Journal Issue
Share
Article

Safeguards Assessments - Review of Experience

Author(s):
International Monetary Fund
Published Date:
November 2010
Share
  • ShareShare
Show Summary Details

I. Introduction and Background

1. This paper reports on the experience of the safeguards policy since 2005, when the policy was last reviewed. In keeping with previous reviews, a panel of experts (the panel) was invited to provide advice for the Executive Board on the implementation of the policy, including through a review of safeguards reports (which are not circulated to the Executive Board).1 The panel’s report is being circulated concurrently to the Executive Board with this paper and the Chair of the panel is expected to attend the Executive Board discussion.

2. The safeguards policy (Box 1) was first introduced on a two-year trial basis in 2000 in the wake of allegations of misuse of IMF resources by recipient countries. In 2002 the Executive Board declared the experimental phase an “unqualified success” and safeguards assessments of central banks became an integral part of most types of Fund financial support. The Executive Board’s 2005 review, which involved a panel of experts that surveyed the views of central banks, re-affirmed the effectiveness of the policy in helping to prevent misuse of IMF resources and reduce misreporting. The key outcomes of the 2005 policy review were as follows:

  • Communications with central banks on safeguards issues should be improved, and the modalities for assessing and monitoring safeguards at central banks should remain flexible.
  • The existing scope of the policy, which does not cover fiscal activities, the framework for assessing central bank operations, and the deadline of the first program review for the completion of a safeguards assessment remained appropriate.
  • The confidentiality of safeguards assessment reports was stressed. Subsequent to the review, it was agreed that safeguards assessments reports could be shared with other international financial institutions on the basis of strict criteria to maintain confidentiality, and with the authorities’ consent.

Box 1.Description of the Safeguards Policy

The safeguards policy was introduced to provide reasonable assurances to the Fund that central banks of member countries using Fund resources have control, accounting, reporting and auditing systems in place to manage resources, including Fund resources, and that are adequate to ensure the integrity of operations. The policy’s main objective is to mitigate potential risks of misuse of these resources and misreporting of monetary program data. The principal instrument in achieving this objective is the conduct of safeguards assessments of five key areas of control and governance within central banks of borrowing countries, namely the external audit, legal, financial reporting, internal audit and control mechanisms (under the acronym ELRIC). If an assessment concludes that the ELRIC framework is inadequate, recommendations to alleviate confirmed weaknesses are agreed with the central bank authorities and implementation is monitored in the context of the member’s financing arrangement with the Fund.

The safeguards policy serves as a complement to the Fund’s other safeguards, which include limits on access, conditionality and program design, measures to address misreporting, and post-program monitoring. A cornerstone of the safeguards policy is that central banks of member countries borrowing from the IMF publish financial statements that have been independently audited by external auditors in accordance with International Standards on Auditing. Safeguards assessments are distinct from audits in that they entail high-level diagnostic reviews of the structures and mechanisms in place rather than a detailed test of transactions or verification of assets. Safeguards assessments are conducted at central banks, which are typically responsible for managing Fund resources and reporting on key statistics used for program monitoring.

The main output of a safeguards assessment is a report that establishes recommendations prioritized to address key vulnerabilities in a central bank’s safeguards framework. The authorities’ comments are included in the safeguards report. The Executive Board receives summaries of the main findings and recommendations of individual member safeguards assessments, but confidentiality requirements limit the circulation of the reports themselves. Since 2005, safeguards reports can be shared with World Bank staff, subject to strict confidentiality restrictions and with the consent of the concerned central bank.

The safeguards policy applies to members seeking financial arrangements from the Fund, except for the Flexible Credit Line (FCL). The modified policy for the FCL calls for the application of limited safeguards procedures, focused on discussions with external auditors of central banks. A safeguards assessment should be completed prior to the first review under a Fund supported program.

3. The current review has been conducted in wide consultation with various stakeholders. Staff from the Finance Department (FIN), which has the primary responsibility for the conduct of safeguards assessments, engaged at an early stage with area and functional department staff on the appropriate scope and focus of the review. In this context, an issues note to guide the conduct of the review was prepared in early 2010, and circulated to the Executive Board in conjunction with a visit to headquarters by the panel of experts in March 2010. The panel has sought the views of Executive Directors, central bank officials, and staff from area and functional departments on the effectiveness and implementation of the safeguards policy. The panel also held informal discussions with senior representatives of some of the international accounting firms on how best to address the perceived inconsistencies in the audits of central bank financial statements.

4. This paper is organized as follows. Section II outlines developments since the previous review. Section III takes stock of policy implementation, including trends in safeguards findings since 2000, the lessons learned over the past five years, and the staff resource envelope. Section IV takes up options for change in the areas of: (i) the scope and focus of safeguards assessments; (ii) instances of budget financing; and (iii) the confidentiality of safeguards reports. Section V outlines issues for the Executive Board’s discussion.

II. Developments Since the Previous Review

A. Activities

5. As of end-March 2010, 165 safeguards assessments of 75 central banks have been completed.2 There are two distinct phases in safeguards work: (i) the conduct of first-time and update assessments in respect of new arrangements, and (ii) the monitoring of the safeguards framework at central banks while credit is outstanding. As reflected in Figure 1, the focus of safeguards work has shifted over the last decade from first-time (initial) assessments towards the conduct of update assessments in respect of successor arrangements. The number of central banks subject to monitoring has been relatively unchanged over the last five years, except for a drop in 2007 reflecting the low credit environment at the time.

Figure 1.Assessments Completed, 2001-March 20101/

Source: FINSA databases.

1/ Excludes 27 transitional assessments completed during 2000-02.

6. Assessment activity has varied in line with both the volume and nature of financing arrangements. The annual output of completed assessments declined from 2005 in line with a lower level of IMF arrangements, but recently increased sharply in response to renewed lending by the Fund. In 2005-2008 a greater proportion of safeguards work was focused on PRGT financing arrangements, where capacity constraints sometimes present a challenge to addressing identified weaknesses in the safeguards frameworks of central banks. Recently there has been a more balanced mix of PRGT and GRA financing arrangements, which has also contributed to an uptick in the number of initial assessments. While the update assessment process is typically more streamlined than that for an initial assessment, the intensity of work for update assessments does vary according to the time elapsed since the initial assessment, the outcome of the monitoring process, and developments at the central bank. Six voluntary assessments were conducted in the last five years.

7. Monitoring activity is closely linked with assessment outputs. Safeguards assessments provide a diagnostic snapshot of a central bank at a point in time. The objectives of monitoring are to follow up on the findings and recommendations of an assessment, and to identify any emerging vulnerabilities at the earliest possible time. A risk-based monitoring methodology was put in place following the 2005 policy review, whereby the intensity of monitoring varies according to the underlying safeguards risks (Annex II). At a minimum, central banks are subject to annual desk reviews of financial statements and related external audit reports, and staff contacts the area department team and relevant officials from the central bank. The results of the monitoring work are a key input into the conduct of update assessments.

B. Evolving Modalities

8. While the core objective of the safeguards policy—reducing the risk of misuse and misreporting—is unchanged since inception of the policy ten years ago, the application of the policy has evolved in keeping with the Executive Board’s guidance from prior policy reviews, the emerging findings of safeguards assessments, and the need to adapt to developments in the Fund’s lending toolkit.

9. Flexibility. The way the staff conducts the various operational aspects of assessment work has been adapted in light of experience and in response to emerging issues and new lending instruments. In particular:

  • Assessments are phased for some countries in light of their post-conflict circumstances. Under this approach, an interim assessment is conducted for the most important elements of the safeguards framework, notably the external audit mechanism and the controls over the management of foreign reserves and monetary program data reporting. Once the central bank’s operations have returned to more normal levels, a full assessment is conducted.
  • The staff’s monitoring work programs are tailored for underlying risk factors and country-specific circumstances. Drawing on past experience, staff has also developed a set of indicators that could signal emerging safeguards issues and trigger a targeted re-assessment of a central bank’s safeguards. The monitoring records are a key input for update assessments and have enabled a streamlining of the safeguards process. In particular, update missions and reports are typically shorter and more targeted than those for initial assessments, and staff has increased the use of headquarters-based work to complete update assessments for lower-risk cases.
  • Known safeguards risks are being addressed at the earliest opportunity in program negotiations. Safeguards monitoring information and discussions with external auditors have proven useful to ensure that any known and high priority safeguards measures are incorporated in Fund arrangements, even when an update assessment has not been completed at the time of Board approval of an arrangement.3 Similarly, and in the context of the emergency financing cases in 2008-09, safeguards procedures were adapted to ensure that potential significant safeguards risks were identified as early as possible, even before the formal initiation of a safeguards assessment.

10. New lending instruments. Safeguards modalities have been modified to accommodate new lending instruments, specifically the Rapid Credit Facility (RCF) and the Flexible Credit Line (FCL).

  • A member’s request for assistance under the RCF requires a commitment to undergo a safeguards assessment and an authorization for Fund staff to have access to the central bank’s most recently completed external audit reports and to hold discussions with the auditors. The timing and modalities of the assessment under the RCF are determined on a case-by-case basis; it is presumed, however, that the safeguards assessment would have been completed before Executive Board approval of any subsequent arrangement to which the Fund’s safeguards policy applies.
  • Safeguards assessments are not conducted for FCL arrangements on the grounds that eligible countries have strong institutional arrangements in place that would mitigate risks of misuse and misreporting. Instead, staff is authorized to hold discussions with the central bank’s external auditors with the objective of ensuring that recent external audit reports do not raise significant issues that could place FCL purchases at risk.4

11. Budget financing.5 Since the last review there has been an increasing number of cases in which IMF resources were disbursed to provide budget financing. In such cases the central bank serves as: (i) the government’s banker and the custodian of the country’s official foreign exchange reserves, which will be a source for servicing IMF obligations, and (ii) the Fund’s depository for its own currency holdings.6 While safeguards assessments apply only to the central bank and in cases of budgetary financing have not been extended to other government entities, they have sought to ensure that appropriate controls are in place for timely servicing of member obligations to the Fund. In particular, safeguards assessments have:

  • Looked for a clear framework between central bank and government for the servicing of IMF lending so that their respective roles and obligations are transparent and understood. In all such cases, the authorities have also agreed that Fund resources be deposited with the central bank.7 The objective has been to replicate as far as possible the situation where Fund resources are on-lent to the government by the central bank.
  • Sought external assurances of the integrity of components of NIR and NDA performance criteria that are sourced outside the central bank. In the absence of well-defined government balance sheets, safeguards assessments have also called for the transparent disclosure of Fund transactions in the notes to the central bank’s financial statements.

