The massive problem with the ratings agencies, a problem that shows up everywhere in the financial crisis, is something that needs a major fix. But it a problem that quickly leads to despair, as every solution seems to bring new problems and for whatever reason it is difficult to get Congress to act on it.
For some background on the topic, Matt Yglesias wrote about NRSROs at length last week that was a good primer for reform.
In order to get into the deep end of ratings agencies reform, with a focus on what we want the end goal to look like, I did an interview with Jerome S. Fons, a former director at Moody’s.
Jerome is a Special Consultant with NERA Economic Consulting and heads his own consulting firm specializing in credit risk applications, expert testimony and litigation support. Until August 2007, he was Managing Director, Credit Policy at Moody’s Investors Service in New York. He was a member of Moody’s Credit Policy Committee and chair of Moody’s Fundamental Credit Committee. You may know Fons as a co-author, with Frank Partnoy, of a New York Times editorial critical of the ratings agencies titled Rated F For Failure.
(Full disclosure. For 1 1/2 years I worked at Moody’s KMV, a well regarded credit risk firm founded as KMV by three old-school quants, including Oldrich Vasicek. Moody’s bought it in the mid-2000s while it was flush with cash from rating all those bonds AAA. Like my friends at Moody’s Economy.com, I often take pains to explain I wasn’t involved in the CDO rating business and that we were our own seperate business line.)
Michael Konczal: What do the ratings agencies do?
Jerome Fons: They play an important role in the capital markets. Their job is to provide an independent assessment of value. They communicate that through a ratings scale. They had a pretty good track record up until the mid-1990s. On the basis of that track record, they became institutionalized in a way that was never intended.
Institutionalized through laws?
Part of it. You can point out some instances of ratings being used by bank regulators in the 1930s. But the real game changer was the SEC’s creation of the NRSRO, or Nationally Recognized Statistical Rating Organization in 1975. That formalized the definition of an officially sanctioned ratings agency. The SEC designated these firms through so-called “no action” letters where they would not prosecute anyone who relied on their ratings. Their original use was for setting broker-dealer capital levels.
Others started to piggyback on this official designation. Numerous laws and designations referenced the NRSROs, and in a parallel fashion, the private sector adopted them in guidelines, financial contracts, loan agreements and swaps, and those sorts of things.
That’s where you get things like pensions needed to hold AAA rated bonds.
That’s another regulation, and that’s from other government agencies piggybacking onto the SEC’s use of the NRSRO. And ultimately the notion was that ratings were there to protect investors, so they’ll also be able to protect taxpayers, pensioners and others. They essentially outsourced the critical function of independently determining risk to these private institutions.
You mentioned the 1990s. When did the problems start to kick in?
People were long aware of the conflicts of the ratings agencies. The ratings agencies didn’t start to charge the issuers until around 1970. Before then, they relied mainly on subscriptions to manuals and consulting type lines of businesses. Moody’s had a service where they’d provide advice to individuals and institutions on how to construct portfolios. That was a massive revenue source.
People were also aware of the barrier to entry posed by the NRSRO designation. The SEC would grant a “no action” letter to those firms that they thought were helpful. There was no clean, clear criterion for who should be designated or not.
So I couldn’t just create my own ratings agency and get designation.
You could create a ratings agency, you just couldn’t get an NRSRO. The classic example is Egan-Jones, who began issuing ratings around 1995. Throughout the late 1990s and early 2000s, they often complained that they weren’t being recognized. They weren’t recognized until late 2007.
By the early 2000s you had a series of problems, the telecoms, Enron, Worldcom were all blowing up with high ratings. Those got Congress’ attention, numerous hearings were held, and this all resulted in the Credit Agency Reform Act being passed in 2006.
So wait, the ratings agencies have already been reformed recently? I take it that isn’t good enough to get us where we need to go.
That act is actually very short, and it simply defines a ratings agency, it defines an NRSRO, and it specifies a process by which a firm can apply. It forced the SEC to develop rules for NRSRO recognition.
So let’s talk a little bit about the conflict as it showed up in the 2000s.
I think for the first time, you had people question the business model of the ratings agencies. And it occurred as a result of the disasters of Enron and Worldcom and other firms that went bankrupt with high ratings. And everyone started to ask, how did the ratings agencies miss this? And people thought maybe that they had an incentive to miss this.
The literature up until that point had argued that reputation concerns would keep these firms in line, but it turns out in hindsight that reputation risk alone didn’t work. In hindsight, Enron and those other disasters might have been good things for the ratings firms. Even though they were sued, nobody successfully sued them. But the disasters focused attention on credit, and they claimed to have beefed up their practices as a result. They hired more people and raised their fees. The disaster didn’t hurt their business at all, in fact it helped it.
Now, Frank Partnoy and others will argue that their reputation concerns didn’t mean a thing. What the ratings agencies were selling was not a quality product but a license. They were selling a regulatory license to those who wanted to issue a bond, especially into a regulated portfolio.
Now there’s two problems people bring up with the ratings agencies’ role in the recent crisis. One is the story of the conflict and the sheer amount of money they were being paid in order to rate these mortgage bonds. There’s another story where they were simply outmatched in technique relative to the task they had to do. Their models, essentially, looked back 30 years and didn’t see a nationwide housing bubble, so assumed that there couldn’t be another one over the the course of the next 10 years. How do you consider these two stories?
People often point this out. I would ask instead “Was it bad models, or was it bad business practices?” Well, if the ratings agencies had better models, would they have blown the whistle and said no? My guess is that the answer is no. They had business pressures that overrode analytic concerns.
