Rethinking Credit Ratings

Now that the summit of world leaders is over, the hard work begins. What will it take to make the global finance system's architecture safer after most of its liquidity drained through a $600 billion hole in the U.S. mortgage market? Thousands of European jobs are in jeopardy as a result.

What European leaders request or require depends on a real understanding of the root causes of the crisis. They may want to follow the lead of voluble EC Internal Markets Commissioner Charlie McCreevy — and start with the rating agencies.

To see why, one first needs to have a clear sense that the current crisis is a failure of investors — and not so much one of borrowers.

Come to think of it, why should we level so much blame at homebuyers — those who actually took out the $600 billion of mortgages that are in trouble?

Plenty of rational borrowers in the United States would do exactly the same thing all over again if faced with the offer of financing a house for a measly 6%, 3% or even no money down. Living in a world of no-recourse loans, they would say, "Thanks, why not?"

The same thinking applies to mortgage brokers, who have come in for a lot of criticism for irresponsible behavior.

It's a fair bet they, too, will continue to do what, in effect, they have done from the beginning of time — or at least the beginning of the banking system. They will continue to originate whatever banks and other mortgage investors will buy.

Next, it makes little sense — even if it's politically convenient — to put much of the blame on Fannie Mae and Freddie Mac.

These institutions have bought mortgages originated by banks, packaged them and sold securities backed by them for over a quarter century, not just the last four years. And Fannie and Freddie were not the first to try to sell sub-prime loans to financial investors.

But once those investors started rewarding borrowers, brokers and originators for what turned out to be crazy loans, Fannie and Freddie panicked and concluded they had to participate in the riskier underwriting that the market seemed to be ratifying — or go out of business.

What's really new this time around — and therefore lies at the core of the problem — is the collapse of credit discipline on the part of investors, and especially the major financial institutions that have suffered losses on securities backed by sub-prime loans.

As these institutions saw it, it was good enough that rating agencies told them the stuff was fine. So the real question is this: Why were the raters wrong — and what must change going forward?

As is well known, the fundamental problem is that ratings are paid for by the institutions that issue securities — and not by investors. That creates all the wrong incentives on the part of rating agencies. But those incentives are easy enough to change.

The more serious problem is that those misguided incentives seem to have eroded sound credit rating practices — and that may lie beyond the reach of simple structural fixes.

The deterioration is serious enough that European leaders may want to start all over again — either demanding new regulation of existing rating agencies or establishing a new one sponsored by Brussels.

A solution at the European level could avoid some of the pressure governments would bring to bear on agencies sponsored by individual countries. And it could repair underlying weaknesses in current rating practices.

Years of research on best practices with performance scorecards and rating systems show that credit analysts at the major rating agencies have lost at least three crucial habits. The first is to be clear about the assumptions they made.

More precisely — they have to be specific about why something new should be as safe as the old stuff.

Credit analysts should at least have explained why, if sub-prime mortgages were riskier than conventional ones, there were such large volumes of risk-free AAA securities backed by them. Such a discipline would have forced rating agencies to explain why risk was evaporating from the new structures.

The second habit is to be skeptical. And that means looking for performance measures that really test the assumptions behind a new product. As humans, we are much more inclined to seek — and pay attention to — affirmation.

We can always find favorable feedback, however, while it is the unfavorable feedback that always provides useful warnings. For example, analysts could have combed the experience of both sub-prime mortgages and sub-prime-backed securities affected by Hurricane Katrina to see whether they were behaving like conventional ones.

You don't need to know everything about a new risk or digest every piece of data. You just need to be skeptical about your assumptions — and test them.

The third habit is to respect experience. For the credit analyst, this means being rigorous about tracking results.

Once one has laid out expectations about risk factors and whatever near-term performance result may reflect credit quality, one needs to compare them with actual results and check the pattern of hits and misses.

This is always a lesson in humility because we are usually wrong. What separates a successful rating or risk analyst from the rest is not how often he or she is wrong, but how aggressively they revise their views with new information.

Humility is rare among risk analysts, because more changes of judgment lead to more volatile ratings. And that is probably unacceptable to the major rating agencies as they are currently constituted.

More volatile ratings, however, would give the rest of us a better sense of the uncertainties in the life of the market.

After all, it did us little good that few AAA mortgage-backed securities suffered downgrades before plunging off a cliff. All of which is why it may be time for Europe to establish its own credit rating competitor.

Establishing a new European rating agency may seem like an extreme solution. However, the competition would be healthy, and the magnitude of the erosion in the main rating agencies' credit practices suggests it may be necessary.