THE TWO PLANES OF JUDGEMENT: Why the African Credit Rating Agency's real test is not only methodological

EXECUTIVE SUMMARY

The African Credit Rating Agency (AfCRA), now moving through its establishment phase under the African Union and the African Peer Review Mechanism, has been framed almost entirely around one question: are African sovereigns underrated by the Big Three. That question is an extremely important question. Nehls and Schmidt, writing for the Konrad-Adenauer-Stiftung, put real numbers behind it, identifying a bias of roughly half a notch to a full notch once methodology and weighting are examined closely. But the question sits on only one plane of the credit rating enterprise, what I have termed the analytical plane, where models, data and disclosure live. It is not the plane on which AfCRA’s fate will be decided.

A second plane governs the conditions under which a credit judgement can be produced at all: committee structure, insulation of analysts, conflict exclusion, and an organisation’s capacity to absorb the political and commercial cost of an unwelcome rating. That plane has two components, rarely separated in the debate but distinct in practice: the internal architecture a credit rating agency builds for itself, and the external ecology, the legal and supervisory environment, within which that architecture is allowed to hold. Credit rating agencies, regulatory bodies, national legislators, and sometimes even the courts regulate this plane in granular detail, not because methodology is unimportant, but because a correct model, applied by an institution that cannot withstand pressure, produces nothing that lasts. As an example, The European Securities and Markets Authority (ESMA) extends the same logic into organisational architecture, treating governance as the precondition for credible analysis rather than an afterthought to it.

A credit rating is more than a piece of information. As for a sovereign credit rating in particular, it is an exercise of delegated authority over a state’s access to capital, and the governance structures built around it exist to protect the legitimacy of that authority when it is exercised against someone powerful. This essay argues that AfCRA’s central challenge is not whether it can build a better model of African creditworthiness. It is whether it can build an institution capable of downgrading a shareholder government, or an AU member with real political weight, and surviving what follows. The precedent of Scope Ratings in Europe, examined in the KAS report and in earlier work on the firm, suggests that governance intentions alone have not been sufficient elsewhere. AfCRA has said the right things about independence and transparency. Whether it builds the committee architecture to make independence operational, and whether the legal and regulatory environment around it can be built to provide the necessary cover, under forms of political and ownership pressure that may differ sharply from those faced by established credit rating agencies, is the question this essay puts on the table.

THE TWO PLANES OF JUDGEMENT: Why the African Credit Rating Agency’s real test is not only methodological

In early 2025, the African Union announced, through the African Peer Review Mechanism, that it would establish its own credit rating agency. Headquartered in Mauritius, AfCRA was presented as a corrective. Marie-Antoinette Rose-Quatre, the APRM’s chief executive, said Africa was no longer content to be a passive observer in the discourse around its own creditworthiness, and would take ownership of its narrative through what she called homegrown solutions. By February the APRM had appointed the Mauritius Commercial Bank as transaction advisor, and technical work on legal, structural and governance frameworks was already under way.

The debate since has centred on one question: whether the Big Three - S&P Global, Moody’s and Fitch - systematically underrate African sovereigns. Currently, only Botswana, Morocco, and Mauritius held investment-grade status on the continent. South Africa, Côte d’Ivoire and Benin sit just below the threshold. The remaining forty-nine states were rated well below or not rated at all. Nehls and Schmidt examined the published criteria behind these outcomes, spanning economic strength, institutional quality, fiscal metrics and external liquidity, and found the surface less neutral than it appears. Even an objective measure like GDP per capita structurally punishes poor countries for being poor, rather than rewarding the trajectory out of poverty: a country would need to raise per-capita income roughly twentyfold, to the level of Mauritius, to secure a single notch of improvement holding other factors constant. Combined with the more subjective weighting given to institutional and governance strength, the KAS-Leibniz analysis puts the average underrating of African sovereigns at somewhere between half a notch and a full notch. Not enormous but not nothing either, when a single notch can move a government’s borrowing cost by tens of basis points across a debt stock worth billions.

It is important to take these findings very seriously. But notice what kind of findings they are. They concern inputs, weightings, model specification and disclosure. They belong to what I call the analytical plane of credit rating, the terrain of methodology, transparency and data quality, where a judgement is produced by asking whether the right variables have been measured and combined fairly. Nearly the entire public conversation about AfCRA sits on this plane. It houses the bias debate, the arguments over sovereign ceilings and unsolicited ratings, the question of whether reserve-adequacy metrics are calibrated fairly across developed and developing economies, and much of the KAS report itself.

