The Neutral Machine: How Credit Rating Agencies Govern Without Governing
Whoever decides what a nation may do holds an office. This one answers to no one.
There is a peculiar sleight of hand at the heart of global finance. Three private firms — Moody’s, S&P Global Ratings, and Fitch — anchored in the New York–centred financial ecosystem that strongly influences global capital conditions, can materially affect the borrowing costs and policy space of sovereign nations. They do this not through force, not through legislation, not through treaty. They do it through notation: a letter grade, an outlook revision, a quiet reclassification from “stable” to “negative.”
Ask them what they do and the answer sounds almost modest: they assess creditworthiness, measure risk, and provide information to investors.
This is true in the same way a thermostat “provides information” about temperature — while also controlling the heating. And it is the second clause the agencies never say aloud. The whole edifice rests on describing an act of control as an act of description.
The Published Criteria and the Unpublished Assumptions
The methodology of sovereign credit ratings is not secret. Agencies publish criteria, run briefings, issue reports, and wrap the whole enterprise in the language of transparency. Formal opacity is not the main problem.
A core problem is what the criteria treat as self-evident.
To score well in a sovereign rating framework, a nation is effectively expected to display a recognisable profile of “policy credibility.” That credibility is often inferred from three broad buckets:
Fiscal posture: deficit and debt trajectories that signal restraint, and can privilege fiscal consolidation under stress.
External posture: openness to trade and capital flows, and a willingness to prioritise debt service within the balance-of-payments constraint.
Institutional posture: central bank independence from political authority, “labor flexibility,” and a governance style that reads as predictable to global creditors.
Presented this way, the move becomes visible: a set of contestable political choices is translated into a risk score that presents itself as non-political. A particular philosophy of economic management can be treated as a baseline for “soundness” rather than as one approach among others.
That is not a semantic quibble. When a normative programme is encoded as empirical technique, the technique becomes governance.
The Wiring: How “Opinion” Becomes Constraint
The rating is not powerful because it persuades everyone. The rating is powerful because it is embedded in plumbing.
Ratings can function as triggers inside the machinery of finance: investment mandates that restrict what pension funds or insurers may hold; index inclusion rules that determine whether large pools of passive capital can buy; collateral frameworks and haircuts that change funding costs; internal risk models and bank constraints that tighten automatically when a rating falls.
The wiring does not transmit an opinion. It converts the opinion into a standing permission — what a fund may hold, what capital may buy, what a bank may carry — a line drawn through the space of allowed action and redrawn each time the grade moves.
Once a metric is embedded into mandates and collateral rules, it becomes not just descriptive but executable.
When the grade changes, it can trigger forced selling under some mandates — by policy, not by deliberation. It can raise borrowing costs through contractual or institutional reflex, not through a fresh debate about fundamentals.
That is governance: constraint built into infrastructure.
The Mechanism of Compliance
Now the feedback loop lands.
A downgrade raises a nation’s borrowing costs — sometimes dramatically, sometimes overnight. For states with limited reserves and narrow fiscal space, this is not an academic inconvenience. It is the difference between refinancing at survivable rates and entering a spiral where every roll-over hardens into austerity.
And austerity is not merely painful; it can be self-fulfilling. Cut investment, depress growth, worsen debt dynamics, trigger further downgrades, repeat.
This is the disciplinary architecture at its clearest: a country can be pushed toward the very “credibility” behaviours that the framework rewards, because the alternative becomes unaffordable.
Industrial policy, capital controls, subsidised domestic production, strategic protection in early-stage sectors — the same kinds of tools used by many now-developed economies on their way up — become harder to sustain under the downgrade threat. Not because these policies are proven to be reckless in all contexts, but because they can be legible to the rating apparatus as “risk.”
The permission line runs through the actual menu of things a state may do and still afford to govern. On one side, the policies the template recognises. On the other, the policies a country must now pay a penalty to attempt. The rating does not merely describe the world. It decides, in practical effect, what a sovereign is permitted to do.
Structural Power: The Agenda That Doesn’t Need to Be Hidden
Here the analysis must resist the gravitational pull of conspiracy.
