SUNANDO ROY – On Banking, Finance and Society

Link : https://sunandoroy.org/fsap-preparedness-assessment-process-and-tools/

Most of what is written about the Financial Sector Assessment Program addresses its content: what the International Monetary Fund and the World Bank assess, against which standards, using what methodology. Comparatively little addresses its management. That gap matters, because the difference between a well-run assessment and a badly-run one is often not primarily technical. Central banks with broadly comparable supervisory frameworks can produce markedly different assessment outcomes, and differences in governance, resourcing and coordination can materially affect how supervisory practice is evidenced and presented to assessors.

This paper treats the management of an assessment as a distinct discipline with three components. The first is governance: who inside the central bank owns the engagement, on what authority they act, and how the board holds them to account. The second is operational: the machinery that runs the exercise, from document control to the handling of draft reports. The third is coordination: internally between directorates, externally with other domestic authorities, and with the mission itself.

The argument advanced here is that these are not administrative matters attached to a technical exercise. They form part of what is being assessed, and they largely determine whether the central bank’s supervisory practice is visible to an external reader at all.

The paper also situates FSAP management within the broader architecture of IMF surveillance. An FSAP is episodic, but its findings are carried forward through the regular Article IV consultation cycle, and much of the management challenge arises from that interaction between a periodic in-depth assessment and a continuous surveillance relationship.

The remainder of the paper is organised as follows. Section I introduces the Article IV consultation and the FSAP, traces the development of the FSAP since its launch in 1999, and explains how the two instruments interact. Section II sets out why the management of an assessment affects its outcome. Section III addresses governance, including board oversight, executive sponsorship and the scoping mandate. Section IV examines operational arrangements: the coordination function, evidence and data, and the handling of draft reports and publication. Section V turns to coordination within the central bank, with other domestic authorities and with the mission. Section VI discusses the accountability loop that follows an assessment, Section VII summarises recurring failure modes, and Section VIII concludes.

I.

Article IV Consultations and the FSAP: Background and a Brief History

Two IMF instruments frame a central bank’s engagement with the Fund on financial sector matters: the Article IV consultation, which is recurring, ordinarily annual and applies to every IMF member, and the Financial Sector Assessment Program, which is periodic and in depth. Both sit within a surveillance relationship between the Fund and its members that is itself continuous. Understanding how each works, and how they connect, is a precondition for managing either well.

1. The Article IV Consultation

The Article IV consultation takes its name from Article IV of the IMF’s Articles of Agreement, under which members undertake obligations regarding their economic and exchange rate policies and the Fund conducts surveillance over them. It is the principal vehicle of the Fund’s bilateral surveillance and is ordinarily conducted on an annual cycle.

The process follows a settled sequence. An IMF staff team visits the member, collects economic and financial information and discusses developments and policies with the authorities, typically including the finance ministry and the central bank. On completing its evaluation, staff present a report to the Executive Board for discussion, and the Board’s views are then transmitted to the authorities, concluding the consultation (IMF, n.d.-c). The consultation covers fiscal, monetary, exchange rate and financial issues (IMF, n.d.-c), and where the member consents the Fund publishes a press release summarising the Executive Board’s assessment, often together with the staff report.

For a central bank, the Article IV consultation is the recurring channel through which the Fund’s view of financial sector policy is expressed in the years between in-depth assessments. Published Article IV statements illustrate the point: they comment on the macroprudential framework, crisis management, supervisory arrangements and the implementation of earlier FSAP recommendations (IMF, 2019a, 2025).

1. The Financial Sector Assessment Program

The FSAP, established in 1999, is a comprehensive and in-depth assessment of a country’s financial sector (IMF, n.d.-b). In advanced economies it is conducted by the IMF alone; in emerging market and developing economies it is ordinarily conducted jointly with the World Bank (IMF, 2010, 2021a), and comprises two components: a financial stability assessment, which is the main responsibility of the Fund, and a financial development assessment, which is the main responsibility of the World Bank. The two components may also be delivered by the two institutions in separate modules (IMF, 2014).

Each FSAP concludes with a Financial System Stability Assessment (FSSA), which focuses on issues relevant to IMF surveillance and is normally discussed by the Executive Board together with the country’s Article IV staff report (IMF, 2014). Recommendations are set out in an aide-mémoire, a confidential and comprehensive document left with the national authorities at the end of the final mission (IMF, n.d.-b). The stability assessment is organised around three pillars — risk analysis, oversight and safety nets (IMF, 2021a) — and assessments of observance of international standards are voluntary for all members (IMF, n.d.-a).