12. Communications. Staff has taken several initiatives to strengthen awareness and improve communications on safeguards issues. Specifically:

  • Raising awareness of safeguards work in a variety of fora. Over 500 representatives from 122 central banks and 125 countries have participated in semiannual safeguards seminars, which are case-study based and equip participants to conduct a self-assessment of their own central bank. The seminars have been adapted in recent years to include presentations from external parties, such as the international accounting firms, professional bodies, and the World Bank. Central banks have also welcomed staff’s guidance notes, including templates, for the implementation of safeguards measures that draw on the staff’s multilateral experience across regions. In partnership with central banks and the international accounting firms, information about the safeguards program has also been disseminated at occasional regional workshops and in ad hoc publications.
  • Regular communication with central banks through a variety of channels (email, tele- and video-conferencing), especially in the periods after a safeguards mission and throughout the monitoring phase. Use of video-conferencing, in particular, can enable central banks to maintain close links with staff. Central banks now provide a designated contact person for safeguards issues, which has facilitated continued outreach, generally timely official responses to safeguards assessment reports, and the follow-up of safeguards findings.
  • Publication on the IMF website. A range of safeguards related material is published, including the annual safeguards activity reports, a summary of findings and recommendations for individual assessments (where the respective staff reports are published), and a Fact Sheet. In addition, staff operational guidelines are now posted on the intranet.
  • Stronger and early collaboration with area department teams. Early notification to FIN staff of potential new arrangements has ensured that any deviation from key safeguards requirements, such as the non-publication of audited financial statements that may become an obstacle for approval of an arrangement, are signaled to the authorities at the earliest opportunity.
  • Increasing interaction with external auditors throughout both assessment work and ongoing monitoring. Experience has shown that regular contact maintains auditor awareness of staff’s interest in central bank safeguards, and allows more frequent and open exchanges of views on safeguards related audit issues and practices observed during assessments.

13. Sharing Reports. Safeguards reports have been confidential documents since inception of the policy. The reports are prepared for an internal purpose and contain confidential information that would not ordinarily be made available to a broad audience. Since 2005, reports have been shared with World Bank staff, subject to certain confidentiality criteria and the consent of the central bank (see Box 2). A total of 44 reports have been shared with World Bank staff, and to date 3 central banks have withheld consent for such sharing.

C. Recent Misreporting and Emerging Governance Cases

14. The preventive, ex ante nature of safeguards assessments precludes definitive conclusions about the extent to which potential misreporting and misuse is being avoided. To reduce the risk of misreporting, safeguards assessments typically recommend a combination of preventive measures (e.g., data reconciliations with the accounting records) and ex post controls (e.g., data audits) to ensure data integrity. However, there is no counter factual on the extent to which misreporting is being avoided by the effectiveness of preventive measures. Also, safeguards assessments are distinct from audits; in particular, the assessments are diagnostic tools, and unlike audits, not a means to provide assurances on the verification of assets and liabilities, or on the effectiveness of controls.

Box 2.Sharing Reports with the World Bank

As follow up to the 2005 review, the Executive Board decision in December 2005 (12/1/05) authorized the sharing of reports with the World Bank staff.

This decision requires that the Managing Director obtains, in particular: (i) satisfactory assurances from the Bank that the report will be kept confidential, its internal distribution restricted and the report will not, either in whole or in part, be quoted from, cited or used in publications; and (ii) the consent of the central bank concerned to the transmittal of the report. In practice, World Bank staff requests for reports are coordinated by, and routed though the office of the Head of the Financial Management Sector Board, Operations Policy and Country Services, and forwarded to the Finance Department, which then seeks the consent of the relevant member central bank to release the report to World Bank staff.

The 2005 decision focused on sharing of safeguards assessments reports upon request by World Bank staff as other large regional development banks did not conduct systematic due diligence assessments similar to the Fund’s safeguards assessments policy—primarily because their financial operations involved mostly project lending that was not conducted through a central bank.

15. Safeguards assessments have continued to play a role in the detection and resolution of cases involving misreporting and governance abuse. Since the last policy review, three instances of misreporting of central bank monetary data have been reviewed by the Board and governance abuses have arisen in three other cases. Safeguards measures played a key role in the detection of two of the misreporting cases and in one of the governance abuse cases. In all cases, safeguards measures were core elements of the action plans to address the weaknesses that gave rise to the incidents.

16. The misreporting cases saw inaccurate reporting of data on international reserves, credit to government, and reserve money. In the case of Mauritania, a safeguards assessment identified significant data issues initially in respect of foreign exchange reserves. A subsequent review of reported program data also led to the authorities notifying the IMF of adjustments to data on net domestic assets of the central bank and net domestic financing of the budget. In the DR Congo, a safeguards assessment found significant safeguards risks and inaccuracies in the reporting of government balances to the Fund due to ongoing weaknesses in the central bank’s financial reporting and controls framework. In the case of Tajikistan (2007) the authorities notified the IMF of inaccuracies in data reported on net international reserves and reserve money that stemmed from central bank financing of the cotton sector. Key elements of remedial safeguards measures in these cases centered on: (i) improving accounting systems, with guidance from Fund-supported technical assistance advisors; and (ii) independent review by external auditors of monetary data as at program test dates.

17. The governance abuses involved fraudulent transactions (Annex III). In Tanzania, the central bank’s external auditors raised a concern with staff about a potential serious fiduciary issue, but client confidentiality prevented them from disclosing the details. Subsequent forensic investigations confirmed significant fraudulent transactions through an account that was administered by the central bank on behalf of government, which were facilitated by high-level override of controls, and weakened central bank Board oversight and internal audit. In Burundi, the Inspector General investigated allegations of misappropriation of government resources and found that fraudulent payments had been made based on false documents submitted by the Minister of Finance and the central bank Governor. In this case controls at both the central bank and Ministry of Finance were overridden, and the earlier safeguards recommendation for the establishment of an independent audit oversight body had not been implemented. In the CEMAC region,8 following media reports of fraud in its Paris Office, the BEAC (the region’s central bank) initiated some special audits, including one by the Audit Committee and another by an external audit firm. The investigations confirmed misappropriation in the period 2004-07 and revealed a deliberate high-level override and breakdown of controls in both the Paris Office and BEAC headquarters, as well as an absence of effective oversight by headquarters. A subsequent special audit of budgetary and accounting practices at BEAC headquarters indicated significant risk that unauthorized outflows from BEAC’s reserves could occur due to poor oversight and inadequate internal controls.

III. Taking Stock

A. Safeguards Findings

18. The findings of safeguards assessments have evolved over the last 10 years, as central banks have strengthened their ELRIC frameworks.9 The initial findings of safeguards assessments revealed significant vulnerabilities in all areas. Notably, some 14 central banks were found to have no external audit at all, while relatively few had adopted a recognized set of accounting and financial reporting standards. In other areas, governance oversight was found to be weak and internal audit often inadequate. Staff has continued to track the findings in eight key areas of vulnerability (Figure 2) and the trends corroborate the general move towards greater transparency and improved control systems by central banks over the last ten years.10 In particular:

Figure 2.Key Vulnerabilities Identified in Central Banks

Source: FINSA databases.

  • Almost 60 per cent of the central banks assessed through end-2002 were found to have no or deficient external audit mechanisms; deficiencies included a lack of auditor independence and inadequate audits conducted by local auditing firms. In contrast, these deficiencies have been identified in only about 25 percent of the central banks assessed since 2002, although the trend is flat.
  • Around 75 percent of central banks are now completing and publishing their full set of financial statements, compared with about 55 percent earlier in the decade.
  • Significant improvements in the adoption of benchmark accounting standards are evident. Recent findings reveal deficiencies in this area in around 30 percent of central banks assessed, compared with 80 percent at inception of the policy.
  • Internal audit practices at central banks are being modernized, and while improvements are still needed at about 40 percent of the central banks assessed, this is markedly less than earlier findings.
  • Reserves management practices and controls continue to strengthen. Staff has seen a more structured approach to reserves management at central banks, drawing on the IMF’s Reserves Management Guidelines. Also, staff has partnered with the World Bank’s Reserves Advisory Management Program (RAMP), which has significantly improved the control and management of reserves through building capacity at a number of central banks.
  • Data management processes have been strengthened more recently. A key development here is that, as internal audit practices improve, program data compilation and reconciliations are increasingly subject to internal audit at key milestones, such as test dates under IMF-supported programs. This development provides added assurances to central bank management on the accuracy and completeness of program data.

19. However, recent findings also show that challenges remain. In some areas, the improvements appear to have plateaued, while progress has slowed in others. Specifically:

  • Staff continues to identify inconsistencies and weaknesses in the external audits of about one in four central banks. The nature of the deficiencies varies from the now very rare absence of an audit at all—which is addressed immediately under an IMF-supported program—to impaired independence of the auditors, which is addressed through a policy for transparent auditor appointment and selection. Another deficiency concerns poor understanding of central banking by the external auditor, which sometimes contributes to the application of inadequate audit procedures. These issues are addressed as they arise, but their continued occurrence points toward the importance of adequate oversight by central banks to ensure the quality of the external audit process.
  • Governance oversight is often provided for through appropriate organizational structures, but is not fully effective in practice. Oversight bodies such as independent audit committees and investment committees are increasingly in place at central banks, but these bodies are not always supported by a substantive work plan and agenda for overseeing the areas covered by safeguards work. Deficiencies are being found in close to half of all assessments, which now focus increasingly on the scope and professional underpinnings of oversight committees, rather than their formal mandate.
  • Slower progress has been made in strengthening governing legislation, mainly because of the practical difficulties in enacting new laws. Where feasible, staff addresses this safeguards risk through temporary measures, such as a Board resolution pending changes in the law, or by a MOU between the government and the central bank that could, for example, clarify an ambiguity in the enabling legislation.

20. The changing nature of safeguards findings has required adaptation of the analysis within the ELRIC framework. With the preponderance of update assessments, where central banks must have an external audit in place, the focus in this area is on developments and any emerging risks since the last assessment. For example, an update assessment may have to review practices of a new audit firm that has been appointed, as well as reviewing the measures adopted by the auditors to ensure the quality and consistency of the audit work and compliance with international standards. In other areas, notably the modernization of the internal audit and financial reporting functions, progress towards the ELRIC benchmarks runs along a spectrum. Update assessments focus, therefore, on the extent to which reliance can be placed on these functions; e.g., is the internal audit function ready to provide assurances in data reporting and if not, does this weakness need to be mitigated by other measures? Furthermore, the controls and audit procedures reviewed under the “C” component of ELRIC now include currency and banking operations in light of staff’s evolving experience with the findings of fraud investigations.