It is tempting to posit a secret agenda, a cabal, a coordinated programme of economic subjugation. But that framing — while emotionally satisfying — understates the problem.
You don’t need a conspiracy when you have a structure.
Susan Strange coined the term structural power for the ability to shape the frameworks within which others must operate — such that your preferences appear as neutral rules rather than as contested choices. This is a clear instance of the kind of power the credit rating system can exercise in practice. It does not need to be coordinated in a back room because it is embedded in the architecture itself. The incentives align without anyone needing to align them deliberately.
Structural power, though, only describes how the power runs. Structures can be benign; this concept alone convicts no one. The wrong needs a second name, and the republican tradition supplies it: non-domination. You are unfree not only when someone interferes with you, but when you stand subject to a power that could interfere on its own terms, owing you no account — an arbitrary power over your choices. The harm is not any particular bad decision. It is the standing exposure to a will you cannot reach.
A permission-power exercised over states that can neither contest it nor remove it is exactly that exposure. Strange names the field the rating system shapes; non-domination names why holding such a power without answerability is not a flaw in an otherwise fair order but domination itself — the costume changed, the structure intact.
The Big Three operate inside a global order shaped by Bretton Woods, refined through the Washington Consensus, and maintained by institutions whose default assumptions converge around creditor confidence, capital mobility, and “market discipline.” The agencies are not aberrations. They are a measurement layer of that system. And measurement layers do not merely observe. They can stabilise a working reality in markets by standardising what counts as credible.
The Office of Permission
To decide what others may and may not do is the defining act of an office. It is the thing a magistrate does, a regulator does, a court does. We have a word for the authority to draw the line of permitted action, and the word is governance, whatever costume it wears.
And offices of permission are not free. The right to bound another’s permissible action is everywhere conditioned on duties — to justify the decision to those it binds, to answer for its consequences, to serve a public rather than a faction, to be removable when it fails. These are the civic duties of office. They are not decoration. They are the price a polity exacts for letting anyone hold permission-power at all, because that is the only thing that converts raw power over others into authority a free people can live under. Strip the duties away and what remains is not authority. It is the will of a private party, enforced.
The rating system holds the power of permission and pays none of the price.
It manages this through a single move, and the move is the neutrality claim. By calling itself opinion, information, measurement, it relocates from one category into another: out of the exercise of governing power, which carries the duties of office, and into speech about the world, which carries none. “We do not govern; we assess” is presented as a description. It is not a description. It is an exemption.
So the indictment was never that the machine has values. Everything has values. It is that the machine draws the line of the permissible — an act of office — and then declines the duties of office on the ground that it was only ever offering a view. One sleight encodes a political programme as neutral technique; the same neutrality then sheds the duties the technique has quietly assumed. The neutrality is not, at bottom, an epistemic posture. It is the mechanism by which the duty is discharged before it can attach.
The Track Record of Neutral Expertise
If authority rests on expertise, the record warrants scrutiny.
These are the same agencies that assigned top-tier ratings to mortgage-linked products that helped precipitate the 2008 financial crisis. The failure was not simply an isolated error. It exposed a predictable tension in a business model where the rated entities paid for the rating — an incentive structure that can distort judgement even when everyone involved believes they are acting professionally. This is a critique of incentive design, not an accusation of individual bad faith.
The aftermath is where the office shows itself. Reforms followed: more oversight, more disclosure, more procedural compliance. The core incentives — issuer-pay dynamics, reputational dependence on the same capital networks, the entrenched embedding of ratings into financial plumbing — were not eliminated. The system did not become neutral. It became more bureaucratically defensible.
That is precisely the signature of permission-power without the duties of office. A magistrate who failed on that scale would face the answerability the office demands — justification, accountability, removal. The agencies faced compliance. They added procedure where a polity would have demanded account. The distance between those two responses is the exact measure of the missing duty.
There is another pattern worth naming: pro-cyclicality. Agencies often downgrade during crises — exactly when a country most needs affordable borrowing — which can deepen the crisis and turn “risk assessment” into risk amplification.
