer Look

43. The effectiveness of the dialogue between staff and the authorities can be judged by the extent to which it is genuinely two-way, open, and substantive, adds to the understanding on both sides, and—where rel-evant—surfaces or addresses issues that need to be followed up in a bilateral or multilateral context. Where no immediate action is called for, and both sides essentially agree, the effectiveness of the dialogue with an individual country arises from the periodic focus given by both sides to the issues discussed, and a shared sense of exploring “what’s new? what’s missing? what if?” Deviations from this ideal could take various forms, including a one-way “lecture” from the IMF (or a perception that this is the case); a lack of sharing/openness by the authorities (of either data or views); set positions presented by both sides in a staid fashion (e.g., with staff unwilling to think outside the box); or simply an implicit mutual pact not to mention the exchange rate or consider contingencies.

44. Evidence from surveys and interviews suggests that, while the dialogue with authorities is considered satisfactory in many cases, there are nonetheless important questions about its effectiveness. Survey responses indicated that the large majority of the authorities generally perceived their discussions as two-way, with staff being seen as both respectful and willing to approach discussions with candor. Similar majorities approved of the frequency of the discussions and their balance between informality, confidentiality, and the reporting requirements to the IMF Executive Board. Staff share these views. Survey responses differed across country groups, with interviews pointing to difficult relationships between staff and the authorities in individual country cases. In interviews, while not all country officials were satisfied with the basis underlying staff advice, most nevertheless appreciated the opportunity to interact, even when the discussions did not change their views.4 Two factors were mentioned in this context: (1) IMF endorsement of certain policies can support decision making within countries by helping to overcome differences in views among different branches of government; and (2) IMF staff can serve as a “sounding board” for policy views, helping the authorities to challenge their own thinking.5

45. However, while there were few obviously negative experiences, the bigger issue appeared to be the tepid enthusiasm expressed privately in several countries. Outward appearances of respectful and polite to-and-fro therefore concealed the authorities’ frustration at a lack of deeper engagement and relevance. Importantly, survey results showed authorities seeking more specific analysis and pointing to other sources of policy advice as useful. These are warning signs that the IMF is seen by some as providing limited value added (and not just in advanced economies, but in other countries too), and that it needs to find a way to reenergize its contribution to members’ ongoing policy discussions.

  • As regards possible shortcomings on the staff side, country officials mentioned several issues, including the frequent changes in mission chiefs, inadequate knowledge of country-specific background and context, and less technical knowledge of the operational aspects of foreign exchange markets than enjoyed by the authorities themselves—a finding that is supported by survey responses. Officials would have welcomed staff having greater familiarity with the experience of other countries, an aspect in which IMF staff should have been expected to have a comparative advantage. In several economies, officials said they would have welcomed more written material to facilitate internal communication of IMF advice.
  • Some authorities admitted to withholding relevant data from the IMF or to excluding sensitive topics from discussion, while staff saw this as a more widespread problem. While the adequacy of data availability is a broader issue that affects analysis, as well as the dialogue (see Box 3.3), it is important to highlight that, in some cases, according to the staff, authorities were unwilling to share relevant data, or excluded certain topics from discussion. These issues are relatively rarely flagged in staff reports, or taken up with management, as forth-rightly as appears warranted. Staff appeared reluctant to risk antagonizing the authorities; and the main reason given by the authorities for not sharing data and for avoiding certain issues, was a concern that information would be passed on, either to the Executive Board or through publication.6

46. Authorities made suggestions on how IMF exchange-rate-related discussions could be improved. There were some notable differences between the views of surveyed respondents who found themselves in broad agreement with IMF advice, and of those who did not. Reasons for a failure to have impact include, in particular, a lack of attention to country specifics, insufficient analytical underpinnings, and an undue sense of urgency on the part of staff—a finding consistent with the IEO’s desk review of 30 countries, particularly in the context of the choice of exchange rate regime (Figure 4.2). The authorities, in turn, have contributed to this state of affairs to the extent that sensitive policy issues have been taken “off the table,” as suggested by the correlation between such instances and different degrees of IMF impact on policy decisions made by the authorities (Table 4.1).

