How Moody's 1909 Letter Grades Became Financial Law

In April 1909 a New York publisher named John Moody issued a thick book about American railroads. Nothing in that was novel; investment manuals crowded the market, and Moody had already published one. What made this volume different was a narrow column running down the side of the tables. Against roughly 1,300 railroad securities he had printed a letter — Aaa, Aa, A, and on down the alphabet — compressing earnings coverage, mortgage seniority and the character of a management into two or three characters of type.

Moody was careful about the claim. He conceded that grading bonds was not entirely original with him, crediting earlier practice he had seen described in Vienna and Berlin, and American mercantile agencies had been coding the creditworthiness of shopkeepers with letters and numbers since the 1850s. His innovation was narrower and more consequential: he published the grades in bulk, for securities that traded on a public market, and sold the book to anyone willing to pay for it.

Within a single working lifetime those letters stopped being one man's published opinion and became a term of art in American banking law. A national bank in 1937 could not buy a bond that carried the wrong letter. A broker-dealer in 1976 calculated its required capital from them. By 2007 a security that had never existed as an asset until a lawyer drafted it could be sold worldwide on the strength of three characters, and the institutions printing those characters were paid by the firms that had drafted it.

The Man Who Went Broke in 1907

Moody's route to 1909 ran through a failure. He had founded the Moody Manual Company in 1900 to publish compilations of financial statistics, and the business grew quickly enough that he borrowed against it. When credit seized in the panic that Pierpont Morgan resolved from his library in 1907, Moody could not refinance, and he sold the manual company to a competitor.

What he came back with was a different product. Statistics he had already proved he could assemble; anyone could. Judgement was scarce, and judgement was what he now proposed to sell. Moody's Analyses Publishing Company produced its first annual volume in April 1909, and the letters were the whole point of it. Richard Sylla's history of the business notes that Moody was selling a service the bond market had previously bought in a personal form, from a trusted broker or a family banker, and that the printed grade worked because the American railroad bond market had outgrown the circle of people who could know an issuer personally (Sylla, 2002).

Competitors arrived fast. The Poor's Publishing Company, descended from Henry Varnum Poor's railroad manuals of the 1860s, began rating in 1916. Standard Statistics, formed in 1906, followed in 1922; the two merged in 1941 to make Standard and Poor's. The Fitch Publishing Company, launched in New York in 1913 by John Knowles Fitch with Henry Clancy and Fabian Levy, introduced in 1924 the AAA-through-D scale that the rest of the industry eventually copied.

Four firms selling graded opinions to subscribers was an ordinary publishing market. It became something else when the government started reading the books.

February 1936: An Opinion Becomes a Rule

Bank regulators had spent the early 1930s looking for a workable test of what a bank could safely hold. Examiners needed a rule that did not require every national bank examiner in the country to conduct independent credit analysis of every bond in every portfolio, and the rating manuals were sitting on the shelf, already written, already updated annually, already trusted by the market.

On 15 February 1936 the Comptroller of the Currency ruled that national banks could not purchase securities that were, in the language of the ruling, distinctly and predominantly speculative. Bonds falling in the top four rating categories could be carried at book value; the rest had to be marked to market or sold. The Comptroller did not write a definition of credit quality. He pointed at the manuals.

Frank Partnoy's account of the industry treats this as the founding moment of everything that followed. A rating had been a piece of commercial speech that investors were free to ignore; after February 1936 it was an input to a legal obligation, and demand for it no longer depended on whether anyone found it persuasive (Partnoy, 1999). State insurance regulators built the same hook into their capital rules, and by 1951 the National Association of Insurance Commissioners was setting reserve requirements for insurers by rating category.

Note what the first five rows share. None of them asks whether a rating is accurate. Each takes the letter as given and attaches a consequence to it, which means the agencies acquired a franchise whose value had nothing to do with their forecasting record.

Who Pays the Rater

For sixty years the agencies charged the people who read the ratings. Subscribers bought manuals; the manuals were expensive; the model worked because a bond investor who wanted Moody's opinion had to buy Moody's book.

Photocopying broke it. By the late 1960s an institution could buy one manual and circulate the relevant pages through a research department at negligible cost, and subscription revenue eroded against a free-rider problem the agencies could not police. Then came a default nobody's letters had flagged.

Penn Central's bankruptcy on 21 June 1970 left about $82 million of commercial paper outstanding and froze a market that had been treated as a cash equivalent. Commercial paper was barely rated at the time. Issuers who wanted to sell paper after June 1970 found that buyers now demanded a grade, and the issuers — not the buyers — were the ones who needed it.

Moody's switched to charging issuers in October 1970; Standard and Poor's followed. Lawrence White's survey of the industry describes the change as the moment the agencies stopped being publishers and became gatekeepers, since the fee now came from the party with an interest in a particular answer (White, 2010). Defenders of the shift pointed out that it funded far deeper analysis: revenue per rated credit rose sharply, and analysts could cover fewer issuers in more detail. Both things were true at once, and the conflict sat unexamined for three decades because the ratings kept working well enough on corporate and municipal debt.