1. A Brief History of the FSAP

Launch, pilot and early evaluation, 1999 to 2006

The FSAP was launched jointly by the IMF and the World Bank in 1999, in the aftermath of the Asian financial crisis (IMF, 2009, 2010). An initial pilot phase was reviewed by the staff of the two institutions in November 2000, in a paper setting out lessons from the pilot and issues going forward, including country coverage, methodology, the prioritisation of work, follow-up and the relationship of the FSAP to other work of the two institutions (IMF & World Bank, 2000).

The Executive Board reviewed the programme again in March 2005 (IMF, 2009). In 2006 the IMF’s Independent Evaluation Office published an evaluation of the programme’s effectiveness from the Fund’s perspective, alongside a parallel evaluation of the World Bank’s role (Independent Evaluation Office of the International Monetary Fund [IEO], 2006).

The 2009 review and the mandatory regime

In September 2009, in the aftermath of the global financial crisis, the Executive Board reviewed the programme’s first ten years and agreed on steps to strengthen it and to integrate financial sector analysis more fully into surveillance (IMF, 2009). The reforms that followed the 2009 review strengthened the analytical focus and coverage of FSAPs, clarified the respective responsibilities of the Fund and the Bank in developing and emerging market countries, and introduced the option of modular assessments, affording greater flexibility on scope and timing (IMF, 2014). At that stage most Directors still considered that participation should remain voluntary, while stressing the importance of participation by all systemically important economies (IMF, 2009).

A decisive change followed a year later. In September 2010 the Executive Board decided to make financial stability assessments under the FSAP a regular and mandatory part of surveillance under Article IV for members with systemically important financial sectors (IMF, 2010). The assessments were expected to take place every five years in 25 jurisdictions; the list was expanded to 29 jurisdictions in 2013, and for all other jurisdictions participation remained voluntary (IMF, 2014).

The 2014 and 2021 reviews

The 2014 review focused on the Fund’s role and responsibilities in the programme, the contribution of FSAPs to surveillance, and the programme’s broader traction and impact, and drew on a special study by independent external consultants on the even-handedness of assessments (IMF, 2014).

A 2019 evaluation of IMF financial surveillance by the Independent Evaluation Office called for further integration of the FSAP with Article IV consultations and for the frequency of assessments to be made more risk-based (IMF, 2021a). The 2021 review, concluded in May of that year, found that the three-pillar approach had worked well, encouraged greater use of flexibility in scoping, and recorded that 157 Fund members had undergone an FSAP since the programme’s inception (IMF, 2021a). The Executive Board also reconfirmed that the 2013 methodology for identifying jurisdictions with systemically important financial sectors remained adequate, with minor adjustments (IMF, n.d.-a).

Recent developments

Two recent developments shape the environment in which central banks now prepare. The first is the 2024 revision of the Basel Core Principles for Effective Banking Supervision, the principal standard against which banking supervision is assessed within an FSAP (BCBS, 2024; IMF, 2024a). The second is a further periodic review of the programme, concluded by the Executive Board on 3 September 2026, which is reported to move the FSAP toward a more risk-based and modular approach, with greater emphasis on system-wide and emerging risks where appropriate (“IMF Revamps Financial-Sector Assessments,” 2026). Both are discussed further in Section III.

1. How the Two Instruments Interact

The FSAP and the Article IV consultation are formally distinct but operationally intertwined. The FSSA is normally discussed by the Executive Board together with the Article IV staff report (IMF, 2014); for jurisdictions with systemically important financial sectors, the financial stability assessment is itself a mandatory part of Article IV surveillance (IMF, 2014); and the 2019 evaluation and the 2021 review both pressed for closer integration of the two (IMF, 2021a).

The most consequential link for a central bank is follow-up. Recommendations arising from an FSAP are revisited in subsequent Article IV consultations, and the Executive Board’s published assessments record progress, or its absence, year by year. An FSAP may therefore be episodic in occurrence, but it is continuous in effect: the Article IV cycle carries its findings forward, often until the next assessment.

This interaction underlies the argument of the rest of the paper. Because FSAP recommendations remain in view through the Article IV cycle, and because the next assessment examines the evidence accumulated since the last, the management of an FSAP cannot sensibly be confined to the months in which a mission is present.

1. Why Management Affects Outcomes

The case for treating management seriously rests on two observations about how assessments work in practice.

1. The Evidence Problem

By the time a mission arrives, assessors are examining evidence of what the supervisor has been doing over the preceding three to five years. Management cannot change what that evidence is. It can determine whether the evidence is retrievable, whether it is presented coherently, whether the central bank’s account of itself is internally consistent, and whether what the mission hears from one directorate matches what it hears from another.