Figure 4.2.Areas of Possible Improvement in IMF Exchange Rate Advice, as Identified by Country Authorities

Proportion of respondents indicating scope for improvement in IMF analysis from country authorities that had agreed/disagreed with IMF advice

(In percent)
Table 4.1.Survey of Authorities: Limits to IMF Impact
Impact of IMF Advice1 onNature of Discussions Between IMF Staff and the Authorities2
Major DecisionsRestrictedIntermediateUnrestrictedTotal
Instrumental7 (29)4 (40)25 (47)36 (41)
Marginal7 (29)6 (60)19 (36)32 (37)
No impact5 (21)3 (6)8 (9)
Limited or no discussions with IMF5 (21)6 (11)11 (13)
Total24 (100)10 (100)53 (100)87 (100)

As judged by the authorities.

Number of respondents (percentages in parentheses); based on authorities’ answers to the survey questions of whether they had “at times excluded certain sensitive policy issues (e.g., foreign exchange market intervention, choice of exchange rate regime) from substantive discussions with IMF staff,” and whether they had “excluded or restrained consideration of certain issues because of concerns about possible dissemination of information.”

As judged by the authorities.

Number of respondents (percentages in parentheses); based on authorities’ answers to the survey questions of whether they had “at times excluded certain sensitive policy issues (e.g., foreign exchange market intervention, choice of exchange rate regime) from substantive discussions with IMF staff,” and whether they had “excluded or restrained consideration of certain issues because of concerns about possible dissemination of information.”

47. Some country episodes called for intense involvement of the staff and management in discussions with authorities, both inside and outside the regular Article IV consultation process. The complexity of such interaction varies according to circumstance. For example, a very rapid response by IMF staff and management is required to contribute to advice in the event of exchange market turbulence. During the two most recent episodes of coordinated intervention among G-7 economies (in 1998 and 2000), the IMF did have views, but the extent to which it expressed them differed (see Box 4.1). In other circumstances, by contrast, an effective dialogue with member countries requires a long-term strategic response by IMF staff and management, involving sustained contacts with country authorities outside the regular Article IV consultations, over several years, attuned to the pace and complexity of the decision-making process.

Box 4.1.Currency Interventions of June 1998 and September 2000

On June 17, 1998, the U.S. and Japanese monetary authorities cooperated in intervening in the foreign exchange markets to support the yen, which had been in an accelerating decline over previous weeks. On September 22, 2000, the European Central Bank (ECB) together with the monetary authorities of the United States, Japan, Canada, and the United Kingdom intervened in the foreign exchange markets to support the euro, which had been in accelerating decline over the preceding weeks, and this intervention was followed in early November by two episodes of unilateral intervention by the ECB. Both episodes of concerted intervention were carried out primarily for signaling purposes and both were moderately successful when judged by subsequent developments.

Did the IMF have a view on the relevant exchange rate developments at the time and the possible actions to be taken? Did it communicate these views to the relevant authorities? And did its views have any impact? In both cases the IMF staff did have a view on exchange rate developments and possible policy measures including the (limited) value of intervention; and it had opportunities to communicate this view to the relevant authorities in the course of Article IV discussions, and at G-7 preparatory meetings. Analysis of staff documents and published accounts, and discussions with IMF officials involved at the time suggest rather different answers to the questions about communication and impact in the two cases.

The June 1998 intervention coincided with annual Article IV discussions with the United States and Japan, and staff appear to have taken the opportunity to discuss the option with both, and to encourage action—which at the time was also seen as important to continued recovery from the 1997 Asian crises. Senior management, apparently, had confidential discussions with both the United States and Japan, separate from the Article IV discussions. Board documents did reveal some discussion in the course of the Article IV consultations, though these documents were not circulated to the Board until after the intervention had taken place.

In 2000, IMF staff had the opportunity to discuss the case for and against intervention with ECB and other euro area officials during a mission to examine “Monetary and Exchange Rate Policies of the Euro Area,” which took place in late June/early July 2000, and to present any advice in the context of their findings at meetings of euro area ministers and officials in mid-July. It seems that they did not do so, possibly in part because at the time they did not see a strong case for action, and in part because they felt this was a matter for the ECB rather than ministers—even though ministers had in fact discussed the option in May and issued a statement designed to encourage it. By September, following further weakening of the euro, the IMF had reached a clearer view on the value of intervention—the World Economic Outlook released in the run up to the September Annual Meetings described the euro as “significantly misaligned”; and at his press conference on September 19, 2000 the IMF Economic Counsellor said, “I think the circumstances in which the major countries would want to use intervention to attempt to influence exchange rates are relatively rare, but they do arise from time to time and one would sort of have to ask if not now, when?” It is clear from subsequent accounts that by that point, following extensive consultations in the G-7 including at a G-7 preparatory meeting on September 13, the decision had been taken to intervene. It is also clear from these accounts that the IMF played no part in these discussions, and was probably unaware of them.