A second plane has received almost no attention. I call it the institutional plane. It does not ask what a rating says. It asks under what conditions a rating can be said at all: who sits in the room when the judgement is made, what protects that person once the judgement is public, how dissent inside the room is handled, who is excluded from voting because they carry a conflict, and what happens to the organisation when the judgement proves unwelcome to the people who fund it, own it, or govern the state it concerns. This is the terrain of committee procedure, conflict management, risk oversight and organisational insulation. It is almost invisible in AfCRA’s public discussion, and it is the terrain on which AfCRA’s fate will actually be decided.

The distinction matters because of where each plane is visible. Reform proposals gravitate toward the analytical plane because it can be published, compared, and argued over in daylight. A methodology paper can be read by anyone. A committee vote, by contrast, happens behind a closed door, and its record is deliberately kept from public view. Governance failures are therefore harder to see coming and harder to diagnose after the fact, which is exactly why they receive less scrutiny than they deserve. Yet it is the institutional plane that determines whether an analytical output remains credible once it becomes expensive for someone with power over the agency’s survival. A perfect model, applied by a committee that quietly softens the number under pressure, produces a rating that looks rigorous and is not.

Consider what Moody’s, as a representative example, actually requires of its own rating committees. A Moody’s credit rating is not, formally, the judgement of an individual analyst. It must be determined by a committee, composed for expertise and diversity of opinion, deciding by majority vote, and structured specifically to encourage the free exchange of dissenting or controversial views. Moody’s will not, the policy states, forbear from a rating action based on its potential economic or political effect on Moody’s itself, on the rated entity, on an investor, or on any other market participant; only analytical factors relevant to the opinion may be considered. No analyst may give assurance of a particular outcome before the committee has met. The confidentiality of the committee’s deliberations, including how any individual member voted and whether an analyst dissented from the majority, is protected as a matter of policy, and personnel with a financial or professional entanglement with the rated entity are excluded from the room, a category defined broadly enough to capture family relationships and recent employment history.

None of this is decoration. Ask why a private company builds such elaborate defences around a single opinion, and the answer follows once the two planes are named. A rating is not merely information. Rather, it is an exercise of delegated authority: a private organisation, with no sovereign mandate of its own, permitted to set the terms on which a government reaches the world’s savers. That authority only functions if it is trusted to be exercised without fear or favour, and the trust is fragile precisely because the judgement is consequential. A downgrade moves borrowing costs, triggers covenant clauses, reshapes investor mandates, and, for a sovereign, can become a matter of national grievance directed at named individuals inside the agency. Moody’s committee structure exists to stop a single analyst from having to personally absorb that grievance. The confidentiality rule exists so no one can be identified and pressured over how they voted. The conflict-exclusion rule exists so no one in the room has a private stake in softening the outcome. Credit rating agencies did not build this architecture because ratings are hard to calculate but because ratings are hard to survive, once issued, unless the institution behind them has been engineered in advance to withstand the reaction.

What that architecture ultimately manufactures is institutional courage: the capacity of an organisation to say something costly, to someone who can make its life difficult, and remain standing afterward, credible enough to be believed the next time. Courage of that kind is rarely a property of any single analyst, however principled. It is a property of the structure around the analyst, built long before the difficult file arrives, so that the person who casts the deciding vote is not the same person who has to answer for it in public.

The International Organization of Securities Commissions (IOSCO) reached the same conclusion by a different route, not as one firm’s internal policy but as an international regulatory consensus built after two decades of observing what happens when this architecture is weak or absent. IOSCO’s Code of Conduct Fundamentals, revised in 2015, organises itself around four objectives, and only one, the quality and integrity of the credit rating process, touches methodology directly. The other three, independence and conflicts of interest, transparency and timeliness of disclosure, and the protection of confidential information, are institutional requirements. A credit rating agency’s decisions, the Code states, should be independent and free from political or economic pressures, and free from conflicts arising out of ownership structure, business activity, or the financial interests of employees. An agency should not delay or refrain from a rating action based on the potential economic or political effect of that action on itself, the rated entity, or any other market participant. That is almost the identical prohibition Moody’s writes into its own internal policy.

The European Securities and Markets Authority (ESMA) extends the same logic into organisational architecture. Its 2021 guidelines on internal control for credit rating agencies ask whether the compliance function reports independently of the business lines it monitors, whether internal audit sits under board oversight rather than management control, whether risk management runs through a defined and continuously updated methodology rather than an ad hoc reaction to crisis, and whether the board retains genuine oversight of management rather than delegating that oversight away in substance while keeping it in name. ESMA calibrates its expectations to the size and complexity of the agency, but the underlying principle does not move: the head of a control function should never report to the person whose activities that function exists to control. This is a rule about where power sits inside an organisation rather than analysis, because a credit rating agency’s greatest vulnerability is not a bad model. Arguably, the biggest risk facing a credit rating agency is a captured committee, or at the least a committee unable or unwilling to give its genuine opinion on the creditworthiness of a particular subject.