The Malta assessment illustrates how directly this bites. Supervisory resourcing was found inadequate not through an argument about headcount but through what had not happened: inspections planned and then postponed, recovery plans and internal capital and liquidity adequacy assessments awaiting review, difficulty maintaining contact with smaller institutions, and stated supervisory priorities that had not been translated into executed action (International Monetary Fund [IMF], 2019c).

That is as much a management finding as a capacity one. A central bank that routinely tracked execution against its own published supervisory priorities would have identified the same gap first, and could have arrived at the mission either having closed it or having documented why it remained open.

1. The Persistence Problem

Assessment recommendations do not expire. They are revisited in subsequent Article IV consultations and recorded in published staff reports and Executive Board assessments. Kuwait’s 2019 assessment published a status table for its 2010 recommendations, attributed by responsible agency, in which items recorded as unimplemented included insurance recommendations addressed to a ministry rather than to the central bank (IMF, 2019b). Saudi Arabia’s 2024 assessment carried an equivalent appendix for its 2017 recommendations, and the subsequent Article IV consultation commented on the progress made (IMF, 2024b, 2025).

Published assessments commonly include an account of what became of prior recommendations. The practical choice for the central bank is therefore between preparing its own account first and responding to one prepared by others.

1. The Implication

If both observations hold, the management of an assessment is not a project that begins when a mission is announced. It is a standing capability whose principal outputs — a maintained self-assessment, a retrievable evidence base, a reconciled data spine and a complete record of prior recommendations — accumulate over years. The governance and operational arrangements discussed below exist to produce and protect those outputs.

III.

Governance

Governance determines who owns the engagement, on what authority decisions are taken, and how the board holds management to account. Two aspects are addressed: the respective roles of the board and executive management, and the scoping mandate.

1. Board Ownership Without Board Interference

The first governance question is where accountability sits. A workable answer separates two things that are easily conflated: the board owns the engagement’s outcomes and holds management to account for them, while management owns what the central bank says.

The board’s proper functions

A board’s role is best confined to a defined set of decisions: approval of any request for a voluntary assessment; approval of the publication decision; receipt of the self-assessment and of the recommendation matrix; periodic reporting on implementation; and approval of any formal decision not to implement a recommendation.

What a board should not do is direct the content of submissions or the central bank’s factual representations to a mission. That line matters because the candour an assessment requires is institutionally fragile. A supervisory directorate that expects its self-assessment to be revised upward before it reaches the board will write for that audience rather than for the external one, and the gap between the central bank’s self-assessment and the assessors’ conclusions is likely to widen.

Why the gap itself is informative

Guidance on the programme treats a self-assessment conducted in advance as an important input, to be prepared in the format prescribed for assessors and provided with the supporting legislation and regulation (IMF & World Bank, n.d.). Assessors form their own view regardless. The distance between the two views therefore becomes a signal about the institution’s self-awareness, and it can colour the narrative sections of the published report, which are generally harder to contest than an individual grade.

A central bank that grades itself compliant across the board, and is then assessed with several materially non-compliant grades, has not protected itself. It has revealed something about its capacity for self-criticism to an audience already looking for it.

A standing executive sponsor

Continuity favours designating a deputy governor as sponsor on a standing basis rather than appointing one for each assessment. The obligations that shape outcomes run between assessments, and a sponsor appointed only when a mission is announced inherits a position rather than having shaped it.

Senior presence during the mission is substantive rather than ceremonial. Assessments examine mandate, independence, supervisory strategy and resourcing — institutional matters that technical staff cannot fully address. Where governors and deputy governors appear only at opening and closing sessions, the mission may draw an inference about the priority attached to supervision, and that inference can find its way into the report.

Fixing approval authorities in advance

An assessment generates a sequence of consequential decisions under time pressure: whether and when to disclose an adverse development, how to respond to a contested finding, whether to publish a detailed assessment. Assigning the approval authority for each before the exercise begins prevents the decision from defaulting to whoever happens to be available.

The decisions that warrant explicit assignment are the request for a voluntary assessment; the scoping position; the self-assessment as submitted; disclosure of adverse developments; comments on draft reports; the authorities’ response; the publication decision; and any formal deferral of a recommendation.

1. The Scoping Mandate

Scoping deserves separate treatment because it is the phase in which the central bank holds the most negotiating leverage, and it is often delegated downward and treated as a logistical exchange.