What Are Other Channels of Impact for IMF Advice?

48. There are several channels in addition to the dialogue with authorities through which IMF surveillance may help to influence policy formulation in member countries (see Figure 4.1): (1) a variant of policy dialogue between staff and authorities (discussed above, which is normally thought of in a bilateral context) are efforts at international policy coordination; (2) the influence of the subsequent Executive Board discussion (including peer pressure from other governments based on the account of the discussions between IMF staff and the authorities); and (3) IMF contributions to domestic policy influences, including market perceptions. The relative importance of these channels varies by country and context, but a few general patterns emerge.

49. The influence of the IMF Executive Board discussions differed according to country grouping. Survey respondents representing the smaller emerging market and developing countries agreed by a 6–1 margin that considerations at the level of the IMF’s Executive Board had provided an important input into the development of policy, with agreement particularly pronounced among those countries that rely on the Board’s approval in the context of IMF-supported programs. In contrast, a majority of respondents from the other country groups disagreed that this had been the case—a perception supported by some 40 percent of staff respondents (see Figure 4.3). Interviews with authorities provided further support for skepticism about the direct role of the Executive Board. No direct evidence was found of peer pressure from other authorities as a result of IMF advice.

Figure 4.3.Authorities’ Views on Policy Inputs Provided by the Executive Board

Proportion of respondents agreeing/disagreeing that the IMF Executive Board had provided important policy inputs

(In percent)

1 IMF-supported program concurrent with at least one of the last two Article IV consultations through 2005.

50. There was some evidence that the Executive Board had indirect influence, which arose, for exam-ple, from staff presenting views that are likely to be endorsed by the Executive Board.7 However, more than 40 percent of staff respondents in the IEO suvey felt that the expectation of publication or the need to preserve close relationships with country authorities tended to dilute coverage of exchange rate policy issues in staff reports. These factors suggest that the influence that can reasonably be exercised by the Board is limited. The opportunity for all member countries to express their opinions through the Executive Board on the policies of individual members is a key part of the governance of the IMF. However, evidence that important information is not conveyed to the Board—or not conveyed to the staff, because it might be conveyed to the Board—raises questions about the accountability of staff in these circumstances, as well as effectiveness of the present setup.8

51. Use of public channels of IMF influence has also been limited in the case of exchange rate policy advice, for obvious reasons. The IMF’s increasing use of public channels to enhance the effectiveness of its surveillance is an important topic, but beyond the scope of this evaluation. Unsurprisingly, many authorities have mixed feelings about the IMF making tactical use of the domestic policy discussion channel to influence exchange rate policy. In countries where the exchange rate is a highly sensitive political issue, authorities may not agree to publication of an Article IV consultation report that delves deeply into exchange rate issues. The option is also available to seek deletions of highly market sensitive information from published Article IV staff reports.9 While generally welcoming IMF staff analysis of topical exchange rate questions, authorities have often been wary of sparking a public debate that might unsettle markets and reduce authorities’ control over the pace and sequencing of reforms. Many authorities nevertheless welcome the opportunity to publicize any IMF endorsement of domestic policies, mindful that endorsement of a country’s exchange rate regime or economic policies can enhance policy credibility and facilitate access to capital markets.10 Other than through this effect, however, the channel linking IMF exchange rate advice to market players appears to be weak—partly because of the IMF’s general caution in dealing with market sensitive information.11