Institutional courage of this kind does not exist in a vacuum. It depends not only on what an organisation builds inside its own walls but also on the legal and supervisory environment that surrounds it, and that environment is not handed down once and left alone. American courts, as an example, treated a credit rating as a form of protected opinion for most of the twentieth century, starting with Jaillet v Cashman in 1921, which held the relationship between a credit rating agency and its reader was closer to that of a newspaper and its subscriber than to any relationship carrying liability, and reinforced through New York Times Co v Sullivan in 1964. That protection meant something concrete for the credit rating agencies’ committee structures, since a court willing to hold an analyst personally liable for a published opinion would have made confidentiality pointless. It also proved conditional. In 1996, in LaSalle National Bank v Duff & Phelps, a court found that once a credit rating agency began offering advisory services alongside its ratings, its output stopped resembling journalism and started resembling commercial speech, unprotected by the same shield. Shelter, in other words, moves when the institution’s own conduct changes what it looks like in law.

Regulatory shelter has followed a similar pattern on both sides of the Atlantic. When the Dodd-Frank Act tried to strip American credit rating agencies of their exemption from expert liability in registration statements, the agencies withdrew consent to be named in those statements at all, freezing the asset-backed securities market within weeks, until the SEC issued a no-action letter, first granted to the Ford Motor Credit Company in 2010, that has stood ever since and left the reform hollowed out in practice. The European Union’s own civil liability regime, under Article 35a, opens credit rating agencies to claims from issuers and investors, yet sets the evidentiary bar so high that no substantial liability has resulted from it in the years since the article was written. Neither legislature failed to try. What both discovered is that an established credit rating agency, woven into the market’s own infrastructure, can absorb or deflect an attempt to strip its shelter in a way a newly formed one cannot. AfCRA will not have that weight on its first difficult file, and whatever legal recognition its opinions receive in Mauritius will be decided without a century of prior case law behind it. What is abundantly clear is that institutional courage requires institutional shelter.

Read together, Moody’s internal policy, the IOSCO Code and the ESMA guidelines, liability-related cases and legislation, and almost any other formal document one cares to review describe an entire regulatory and institutional apparatus, built over two decades by the world’s most established credit rating agencies and their supervisors, for one purpose: making it structurally difficult for an agency to flinch. That apparatus is not a byproduct of credit rating agency governance but rather it is the whole point of it. An agency without it does not produce inaccurate ratings so much as unreliable ones, judgements that hold in calm conditions and buckle the moment they become expensive for someone powerful.

This is the frame an ambitious and potentially important endeavour like AfCRA deserves, and it changes the questions worth asking. Suppose AfCRA builds an excellent model, correcting the GDP-per-capita distortion Nehls and Schmidt identify, weighting institutional quality using metrics attuned to African administrative realities, drawing on closer government contact to assemble a genuinely richer information set than the Big Three currently manage from further away. Suppose AfCRA wins comprehensively on the analytical plane. Imagine, then, that its model concludes one of the continent’s largest economies faces mounting refinancing risk and deserves a downgrade into deeper speculative territory. The state in question sits on the African Union’s leadership rotation, has been a vocal political sponsor of AfCRA’s creation, and is linked, through a sovereign wealth vehicle or a state-owned bank, to the capital base that funds the agency itself. Does the committee’s vote hold? Does publication get delayed pending further review? Does the methodology suddenly require reconsideration, on the grounds that African-specific factors were not fully captured the first time? Every one of those outcomes is available to an institution that wants to avoid the collision without admitting that it has done so, and every one of them would be indistinguishable, from the outside, from ordinary analytical diligence. That is what makes the institutional plane so hard to audit, and so easy to neglect until the moment it fails.

Nehls and Schmidt are unambiguous that the answer to this question determines everything else. A basic requirement for the new agency, they write, is strict independence from the national governments it rates, which is why they endorse a private ownership structure modelled on Scope Ratings in Europe, spread across private individuals and institutional investors such as banks and insurers, rather than financing through the AU or through development aid, which they warn would create new dependencies and impose an unrealistic timeline. This is the KAS report moving, briefly, from the analytical plane onto the institutional one. It recognises that a reputation for objectivity cannot be manufactured through a better formula. It has to be earned through a capital and governance structure that makes it costly, or impossible, for any single stakeholder to lean on the committee.

