"Ratings Agency Primer: Understanding the “Nationally Recognized Statistical Ratings Organizations” Rule"
Pretty much everyone agrees that we have a ratings agency problem. But I think the conventional way of describing it as a “conflict of interest” that’s created by the fact that “banks pay them to rate their securities” leaves some crucial steps out.
Like suppose I was opening a salad joint in downtown Washington and wanted to convince people of the high quality of my salads. This might be tricky, especially if my business model was to buy a bunch of crappy ingredients and tell people they’d been mixed together in just such a way as to be delicious. One thing I definitely couldn’t do is just turn to Moody’s Salad Rating Agency and pay them to say “these ingredients may be gross individually, but combined they’re delicious—AAA!.” And the crux of the problem isn’t that regulators would stop me, it’s that it would be pointless—nobody would take Moody’s Salad Rating Agency seriously if it operated on that business model.
A plan that actually might work is that I could go to The Washington Post and promise to pay them under the table in exchange for a favorable review. But it’s important to note that that kind of fraud isn’t what’s going on with the ratings agencies. Everyone understands that this conflict of interest exists, but it doesn’t kill them off.
That seems mysterious, and at first glance you might reach for a free market approach. After all, why does it make sense for me to pay Moody’s SRA in the first place? Well, presumably because salad-eaters deem Moody’s to be credible. Consequently, even though Moody’s does< want as much business as possible, Moody's knows that in the long run it maximizes revenue by only giving the AAA-rating to really good salads. If Moody’s engaged in some giant salad screw-up, salad-eaters would lose faith in Moody’s-rated salad shops and reputable salad vendors would start shifting their business to other salad raters. In theory, a salad-rater could be perfectly credible notwithstanding the conflict of interest. In practice, however, consumers are likely to put more faith in third-party reviews in reputable media outlets or just rely on recommendations from friends.
The real question about the ratings agencies is why this kind of process doesn’t work. To understand the answer, you need to learn the term “Nationally Recognized Statistical Rating Organizations”. What’s an NRSRO? In short, it’s a ratings agency. But NRSRO isn’t just the name of a line of work, it’s the name of a special regulatory status. To be an NRSRO you need to be certified as one by the Securities and Exchange Commission. And becoming an NRSRO isn’t just a matter of having a fancy label, it’s embedded in a web of other regulations. For example, regulatory assessment of bank and insurance company capital reserves requires assets posted as reserves to have an NRSRO-certified rating. NRSRO-certified ratings are also relevant to regulations about money market funds and some pension funds.
This does two things. One is that it reduces competition, since it’s hard to get that NRSRO label. But that’s probably the less-important consequence. The more important consequence is that it undermines buyer-side incentives for ratings to be accurate. The “major” ratings agencies all use the widely criticized issuer-pays model. But if, qua buyer, you don’t like that model you can instead choose to rely on a smaller rater like Egan-Jones that doesn’t use that method. That, however, assumes you actually care about how risky the assets you’re buying are. But in many cases you care less about actual risk than about how your risk plays for regulatory purposes. It’s as if you’re not actually planning to eat the salad, you just need to tell someone else that you bought a AAA-certified tasty salad. And if you care more about the rating than the actual quality, then you don’t care whether or not Moody’s SRA has a sound methodology, you just care that it’s cheaper and more convenient to let the salad vendor go through the hassle of getting the rating.
Note that this is an issue where adopting the kind of discretion-free “Roman” approach that many liberals have been promoting in other contexts is part of the problem. The rigid rule may have made sense when it was initially adopted, but over time what started as a rule business practices have evolved that are based on knowledge of how the rule works that undermine its spirit.