21. The implementation rate of safeguards recommendations has remained high (Annex IV). The overall implementation rate for some 1006 recommendations as at end-March 2010 was 77.1 percent, while implementation of priority safeguards recommendations has averaged around 95 percent.11 Safeguards monitoring plays an important role in this regard through the follow-up of prior recommendations. Safeguards reports typically recommend that conditionality is warranted for high priority findings, but the need for conditionality varies according to country specific circumstances and the timing for implementation of safeguards measures. Formal conditionality also becomes moot when central banks have already addressed a safeguards finding prior to the time of program approval (or the first review).

22. The introduction of conditionality reforms in March 2009 does not appear to have had a significant impact on implementation rates. Early indications are that implementation rates have not declined despite the removal of structural performance criteria. While the absence of any needed waivers may somewhat reduce the visibility of delayed implementation of safeguards measures, the introduction of review-based conditionality is helping to promote country ownership. It also provides greater flexibility in elevating commitments to structural benchmarks, as needed, in a later stage of an arrangement. Critical safeguards recommendations can continue to be considered for inclusion as prior actions under program conditionality.

23. There has been some limited interest in a wider dissemination of safeguards findings. In 2008 staff received requests in two country cases from the donor community for details about the findings of safeguards assessment mission teams. Staff could point towards the expected publication of summary findings through the publication of staff reports and the annual activity papers prepared for the Executive Board, but was not in a position to provide a comprehensive briefing on the mission’s detailed findings due to the confidentiality of the reports.

B. Lessons Learned

24. The conduct of safeguards assessments provides an important signal to both member country central banks and other IMF stakeholders, including the general public, that the IMF views the basic due diligence of safeguards assessments as an essential component of overall program design. The underlying premise is that strong governance and control frameworks at central banks should a priori help to better safeguard Fund resources and reduce opportunities for misreporting. Against this background, this section looks at the lessons learned through the safeguard experiences of the past five years.

25. Control override by executive management is a common theme of the high profile governance abuse cases. Box 3 highlights some of the general attributes of the fraud cases, but a common theme was control override at the highest level. Preventing high-level control overrides is not achieved through more operational controls, but rather through effective governance and independent oversight. In particular, while improved capacities in areas such as operating controls, internal audit review, and external audit and financial reporting help mitigate safeguards risk, they need to be complemented by effective oversight mechanisms, such as an independent audit committee, and a sound overarching governance framework. In this regard, the panel has emphasized the importance of a sound governance framework that sets an ethical tone at the top of an organization, and governs the entire organization.

Box 3.Observations from the High Profile Governance Abuse Cases

  • It is important that the substance of governance and control measures is present, and not just the legal form. For example, oversight bodies need to be effective, and not just exist in form.
  • The control environment is dynamic and may deteriorate over time. No control framework can guarantee against the willful override of controls, but it may facilitate earlier detection.
  • Ex-ante and ex-post controls are important, especially over government accounts and accounts maintained abroad.
  • Frauds can be perpetrated in a non-core function of the central bank, where transactions are rare or exploitation of gaps in controls may go unnoticed.
  • Central bank staff needs to have access to appropriate channels to report fraud or questionable behavior.
  • A quality external audit can play a key role in uncovering the fraud. Staff needs to remain vigilant to identify cases where external auditors face any restrictions of scope.
  • Short term remedial measures are important but cannot substitute for the restoration of the integrity of control mechanisms and effective governance oversight.

26. Update assessments are showing that challenges remain in sustaining the effectiveness of ELRIC frameworks at central banks. Progress has been somewhat uneven across countries and areas covered by safeguards assessments; for example, improvements in internal controls have generally lagged behind the progress made in strengthening financial reporting frameworks. Also, experience has shown that new measures, such as the introduction of an internal audit function, have proven vulnerable to the departure of key personnel. Similarly, audit committees are not always adequately equipped to conduct effective oversight of the system of internal controls and to ensure that external and internal audit findings are being adequately addressed. These challenges point towards the need for continued vigilance, highlighting the importance of the risk-based monitoring framework. Technical assistance from the Fund, and other sources, that is tailored to match specific safeguards priorities can also play an important role in implementing and sustaining necessary reforms.

27. Ownership of safeguards measures is key for sustaining their effectiveness. Beyond the Fund’s narrow objectives, central banks themselves see advantages in strengthening their safeguards framework because of the positive impact across the broader institutional setting of a central bank. As indicated in the feedback received by the panel of experts, central banks continue to welcome safeguards findings, especially when they see the recommendations as helpful for achieving their own reform objectives. The official responses from central banks to safeguards assessments, which are included in the reports, are also generally positive. However, not all central banks have fully embraced safeguards findings. A very few object to policy requirements, such as an external audit or publication of financial statements, while others sometimes question the prioritization and urgency of a report’s recommendations. Area department teams sometimes echo these sentiments, noting that safeguards reports should remain tightly focused on critical vulnerabilities for safeguarding Fund resources, and resist drifting in the direction of technical assistance.

28. External auditors remain important partners in the safeguards process, provided quality is maintained. External audits provide an independent “first line of defense” in obtaining assurances on the integrity of a central bank’s financial reporting processes. A significant benefit of the annual external audit is the independent verification of a central bank’s assets and liabilities. Issues identified in this work by external auditors have exposed deliberate and inadvertent misreporting cases, and have provided indicators of potentially fraudulent activities. However, external audits do not provide a guarantee against fraud due to their relatively narrow scope.12 The BEAC case, for example, involved fraudulent payments that were well below the materiality thresholds for an audit of the financial statements, and hence were not readily detected by audit sampling of transactions. Also, as noted earlier, audit quality is sometimes uneven and can deteriorate quickly (e.g., through appointment of a new audit firm or changes in key personnel), and auditors can avoid accountability by long delays in releasing signed audit opinions, even though the audit work and the draft audit opinion have not identified significant issues. While central banks have primary responsibility for audit oversight, these issues point towards the importance of continued vigilance by staff over the status and timing of external audits, including through periodic follow up with the central banks and the auditors.

29. Raising awareness of safeguards risks among a variety of stakeholders is a key driver in meeting the policy’s objectives. Previous policy reviews have emphasized the importance of stronger communications on safeguards issues across all stakeholders. Central bank staff is now more aware of the importance of accurate program data, as evidenced by the increasing number of central banks that have introduced data reconciliations with the audited financial statements as a standard operating procedure. Raised awareness by external auditors of the importance that the IMF places on central bank audits has played a role in their bringing issues to staff’s attention, which can then be addressed in partnership with the authorities. Equally, raised awareness of safeguards risks among area department teams can help avoid or deflect subsequent issues, especially misreporting.13 The publication of the special investigations into high-profile cases has also raised awareness of the unfortunate consequences of fraudulent activities, which may help deter such events in the future.

30. The emergence of high profile governance abuse cases can both harm and enhance the Fund’s reputation as a prudent lender. Fast, decisive action by country authorities, with assistance and advice from the Fund, to: (i) diagnose the problem; (ii) publish the findings of investigative reports; and (iii) take remedial actions, positions the Fund as a respected partner for the international community in identifying and resolving such issues. In contrast, the unexpected emergence of such issues into the public domain, without direct IMF involvement in the identification and resolution, can create the impression that the Fund was “asleep at the wheel.” A key consideration for the Fund, therefore, is to stay engaged as far as possible in a trusted advisor role with central banks on safeguards issues. On the latter point, central banks also acknowledged to the panel that safeguards findings have enabled them to better align with best practices.

31. In summary, experience has shown that safeguards assessments have helped to identify emerging governance abuse and misreporting issues at central banks, but cannot be a panacea for control overrides. While improved capacities in areas such as operating controls, internal audit review, and external audit and financial reporting help mitigate safeguards risk, they need to be complemented by sound and effective governance oversight and independent mechanisms to help deter control override. Other factors that have become evident during the conduct of safeguards assessments include the importance of: (i) the monitoring framework to help ensure the sustainability of safeguards measures implemented by central banks; (ii) maintaining a focus on the underlying objectives to ensure ownership by the various stakeholders; (iii) raising awareness of safeguards issues to help deter and detect opportunities for misreporting and misuse of IMF resources; and (iv) remaining vigilant on external audit quality and reacting to emerging indicators of vulnerability.

C. Staff Resources

32. Safeguards assessments work is mainly conducted by a specialized staff team of 8-10 professional accountants (Annex V).14 Staff resources are not easily scalable in light of the specialized expertise needed, but the overall resource envelope has stayed quite stable over the last 5 years (Figure 3). This reflects the continuum of the safeguards work, whereby update assessments are an off-shoot of monitoring rather than de novo assessments. Also, the relatively high demand on staff time for missions requires a minimum resource base to ensure operational flexibility is retained. The recent increase in assessment volume was accommodated by hiring two additional accountants on limited term positions, internal redeployment on a temporary basis from OIA, and a significant increase in uncompensated overtime.

Figure 3.Safeguards Activity

Source: TRS and FINSA databases.

33. Safeguards work continues to be a relatively small portion of Use of Fund Resources (UFR) work. Overall, and using data in the Time Recording System (TRS), direct safeguards assessment work by the staff team represented less than three percent of total staff time allocated to UFR over 2005-2009. In dollar terms, and using FY11 standard costs, the direct annual personnel cost of the safeguards team is around $3m, or about $45,000 per central bank monitored. On a country basis, the resource needs can vary substantially, especially for the high profile cases. The information in the monitoring database has also enabled a larger number of assessments to be undertaken at headquarters and has therefore reduced travel costs.

IV. Looking Ahead: Options for Change

34. The foregoing discussion suggests that the safeguards policy continues to serve the Fund well and helps to maintain the Fund’s reputation as a prudent lender. The strengthened governance and control frameworks at central banks de facto help to mitigate misuse of Fund resources. Central banks have generally embraced the safeguards findings and continue to benefit from the diagnostic assessments and monitoring process. The resource cost is relatively modest in relation to the total investment of staff resources in all lending activities. The panel has also affirmed the usefulness and value of the policy, noting that they found among stakeholders broad approval for the policy and acknowledgement of its value.

35. Nonetheless, some changes to the safeguards policy could be considered. Against the background of the evolving findings of safeguards assessments and lessons learned in applying the policy, staff has examined: (i) the appropriateness of possible changes in the scope and focus of safeguards assessments at central banks; (ii) options for providing safeguards assurances of fiscal fiduciary risks in cases of budget financing; and (iii) issues surrounding the confidentiality of safeguards reports. These aspects and the corresponding views of the panel are discussed below.