David Lubin’s scepticism as articulated in a Chatham House event launching the report, goes to the same point from the opposite direction. Lubin, Senior Research Fellow at Chatham House, asked whether AfCRA risked being Africa marking its own homework, and why pension fund managers, corporate treasurers and international investors would trust an African agency to rate African sovereigns objectively. That is an institutional question instead of a methodological question. Lubin was not questioning whether AfCRA’s model would be well built. Instead, he was questioning whether any structure of ownership, governance and committee process could credibly separate AfCRA’s analytical output from the interests of the continent, and the states, that gave it life. That is precisely the question the credit rating agencies’ committee architecture, the IOSCO Code and the ESMA/SEC control guidelines were built to answer for the Big Three, over decades, through repeated tests, including institutional failures along the way. AfCRA does not have decades. It has, in effect, one or two early, highly visible tests, in which the world will watch to see whether the agency downgrades when its own model says to, or finds a reason not to.

The precedent Nehls and Schmidt cite, Scope Ratings, is instructive precisely because it shows that reasonable capital and sound intentions are not sufficient on their own. Scope was founded in the wake of the European sovereign debt crisis to correct the perception that the Big Three were insufficiently attuned to European specifics, the same motivation now driving AfCRA. Backed by a genuinely diversified base of European insurers, banks and private investors, and eventually recognised by the European Central Bank as an accepted external credit-assessment institution under the Eurosystem’s collateral framework, Scope built the kind of independent, diversified ownership structure the KAS report recommends AfCRA emulate. Yet, as Nehls and Schmidt record, more than a decade after its founding, Scope’s global market share remains below one per cent, and it has not lowered the credit cost of the European issuers it set out to serve. Earlier work examining Scope’s growth strategy traced a similar pattern: an entrant with reasonable institutional ambitions, hampered less by analytical capacity than by the sheer weight of incumbency, regulatory designation requirements, and a market that defaults to trusting the names it already knows, including, that same analysis noted, the risk that legal exposure and personnel drawn from the very culture a challenger sets out to correct can quietly reproduce the practices it was founded to displace. If a European challenger, operating inside one of the world’s most developed and rule-bound securities markets, with none of the added burden of proving political independence from the governments it rates, could not translate sound governance intentions into market share or borrowing-cost relief, AfCRA’s task is harder still. This is because the credibility test on the institutional plane is more severe than the one Scope faced, not because its analysis will be any weaker than Scope’s.

None of this argues against AfCRA. Rating coverage across the continent is thin, with only thirty-two of fifty-four states publicly rated at all, and the case for an agency capable of extending coverage to sub-sovereigns, state-owned enterprises and mid-sized corporates, entities the Big Three have little commercial incentive to cover given the fees involved relative to the workload, stands on its own terms, independent of the bias question entirely. Nehls and Schmidt are right that closer government contact could plausibly improve the information base underlying African sovereign analysis, provided that closeness does not curdle into capture. The point is narrower, and more consequential: AfCRA’s success will not be determined by whether its ratings differ from the Big Three’s. It will be determined by whether the institution can survive the first moment its ratings become genuinely inconvenient to someone who matters to its survival. There is also a need not to ask too much from the fledgling institution, with undertones suggesting that the agency can also take the role of a development institution at the same time as being a rating agency.

This is where the practical questions live, and where the public debate has been almost entirely silent. Who will sit on AfCRA’s rating committees, and by what process will they be selected, insulated and rotated. Will dissent inside a committee be recorded and protected, kept confidential even from the rated entity itself in the manner Moody’s requires of its own personnel, so no single analyst can be identified and pressured after the fact? Will AfCRA build a compliance function that reports to its board rather than to the management whose activities it exists to police, the baseline condition ESMA’s guidelines impose on European agencies as a matter of registration? Will its ownership structure, whatever mix of private African and international capital it eventually settles into, carry enough diversification to make it structurally difficult for any single investor, or any coalition of African governments acting through investor proxies, to lean on a specific sovereign file? Will the agency publish, as the KAS report itself suggests, the substance of its committee deliberations with the delay and rigour some central banks apply to monetary policy minutes, so the market can audit the process that produced the rating and not merely the rating itself. That last question returns the essay to where it started. Investors want the deliberative process opened up, since transparency on the analytical plane is what lets a market trust a model. But the protection the traditional credit rating agencies build around their committees, the confidentiality of the vote, the anonymity of the dissent, exists on the institutional plane precisely to shield the room from the pressure that publication would invite. The more fully AfCRA opens its committee process to public view, the harder it may become to preserve the protected space in which a difficult judgement gets formed in the first place. The two planes do not simply sit side by side. On the question of disclosure, they pull against each other.