The basis for a negotiating position

The Fund’s own framework supports engagement. The periodic review concluded in 2021 endorsed the three-pillar structure of risk analysis, oversight and safety nets, and encouraged greater use of flexibility in scoping, with continued tailoring to country specifics, effective prioritisation and close consultation with country authorities (IMF, 2021a). Assessments of compliance with international standards are, separately, voluntary for all members, including those for which the stability assessment is mandatory (IMF, n.d.-a).

A further periodic review, concluded by the Executive Board on 3 September 2026, is reported to move the programme toward a more risk-based and modular approach, with the scope of individual assessments shaped by stronger initial diagnostics of the most material vulnerabilities, greater emphasis on system-wide and emerging risks where appropriate, and coverage tailored to national circumstances (“IMF Revamps Financial-Sector Assessments,” 2026). That direction increases rather than diminishes the value of a prepared scoping position: the more assessments are tailored to country-specific diagnostics, the more the central bank’s own diagnosis matters.

What is negotiable

Several matters are open and should be settled in writing: whether a standards assessment is undertaken and, if so, whether full, targeted or none; against which standard, where more than one is available; the institutional perimeter and architecture of stress testing; which technical notes are produced, each being a separate publication; the scope of any development module where the World Bank participates; and the timing of the exercise relative to the central bank’s own reform pipeline.

The choice of standard is more consequential than it may appear. Saudi Arabia’s assessment combined an assessment against the Basel Core Principles with a focused review against the Core Principles for Islamic Financial Regulations, and concluded that the materiality of several of the Islamic principles was low given the jurisdiction’s actual product and funding mix (IMF, 2024b). A central bank supervising a significant Islamic banking sector has a real decision to make here, and the evidence for its position is quantitative.

Preparing the position

A position paper cleared at governor level before the scoping engagement is the practical mechanism. It should state what the central bank wants in scope, what it wants out, and why; identify the findings that would be most consequential if published, with a decision on whether the strategy is to close, contextualise or accept each; and record the read-across from published assessments of comparable jurisdictions.

That last element repays the effort. Where a central bank believes a standard is being applied to it differently from comparable jurisdictions, the argument is legitimate — but it works only where specific treatment of specific principles in a published assessment can be cited. An unsupported assertion of inconsistency is counterproductive.

The cost of delay

Timing is a legitimate consideration, and there is nothing improper in proposing a window in which an assessment will observe reforms in operation rather than in passage. The United Arab Emirates took precisely this position: in 2019 the IMF mission welcomed the authorities’ request for an assessment once the new banking law had been fully implemented (IMF, 2019a).

The counterweight is that the longer the interval since a previous assessment, the wider the eventual scope is likely to be. A jurisdiction whose last full assessment predates the 2024 revision of the Core Principles will be assessed against a standard for which no transitional period was provided, and which expects demonstrable supervisory practice in areas — climate-related financial risk, operational resilience, non-bank financial intermediation and business model sustainability — where many authorities hold only policy statements (Basel Committee on Banking Supervision [BCBS], 2024; IMF, 2024a).

IV.

Operational Arrangements

Governance assigns accountability; operational arrangements determine whether the exercise actually runs. Three sets of arrangements are considered: the coordination function, evidence and data, and the handling of outputs.

1. The Coordination Function

A single channel

The foundational operational decision is that all communication with the mission passes through, or is recorded by, one function. Where mission members can approach individual analysts directly for data, the central bank loses control of its own evidentiary record: material reaches the team under the institution’s name that the institution cannot subsequently identify, reconcile or stand behind.

A coordination unit of one named coordinator and one named deputy, reporting to the sponsor, is sufficient for most exercises. The realistic resourcing assumption is that the coordinator will lose the substantial majority of their existing role for the duration, and that the role must therefore be backfilled. Appointing a coordinator on the assumption that the work can be absorbed alongside a full existing workload is a recurring resourcing error, and it tends to surface as missed deadlines in the inter-mission period when attention has already moved on.

Document control

Every item provided to the mission should carry a reference, a date, a title, a version and — for data — the reporting date and vintage, an owner, a recipient and a retained copy.

The version and vintage field is frequently omitted and is particularly important. The scenario it protects against is specific: months after the mission, a published technical note quotes a figure back at the central bank. Establishing which submission that figure came from, and what vintage of data it reflected, is either a quick lookup or a reconstruction from memory across a team that has partly turned over.

The outstanding items log

A log of what was asked and what was promised, with owner and due date, should be worked to zero before each mission departs. Outstanding commitments at the end of a mission are a frequent source of friction in the inter-mission period, and an item that cannot be delivered is better explained while the team is still present than left open in their record.