A. The Scope and Focus of Safeguards Assessments

36. Overall, staff believes that the inherent flexibility of the ELRIC framework continues to make it an effective tool for assessing safeguards in central banks. The scope of safeguards assessments provided through the five ELRIC pillars are sufficiently encompassing to accommodate new control and related benchmarks, while none of the components is redundant or no longer relevant for the assessment objectives. In its work, the panel endorsed the applicability of the framework, and made several recommendations for continuous improvement over time.

37. The descriptions of the five ELRIC pillars could be updated for general developments in control frameworks. The ELRIC pillars were developed in 2000, drawing on standards and codes in place at that time. The past 10 years have seen an evolution in corporate governance and control frameworks, and a marked trend towards increased transparency and accountability. The Sarbanes-Oxley Act of 2002, for example, introduced the requirement for public companies in the U.S. to maintain systems of internal controls, and for management to certify and the independent auditor to attest to the effectiveness of those systems. National legislations have strengthened corporate governance through increased responsibility and independence of audit committees. The scope and coverage of international auditing and accounting standards have also increased over the past ten years. Most recently, the OECD and Group of 30 have released further guidance on governance principles in the wake of the global financial crisis. The UK Walker Report of July 2009 also has a number of important recommendations that will come to impact good governance practices of central banks.

38. In its report, the panel calls for a sharpened focus on governance in the ELRIC framework. Staff notes that the framework already captures the various attributes of governance within each of the five separate pillars, but does not explicitly emphasize governance as a common theme transcending the pillars. Mindful of the control overrides in the high profile cases and the importance of effective oversight, staff believes that explicitly defining governance principles enunciated by the panel as overarching the ELRIC pillars would help in better focusing the diagnostic analysis of the governance structures at central banks. Further, these governance principles are cross-cutting elements that can be found in each of the ELRIC pillars. To reflect recent developments and the panel’s recommendation, staff proposes updating the description of the five ELRIC pillars and adding the key principles of governance as a chapeau (Box 4).

39. The panel also recommends that more attention be given to integrated risk management frameworks (IRMF) within the safeguards framework. Staff agrees with the panel’s assertion that “the effectiveness of central banks in managing their own risks ultimately affects the degree of risk exposure for the Fund.” In this regard, assessments currently ascertain whether an IRMF is in place, but in practice staff has encountered few central banks that have completed the process of formalizing and implementing an IRMF across the various functions of the bank. Some banks have embarked on the centralization of risk management and established a Risk Management Committee of the Board, but the process is not yet fully integrated across the entity. Most central banks continue to manage risks within functional areas, under the broad oversight of the management and Board, and there are capacity constraints in implementing IRMF in substance rather than just form. The importance of appropriately sequenced reforms in the risk and control environment is another key consideration for recommendations in this area; for example, staff’s experience suggests that it would be premature to attempt an IRMF before having the basic ELRIC building blocks in place.

Box 4.Changes to the ELRIC Framework

The description of the five ELRIC pillars (Annex VI) has been updated to reflect key external developments and ten years of safeguards experience as follows:

  • The descriptions of the external audit (“E”) and financial reporting (“R”) refer to International Standards on Auditing (ISA) and International Financial Reporting Standards (IFRS) as accepted international benchmarks, which remain appropriate. The description of the internal audit mechanism (“I”) has been revised to reflect the emergence of International Standards for the Professional Practice of Internal Auditing issued by the Institute of Internal Auditors.
  • The description of the legal structure and autonomy (“L”) has been revised to reflect improved central bank governance and autonomy and lessons learned on common avenues for limiting outside interference in central bank operations, besides government credit. Improved central bank transparency and accountability has reduced the need to focus on statutory audit, financial reporting, and publication requirements. The shift towards monetary policy as the principal objective of a central bank has enabled greater independence and reduced the role of credit to government as a source of impaired autonomy.
  • The description of the system of internal controls (“C”) has been updated to reflect evolution in control models and good governance. Newer control models consider the control environment established by top management as the most important component in a robust system of internal control. Accordingly, the assessment of the “C” pillar places greater emphasis on assessing the quality of the control environment and the degree of governance and oversight exercised from the highest level.

A governance chapeau has also been added to the description of the ELRIC framework (Annex VI). As noted by the panel, the attributes of good corporate governance cover: (i) discipline, represented by senior management’s commitment to observe and promote good governance; (ii) transparency; (iii) autonomy; (iv) accountability by key decision-makers; (v) responsibility; (vi) fairness; and (vii) social responsibility. For central bank purposes, the relevant attributes of fairness and social responsibility have been subsumed under the responsibility heading (see Annex VI).

40. In the context of the goal of continuous improvement in the safeguards framework that underlies the panel’s recommendations, staff sees merit in giving greater prominence to risk management. Safeguards assessments could explicitly consider how central banks are embarking on, or progressing along, the path towards an IRMF. Depending on the stage reached, assessments could ascertain how well central banks manage their own risks. This could include analysis of whether a central bank has established a benchmark and methodology for assessing risk and, if so, whether the risk management system operates effectively. For the immediate future, staff believes that central banks’ approach to, and progress with, risk management can be assessed within the “C” pillar of the ELRIC framework.

41. Looking ahead, the existing modalities for conducting safeguards assessments at central banks need to remain flexible, within the policy’s existing mandate. The lessons learned from experience so far emphasize the importance of continued vigilance in the monitoring process, quick and appropriate reaction to emerging issues, maintaining a focus on those areas that are critical to safeguarding Fund resources, and adapting remedial measures to country circumstances to ensure ownership. At the same time, policy requirements such as: (i) the publication of financial statements that have been independently audited in accordance with international standards; and (ii) the deadline of the first program review for completion of a safeguards assessment remain appropriate and should continue to be applied consistently.

42. The monitoring modalities remain broadly appropriate. The panel has affirmed the effectiveness of the risk-based monitoring framework, which is largely dependent on the continued cooperation of central banks and their external auditors. Monitoring continues for as long as Fund financing is outstanding, but staff has sometimes found that it is difficult to obtain the required monitoring documentation once an arrangement has concluded.15 Shortening the monitoring period (e.g., to one year after the conclusion of an arrangement) or adjusting the monitoring period based on the underlying risk could help address this difficulty, but would also negate the advantages of staying engaged to facilitate potential update assessments for successor arrangements. On balance, staff proposes that central banks should continue to provide monitoring information in accordance with current requirements, but that instances of non-receipt of such information be explicitly flagged in staff reports.

43. Safeguards modalities also need to be adapted in response to issues identified by stakeholders during this review. In particular, and as reported by the panel, stakeholders have emphasized the importance of collaborating more closely with central banks and addressing more systematically the concerns identified in the quality of external audits. In this context:

  • Opportunities for harnessing technology to make the safeguards process more open and collaborative could be explored. The improved communication with central banks since the 2005 policy review has helped raise awareness and reduce misperceptions of the safeguards policy, but feedback from central banks to the panel suggest a renewed appetite for additional information and knowledge sharing, especially on cross-country experience, and more frequent dialogue between central bank officials and the safeguards team. New technology platforms for collaboration, such as extranet websites and social networking tools could help achieve these objectives and contribute towards more effective and efficient monitoring. They could also play a role in promoting self-assessment techniques, as recommended by the panel.
  • Improved interaction with the large external audit firms at the global level to discuss audit quality issues. Staff sees merit in complementing the existing “bottom up” approach in dealing with these issues with a “top down” approach by conducting an annual forum at the global firm level, as proposed by the panel.

44. Safeguards policy requirements may also have to be adapted for possible developments in the Fund’s future financing role. The limited safeguards procedures applied for FCL cases appear to have been effective in the context of an insurance-type facility and the presumptive strong institutions. Similar adaptations in scope may prove feasible for further lending reforms, depending on the nature of the underlying instruments and the practical challenges in conducting a safeguards assessment.

45. Finally, staff believes existing modalities can be adapted in several other areas noted by the panel. These include (i) improving internal collaboration, particularly through raising staff awareness of the relevance of safeguards activities to program issues and technical assistance work; (ii) examining peer review or regulator’s reports in the assessment of the quality of external auditors, and (iii) incorporating in the safeguards toolkit an assessment of the need for, and appropriateness of, a “whistle blowing” mechanism.

B. Budget Financing

46. The current safeguards framework does not provide assurances about entities beyond the central bank. Safeguards assessments were introduced to demonstrate a level of due diligence, but the mandate of the policy has never covered entities other than the central bank. Therefore, in cases where IMF resources are provided directly to the government, the safeguards assessment does not provide the same level of assurance about minimizing the risk for misuse of IMF resources and about the quality and reliability of fiscal data and other information related to performance criteria used in IMF-supported programs. The panel notes that the shift in the volume of cases of budget financing has come with an associated shift in the nature of the risk to Fund resources.

47. Consideration was given in past reviews of the safeguards policy to extending the mandate beyond the central bank. In the 2002 review, the Board noted that safeguards assessments should not be applied to fiscal issues since that would require a new mandate for staff; this view was reaffirmed in the 2005 review. The main obstacles for replicating safeguards assessments in the whole of government sector identified by these reviews were practical:

  • Significant informational obstacles with regard to the quality of, and delays in finalizing, government accounts and the wide dispersion of practices in this area could preclude the application of a uniform approach to such assessments; and
  • Obtaining safeguards assurances beyond the central government to the activities of state enterprises and lower tiers of government would not be feasible. Public Financial Management (PFM) systems in developing countries, especially at the line ministry and agency level, are often very weak; many countries with which the Fund engages in program support would not meet criteria for reasonable assurances for the whole of government. The attainment of such standards would require a multi-decade development effort.

48. A safeguards assessment of the whole of government remains unfeasible. The practical considerations from past policy reviews remain relevant. While many countries have strengthened their PFM systems, an assessment across the whole of government would be time and resource consuming, which could lead to long delays in its completion and thereby negate the objective of providing assurances at an early stage of program engagement. Also, such assessments would overlap to some extent with PFM work conducted by other multilateral partners, notably the World Bank.