Fifty-eight American companies carried the top grade in the late 1970s. Leveraged takeovers took out a first wave in the 1980s, as acquirers loaded acquired balance sheets with the debt that had paid for them, and a second wave went voluntarily when finance directors concluded that a fortress balance sheet was an inefficient one. Berkshire Hathaway lost its AAA in February 2010 while raising $8 billion for the Burlington Northern purchase. Exxon Mobil lost a rating it had held since at least 1949 on 26 April 2016, leaving two.

The Letters Become a Licence

The Securities and Exchange Commission needed, in the early 1970s, a way to set capital charges on broker-dealer inventories that varied with credit risk. It reached for the same shortcut the Comptroller had used in 1936, and in 1975 wrote into the net capital rule a category it invented for the purpose: the nationally recognised statistical rating organisation.

The Commission never defined the term. Recognition came by staff no-action letter, granted case by case on the basis that a firm's ratings were already widely used — a test that only an incumbent could pass. Three firms held the designation for most of the next thirty years. Cantor and Packer's survey for the Federal Reserve Bank of New York laid out the resulting structure plainly: a regulatory barrier to entry had been created accidentally, in a rule about capital haircuts, by an agency that had no mandate to regulate the rating business at all (Cantor and Packer, 1994).

What the franchise was worth became visible once sovereign borrowers were inside it. Thomas Friedman put the point on American television on 13 February 1996, talking to Jim Lehrer:

"There are two superpowers in the world today, in my opinion. There's the United States and there's Moody's Bond Rating Service. The United States can destroy you by dropping bombs, and Moody's can destroy you by downgrading your bonds. And believe me, it's not clear sometimes who's more powerful."

What the Letters Stopped Measuring

Enron's senior debt was rated investment grade until four days before the company filed for bankruptcy on 2 December 2001. Congress held hearings titled Rating the Raters, and the agencies' defence — that they had been deceived by the same disclosures that deceived everyone else — was largely accurate and entirely beside the point, because the collapse of a company built on off-balance-sheet vehicles was exactly the case in which an independent gatekeeper was supposed to earn its franchise.

Structured finance was a harder problem and a much larger business. Rating a railroad meant assessing a going concern with a history. Rating a securitisation meant modelling a pool of loans against assumptions about correlation and house prices, on a security that the arranger could redesign in response to the model's output. The arranger knew the model. The arranger could add or remove loans, resize tranches and adjust credit enhancement until the senior piece cleared the threshold for AAA — which, because of the 1936 hook and its descendants, was the grade that determined whether banks, insurers and money funds could buy it at all.

Volume followed. The market that had grown out of Ginnie Mae's pass-through experiments of the late 1960s was by 2006 producing rated tranches faster than any rating committee could scrutinise them individually, and the agencies were competing for the mandates. Two Standard and Poor's staff discussed a deal by instant message on the afternoon of 5 April 2007. One said the model did not capture half the risk. The other replied:

"It could be structured by cows and we would rate it."

The Financial Crisis Inquiry Commission reached its verdict in January 2011. Its report called the three agencies key enablers of the financial meltdown and observed that the mortgage-related securities at the heart of the crisis could not have been marketed and sold without their seal of approval. The Commission's own estimate was that around 83 per cent of the mortgage-backed tranches Moody's rated Aaa in 2006 were later downgraded.

After the Reckoning

Congress passed the Credit Rating Agency Reform Act in 2006, replacing the no-action-letter path to recognition with a registration system and opening the category to new entrants. Section 939A of the Dodd-Frank Act in 2010 went further and directed every federal agency to remove references to credit ratings from its rules and substitute standards of creditworthiness of its own. Regulators found the second instruction much harder to obey than to legislate, because the thing being removed was a shortcut that no agency had the analytical staff to replace.

Money changed hands eventually. Standard and Poor's parent settled with the Department of Justice and nineteen states on 3 February 2015 for $1.375 billion over ratings issued between 2004 and 2007. Moody's settled with the Justice Department, twenty-one states and the District of Columbia in January 2017 for roughly $864 million. Neither settlement included an admission that any specific rating had been knowingly false, and no individual analyst at either firm was charged.

The business emerged larger. Sovereign downgrades during the eurozone's long Greek emergency moved bond markets in 2010 and 2011 much as they had before, and Standard and Poor's cut the United States itself from AAA on 5 August 2011 without any consequence to its own franchise. Flandreau and Mesevage's archival work on the nineteenth-century antecedents makes the structural point that survives every scandal: rating agencies are durable not because their opinions are unusually good but because a market that has organised its rules around a shared vocabulary cannot easily agree on a different one (Flandreau and Mesevage, 2014).

Two American companies still carry the grade John Moody invented — a pharmaceutical maker founded twenty-three years before his book appeared, and a software company founded sixty-six years after it. Everything else in the alphabet he printed in April 1909 has been repriced, litigated and written into statute, and the column of letters still runs down the side of the page.