The Report That Disappears Before It's Written
Picture yourself sitting across from an analyst on the fourteenth floor of a midtown office building, watching her close a spreadsheet she has just spent three weeks building. The numbers are clean. The methodology is sound. The conclusion is uncomfortable: a mid-sized emerging-market government has borrowed beyond its capacity to service debt without a major external shock, and the current investment-grade rating is a fiction the market has politely agreed to maintain. She knows this. Her supervisor knows this. Then the committee meets, and the rating doesn't move.
This is not corruption in the cinematic sense. No envelope changes hands. What happens instead is quieter, more durable, and far harder to legislate away: an internal escalation culture that determines, long before any formal vote, which sovereign downgrades are politically survivable inside the agency itself.
The Architecture of a Rating That Never Leaves the Room
Credit rating agencies use committee structures specifically to distribute accountability. A lead analyst builds the case; a ratings committee, typically composed of senior analysts and a chair who may sit several management layers above the lead, votes on the final action. Moody's, S&P, and Fitch each run variations on this structure, but the logic is consistent. No single person owns the outcome.
The problem is that distributed accountability tends to suppress dissent rather than aggregate it honestly.
Consider the mechanics. An analyst covering a Gulf state whose sovereign debt metrics have deteriorated over several years must first convince her direct supervisor to escalate the case to committee. That supervisor is weighing the reputational exposure of being the person who initiated a downgrade of a government that has historically generated substantial fee revenue from rated bond issuances, structured finance vehicles, and subsidiary entities. The rational move, absent strong institutional pressure to escalate, is to request more data, schedule a follow-up sovereign review, and wait.
If the case does reach committee, the chair controls the agenda framing. A case presented as "should we downgrade" is evaluated differently from one presented as "is the current rating still defensible." The first framing invites inertia; the second demands justification for the status quo, which is a much harder posture to sustain in a room full of people who approved that status quo last year. Chairs who consistently bring uncomfortable cases rarely hold the chair for long.
The result is a system structurally biased toward what analysts inside the industry sometimes call "ratings stability," a phrase that sounds like a virtue until you notice it applies asymmetrically. Upgrades face scrutiny too, but a missed upgrade costs no one a relationship. A downgrade of a sovereign that then calls the agency's country head to complain creates a paper trail, and paper trails have a way of following people.
What the Gap Between Private Assessment and Published Rating Actually Looks Like
The agencies are not entirely without internal honesty. Each maintains what amounts to a shadow rating: the analytical team's working assessment, circulating in internal memos and committee pre-reads and, occasionally, in the frank conversations analysts have with each other in the elevator. The gap between that internal view and the published rating is where the real story lives.
Take a plausible but invented scenario. An analyst named Carla has been covering a fictional country, call it Valdoria, for four years. Valdoria carries a BBB-minus rating, the lowest investment-grade tier. Over two annual review cycles, Carla's internal model has produced a score consistent with BB-plus, one notch below investment grade. The committee has twice opted to affirm BBB-minus, citing policy trajectory and expectations of multilateral support. Nothing has been published. No negative outlook, no review for downgrade. From the outside, Valdoria looks stable.
Now consider what happens to investors relying on that published rating. Pension funds with investment-grade mandates hold Valdoria bonds. The spread between Valdoria's bonds and comparable BB-rated issuers has quietly widened in the secondary market, because sophisticated traders can read sovereign balance sheets too. The market has partially priced in what Carla's model has been saying for two years. The retail investor holding a fund benchmarked to investment-grade sovereign indices has not.
That spread widening is the tell. It is the market's way of publishing the rating the committee refused to issue, and it is the number that should make regulators uncomfortable.
These dynamics were visible, in various forms, in the lead-up to several sovereign stress episodes across the past three decades, where agencies moved ratings sharply after market dislocations that their own internal analytics had been flagging for months. The downgrade, when it finally came, landed like a diagnosis delivered at the funeral.
What People Get Wrong About Agency Capture
The standard critique of rating agencies focuses on the issuer-pays model, where the entity being rated pays the rating fee. That critique is legitimate. It is also incomplete, and it causes people to mislocate the problem.
Sovereign ratings are not paid for directly by governments in the way corporate ratings are. Many sovereigns are rated on an unsolicited basis, meaning the government didn't commission the rating at all. The financial incentive to flatter a sovereign is therefore less direct than in structured finance, where a bank might withdraw its mortgage-backed security business if the rating is unfavorable.
The actual pressure is relational and reputational, not transactional. Agencies need sovereign governments to grant access: access to finance ministry data, to central bank officials, to the off-the-record briefings that inform the qualitative judgment layers of the model. A government that feels aggrieved by a downgrade can revoke that access, refuse to engage with the annual review process, and publicly dispute the methodology. That is a headache for the analyst team, an embarrassment for the country head, and a problem that escalates upward fast. The rational institutional response is to avoid creating the headache in the first place.
So ask yourself: if the suppression of a sovereign downgrade requires no villain, no bribe, no explicit instruction, only a series of individually rational choices by people protecting their working relationships, how exactly does a regulator write a rule against it?
This is the mechanism most critics miss. The suppression doesn't require anyone to behave badly. It requires only that everyone behaves sensibly within an incentive structure that rewards relationship preservation over analytical candor.
The Analyst Caught in the Middle
For the individual analyst, the situation is genuinely difficult. She has professional standards, a methodology she believes in, and a conclusion she stands behind. She also has a career, a manager who controls her performance review, and an institutional culture that signals, without ever quite saying so, that certain conclusions require more evidence than others.
Agencies have formal protections for analytical independence, committees specifically designed to prevent any one relationship manager from killing a rating action. Those protections are real. They are also navigated by people who understand that the committee chair's skepticism about a politically sensitive downgrade is not a formal veto but functions like one. The difference between a rule and its enforcement is the whole story.
The honest answer to whether any individual analyst can fix this is probably not, at least not unilaterally. The escalation culture is self-reinforcing. The analyst who pushes hardest against it tends to find herself reviewing less prominent credits.
Regulatory frameworks in both the European Union and Washington have required agencies to maintain internal independence policies, document rating rationales, and submit to periodic audits. These requirements have improved transparency at the margins, and the marginal improvement is real but modest. They have not changed the fundamental fact that the people deciding which sovereign downgrades clear committee are the same people whose professional standing depends on institutional relationships that a downgrade would strain.
Ratings that never publish don't leave a record. That's the point. And the absence of a record is precisely what makes the escalation culture so difficult to reform: you cannot audit a decision that was never formally made. The next sovereign stress episode will not announce itself in advance, but somewhere on a fourteenth floor, a spreadsheet will already have the answer.