49. Limited assurances on fiscal safeguards risks are already provided by existing practices. In 1998, the IMF introduced a Code of Good Practices on Fiscal Transparency (“the Code”), which led to a voluntary program of fiscal transparency assessments, called fiscal transparency modules of the Reports on the Observance of Standards and Codes (“fiscal ROSCs”), covering now more than 90 countries. The first version of the Manual on Fiscal Transparency (“the Manual”) was issued the same year. The Code and its corresponding manual were updated in 2007. With a particular focus on tracking of aid funds (and debt relief benefits), the multi-donor Public Expenditure and Financial Accountability (PEFA) Performance Measurement Framework published in 2005 includes a series of 28 high level performance indicators covering public financial management effectiveness. Unlike fiscal ROSCs, PEFA evaluations are quantitative assessments, scoring the performance of the country against the 28 indicators. Program design is also a key safeguard in the case of budgetary support. IMF budget support is for financing the fiscal deficit and not linked to a specific developmental policy or expenditure. Elements of program design are therefore focused on the quality of budgetary processes, including fiscal transparency and accountability, rather than the quality of budgetary spending.

50. Building on existing practices and information, additional steps could become a standard part of budget financing cases:

  • The existing practices for central bank involvement in budget financing arrangements could be explicitly incorporated under the umbrella of the safeguards policy by formalizing requirements as follows: (i) the maintenance of a specific government account at the central bank to receive Fund resources;16 (ii) the government should hold foreign exchange balances only with the central bank; and (iii) the establishment of a clear framework agreement between central bank and government on responsibilities for servicing obligations to the Fund. As noted earlier, these arrangements seek to replicate, as far as possible, the situation where Fund resources are on lent to governments by central banks.
  • In addition to current practices, a more explicit reporting approach to the Board could be considered whereby staff reports in a new arrangement involving budget financing would specifically report on known fiscal safeguards risks (drawing on available information, such as ROSC reports and other diagnostic instruments (PEFA)). It would be important to recognize that these diagnostics do not go into the depth needed to assess risks at the entity level. Accordingly, care would be needed to avoid creating an expectation (in the Board and elsewhere) that such procedures provide safeguards assurances.

51. A more far reaching change would be to initiate an assessment of fiscal safeguards risks at the level of the state treasury. The treasury is the recipient of Fund resources and is responsible for control of budgetary and spending processes. Staff has identified an approach for a narrow assessment, which could focus on criteria (Box 5) derived from the Fiscal Code. Evaluating the state treasury and the core budget/spending process would provide some additional but still limited assurances on the appropriate use and reporting of government resources, including Fund resources. It should be emphasized, however, that the proposed framework for state treasuries and its implementation would need to be further developed and piloted to ensure that the value-added from injecting safeguards assessments is not outweighed by the costs. Many Fund members, particularly low-income countries, may face difficulties complying fully with the criteria, and remedial actions may not always be easy to implement. Also, while the Fund is clearly the lead institution among the donor community on central banking issues, this is not the case in PFM work.

52. Almost all of the information needed for a targeted assessment should be available from recent fiscal ROSCs. The latter generally provides most of the answers to the criteria that would need to be evaluated for a targeted assessment at the state treasury level. An important consideration is that the ROSC is broader in scope than just state treasuries; it does not review all necessary issues systematically so additional information may be needed. PEFA and other diagnostics could provide supporting information on issues of existence of a treasury single account, bank reconciliation, production of final accounts, and external audit. If fiscal ROSC and other information are not sufficient or up to date, staff would need to consider additional steps, including mission work, to collect necessary information.

53. Going beyond the central bank to the state treasury would require a parallel assessment mandate and product. Safeguards assessments are conducted in collaboration with just one entity and typically one external audit firm. The assessments require an authorization for the external auditor to discuss findings with Fund staff and the reports include an official response from the central bank. Adding the state treasury would raise new issues, including how best to engage with other parties such as government auditors, and whether staff would be given the requisite access to relevant information and documentation. Given the confidentiality of safeguards reports and their coverage on one entity only, separate reports would also likely be needed. The key findings of both would need to be consolidated in the relevant country paper prepared for the Executive Board. From the staffing perspective, the existing FIN team of accountants would not have the requisite expertise.

Box 5.Criteria for a Possible Targeted Assessment Covering the State Treasury

The criteria below should be seen as a first, tentative outline of an assessment framework for state treasuries. It would require further development and piloting in countries receiving budget financing. The criteria would be used as end-point objectives, which most countries would gradually and progressively achieve.

  • Does the State Treasury operate a main bank account at the central bank, through which all government transactions are processed? There should be a legal requirement that enforces the use of such an account, and the legal basis should be followed in practice. If there are alternate central government bank accounts from which budget expenditure takes place, these accounts should also be under the control of the state treasury, be consolidated on a regular basis, and also held at the central bank. If there are commercial bank accounts through which budget expenditure is executed in part, these should preferably be zero-balanced, but at least consolidated on a regular basis.
  • Is all central government expenditure authorized by the budgetary appropriation or other law? The annual budget and its formal amendments, as approved by parliament, should define all uses of resources from the main central bank account. This principle may be enshrined in the Constitution or Financial Management Act, and applied in practice.
  • Is the main bank account at the central bank (and other possible authorized accounts) reconciled regularly? Reconciliation of expenditure, revenues and financing flows with account statements would provide assurances on the accurate accounting of transaction over the government’s bank accounts.
  • Are all central government transactions executed through the state treasury properly accounted for using clearly defined national accounting standards that are on major issues aligned with international standards such as International Public Sector Accounting Standards, and are all final accounts under the responsibility of the state treasury produced in a timely fashion? Where there are departures from international standards are the relevant circumstances and the effects of such departures fully disclosed in the accounts? The accounting framework should in any case record both above and below the line transactions. A minimum standard for timely preparation final accounts of developing countries could be, for example, within 6 months of the end of the year.
  • Are the final accounts prepared by the state treasury audited by an independent commercial or state audit institution, according to national standards aligned (on major issues) with international standards such as the International Standards on Auditing (ISA) and/or standards issued by the International Organization of Supreme Audit Institutions (INTOSAI), and are audit reports on the final accounts provided to parliament in a timely manner? This would provide reasonable assurance on the qualitative characteristics of government accounts and their consistency with the financial reporting framework. A minimum timeliness standard for developing countries providing audited final accounts could be, for example, twelve months.
  • Are observations by the commercial and/or the state auditor addressed to the state treasury to improve financial management processes taken onboard systematically? Deficiencies in budget management and accounting process should be addressed by the State treasury as necessary. This process should be monitored by the external auditor or Public Accounts Committee.
  • Are adequate internal control procedures in place at the state treasury to safeguard public resources, generate reliable financial reports, and comply with applicable laws and regulations? The system of internal control in the state treasury should be documented and subject to regular review.

54. The panel supports adoption of a framework for targeted assessments of state treasuries. The panel, in particular, found merit in the proposed criteria for conducting such assessments, and noted that existing data could be leveraged so long as due diligence by staff ascertains the quality of data. It also notes that, consistent with existing practices for central banks, the assessments would need to be performed independently from other IMF operations with the state treasury.

55. If Directors see merit in targeted assessments of state treasuries, staff will revert with a specific proposal. The next steps could be to develop a prototype assessment, identify gaps in the availability of fiscal ROSCs and other information relative to recent budget support financing, propose specific modalities for the conduct of targeted assessments, and estimate the resource implications.

C. Confidentiality of Safeguards Reports

56. Staff believes that the confidentiality of the reports has been a key factor in the successful implementation of the policy. In particular, staff has been freely granted access to sensitive information (including those classified as state secrets in some countries), secured physical areas at central banks, and confidential client information contained in external auditor management letters. Further, the limited availability of safeguards reports encourages candor in assessment reports that might not otherwise be possible if reports were more widely disseminated. Publication could also undermine the Fund’s role as a confidential advisor.

57. The Fund’s transparency policy (recently amended) calls on the Fund to disclose documents unless there are “strong and specific reasons” not to. As noted by the panel, some central banks see publication as potentially beneficial, especially since transparency surrounding the actions taken by central banks to strengthen their ELRIC frameworks could enhance their accountability more generally. As noted, there has also been some limited interest shown by outside parties in the safeguards process and the results of safeguards assessments.

58. On balance, staff believes that safeguards reports should remain confidential at this time. It is true that publication of the reports could reduce the risk of misperception and misunderstanding of the nature and scope of the diagnostic assessments and promote a “race to the top” among central banks. Publication would also help promote accountability for the reports themselves and fulfill a need for the public good element of the reports. However, the benefits need to be weighed against the primary due diligence focus of assessment reports, which reflect staff appraisals used as internal inputs for IMF decision making. Also, some downside risks for publication of full reports remain:

  • Redactions would be needed for potentially market sensitive information. Reaching agreement on such deletions would deflect attention from the underlying objective of the assessments, and require an additional round of discussions with country authorities, which could delay reports and unnecessarily consume staff resources.
  • The reports are quite candid and include precise risk assessments by ELRIC category, which could be misinterpreted by external readers, who have different needs and objectives from the Fund.
  • Equivalent comparator reports (e.g., the “management letters” issued by external auditors and due diligence reports produced in the private sector for investment decision making) are typically not made publicly available.17
  • Collaboration with external auditors would likely deteriorate. Reports include an assessment of the quality of the most recent external audit and the international audit firms could react adversely to public criticism, which would impair the effectiveness of the policy in this area. It could also have potential legal implications if safeguards reports are seen as the reason for loss of clients, or if safeguards reports are used as evidence by litigants against the international firms.
  • Voluntary publication on a case-by-case basis, and with the consent of the central bank, has the potential for creating negative perception issues when a member chooses not to publish.

59. Nevertheless, there is scope for wider dissemination of safeguards findings, as a part of a phased move to enhanced transparency. While safeguards assessments are conducted for an internal purpose and contain staff judgments specific to the IMF context, the factual findings of the assessments have a public good element. Building on current practices, options for broader dissemination of the findings of individual assessments both internally and externally could be explored, for example:

  • As recommended by the panel, the safeguards assessment report format could be adapted to include a bulleted list at the beginning that identifies the top safeguards-related risk factors for lending. This list could also form the basis for what is included in staff reports, bringing a more structured reporting format of safeguards findings to the Executive Board and to the public, when staff reports are published.
  • Staff reports could be supplemented with an expanded standalone annex that describes the safeguards measures that are in place at the central bank, as well as a substantive summary of safeguards issues and measures agreed with country authorities.
  • The annual safeguards activity papers released to the general public could be expanded to include findings in an aggregate, anonymous form that reveals issues, not central banks, and indicate corrective actions taken, as recommended by the panel.
  • Formal confidential briefings could be provided to donors, if requested and with the consent of the central bank.
  • Extension of the current arrangement for sharing reports with the World Bank staff could be extended to other IFIs (e.g., the ADB, AfDB, EBRD, and IADB) could also be considered.

