Bust up the Ratings Cartel

The other Wall Street research problem. 

Last week’s grand settlement between New York Attorney General Eliot Spitzer and all the major Wall Street firms may or may not result in better or more honest stock research. But as the era of corporate scandals draws to a close (its conclusion heralded by the perennially lagging indicator of a Time “Persons of the Year” cover), it may be worth lingering over another area of securities analysis as yet untouched by reform: credit ratings. This sleepy, little-understood area of the financial-services sector remains significantly profitable. And it’s essentially a government-declared cartel that delivers poor service to the investing public.

Every year, far more debt than equity is sold to investors.In the first quarters of 2002, according to Fitch, $482 billion in corporate bonds were issued, plus hundreds of billions more in mortgage-backed securities and other securitized loans. Each such issue is assigned a rating by credit-rating agencies like Standard & Poor’s, Moody’s, or Fitch. These ratings don’t function as “buy,” “sell,” or “hold,” recommendations. Rather, the complex system of letter grades and pluses and minuses—an issue might be rated BB- or CCC+—indicates a company’s creditworthiness, and hence dictates how much interest it should have to pay. The ratings agencies also wield power over sovereign nations, since they rate government bonds.

In the late 1990s, the credit-rating agencies performed just as poorly as the stock analysts who maintained “buy” ratings on Enron and Global Crossing even after they had fallen more than 99 percent and were about to enter bankruptcy. Bond-rating agencies are supposed to be the canaries in the coal mines, since they traffic more in the hard data that matters to bond investors than in concepts crafted to appeal to stock investors. But in large measure they failed to sniff out the problems of off-balance-sheet financing at Enron and the rest of the energy sector, the bogus capacity swaps at Qwest and Global Crossing, and the plain fraud at WorldCom. Enron somehow maintained its investment-grade credit rating until Nov. 28—long after its balance sheet was revealed to be a house of cards.

Credit-rating agencies have been around since before the Civil War. In the late 19th century, R.G. Dun & Co. employed a network of 2,000 correspondents who would provide data—much of it gossip—on the creditworthiness of dry-goods merchants, traders, and companies throughout the United States. Dun’s handwritten, meticulous ledgers, now housed at Harvard Business School’s Baker Library, make for fascinating reading.

The business grew and expanded as government regulation of the financial industry increased. Companies like A.M. Best began to rate insurance companies, while others specialized in rating railroad bonds. As the capital markets developed, the corporate credit ratings evolved into a proxy for creditworthiness, and hence into a guide to investors as to how much interest they should charge.

In 1975, a Securities and Exchange Commission regulation created the modern credit-rating cartel. The agency declared that financial institutions that fell under its purview—asset managers, mutual funds, etc.—who wanted to keep bonds or other debt instruments in their portfolios could purchase only assets that had been rated by so-called Nationally Recognized Statistical Ratings Organizations. The theory: Doing so would help ward off investment-company failures and give regulators and investors a better sense of the risk level of an asset manager’s portfolio.

But the SEC’s attitude toward the NRSROs was like that of Justice Potter Stewart toward pornography. It couldn’t offer a specific definition, but it knew one when it saw it.  At first, the SEC recognized three existing credit-rating companies as NRSROs:  Moody’s, Standard & Poor’s, and Fitch. In later years it granted the status to four other firms, including Duff & Phelps. But because of mergers and consolidation—Fitch acquired Duff & Phelps in 2000—only the original three remain.

For the past 27 years, then, any company or institution wanting to sell debt has essentially been required to obtain a credit rating from one—or more—of the Big Three. Yes, the NRSROs compete. But there’s plenty of work to go around. And frequently companies like—or need—to have their bonds rated by two or more agencies.

To a degree this business creates the same principal-agent problem as stock research, in which companies essentially pay investment banks for ratings by funneling investment-banking fees their way. In the credit sector, it’s more upfront: Companies pay an agency for the time and resources it takes to issue a credit rating. The ratings agencies also sell their services to large investors, who rely on them to provide honest assessments of the company’s balance sheet, cash flow, and ability to pay back loans.

In practice, however, the rating agencies display the classic characteristics of an entrenched cartel: They’re lazy, unresponsive, and ultimately unhelpful. They tend to play catch-up, downgrading ratings only after financial weaknesses are revealed or ferreted out by more enterprising researchers.

The answer, of course, is more competition. And there are plenty of companies that would like a piece of the ratings action. But the SEC has been less than responsive to potential entrants. It hasn’t recognized an NRSRO since 1992. And the application process sounds like a bureaucratic nightmare. LACE Financial, a company in Frederick, Md., that rates banks, first applied to the SEC for NRSRO status in 1992. But it didn’t get a ruling on its application until 2000. And when it did, it was negative.

New entrants alone wouldn’t guarantee better performance. But the existing complacent crew—with a guaranteed stream of business and no competition—isn’t likely to provide the marketplace with what it needs. Last summer’s Sarbanes-Oxley Act asked the SEC to conduct a study on credit-rating agencies and examine to the barriers to entry. Let’s hope the SEC doesn’t take eight years to respond to this request.