Resident Fellow Alex J. Pollock

In that wonderful 1940 book on investing, Where Are the Customers' Yachts?, Fred Schwed pointedly and so rightly observed that what everybody in financial markets wants to know is the one thing nobody can know: the future. So Wall Street invents lots of ways to give opinions about the future, which will make people feel confident enough to buy or sell something.

For the fixed income markets, the most important types of such opinions are those of the credit rating agencies on the probability that obligors will be able to pay as called for in the debt instruments. Indeed, the rating agencies describe themselves as being in the business of publishing opinions on credit.

In this I believe they are right, and I have a good deal of sympathy with the thought that in the course of financial events, some such opinions will inevitably prove to have been mistaken, even disastrously mistaken. Of course, sources of opinions which are too often mistaken will have little value to investors, except that they may have regulatory, as opposed to predictive, value. Of this, more in a moment.

Beset and Beseiged

When credit opinions turn out to have been wrong, and many people have lost a lot of money, let's say tens of billions of dollars, naturally there is a search for the guilty. In the midst of our current subprime mortgage bust, the credit rating agencies find themselves uncomfortably in the glare of accusation. I cannot help feeling a sneaking sympathy for them, as they are attacked from so many sides at once.

The U.S. Congress has held hearings, in which some of our distinguished colleagues on the panel have raised thoughtful, well-informed and important criticisms of the dominant rating agencies' structures and practices with respect to structured securitizations of subprime mortgages. The President of the United States has asked the Secretary of Treasury to examine their role through the President's Working Group on Financial Markets. The SEC is investigating whether they were unduly influenced by issuers of the securities they rated. The Attorney-General of Ohio, which has especially high foreclosures from subprime loans and also investment losses in its state pension fund, wants to sue them.

Internationally, the Chancellor of Germany, the President of France, and members of the British Parliament have singled the rating agencies out for criticism. The President of the European Central Bank has suggested possible regulatory actions.

Meanwhile, the stocks of Moody's and of McGraw Hill, S&P's parent company, are off over 40% and over 30%, respectively, since June.

So beset and besieged, no wonder all the major rating agencies declined our invitation to participate in this conference.

Risk Models

"Investors across the globe have lost confidence in the ratings," the Wall Street Journal has opined. This is certainly part of the large uncertainty premium now affecting the prices, or lack of any clear price, of structured mortgage securities.  In my view, one should be skeptical of all opinions about the future, even when they are formed with thought, care, the analysis of massive amounts of data, complex models, and good intentions.

Complex risk models themselves risk becoming outmoded as financial market behavior changes to take account of the use of the models. In other words, the recursive nature of financial markets may cause the model ultimately to fail. This obviously happened in the subprime mortgage boom and bust. How ironic that the banking world is moving to the "Basel II" capital standards, with their heavy reliance on credit ratings and models, just as credit ratings and models have been embarrassed by the subprime disaster.

Of special interest in the ratings models is the role of House Price Appreciation or "HPA," because the mortgage boom reflected and fed a remarkable house price inflation. A reliable factor in mortgage credit behavior is that rapid increases in house prices will keep defaults and losses low--if you get in trouble, sell the house at its new, higher price. If you are short of cash, refinance and withdraw the new equity. Going the other direction, falling house prices increase defaults--why keep paying on the mortgage when the house is worth less than you owe?--and increase the losses to mortgage investors as they sell foreclosed properties.

As recently as last year, it was common to say that although there could of course be regional declines in house prices, U.S. house prices could not fall on a national basis--well, they hadn't done so since the 1930s. But as the first years of the 21st century brought an unprecedented inflation in house prices, there now follows the inevitable correction. House prices are falling on a national basis, and by some forecasts may fall 15% or more. So HPA has become "HPD": House Price Depreciation.

What would happen if all structured mortgage securities were re-rated with the current assumption of falling house prices? Presumably it wouldn't be pretty. How should the original ratings have taken account of the future possibility of nationally falling prices? When should possible HPD have become a factor?