60. The panel also recommends that central banks be encouraged to voluntarily publish their own safeguards action plans and accomplishments. This step could serve to provide a “middle ground” for voluntary public disclosure that would not carry the same risks associated with voluntary publication of full reports. Since such information would be sourced from a confidential Fund document, it would be necessary for staff to agree on the modalities and a template for such disclosures to reduce the resource needs going forward for clearing individual disclosures.

V. Issues for Board Discussion

61. Directors may wish to focus comments in the following areas:

  • The appropriateness and effectiveness of the policy in helping mitigate the risks of misreporting and misuse and maintaining the Fund’s reputation as a prudent lender.
  • The proposals for updating the scope of the safeguards framework through a sharper focus on governance and more attention to risk management as central banks develop capacities in this area.
  • The proposed adaptations to safeguards modalities to enhance collaboration with stakeholders, including through engagement with the large international accounting firms to address audit quality issues.
  • The sufficiency of the steps proposed to be a standard part of budget financing cases (paragraph 50) and the merits of an assessment mandate in the context of budget financing that targets risks at the level of the state treasury.
  • The confidentiality of safeguards assessment reports and the options for wider dissemination of safeguards findings as part of a phased move to enhanced transparency.
Annex I: Assessments Completed by Country (Calendar Years 2001-March 31, 2010)18
2001200220032004200520062007200820092010
1622241912111215286
Albania

BEAC

Ethiopia

Kenya

Latvia

Lesotho

Lithuania

Madagascar

Malawi

Nicaragua

Nigeria

Pakistan

Peru

Serbia

Sri Lanka

Tajikistan
Albania

Argentina

Armenia

Azerbaijan

Bangladesh

BCEAO

Bosnia and

Herzegovina

Brazil

Brazil

Bulgaria

Cape Verde

El Salvador

Georgia

Guatemala

Guinea

Kyrgyz

Republic

Moldova

Mongolia

Nepal

Romania

Sierra Leone

Turkey
Bolivia

Colombia

Congo

Dem. Rep.

Croatia

Dominican

Republic

ECCB

Ecuador

Ghana

Guyana

Jordan

Kenya

Lao PDR

Lesotho

Macedonia

Malawi

Mongolia

Nicaragua

Paraguay

Rwanda

Sri Lanka

Tajikistan

Tanzania

Uganda

Uruguay
Argentina

BEAC

Belarus

Bolivia

Brazil

Bulgaria

Burundi

Cambodia

Georgia

Honduras

Mauritania

Mozambique

Nepal

Peru

Romania

Sao Tome & Principe

The Gambia

Ukraine

Zambia
Armenia

Bangladesh

BCEAO

Bosnia and

Herzegovina

Colombia

Croatia

Dominican

Republic

Haiti

Kenya

Kyrgyz

Republic

Turkey

Uruguay
Afghanistan

Albania

Burundi

Croatia

Iraq

Macedonia

Madagascar

Malawi

Moldova

Paraguay

Sierra Leone
Comoros

ECCB

Guinea

Guyana

Haiti

Liberia

Mauritania

Nicaragua

Peru

Rwanda

The

Gambia

Uganda
Afghanistan

Burundi

Cape Verde

Congo Dem.

Rep.

Georgia

Haiti

Honduras

Iraq

Lebanon

Liberia

Madagascar

Malawi

Mozambique

Seychelles

Tanzania
Armenia

BEAC

Belarus

Bosnia and

Herzegovina

Costa Rica

Djibouti

El Salvador

Ethiopia

Ghana

Guatemala

Hungary

Iceland

Kenya

Kyrgyz Rep.

Latvia

Lebanon

Mongolia

Nicaragua

Pakistan

Romania

Sao Tome & Principe

Serbia

Sierra Leone

Sri Lanka

Tanzania

The Gambia

Ukraine

Zambia
BCEAO

Cambodia

Georgia

Maldives

Mozambique

Pakistan
Annex II. The Safeguards Monitoring Framework

A risk-based monitoring methodology was put in place following the 2005 policy review. The safeguards policy calls for monitoring of central banks previously subject to an assessment for as long as the member country has Fund financing outstanding. Recognizing that safeguards assessments are a diagnostic tool that provides a snapshot of a central bank at a point in time, monitoring is the key tool for obtaining information on, and gauging the progress or sustainability of the measures being implemented by the authorities. The number of central banks subject to ongoing monitoring has increased over time. Some 72 central banks were subject to safeguards monitoring work at end-March 2010, of which 44 were low income countries.

The monitoring methodology aims at identifying emerging vulnerabilities in a central bank’s safeguards framework at the earliest possible stage. Experience has shown that vulnerabilities can emerge after the completion of an initial assessment. In this context, monitoring focuses on ex-ante prevention of misuse and misreporting through follow-up on the status of safeguards recommendations, and the identification of emerging new risks that may impair a central bank’s safeguards framework. Such monitoring is especially important during the term of an arrangement, when action can be initiated in the context of conditionality, or through the authorities’ commitments in LOI of MEFP documents to address any delays or obstacles in the authorities’ action plans.

The risk-based methodology assigns monitoring intensities to central banks and adopts uniform work plans based on a number of criteria. The starting point for monitoring activities is the existing information from earlier safeguards assessments, supplemented by information provided by the authorities and external auditors. From this, a Risk Mitigation Strategy (RMS) is developed that determines the intensity of monitoring based on an evaluation of individual central bank risks across a range of objective criteria.19 In the process, central banks are divided into four groups that determine the intensity of work programs and the length of the monitoring cycle.20 At a minimum, central banks are subject to desk reviews of financial statements, management letters and staff reports, along with ongoing contact with area department staff and central bank designated contacts.21

Monitoring can result in a reassessment of the safeguards in place. If during the course of monitoring a significant or potentially significant risk event arises, which may pose a threat to a central bank’s safeguards framework, a reassessment of the safeguards in place may be necessary or an update assessment may need to be initiated. In the absence of such developments, staff reviews the monitoring intensity for a central bank at the end of each cycle and a new work program is formulated. Where implementation rates for safeguards measures are high, monitoring intensity typically declines. This, in turn, releases staff resources to focus on high profile cases, such as those involving fraud or misreporting, and which require intensive monitoring and analysis of special or forensic audits and remedial actions agreed with the authorities.

Safeguards monitoring has been supported by an automated database since May 2008. The software, developed initially for monitoring compliance with Sarbanes Oxley, has been streamlined to facilitate safeguards specific requirements. Known as Paisley, the database provides a structured platform for guiding and documenting all safeguards work—from initial assessments to ongoing monitoring and any subsequent update assessments. It enables custom reporting on user selected data, and also supports a library of guidelines, best practice notes, case studies, and an archive of safeguards reports. A planned upgrade in 2010 will provide further “safeguards user-friendly” streamlining, and enhanced reporting for management oversight of progress with individual assessments, and their results.

Key aspects of recent ongoing monitoring activities have included:

  • A focus on high-risk areas to ensure adequate management of resources, in particular for identifying countries where there is a high likelihood of future Fund resources being provided under active financing arrangements with the Fund.
  • The identification and follow up of “risk triggers” that, based on previous safeguards experience, may indicate emerging safeguards issues.22 The flexible modalities help to ensure responsiveness to emerging issues and targeted reviews can be undertaken to identify or resolve safeguards issues that could impact the adequacy of the safeguards framework at the central bank. For example, the appointment of the Central Bank of Iraq auditor for 2008 and the commissioning of a special audit of NIR as of end-June 2009 was achieved through monitoring well in advance of a new program being in place and an update assessment being initiated.
  • During the response to the global financial crisis, update assessment work benefited from the existence of information readily available from the monitoring work as well as earlier assessments. For example, the design of the Stand-By Arrangement with Iraq, approved by the Board in February 2010, already included three safeguards-related prior actions to better protect the Fund.
  • Ongoing contact with internal and external key stakeholders, including area department mission teams, resident representatives, central bank officials, and external auditors.

The Safeguards Cycle

Annex III. High Profile Safeguards Cases: Experience and Remedies

Bank of the Central African States (BEAC)

The Bank of the Central African States (BEAC) is the regional central bank of the Central African Economic and Monetary Community (CEMAC).23 As with other regional central banks that handle monetary transactions for members who have an arrangement with the Fund, the BEAC is subject to a safeguards assessment every four years; the most recent assessment was completed in July 2009. While the assessment was underway, Fund staff received information indicating possible instances of unauthorized expenses in the BEAC’s Paris Office.24 A number of special audits were initiated by the BEAC, including one by the Audit Committee and Internal Audit, and one by an external audit firm that focused specifically on the Paris Office operations.

The audits revealed fraudulent transactions at the Paris Office, which involved a deliberate high-level override and break down of controls at both the Paris Office and BEAC headquarters. An Action Plan was approved by the BEAC Board in December 2009 to strengthen governance and internal controls, which included inter alia an audit of BEAC headquarters focusing on budgetary and accounting practices. This audit (completed in May 2010) revealed significant risk that unauthorized outflows from BEAC’s reserves could occur due to poor oversight and inadequate internal controls.

What went wrong:

  • The functions of the Paris Office expanded beyond its official role and were not subject to effective oversight by independent bodies. Over time the Paris Office became involved in certain operational activities that served as a backdrop for different schemes, consisting principally of fraudulent invoices and payments, and forging of documentation. The office was assigned directly to the Governor’s office, and neither the internal nor external auditors had access to it.
  • Earlier safeguards recommendations had not been implemented. Only limited progress was observed in the implementation of the recommendations of the 2001 and 2004 assessments. Key recommendations which were not acted upon aimed at strengthening Board oversight, increasing communication between external auditors and the audit committee, and implementing risk-based audit planning.

Remedial measures focus on implementation of time-bound action plans based partly on special audit findings, and heightened staff monitoring of progress and developments. Staff has adopted an approach that would allow program reviews and new programs for CEMAC countries to proceed as long as BEAC makes satisfactory progress in addressing the safeguards issues. Board consideration of program reviews for the Central African Republic and Republic of Congo scheduled for late May were postponed because of BEAC’s delays in implementing some actions. Staff has been engaged with BEAC and CEMAC authorities to assist them in addressing the underlying issues that have given rise to safeguard risks.