Records of meetings

A note-taker should attend every session, with notes circulated internally within the working day. The purpose is not archival. The aide-mémoire and each draft report should be read against the central bank’s own contemporaneous record rather than against recollection, so that factual errors are identified before they harden into published text.

Correcting an error in the room, and confirming the correct position in writing the same day with the supporting extract, is straightforward. Correcting the same error at the draft comment stage is generally harder.

1. Evidence and Data

Two operational disciplines account for a disproportionate share of assessment outcomes, and neither can be established within an assessment cycle.

Evidence as a byproduct of supervision

The Core Principles are assessed criterion by criterion, and the assessment considers both the formal legal and regulatory framework and evidence that supervisory powers and practices are implemented effectively. Many central banks can evidence the first comprehensively and find the second harder — not because they do not supervise, but because their supervision is not readily retrievable.

A practical test is simple. If an assessor selects three inspections at random from the past two years, does each file show a documented rationale for scope, a record of the judgments made and the reasoning behind them, the findings as communicated, the remediation obtained, evidence of follow-up and verified closure? Files that record conclusions without reasoning are difficult for an external reader to distinguish from assertions.

The remedy is structural rather than documentary: inspection templates that require reasoning fields as well as conclusion fields, supervisory systems that capture judgments as they are made, and committee secretariats that minute deliberation rather than only decisions. All three are worth doing for internal quality assurance, staff induction and the defensibility of decisions on appeal. The assessment benefit is incidental, which is precisely why it is reliable.

A reconciled data spine

A finding on data quality is particularly damaging because its effects are rarely confined to one workstream. Once a mission loses confidence in the figures, subsequent submissions are read with caution and the analytical sections acquire qualifications that are difficult to remove at the comment stage.

Saudi Arabia’s assessment is explicit about the consequences. Default rate series were volatile and did not cover a major recession, household and corporate micro data were unavailable, and missing institution-level exposures between banks, state-owned financial institutions and non-banks limited the assessment of contagion and concentration risk. Six of the twenty-eight key recommendations fell under systemic risk analysis and monitoring (IMF, 2024b). Kuwait was separately directed to fill gaps in household debt data and to develop a real estate price index (IMF, 2019b).

The operational discipline is a single authoritative source for each key series with a named owner, and a written protocol reconciling those series across the prudential returns, the published financial stability report, Financial Soundness Indicator submissions and the data annexes to Article IV staff reports. Divergence is not itself a problem: definitions, perimeters and consolidation bases legitimately differ. Divergence the central bank has not identified, cannot explain and discovers under questioning is the problem.

1. Handling Outputs

Reviewing draft reports

Comments on draft reports are most effective when separated into three categories and kept separate: factual corrections, with the evidence establishing the correct position; reasoned disagreement, each stated as a proposition, its basis and its supporting evidence; and accepted findings, with an owner, a remediation plan and a date.

Mixing the categories dilutes all three, and a response consisting of general assurances carries little weight. A further discipline is to contest facts rather than grades. Arguing that a materially non-compliant grade should be largely compliant is an argument about the assessor’s judgment. Demonstrating that a specific essential criterion is met, with the legal citation and the inspection evidence, is an argument about facts — and the grade may follow.

The authorities’ response

Saudi Arabia’s published response is instructive as a model of form. The authorities welcomed the recognition of progress since the previous assessment; argued that the Core Principles assessment should better reflect the strengths and performance of the supervisory approach in practice; argued that certain macroprudential recommendations could have been better tailored to the conjuncture and to mitigants arising from fiscal buffers; set out a position on the prudential supervision of development funds with a stated condition and sequence; and identified a specific legal power that mitigated a limitation the assessment had perceived in the resolution law (IMF, 2024b).

That is specific, reasoned disagreement placed on the public record. It preserves both the relationship and the position.

In-flight remediation

Where a finding is accepted and closable, closing it — or demonstrably beginning to — during the assessment period is one of the more effective ways to influence an outcome, because assessors can take account of action taken during the exercise. This argues for triaging emerging findings immediately after the first main mission into those closable within months and those that are not, and directing resources to the first category while it can still count.

Publication

Publication of the stability assessment is presumed but remains voluntary; publication of technical notes and detailed assessment reports is voluntary (IMF, n.d.-a). The arguments run in both directions and should be weighed explicitly rather than by default.