52. There were no significant examples of exchange rate policy coordination during the period covered by this evaluation; 12 the question is whether the IMF missed certain opportunities to have greater impact. Past episodes of (regional) collective action by member countries (see Box 4.2) suggest that IMF contributions to coordinated policy responses have varied greatly, depending on factors including the degree of staff expertise, the extent to which the countries concerned rely on IMF support, and the manner of involvement of the Managing Director (which was clearly a factor in the 1994 CFA franc devaluation). A key question, against this backdrop, is whether the IMF should and could have done more and earlier to facilitate the analysis and resolution of global imbalances. As already noted (Box 3.4), integration of multilateral considerations—and, specifically, concerns about global imbalances—into country-level analysis was often lacking. Partly as a result, messages conveyed to key countries in the course of bilateral consultations were given insufficient emphasis, with limited management follow-up at higher political levels. Scope for more active IMF engagement at an early stage in seeking to promote collective policy action was thus not exploited effectively. This may have contributed to delays in policy action, and it began to hurt credibility as the IMF was subsequently perceived as reacting to outside pressures.

Box 4.2.IMF Involvement in Selected Episodes of Exchange Rate Policy Coordination

The IEO examined the extent and nature of the IMF’s involvement in two major exchange rate events in the early 1990s: the crises in the European Exchange Rate Mechanism (ERM) in 1992–93; and the CFA franc devaluation agreed in January 1994. Examination of internal documents of the time was supplemented by discussion with some of those in the IMF and country officials who were most closely involved with these two episodes.

1992–93 crises in the ERM

Most retrospective analyses of the events of 1992–93 draw similar conclusions. With today’s global financial markets, systems of fixed but adjustable exchange rates (short of monetary union) between internationally traded currencies can provide markets with effective one-way bets when under strain. The ERM in 1992–93 suffered major strains resulting from an imbalance in fiscal and monetary policy in Germany, the anchor currency country, following unification. Other ERM countries were forced to hold interest rates higher than warranted by domestic circumstances to defend their central rates. When markets started to doubt the political sustainability of such action, the game was up. In September 1992, sterling and the lira were forced to exit from the arrangement and, at the end of July 1993, the ERM countries were forced to agree to the adoption of widened, 15 percent margins—marking the end of the ERM, as originally set up.

The IMF’s role in these events was limited. Staff did not attend any of the critical meetings of the European Monetary Committee or European finance ministers, and were not invited to do so or to offer written advice. The IMF did of course have opportunities to develop its analysis and communicate views and policy suggestions, both in the course of regular Article IV discussions with member countries and through ad hoc management contacts. All the evidence is that the IMF’s analysis at the time was partial, missing the financial market dimension, and that such messages as were passed were limited to suggestions about policy actions that would help countries live within the system. Staff did, in 1992, identify the major underlying policy source of strain in the ERM, but they did not appear to recognize the market dynamics creating the fundamental vulnerability of the mechanism until after the forced widening of the ERM bands in 1993. Some confidential high-level messages were passed, notably to the German authorities in the summer of 1993, about policy changes that might ease the underlying tensions, but they were not well received. There was no staff or management discussion with European officials of options for handling crises should they occur.

After the event some useful lessons were drawn for the future. Staff now regularly interacts with key groups of European and euro group officials and ministers. A number of steps were taken to strengthen the IMF’s knowledge of and links with private financial markets, and its work on equilibrium exchange rates was beefed up by starting CGER. Management also instituted regular internal meetings of a “Surveillance Committee” (meetings of which continued for the rest of the decade) to discuss global exchange rate developments, and the Board instituted its continuing regular discussions of World Economic and Financial Market Developments.

1994 CFA franc devaluation

From the 1980s, a loss of competitiveness became increasingly evident for the members of the two CFA zones. Some time in 1990 or 1991, IMF staff and management concluded that a substantial devaluation of the CFA franc was needed. There followed three years of quiet diplomacy and persuasion before a 50 percent devaluation was finally announced in January 1994. During this period, while IMF staff and management were extremely active, almost no signal was given to the Board of what was afoot. Several factors led the IMF to adopt this approach. First, it was the IMF’s view that the CFA franc zone arrangements had, on balance, served its members well and there was no wish to disband the system—implying a one-off devaluation. Second, it was felt that there would only be one chance to get it right and that the devaluation would therefore have to be substantial. Third, nothing could be achieved without convincing all the member countries and also the French authorities, and the Managing Director judged that this would take time and could be done only in private; and fourth, there was the need to avoid the damaging capital outflows that rumors would induce.