I find it helpful to make a distinction between the business of ratings, and the craft of ratings.
So to whatever extent there was a craft problem, there was certainly a business problem.
And what’s unfortunate is that the craft and profession of ratings has been sullied by this. I got this when I was on the Hill, people would say to me “it’s all bulls**t. Ratings are crap, there’s nothing there.” And I would say look, there’s a history of nearly 100 years of doing a good job. It can be done. Don’t throw the baby out with the bathwater. There is some substance to the practice, but in this case it was overwritten by the business pressure.
So, what can we do about the business pressure?
That’s the $50,000 question. Some of the reforms get to it. For example, there were two things that stood out in this last crisis. The first is the forum shopping, or what I call rating shopping. The idea that the issuer gets to pick which agency to use, and exclude others. This leads to cherry-picking and contributed to the race to the bottom on standards. So one of the things that the SEC did was that, for any structured product, you have to make the underlying data available to every otter NRSRO. The implication is that rating agencies having access to that information will call out shoddy ratings by a competitor. I’m not sure they will do this.
They should go further and require that it be disclosed to the market. Anything that privileges the ratings agencies shouldn’t be done.
Why don’t investors pay for the ratings? Doesn’t that make more sense?
There are two things here. One is that the value of a rating accrues mainly to the issuer. The value comes in the form of better access to capital and lower cost of funding. Second, investors have historically been unwilling to pay for ratings. There’s a free rider problem, and there’s also an institutional bias against it.
So one regulatory suggestion is to randomized selection of ratings agencies. The SEC picks a ratings agency at random, instead of the person who wants to issue a bond.
Rotating agencies. One problem is that for much of the US, it’s a 2 or 3 ratings market. Most of the structured transactions had 2 rating agencies assign ratings and some had 3. Likewise with rated firms. It’s a very small field and there tends to be multiple ratings agencies on a deal anyway.
So the next step would be to get more ratings agencies into the playing field.
Right, except that the more ratings agencies you have, the more they compete not on the accuracy of their ratings, but on the ratings they are willing to issue and on other aspects of “service.” Remember the agencies aren’t selling accuracy, they are selling this regulatory license.
So we need to decrease the value of this license.
Right. You need to disentangle the ratings agencies and the regulation. One of the nice things about the House bill is that it raises rating agencies’ legal liability and it instructed every federal agency to remove references to NRSRO. That did not make it into the current Senate bill. In the Senate bill, the GAO will study it and give a report.
Currently, the ratings agencies stand behind the first amendment shield, claiming that ratings are nothing more than opinions, and opinions are protected by free speech. That’s plausible in the corporate bond market, they often are providing opinions like a consumer reports. In the structured finance market, it seems that that is a more tenuous argument. They weren’t just impartial observers at arm’s length. They were intimately involved with the creation of these things. Many of them wouldn’t have been created without the ratings agencies, and the ratings agencies had a financial stake in marketing them.
Rating a corporation is more of an arms-length affair: the rating agency has little impact on the corporation’s business, the rating fees are modest and there is typically enough information in the public domain to expose a shoddy rating. In the structured area, the fees were quite high and the ratings agencies were deeply involved with the creation of these things. So for some folks, it crossed the line from free speech to commercial speech.
Let’s throw out some random ideas I’ve heard. One is that we use bond spreads and/or CDS spreads and/or other market information to do the ratings for us.
The CDS market is not as liquid as you’d think. They are limited in the number of names they cover, maybe 600 or 700 are actively traded. I don’t know what really good information is in them. They are also really volatile. For better or worse, investors want fairly stable ratings. So if I have a rating jumping from single-A to B to single-A twice in a day, investors are not going to rely on them for portfolio governance purposes– not that that is a bad thing. The same is true for bond spreads. However, they may make good (though not great) early warning indicators.
This also doesn’t help us for the problem of new issuances. Which is a big problem.
Exactly. I’m a big fan of using markets, though I’m not a huge fan of market efficiency or market omniscience. I’m not a Chicago School person. I intellectually buy into a lot of their stuff, but for that I don’t.
What about not paying the agencies upfront? What about forcing them to get paid over the life of the deal, so there’s a de facto clawback mechanism?
That’s an interesting concept. The bankers and lawyers on the deal are also paid up front too. So you’d have to do that in a fair way.
What about a public option for ratings agencies? The SEC has a division of ratings, and then the market is completely deregulated and everyone else can do whatever they like.
Interesting. Could it truly be independent? And get the trust of investors? That would be a challenge. Would it become politicalized? Would it not be able to favor domestic firms over foreign firms? I don’t think so.
I agree. Now what about something like what Josh Rosner focused on at Make Markets Be Markets for securitization reform, to go with what is in the House, a focus where you ruthlessly focus on the transparency and accessibility of the information ratings agencies use.
I like that approach. Here’s a bolder concept, and I’m not sure if its going to happen in our lifetime. Imagine a central repository of ratings methods and data, accessible to all. Imagine a world of transparent data for all transactions and all securities. What we could then do is empower users, train them in the use of the methods. And evolve methods as needed. You could broadly democratized the process.
In this world you would do your own work, and it isn’t outsourced to clearly conflicted people. I think that’s where we’ll evolve to, but it’ll take a huge shift in our thinking and infrastructure.
If you can get the information out there, people will find the methods. So get the information out there, reduce the value of the license, make firms accountable in some real way.
The major ratings firms could improve transparency today by simply refusing to rate an obligation unless all relevant data is made public. I think that would be an important first step.
The accountability thing is also important. Right now people feel they aren’t accountable.