In favour: a published assessment with credible grades and a clear reform agenda is a reputational asset, supports correspondent banking relationships, and provides external validation for reforms the central bank wishes to pursue. Against: a published detailed assessment containing several materially non-compliant grades is durable, and may remain relevant for years to foreign supervisors assessing cross-border applications and to counterparties conducting jurisdiction reviews.

A consideration that is easily underweighted is that non-publication is itself a signal. In a jurisdiction that presents itself as committed to regulatory transparency, declining to publish invites inference. The decision is therefore between two communications rather than between communicating and not.

V.

Coordination

Coordination operates at three levels: within the central bank, with other domestic authorities, and with the mission itself. Failures at each level are visible to assessors in different ways.

1. Coordination Within the Central Bank

Internal coordination has a specific failure mode that assessments tend to detect: the central bank speaks with more than one voice.

Workstream ownership

Assigning a named lead at deputy director level or above to each workstream — banking supervision, insurance, capital markets, financial market infrastructures, systemic risk and stress testing, macroprudential policy, systemic liquidity, crisis management and resolution, anti-money laundering, and any thematic area in scope — establishes accountability for the accuracy of submissions, for the workstream’s portion of the self-assessment, and for briefing every member of staff who will attend a mission meeting.

The distinction worth preserving is that directors retain responsibility for the conduct of their functions. A workstream lead’s role is to evidence and present that conduct, not to assume ownership of it.

The exposure of integrated regulators

A central bank that supervises more than one sector is assessed partly on whether integration delivers anything. Assessors look for a consistent approach where consistency is appropriate, differentiation where differentiation is appropriate, effective information flow between directorates, and a group-wide view of financial conglomerates operating across licensed sectors.

Fragmentation inside an integrated regulator is a finding in its own right, and it surfaces in ordinary ways: two directorates supervising related entities of the same group that have not exchanged information, or inconsistent characterisations of the same licensee in different submissions.

Briefing without scripting

Every participant in a mission meeting should be briefed on the subject, the likely lines of questioning, the relevant submissions and what is already in the public domain. Supervisors who improvise in front of assessors can create findings.

The line that should not be crossed is scripting. Preparing staff to discuss their own work accurately is legitimate; rehearsing answers is not, and experienced assessors are likely to notice the difference. The same applies with greater force to the industry: briefing licensees on the process and timetable is appropriate, briefing them on the substance of answers is not.

1. Coordination With Other Domestic Authorities

An assessment reaches well beyond the central bank. The mission will ordinarily engage the ministry of finance, any separate securities or insurance regulator, deposit protection arrangements, payment system operators, the statistical authority, state financial institutions and, where they exist, regulators of financial free zones.

What assessors actually test

Coordination is tested by joint work rather than by memoranda. Hong Kong’s assessment encouraged domestic sectoral regulators to conduct joint examinations of cross-sectoral business to promote consistency (IMF, 2021b). Saudi Arabia’s found that cooperation arrangements should be made more comprehensive and should promote proactive periodic sharing and exchange of data rather than remaining reactive (IMF, 2024b).

A memorandum of understanding is not, by itself, evidence of coordination. Joint examinations, scheduled data exchange and recorded escalations are.

Practical arrangements

Named counterparts at each authority, agreed in writing before the first mission, together with an understanding of which authority speaks to which subject, helps prevent a damaging outcome: inconsistent representations between domestic authorities, which assessors may record as a finding about coordination rather than about either authority’s substantive work.

Recommendations owned elsewhere

Published assessments suggest a pattern in which recommendations fail to close. They appear disproportionately to be those addressed to bodies other than the supervisor, and those requiring primary legislation. Kuwait’s status table for its 2010 recommendations illustrates the first category (IMF, 2019b).

The operational response is to record such recommendations in the central bank’s own tracker with the external authority named as owner, and to agree an escalation route at the outset. A recommendation with no owner in any institution is unlikely to close by default, and may be raised at subsequent Article IV consultations.

The development mandate question

For central banks in jurisdictions where financial sector development is an explicit national priority, a further coordination issue arises. Hong Kong’s assessment observed that, as at other international financial centres, ongoing diligence is needed to ensure that developmental and competitive considerations — particularly those connected with new initiatives — do not conflict with prudential mandates, and recommended regular review of the checks and balances so that monetary and financial stability is not undermined by developmental initiatives (IMF, 2021b).

A defensible position is a documented separation: how development initiatives are prudentially assessed, who within the central bank can object, and what the record shows about occasions on which an objection was raised.