In the event, the negotiations were long and complex with several setbacks before agreement was eventually reached. While some member countries agreed on the need to devalue at an early stage, others, including those with smaller losses of competitiveness, were more reluctant and for a while persuaded their colleagues that adjustment could be achieved with internal measures alone. Governments were also concerned about the impact of a devaluation on living standards, and it was an innovative feature of the IMF-supported programs eventually put in place that they contained measures to protect the poor from the worst impact of devaluation. One advantage of the time lag was that IMF (and World Bank) staff were able to agree to support detailed programs for most members, which were announced very quickly after January 1994. The Managing Director played a key role, both with CFA zone country leaders and in persuading the French authorities and senior politicians. The analytical work that preceded the operation had extended over a long period of time, with the analysis done by the IMF matched by parallel work by the Bank on the “real” economy.

The IMF’s role in securing exchange rate policy action in this case was central. The task was exceptionally complex, because of the number of parties involved (13 member countries plus France). One conclusion is that the effective role played by the Managing Director was critical in securing agreement. Another conclusion is that, particularly in the final run-up to devaluation and in the phase after the event, cooperation between the two Bretton Woods institutions, the African governments concerned, and the French government was excellent, and this was an important factor in the success of the operation. A final conclusion is that, in today’s circumstances of more open and efficient capital markets, action might well have been forced sooner, with less time for preparation.


53. Improvements in the effectiveness of dialogue and other channels to maximize the impact of IMF advice, including policy coordination efforts, require strategic planning and intense management focus. The episodes considered in this evaluation suggest that the following elements are needed: formal planning of strategic focus, both bilaterally, and multilaterally including proactive involvement in the various fora of country officials; assembling the right staff expertise and analysis in advance, and integrating the best advice from inside and outside the IMF; and a clear sense of whom to talk to among the authorities, and how to convey the message most effectively.


See Background Document 7 for details.


The concluding statement of the mission does not receive input from the Executive Board. In some countries, a media event at the end of the staff visit, at which the concluding statement might be released, takes place up to three months before the Executive Board discusses the staff report and provides the formal IMF view.


A review of Executive Board minutes (EBMs) for the IEO’s 30 in-depth sample economies suggests that changes made by the Executive Board to staff positions largely relate to issues of urgency and sequencing, with Directors generally favoring a more cautious approach to implementing exchange-rate-related policy advice.


With regard to discussions between staff and authorities, 70 percent of the country authority respondents indicated that—in areas that had been a focus of policy attention—the authorities and the IMF agreed on the analysis (suggesting either that IMF staff were convinced by the authorities; that IMF advice—to the extent it was given—convinced the authorities; or that the authorities did not need any further convincing to pursue a particular course of action). Staff respondents also reported widespread agreement, but noted a greater level of disagreement on important details such as emphasis, timing, or political feasibility.


See Chapter 2, section on “Perceptions of Country Authorities and IMF Staff” for survey evidence.


In interviews, another concern voiced was that information conveyed might be passed to the markets; and relatedly, that some senior staff members have left the IMF and subsequently taken jobs in the private sector—something that reduced the trust the authorities would have in the IMF as confidential advisor.


Some support for this view has emerged in interviews with mission chiefs to selected countries, who also described specific cases where Executive Directors for those countries had accompanied missions and helped to resolve disagreements between staff and authorities, in part by communicating to authorities what the sense of the Executive Board was likely to be on the issues of contention.


One, often overlooked, indirect channel of influence is that staff take note of Board discussions and try to reflect these in subsequent discussions with the authorities of the same or another country. Article IV consultations for Japan are a case in point. After complaints by individual Executive Directors in 2003 and 2004, the 2005 staff report did contain an analysis of the intervention episode in 2003/04.


In the majority of country cases reviewed, little or no material change related to exchange rate issues was made to the public versions of recent staff reports by way of deletions and corrections.


Most country authorities’ survey respondents agreed that the IMF had appropriately played its role as provider of credibility (see Chapter 2).


Market participants interviewed by the IEO suggested that the IMF has a constructive role to play in causing more and better data to be made public. Strong statements by the IMF were welcomed by some, but views differed on when and under which circumstances IMF views would influence markets.


As noted earlier, the evaluation does not cover the multilateral consultation announced in late 2005.

    Other Resources Citing This Publication