Reading this you may think these are exotic questions, or academic questions. In reality, they are the ordinary questions any agency seeking to be a part of the international framework must already answer as a condition of operating credibly. AfCRA faces them without decades of incremental institution-building behind it, and without an established regulator playing the disciplining role ESMA plays for European agencies.

A further set of questions concerns not the committee room but the jurisdiction around it. The APRM selected Mauritius as AfCRA’s host and primary supervisory jurisdiction in September 2025, with the first ratings expected in 2026. What form will Mauritian supervision of AfCRA take, and will its findings be published with anything like the frequency and detail ESMA applies to European credit rating agencies? What happens, in practice, when a sovereign believes AfCRA has downgraded it unfairly, and does the legal environment surrounding AfCRA extend to its committee the same recognition that a rating is an opinion, produced through a defined process, rather than a guarantee open to challenge on its substance? That recognition, as the American case law shows, is never simply assumed into existence. It took decades of litigation to establish, and moved again the moment the agencies in question changed their own conduct. Mauritius is rightly regarded as one of Africa's key financial centres, but this is an internationally important project with the world's eyes focused on the agency and the country. Mauritius’ has experience hosting the African subsidiary of Indian firm CARE ratings, but there may be several significant tests facing the jurisdiction in the coming years. ESMA’s findings carry weight because markets already believe ESMA capable of disciplining the agencies it supervises, a belief built over more than a decade of registration decisions, on-site inspections and public guidance. Supervisory legitimacy, like rating legitimacy, is a form of authority that has to be earned rather than assumed, and Mauritius has not yet had the occasion to earn it, although the experience with CARE should help. None of this suggests Mauritius is unfit for the task. It is a small jurisdiction being asked to supervise an institution whose judgements may move the borrowing costs of the continent’s largest economies, and the honest question is whether any single small jurisdiction can, on its own, supply the weight of legal and supervisory recognition that took the United States, and Europe decades and an entire union of member states to build. AfCRA’s founding narrative, correcting a bias against African states, also sits in permanent tension with any perception that it might, even occasionally, soften a judgement out of the same continental solidarity that motivated its creation. That tension is not a flaw in the project. It is the central governance problem the project must solve, and solving it has almost nothing to do with getting the GDP-per-capita weighting right.

CONCLUDING REFLECTIONS

The debate around AfCRA has organised itself around the most visible and most measurable question available: is there bias, and how large is it. Nehls and Schmidt have given that question empirical shape, and the answer they provide, a bias of roughly half a notch to a full notch, deserves to inform how African governments and international investors read existing sovereign ratings. But an agency capable of a better answer to that question is not thereby an agency capable of surviving the consequences of its own answers. The traditional credit rating agencies’ committee architecture, the IOSCO Codes, and SEC/ESMA’s internal control guidelines were not built by institutions preoccupied with getting the analysis right. They were built by institutions that had already learned, at real cost, that getting the analysis right is the easier half of the problem. The harder half is building an organisation that can say something true and unwelcome, to someone powerful, and still be standing, and still be believed, the following year.

AfCRA’s founders have said the right words about independence and transparency, and the KAS report is correct to press them toward private ownership and external evaluation as the means of holding those words to account. But committee design alone will not carry the institution through its first genuinely costly file. The Big Three did not become durable on the strength of their internal procedures. They evolved inside legal and supervisory systems that, imperfectly, gave their opinions somewhere to stand once the opinions became unwelcome. AfCRA has to build that internal architecture and find, or help construct, an equivalent external shelter, in a jurisdiction that has never had to provide one before, at least at this level. Every credit rating agency eventually faces the same examination: whether it is prepared to issue a judgement that damages someone with power over its own survival. The agencies that matter, in the end, are the ones that do it anyway. Whether AfCRA becomes one of them will not be settled by any methodology paper but will be settled in the room where the vote is taken, by whether the institution built around that room, and the jurisdiction built around that institution, has created sufficient distance between the rating and the relationship.

None of these questions are unique to Africa. Every major credit rating jurisdiction has had to confront them, and many of the institutional protections now taken for granted in the United States and Europe were themselves products of crisis, litigation, and decades of institutional learning. The African Union is proposing something bold and is at the stage of being relatively unprecedented, which is precisely why so many eyes are on it. The appropriate response is therefore neither uncritical celebration nor premature scepticism, but serious engagement with the institutional conditions that will determine whether AfCRA succeeds.