So the question sharpens:
Whose risk is being managed here? The debtor nation’s — or the creditor’s? Whom does the office serve?
The Global South Objection (and Why It Matters)
The objection raised by the African Union and echoed by economists across the Global South is not that creditworthiness should not be assessed. It is that the criteria, as operationalised, can structurally favour a particular model of political economy — one aligned with the preferences of many large creditors that play an outsized role in international capital markets.
The objection is not, at root, a complaint about accuracy. It is a complaint about whom the office serves — raised by the parties the permission-power binds, who have no standing to contest it. Even when methodologies are published, the judgement calls inside them — what counts as “credibility,” what counts as “institutional strength,” what kinds of state capacity count as stabilising rather than distortionary — can contribute to persistent disadvantages for countries pursuing development trajectories that do not mimic post-industrial Western norms.
This is what structural power looks like from beneath: an evaluation frame that presents itself as universal while quietly penalising deviation from the dominant template. And it is what domination looks like from beneath: a line drawn through your permitted choices by a hand you cannot reach.
What Would an Alternative Look Like?
This is the harder question, and the more necessary one. Once the office is named, the reforms reduce to a single entailment: if a function governs, it owes the duties of government. Permission-power of this reach must face governance-grade standards — transparency not only of criteria, but of assumptions, incentive structures, and accountability pathways; answerability to the parties it binds; some pathway by which sustained failure carries consequence. This is not one reform among several. It is the thing the other reforms are for.
Two tactics follow:
Regional or developmental rating institutions that explicitly model structural constraints — commodity dependence, legacy debt burdens, external vulnerability — and treat state-led development capacity as a variable rather than a deviation. These matter not because plurality is pleasant, but because a single unanswerable arbiter is the precise structure non-domination forbids.
Methodological reform that incorporates long-horizon resilience: sustainability, inequality, institutional learning capacity, productive investment — not only short-horizon debt-service comfort. This matters because an office that serves a public must at least be able to see the public’s interest in its instruments.
Both are downstream of the duty. Neither substitutes for it. You can pluralise the raters and refine the models and still have left the central wrong untouched, if the power to permit remains a power that answers to no one.
None of this is simple. Network effects are real: once ratings become a coordination device embedded everywhere, the incumbents acquire a kind of inertia that looks like inevitability. But inevitability is often just the name we give to a path-dependent arrangement that benefits the already-positioned.
The Craft of Appearing Objective
There is a broader lesson here, well beyond finance.
Power is most durable when it is most invisible. Not invisible in the sense of hidden — the criteria are published, the downgrades are announced, the consequences are measurable. Invisible in the sense of naturalised.
When a particular arrangement of interests succeeds in presenting itself as “just the way things are,” as the only rational approach, as common sense rather than ideology — that is when critique becomes most difficult and most essential.
The credit rating system is not a conspiracy. It is something more resilient than a conspiracy: a structure that can produce compliance without overt coercion, govern without democratic mandate, and discipline without accountability. It is, in the most precise sense of the word, a technology — one designed not merely to measure the world as it is, but to shape the world toward a particular vision of how it should be.
But seeing the hand inside the machine is not the end of the matter. It is the precondition for the only thing that changes it.
A power that has named itself neutral has, in the same gesture, excused itself from duty. It can be held to the duties of office only once it is made to wear the name of office again — and that re-naming is not rhetoric. It is the civic act on which every later reform depends, because none of them can be demanded of a thing we have agreed to call mere opinion.
So the obligation falls to the reader as squarely as it was shed by the firm. The agencies’ civic duty is the one they have escaped: to answer for the permissions they grant and withhold. The citizen’s civic duty is the one that makes the first enforceable: to refuse the exemption — to deny that any permission-power is ever merely neutral, and so to keep it within reach of the duties it would rather not bear.
The neutral machine is never neutral. To say so is not an observation. It is the first act of holding it to account.
If you found this useful, consider subscribing to nodepunk for more writing at the intersections of philosophy, poetry, thought leadership, and the structures that shape how we think.