Burundi

The 2008 assessment was conducted amid allegations of a high-level override of controls at the central bank and the Ministry of Finance that facilitated significant misappropriations of government resources estimated at 1.6 percent of GDP, based on seemingly fraudulent documents submitted by officials at the highest levels of the Ministry of Finance and the central bank. The Inspector General’s report on the incident pointed to substantial irregular government payments made to a third party and highlighted improprieties that may have occurred in 2007.

What went wrong:

  • Controls were overridden at a high level. Apparent by-passing or override of internal controls at the central bank as well as the Ministry of Finance allowed for alleged significant misappropriations of government resources held at the central bank. Ex-ante controls over Fund resources and transfers at the time were found to be poorly applied or lacking, and authorized signatory protocols fell short of good banking practices.
  • Earlier safeguards recommendations had not been implemented. The fraudulent transactions occurred before the establishment of an effective and independent audit oversight mechanism—a priority recommendation of the 2006 safeguards assessment. Further there was no mechanism in place, such as incident reporting, to alert relevant stakeholders such as the Board of possible unauthorized transactions.

Remedial measures focused on restoration of ex ante and ex post controls. This case highlighted the importance of both ex ante controls (appropriate authorized signatory protocols requiring additional senior approvals as amounts increase) and ex-post controls (a risk reporting framework that regularly informs the audit committee, Board and senior management of significant control and audit concerns, and a risk based internal audit function). As a remedial measure special (semi-annual) audits of the controls over significant Fund resources and transfers were established until such time as the restoration of controls could be confirmed by the audits.

Tajikistan

In late 2007, the authorities provided information to the IMF that was subsequently confirmed as giving rise to misreporting in the context of the PRGF arrangement that concluded in February 2006 (the so-called cotton financing fraud). During 2007, the National Bank of Tajikistan (NBT) had pledged substantial portions of its foreign reserves, and issued numerous guarantees in favor of foreign commercial banks to facilitate loans to the cotton sector. The authorities commissioned a special investigative and forensic audit of the underlying transactions that gave rise to the misreporting, and to verify revised data provided by the authorities.

What went wrong:

  • The cotton financing fraud was fuelled by weaknesses at the central bank such as fragmented accounting, silo organizational structure, and inadequate internal audit. In addition, manual procedures for compiling program data posed a high risk of inadvertent error. The subsequent special audit uncovered repeated high level overrides of controls, widespread accounting fraud and forging of documents that included bank confirmations provided to the external auditors, and a lack of independent oversight.
  • The misreporting by Tajikistan came to light only upon notification to the IMF by the authorities. It appeared, therefore, that a key safeguards measure, namely the external audit of the financial statements, had failed to detect the underlying misreporting. Subsequent follow up, however, showed that the accounting records had been extensively falsified, information had been withheld from the auditors, and that controls were overridden at the highest level. It also transpired that the external audit firm had never in fact signed off on the purportedly final audit reports presented to staff.
  • There were protracted confirmed weaknesses in the safeguards framework of the NBT. Previous recommendations to strengthen the NBT’s governance and control framework had not been addressed. In particular, both external auditors and safeguards assessments repeatedly observed weaknesses that included: (i) absence of an independent Board; (ii) inadequate internal audit authority, coverage, and capacity; and (iii) absence of independent audit oversight directly accessible by the internal and external auditors.

Remedial measures focused on actions to rebuild the credibility of the NBT and to establish independent audits of financial statements and program data. The replacement of the NBT Chairman and the publication of the executive summary of the special audit report were helpful initial steps in addressing previous control overrides and rebuilding credibility. The bank also adopted an action plan to implement the recommendations of the special audit report, and appointed an international firm to audit the 2008 and 2009 annual

financial statements. On program data, independent reviews of the net international reserves data were introduced for each test date under the ECF arrangement. As a separate but related safeguard, the internal audit of the bank’s foreign exchange lending and liquidity support activities was outsourced to an external consulting firm to ensure adequate audit coverage in the near term and to build internal capacity over time.

Tanzania

A safeguards assessment of the Bank of Tanzania (BoT) was conducted in 2003. The assessment concluded that, at the time, the BoT had a relatively sound control environment, but also recommended that the central bank implement IFRS and appoint an external auditor with expertise in IFRS, ISA, and information systems audits.

During the course of the FY 2005/06 audit, the external auditor of the BoT raised concerns about a potentially serious fiduciary issue at the BoT; the auditor’s contract was then terminated by BoT management. Client confidentiality provisions prevented the audit firm from disclosing the details of the matter, but on the basis of the concern raised, the scheduled sixth review under the PRGF arrangement was delayed. The review was subsequently completed on the basis of a supplementary letter of intent from the Minister of Finance that explained the allegations of fraudulent transactions involving the External Payments Arrears account (EPA) and steps that would be taken to address them—in particular a forensic audit by an international audit firm on the basis of a terms of reference agreed with Fund staff. The forensic audit was completed in late 2007, identifying fraudulent transactions of about 1 percent of GDP.

What went wrong:

  • Controls were deliberately overridden at a high level. Excessive authority was vested in the Governor of the BoT who served as Governor, Chairman of the BoT Board, and Chairman of the Audit Committee.
  • Checks and balances in the control environment failed. While the BoT Board was informed of the auditors’ concerns, there was limited follow-up. Further, for much of the period during which the fraudulent transactions occurred, there was no head of the Internal Audit Department (IAD) with the staff of IAD both denied access to information and denied access to the Audit Committee.

Key measures to address governance and internal control weaknesses were taken upon receipt of the forensic audit report and subsequently strengthened via an updated safeguards assessment. In January 2008, the authorities issued a press release detailing the conclusions of the forensic audit and the steps that were underway to address these issues. Key steps taken were the dismissal of the then-Governor of the BoT, the launching of criminal investigations against all those involved, the appointment of a new Board of Directors of the BoT, and the reconstitution of the Audit Committee, which is now chaired by an external member of the Board with the Governor no longer a member.

Annex IV. Implementation of Safeguards Recommendations for the period March 2000 to end March 2010
Total number of recommendationsImplementation Rate (in %)
1. Recommendations with formal commitment from the authorities
a. Under program conditionality10596.2
Of which: Implemented101
Not Implemented4
b. LOI/MEFP commitments7792.2
Of which: Implemented71
Not Implemented6
2. Other recommendations824
Of which: Implemented60473.3
Not Implemented220
3. Total recommendations (1+2)100677.1
Of which: Implemented776
Not Implemented230
Of which: Overdue less than 3 months49
Overdue 3 months to one year94
Overdue more than one year39
Overdue more than two years48
Annex V. Staff Resources

Safeguards assessments work involves a mix of specialist professional staff that has averaged less than 15 FTE’s in recent years (Table 1). Mostly Chartered Accountants or Certified Public Accountants, the staff has been drawn from the professional accounting firms with external audit experience, as well as from member central banks. The safeguards assessment division in the Finance Department (“FINSA”) has historically been comprised of a Chief, Deputy Chief, 8-9 accountants, a research officer and three assistants. This composition changed during 2009 and 2010 as two additional accountants were hired on limited term positions to accommodate assessment activity emanating from the global financial crisis, and one assistant position was eliminated as part of the Fund-wide restructuring exercise.25 Staff resources for the evaluation of central bank legal frameworks are also provided by staff from the Legal Department (LEG).

Table 1FINSA staff in terms of FTE’s
Calendar Year200520062007200820092010 (March)
Total FINSA131314151717

Staff resources are invested in both fixed and variable elements of the safeguards work. Fixed elements, for example, represent work undertaken to perform ongoing safeguards monitoring, the generation of annual updates to the Board on safeguards activity, the delivery of semi-annual safeguards seminars, and incorporation of ongoing policy developments into safeguard procedures (e.g., those arising from the 2009 low income lending reforms, and FCL arrangements). The variable element reflects safeguards assessments of central banks, as evidenced by the temporary adjustment in staff levels to handle recent peaks in activity. Together, safeguards assessments and ongoing monitoring of central banks accounted for around 55 percent of total regular FINSA time over 2009-2010.

Safeguards work continues to involve a relatively small portion of total time reported by Fund staff in TRS terms (Table 2). The FINSA division team time, for example, represents about 0.4-0.6 percent of total Fund time reported in TRS. Also direct safeguards assessment work by the FINSA division represented 2.6-3.7 percent of total staff time allocated to the Use of Fund Resources over 2006-2010.

Table 2.FINSA and Fund UFR Time: TRS regular hours as percent of total
Calendar Year20052006200720082009Mar-10
FINSA as percent of Total Fund TRS0.40.40.40.50.60.6
Safeguards (total) as percent of Fund UFR2.32.62.73.43.73.7
Safeguards (country specific) as percent of Fund UFR1.61.51.92.72.92.9
FINSA overtime as percent of regular hours91011131521

Activity in 2009 and 2010 saw increases in safeguards assessments time in both absolute and relative terms. In addition to the time required to physically undertake the increased levels of assessment work, other factors also come into play:

  • Work pressures were addressed in a number of ways, as the number of central banks under assessment or monitoring increased from 55 to 72 (Figure 1) over 2009-2010. This included seven first time assessments, and update assessments of several higher risk countries. In addition to meeting work pressures by newly hired staff, as noted earlier, uncompensated overtime increased to around 15 percent of total regular hours in 2009, and 21 percent in the first three months of 2010. Streamlining and other modalities were also used to complete assessments.
  • Assessments and monitoring of high risk or “problem” cases, which can be particularly time consuming, has been a feature of recent safeguards work. During 2005–2009, for example, the proportion of total safeguards assessments time (regular and overtime) consumed by five central banks consumed ranged from between 30 to 40 percent (Table 3). The proportion jumped significantly in the first quarter of 2010 when number of high intensity country cases was underway at the same time.
    Table 3.Assessments and Mission Data 2005-March 2010(in days)
    Calendar Year20052006200720082009Mar-10
    Average number of mission days985766
    Number of initial assessments completed132341
    Number of update assessments completed1181012245
    Number of assessments completed (initial and update)12111215286
    Number of missions completed781012213
    Percent of total country time spent on the 5 most resource intensive central banks324041292360

The changing nature of safeguards work facilitated higher levels of assessment activity in 2009 (Table 3). The predominance of update assessments, supported by the monitoring database, has generally seen a streamlining of safeguards work. At the same time, access to information as a result of improved central bank transparency has led to lesser demands on country authorities for update information. The experience gained from earlier assessments coupled with information in the monitoring database has also enabled a greater number of update assessments to be undertaken at headquarters through a desk based review of information. Where missions have been necessary, those for both initial and update assessment are now smaller (2-4 people) and shorter than at the time of the last review. Update assessments also typically result in shorter reports, unless significant safeguards issues arise during the conduct of an assessment, or if the assessment is triggered by the unearthing of either misuse or misreporting.