The "NRSRO" System

Since all opinions are liable to error, and opinions based on models are liable to systemic error of vast proportions, as we now can see, why would the U.S. government want to enshrine certain opinions as having preferred, preferential, indeed mandatory status? It should not want to, but it did, under the SEC's old system of "NRSRO" designation for a few favored credit rating agencies.

By requiring all regulated investors--banks, insurance companies, pension funds, thrifts, mutual funds, or the vast majority of institutional investors--to use NRSRO credit ratings, government regulation created a mandated demand for these favored ratings. Whether or not they had any predictive or informational value, i.e. competitive market value, they had and have great regulatory value.

While mandating demand, regulation also restricted supply. What do you suppose happened to price? This combination powered the remarkable earnings, return on investment, and until lately, stock prices of the dominant rating agencies.

A central objective of the Credit Rating Agency Reform Act of 2006 was to increase competition in the rating agency sector. It took away the complete discretion the SEC staff had exercised to designate NRSROs based on never-articulated criteria, and replaced it with a registration regime, opening the market for credit ratings in the regulated sector to competitors who would meet the registration requirements.

Greater competition remains a key objective. When it comes to opinions about the future, the more, the merrier. Having more competitors increases the chance that new insights into credit risks and how to conceptualize, analyze and measure them will be discovered and made available to investors, advisors and other actors in the debt markets. It will also reduce the economic rents (monopoly profits) granted to the old government-sponsored cartel and paid for by everybody else.

The Investor-Paid Rating Agency Model

A particularly desirable form of increased competition would be from rating agencies paid by investors, as opposed to those paid by the issuers of securities. It needs no argument to see that the alignment of incentives in the investor-paid case is superior to that in the currently dominant issuer-paid model.

The major rating agencies have received much criticism of late for the potential conflict of interest represented by being employed by those who obviously want the highest possible ratings and the lowest possible subordination in structured securitizations. My idea is not to restrict the issuer-paid model, but to induce a meaningful competition between the two payment models and let investors and other users of ratings make the choice of which they prefer.

It has been suggested that ratings should include a disclosure of whether they were paid for by issuers or investors. This seems a reasonable idea. Also Congress might require all the regulatory bodies whose regulations supported the government-sponsored credit ratings cartel to study how they might promote the competitive objective of the 2006 act and to consider facilitating the investor-paid model. Certainly the SEC should get more investor-paid rating agencies registered under the new NRSRO definition of the 2006 act.

Is there a way that a more robust presence of the investor-paid ratings could be created? One way not to achieve this would be to have heavier SEC regulation of the rating agencies. But why couldn't a group of major institutional investors set up their own rating agency, capitalized by and paid for by the investors, working from their point of view, and supplied with top talent and technology?

I think it likely that the market would reflect a preference for ratings from such an agency.

A number of good proposals for improving the rating agency system have been made, by members of this panel, among others. These include:

• Distinguish the ratings of corporate debt from ratings of structured securitizations, perhaps using a different range of symbols.
• Accompany ratings of structured debt with calculated probabilities of default and loss-given-default, as well as the key assumptions.
• Have much more emphasis on monitoring and updating ratings over the life of securitized asset pools.
• When new values for key parameters are adopted, for example, when HPA becomes HPD, have existing structured securities re-rated using the revised assumptions.
• Have ratings reflect the risk of increased complexity and opacity of structures.

If major investors took up the suggestion to set up their own rating agency, they could instruct it to carry out all these ideas, which would make its ratings more useful and competitively robust. This would certainly be a more direct and controllable process than trying to get the Congress to redesign rating agency practices.

We were treated once again in this morning's papers to discussions of multi-billion dollar write-offs resulting from the subprime mortgage bust. With the rating agencies so tangled up with the troubles of subprime securitizations--making those of us with a certain kind of education think of Laocoon and his sons intertwined with the serpents--we can say with assurance at that the issues and debates about the rating agency system are not going away any time soon.

Alex J. Pollock is a resident fellow at the AEI.