Credit-Rating Agencies in Structured Finance

A rating was both an opinion and a production requirement

Mortgage-backed securities and collateralized debt obligations divided pools of loans into tranches with different claims on cash flow and loss. Rating agencies applied models, data assumptions, analyst review, and committees to assign grades. Investors used those grades as a compact signal, while laws, regulations, and investment mandates often made high ratings a practical condition for purchase. The rating therefore did more than describe a finished product. It helped determine which product could be built and who could hold it.¹

Under the issuer-pays model, the bank or other sponsor arranging the security selected and paid the agencies. Issuers could present proposed structures, ask how much credit enhancement a desired grade required, and decide whether to take the deal elsewhere. The agencies' professional judgment remained real; it operated inside a market where the subject of the judgment was also the customer.

Models compressed uncertainty into a tradable grade

Structured-finance ratings depended on assumptions about mortgage defaults, correlation, housing prices, recovery, and the behavior of newly originated loan types. A letter grade concealed those moving parts from many downstream users. The model could be internally consistent while its historical data and stress range failed to represent a nationwide housing decline or deteriorating underwriting.²

This is a problem of bounded judgment, not simply dishonest arithmetic. Specialists faced novel products, incomplete loan data, high volume, commercial deadlines, and competitors. Those conditions do not excuse inflated ratings. They explain why governance needed robust challenge, model uncertainty, surveillance, and authority to stop issuance instead of treating a calculated output as truth.

Raymond McDaniel, then Moody's chief executive, wrote internally that market share and ratings quality could conflict; former managing director Jerome Fons later testified that the issuer-pays model and pursuit of revenue weakened analytical independence. Fons was a participant and critic, not a neutral measurement instrument, but his testimony supplies an inside account of how commercial and professional authority interacted.³

Deal volume overran the evidence loop

The Senate Permanent Subcommittee on Investigations reconstructed how Moody's and Standard & Poor's competed for structured-finance business, adjusted models, handled issuer pressure, and struggled with staffing and surveillance. Internal messages recorded employees warning that the market was creating risks the ratings process could not contain. The report found that conflicts of interest and failures at the agencies contributed to the crisis; it did not claim that ratings alone created weak mortgages, bank leverage, investor demand, or the housing bubble.²

The distinction matters because the operating coalition was broad. Originators made loans, arrangers pooled them, agencies rated them, investors bought them, insurers guaranteed some, and regulators embedded ratings in rules. Each actor could point to another actor's evidence. An investment committee relied on the grade; the agency relied on supplied loan data and assumptions; the issuer relied on investor demand; the originator no longer held much of the risk. Distributed expertise became distributed permission.

Downgrades arrived after exposure had scaled

As mortgage performance deteriorated in 2007, agencies placed large numbers of securities on watch and downgraded them, sometimes by many notches. The changes conveyed new evidence, but they also forced institutions governed by rating thresholds to sell or recognize losses together. A grade designed as a stable coordination device became a source of synchronized instability.¹

The Financial Crisis Inquiry Commission concluded that failures of the rating agencies were essential cogs in the machinery of the crisis. That is the commission majority's causal judgment, supported by documents and testimony but not the only interpretation of the crisis. The narrower organizational finding is strong: vast issuance depended on ratings that did not hold, and the feedback capable of correcting them arrived after the rated products had spread through the system.

Independence requires changing who can stop the line

Disclosure of a conflict helps users interpret a rating, but it does not remove the issuer's selection power or the grade's regulatory role. Better governance would separate commercial negotiation from analysis, preserve analyst dissent, publish assumptions and performance histories, require uncertainty and stress ranges, and let surveillance revise methods before deal volume becomes a reason not to act.

Most importantly, ratings should inform judgment rather than replace it. A public rule or investment mandate that treats one private grade as permission delegates authority without a matching remedy. The lesson for knowledge and professional autonomy is that expertise needs institutional protection from the customer whose product cannot proceed unless the expert says yes.