Operating and administrative efficiencies have also been achieved. Desk based assessments, have been facilitated by utilizing tele and video conference facilities to meet with external auditors and central bank staff, thereby reducing both time in-field and travel costs per assessment. Translation budgets have also been controlled through use of machine based translation for documentation received from the authorities, and use of draft quality as opposed to “glossy” documents for final reports.

Annex VI: An Updated ELRIC Framework

Governance is an overarching theme of the ELRIC Framework. Governance has different meanings, depending on the context, but in the safeguards context is most closely associated with the good governance principles that are relevant for the structure of central bank entities.26

The attributes of good governance are cross cutting elements of all ELRIC pillars, and can be found in the use of the framework to assess whether a central bank’s auditing, financial reporting, internal control systems, and its legal structure and autonomy are adequate to ensure the integrity of operations and minimize the risk of misuse of resources or misreporting of data. In this regard, key attributes of good governance reflected in the ELRIC framework cover: (i) discipline, represented by senior management’s commitment to observe and promote good governance; (ii) transparency, which is necessary to facilitate effective communication to, and meaningful analysis and decision-making by, external parties; (iii) autonomy, which is essential for the top decision-making body such as the Board to operate without risks of undue influence or conflicts of interest, and is achieved through the application of objective criteria for the composition of, and appointments to, such bodies; (iv) accountability, whereby decision-makers such as the Board, its committees, and top management have effective mechanisms in place to report to, and as necessary discuss their actions with, the Parliament; and (v) responsibility, which should entail giving a high priority to observance of ethical standards and permit corrective action, including for mismanagement where appropriate.

The five ELRIC pillars and the main safeguards assessment objectives for each of these are as follows:

1. External Audit Mechanism. This encompasses the practices and procedures in place to enable an independent auditor to express an opinion on whether the financial statements are prepared in accordance with an established financial reporting framework, such as IFRS. This mechanism is important for the credibility of a central bank. The objective of assessing the external audit mechanism is to establish whether an independent external audit of the central bank’s financial statements is conducted regularly in accordance with internationally accepted auditing standards such as ISA, previous audit recommendations have been implemented, and to ensure that the external audit opinion is published with the full audited financial statements.

2. Legal Structure and Autonomy. Government interference in central bank operations undermines central bank autonomy and could increase the risks facing the central bank. A sound legal framework enshrines central bank autonomy and complementary transparency and accountability. The objective of assessing the legal framework and its application in practice is to: (i) establish whether the legal framework provides the central bank with an appropriate level of autonomy (including institutional and operational autonomy) along with adequate internal and external checks and balances; (ii) ascertain whether key legal requirements are complied with without interference or override; (iii) clarify if other legislation exists that could impair central bank autonomy; (iv) determine whether the respective roles and responsibilities of the central bank and other agencies are transparently and explicitly defined in cases of shared monetary authority; and (v) ascertain that the legal framework supports the other four ELRIC pillars.

3. Financial Reporting Framework. Adequate financial reporting principles and practices are essential elements of effective central bank operations that encompass the provision of financial information to both central bank management and to external parties, the latter typically through published interim and annual financial statements. For such information to be useful, it must be relevant, reliable, timely, readily available, consistent in presentation over time, and based upon recognized standards, such as IFRS. Non-adherence to accepted international practices could indicate a lack of transparency or accountability. The objective of assessing financial reporting is to ensure that the central bank adheres to international good practices in the adoption of accounting principles for internal reporting to management, and the published annual and interim financial statements.

4. Internal Audit Mechanism. Internal auditing is an independent, objective assurance and consulting activity designed to add value and improve an organization’s operations. It helps an organization achieve its objectives by bringing a systematic, disciplined approach that adheres to international standards, to evaluate and improve the effectiveness of risk management, control, and governance processes. The objective of assessing a central bank’s internal audit function is to determine whether internal audits are performed in accordance with international standards, such as International Standards for the Professional Practice of Internal Auditing, and whether the function is assigned sufficient independence and authority to fulfill its mandate effectively, including whether procedures exist for communicating results without interference.

5. System of Internal Controls. Internal control is a process put in place by an entity’s board, senior management, or staff with the intent to provide reasonable assurance regarding the achievement of objectives in the following categories: (i) effectiveness and efficiency of operations; (ii) reliability of financial reporting; and (iii) compliance with applicable laws and regulations. A sound internal control system is a key component of an entity’s risk management framework and essential in safeguarding the integrity of its operations, resources, and reputation. The objective in assessing the internal control system in a central bank is to ascertain the quality of high level governance and oversight, employee integrity, and the bank’s commitment to building and maintaining internal competence. The assessment looks in particular at areas of high importance for central bank operations and of significant relevance to a Fund arrangement, including controls in the areas of reserves management, accounting, and currency and banking operations, and measures in place to ensure accurate and timely reporting of monetary program data.

1

The panel comprises Mr. Lynn E. Turner (Chair), current Senior Advisor to an international forensics and economic consulting firm, former Chief Accountant of the Securities and Exchange Commission, and member of the 2000 expert panel for the IMF safeguards policy; Dr. Ranee Jayamaha, current Advisor to His Excellency the President of Sri Lanka and former Deputy Governor of the Central Bank of Sri Lanka, and Dr. Len Konar, current Director and former Chair of the Audit Committee of the South African Reserve Bank and former Chair of the External Audit Committee of the IMF. Panel members were subject to confidentiality obligations.

2

A further 27 abbreviated assessments of the external audit mechanism only were conducted in 2000-2002 under transitional procedures during the experimental phase. Annex I lists completed assessments.

3

A recent example is the SBA for Iraq where known safeguards measures were addressed at an early stage, including through conditionality at the time of the approval of the arrangement, and before completion of the update safeguards assessment.

4

Such discussions are held as early as practicable. In four of the five FCL arrangements approved so far, the discussions were concluded prior to the Executive Board’s consideration of the FCL arrangement. In the fifth case, the timing of the audit cycle was such that staff was able to hold such discussions only after the approval of the arrangement.

5

The term “budget financing” in this paper is used to refer to cases in which a member uses the domestic counterpart of resources received from the Fund to finance the budget deficit of the government. This type of use of Fund resources is consistent with the Fund’s legal framework to the extent that the member has balance of payment problems and it is implementing a program that will assist it in resolving such problems.

6

In previous safeguards policy reviews, the Executive Board has re-affirmed the focus on central banks, even though Fund lending can involve budgetary financing. Section IV B (below) takes up this issue in the context of the current review.

7

Framework agreements between a central bank and the MOF have typically provided for the maintenance of a government account at the central bank to receive incoming loan disbursements, controls (authorizations) for operating the account (i.e., budgetary use of the funds), and responsibilities for servicing and repaying obligations to the Fund.

8

CEMAC is the Central African Economic and Monetary Community, whose membership comprises Cameroon, Central African Republic, Chad, Republic of Congo, Equatorial Guinea, and Gabon.

9

Detailed findings have been reported to the Board in the annual activity papers.

10

Eight categories of vulnerabilities were identified in 2002, drawing on the assessment work in the initial two years of the policy. These same categories were reported in 2005 (see Table 2 of The Acting Chair’s Summing Up (04/28/05).), and have been updated for 2010 with two changes: the category dealing with the reporting of Fund transactions is now subsumed within the financial reporting category, given the move towards IFRS by many central banks; and a new category covering data reporting has been developed.

11

In this context, priority recommendations are defined as those that have been included under program conditionality, as well as commitments by country authorities in their in Letters of Intent (LOI) and Memoranda of Economic and Financial Policies (MEFP).

12

The scope of external audits is restricted to controls surrounding the preparation of financial statements, and the audit evidence collected by auditors does not necessarily require a testing of those controls.

13

For example, the safeguards teams have been able to address potential misreporting issues by advising on inconsistencies found in the data regarding TMU definitions before program approval.

14

The team has averaged around 14 FTEs since 2005, including the divisional management and support team. Other departments (area, MCM, LEG, SPR, STA, and OIA) are involved in the review of safeguards reports. LEG also conducts desk reviews of central bank laws to identify safeguards issues and occasionally participates in safeguards missions.

15

The documentation comprises financial statements, management letters issued by the external auditors, and authorization to hold discussions with the external auditors.

16

Some exceptions may be warranted, e.g., for embassy working balances.

17

The World Bank’s fiduciary assessments for DPL lending are publicly available, although these are not a direct comparator given the breadth of entities covered and the more general analysis.

18

In addition, 27 transitional assessments were completed during 2000-02 for: Argentina, Bolivia, Bosnia and Herzegovina, Bulgaria, Cambodia, Colombia, Djibouti, Estonia, Ghana, Guyana, Honduras, Indonesia, Jordan, Latvia, Lithuania, Mauritania, Mozambique, Panama, Papua New Guinea, Romania, Sao Tome & Principe, Tanzania, Turkey, Ukraine, Uruguay, Yemen, and Zambia. In 2009, three central banks were subject to FCL procedures: Colombia, Mexico and Poland.

19

Classification criteria used to assign risks to monitored central banks include, for example, Fund exposure, country corruption index, past experience, status of recommendations, audit quality, governance, and legal changes, and the results of any update assessments.

20

Monitoring intensities that are rated as High or Medium-High have a 12 month cycle; those rated Medium-Low and Low have an 18 month cycle.

21

Accompanying steps may include discussions with the external auditors, review of internal audit reports, discussions with functional departments and technical assistance experts, meeting the authorities during IMF Spring or Annual meetings, and on site staff visits. Staff may also request additional documents including the use of periodic questionnaires to obtain relevant information on the status of the safeguards framework at the central bank.

22

Indicators of emerging safeguards issues include: (i) unexpected changes in external audit mechanisms, such as dismissal of the auditor; (ii) governance, control or financial reporting issues raised by internal or external auditors; (iii) delayed publication of audited financial statements; and (iv) unexpected changes in the governance structure or legal framework.

23

The member countries of BEAC are: Cameroon, Central African Republic, Chad, Republic of Congo, Equatorial Guinea, and Gabon.

24

Officially the Paris Office serves as the representative office for BEAC and its members.

25

During 2009 for example, internal staff secondment through mobility was utilized.

26

In this regard the governing framework of central banks can be guided by general principles of good governance (such as those enunciated in the OECD Principles of Corporate Governance), appropriately modified to fit the situation of central banks operating as autonomous agencies of the State.

Other Resources Citing This Publication