1. Coordination With the Mission

Independent access

Missions meet licensees, external auditors, industry associations, rating agencies and other authorities, frequently without the central bank present. This is standard practice, and resisting it is likely to be read as defensiveness about what might be said. Facilitating it without attempting to shape it is the appropriate posture.

The no-surprises principle

The principle runs in both directions. A well-run mission will usually flag emerging concerns before the aide-mémoire, and a mid-mission read-out should be requested if one is not offered. Equally, where the central bank knows of a material adverse development — a licensee in difficulty, a supervisory failure, a significant data error, a control weakness — it should disclose it itself and early.

Discovery is generally more damaging than disclosure because it converts a substantive issue into a question about candour, and a question about candour can affect the reading of the entire assessment. The operational precondition is that the coordination function itself knows. A supervisory directorate managing a difficult situation quietly, while the coordination unit assures the mission that nothing is outstanding, is a failure mode worth guarding against explicitly.

Channels

Engagement through the mission chief is the primary channel. The office of the Executive Director representing the jurisdiction at the Fund is a legitimate parallel channel, particularly on scoping and on questions of even-handedness, where institutional considerations may be better understood than in a purely technical exchange.

VI.

The Accountability Loop After the Assessment

The recommendation matrix becomes the operative document for the years that follow, and the management arrangements that matter most are those that survive the end of the exercise.

A tracker should give each recommendation a named owner at director level, a definition of what completion means, a target date and a status. The definition of completion is frequently omitted, and its absence is a common reason recommendations stall: nobody agreed in advance what done looked like.

Regular reporting to the governor and the board is a particularly effective discipline for sustaining implementation. Momentum tends to decay once reporting upward stops.

Some recommendations require legislation the central bank does not control, budget it will not receive, or headcount that will not be approved. Others are overtaken by events or by a change in the international standard.

The appropriate response is formal deferral at board level with a documented reason, rather than an item left recorded as in progress indefinitely. A deferral taken deliberately and explained can be defended to Fund staff at the next consultation. An item recorded as in progress since the last assessment is harder to defend, and is likely to be raised again at subsequent consultations.

The final step is to feed every finding back into the standing self-assessment: each accepted finding becomes a criterion recording a gap and a remediation plan, and each closed recommendation becomes a criterion recording evidence.

Done consistently, the central bank arrives at its next assessment with a documented, continuous and evidenced account of what changed and why — which is what assessors most want and rarely receive.

VII.

Red Flags and Recurring Failure Modes

Across published assessments and the management practices they describe, a recurring set of failure modes can be identified. Most are organisational rather than technical, and most are correctable well in advance.

• Treating the exercise as a compliance matter. The Core Principles test supervisory effectiveness; a jurisdiction with a comprehensive rulebook and thin supervisory capacity is unlikely to score well.
• Preparing by rehearsal rather than by remediation. Mock missions and scripted answers may produce a smooth first week but not a better outcome.
• Fragmented data ownership, with multiple directorates producing different versions of the same series.
• Senior disengagement, with leadership appearing only at ceremonial sessions.
• A self-assessment delivered too late to be used by assessors, or graded so defensively that it undermines credibility for the whole exercise.
• Inspection files that record conclusions without reasoning.
• Stated supervisory priorities with no corresponding executed actions.
• A rulebook provision that has never been enforced, or a supervisory practice with no legal basis.
• Vacancy rates in core supervisory functions that the central bank has not quantified, leaving it unable to answer a resourcing question with evidence.
• A prior recommendation still open and unexplained — among the most consequential of the set, because the fate of prior recommendations is a question a mission can be expected to raise early.

VIII.

Conclusion

The management of a Financial Sector Assessment Program engagement is a governance discipline in its own right, and one in which central banks often tend to under-focus/ under-invest precisely because it appears administrative.

Three propositions follow from the argument advanced here. First, the outputs that shape assessment outcomes — a maintained self-assessment, retrievable supervisory reasoning, a reconciled data spine and a complete record of prior recommendations — accumulate over years and cannot be produced within an assessment cycle. Management arrangements should therefore be standing rather than project-based.

Second, the coordination failures that assessors tend to detect are internal and inter-agency rather than technical: inconsistent representations, fragmented data ownership, a group-wide view that no one holds, and recommendations owned by nobody. These are addressable by governance rather than by resource.

Third, a self-assessment written defensively, an adverse development concealed rather than disclosed, or a recommendation left nominally in progress rather than formally deferred each tends to produce a worse outcome than the position it was intended to protect.

Authorities that emerge poorly from an assessment are not necessarily those with the weakest frameworks. Weak outcomes can equally reflect disorganisation, defensiveness and surprise — none of which is a function of supervisory capacity, and all of which are within the institution’s own control.

References

Basel Committee on Banking Supervision. (2024). Core principles for effective banking supervision. Bank for International Settlements. https://bis.org/bcbs/publ/d573.htm

Independent Evaluation Office of the International Monetary Fund. (2006). Independent Evaluation Office announces release of report on the evaluation of the Financial Sector Assessment Program (FSAP) (IEO Press Release No. 06/01). https://ieo.imf.org/-/media/ieo/files/evaluations/completed/01-05-2006-report-on-the-evaluation-of-the-financial-sector/pr0601.pdf

International Monetary Fund. (2021). Financial Sector Assessment Program: Frequently asked questions. Retrieved September 27, 2026, from https://www.imf.org/en/publications/fssa/financial-sector-assessment-program-faq

International Monetary Fund. (n.d.-b). Financial Sector Assessment Program (FSAP). Retrieved September 27, 2026, from https://www.imf.org/en/publications/fssa

International Monetary Fund. (n.d.-c). IMF policy advice [Factsheet]. Retrieved September 27, 2026, from https://www.imf.org/en/about/factsheets/imf-surveillance

International Monetary Fund. (2009). IMF Executive Board reviews experience with the Financial Sector Assessment Program, options for the future, and complementary reforms in surveillance and the assessment of standards and codes (Public Information Notice No. 09/123). https://www.imf.org/external/np/sec/pn/2009/pn09123.htm

International Monetary Fund. (2010). IMF expanding surveillance to require mandatory financial stability assessments of countries with systemically important financial sectors (Press Release No. 10/357). https://www.imf.org/en/news/articles/2015/09/14/01/49/pr10357

International Monetary Fund. (2014). IMF Executive Board concludes review of the Financial Sector Assessment Program (Press Release No. 14/447). https://www.imf.org/en/News/Articles/2015/09/14/01/49/pr14447

International Monetary Fund. (2019a). IMF staff completes 2019 Article IV mission to the United Arab Emirates (Press Release No. 19/398). https://www.imf.org/en/News/Articles/2019/11/06/pr19398-united-arab-emirates-imf-staff-completes-2019-article-iv-mission

International Monetary Fund. (2019b). Kuwait: Financial system stability assessment (IMF Country Report No. 19/96). https://www.imf.org/-/media/files/publications/cr/2019/1kwtea2019002.pdf

International Monetary Fund. (2019c). Malta: Financial system stability assessment (IMF Country Report No. 19/347). https://www.imf.org/en/Publications/CR/Issues/2019/02/27/Malta-Financial-System-Stability-Assessment-46636

International Monetary Fund. (2021a). IMF Executive Board concludes periodic review of the Financial Sector Assessment Program (Press Release No. 21/143). https://www.imf.org/en/News/Articles/2021/05/27/pr21143-imf-executive-board-concludes-periodic-review-of-the-financial-sector-assessment-program

International Monetary Fund. (2021b). People’s Republic of China—Hong Kong Special Administrative Region: Financial Sector Assessment Program—Technical note on banking sector supervision and regulation (IMF Country Report No. 21/118). https://www.imf.org/-/media/Files/Publications/CR/2021/English/1HKGEA2021006.ashx

International Monetary Fund. (2024a). 2024 revised Basel core principles for effective banking supervision (IMF Policy Paper No. 2024/037). https://www.imf.org/en/Publications/Policy-Papers/Issues/2024/08/02/2024-Revised-Basel-Core-Principles-for-Effective-Banking-Supervision-553136

International Monetary Fund. (2024b). Saudi Arabia: Financial Sector Assessment Program—Financial system stability assessment (IMF Country Report No. 24/281). https://www.imf.org/-/media/files/publications/cr/2024/english/1sauea2024002-print-pdf.pdf

International Monetary Fund. (2025). IMF Executive Board concludes 2025 Article IV consultation with Saudi Arabia (Press Release No. 25/275). https://www.imf.org/en/news/articles/2025/08/03/pr25275-saudi-arabia-imf-executive-board-concludes-2025-article-iv-consultation

International Monetary Fund & World Bank. (n.d.). Financial sector assessment: A handbook. Retrieved September 27, 2026, from https://www.imf.org/en/publications/manuals-guides/issues/2017/01/01/financial-sector-assessment-a-handbook-18600

International Monetary Fund & World Bank. (2000). Financial Sector Assessment Program (FSAP)—A review: Lessons from the pilot and issues going forward. https://www.imf.org/external/np/fsap/2001/review.htm