Let’s not stop there with the clichés.* Here’s a great one: “never attribute to malice that which can be adequately explained by stupidity.” In applied form: your model of all the AAA mortgage CDOs that were maybe not so AAA could be “ratings agencies were paid by banks so they were venal and corrupt and sold the banks good ratings on products they knew were bad.” Or it could be “ratings agencies created medium-dumb criteria to make a thing be AAA, and bankers who were smarter than medium-dumb arbed those criteria to make more things be AAA than should have been AAA.” The incentives model has good economic theory behind it, and some suggestive evidence; the stupidity model has that lovely cliché but also some evidence, about which more later.
But first hilarious contrarian ratings agency Egan-Jones is in trouble:
The U.S. Securities and Exchange Commission is planning to vote on Thursday on whether or not to charge credit-rating firm Egan-Jones with making intentional misstatements to regulators when applying to be a “nationally recognized” rating agency, people familiar with the matter said.
The possible charges, which would require a majority commission vote, pertain to issues such as misrepresenting the firm’s rating experience, conflict-of-interest policy issues, and a failure to keep certain books and records, the people said.
So, look, I’ve teased Egan-Jones a bit before, because the things that have gotten them a lot of attention recently – being first to downgrade a sovereign issuing 2%ish debt in its own currency that happens to be the world’s reserve currency; downgrading Jefferies possibly without reading its balance sheet – have all looked, to this casual observer, pretty silly.** So maybe that’s due to a lack of rating experience, conflict-of-interest policy issues, and/or a failure to keep certain books and records.
Or maybe it’s due to stupid. The Journal has more background on the controversy around Egan-Jones’s approval as an NRSRO in 2007:
[Much lamented former SEC inspector general David] Kotz said in the report that SEC staff had expressed “concerns about the adequacy of the [firm's] managerial resources, suspicions regarding the accuracy of the financial information provided in its application and concerns about the authenticity of a number of certifications,” but the agency ultimately approved the application.
This seems closer; “concerns about the adequacy of managerial resources” is not unsynonymous with “stupid” and getting your own financials wrong doesn’t bode well for getting Jefferies’s right.
Still, the discussion here has to leave you with a bit of unease that ratings agencies live in a world not of market discipline – get it right and customers will trust you, get it wrong and they won’t – but of regulatory form-checking, where the way to get the (still important) public blessing as a “nationally recognized statistical rating organization” is to fill out certain forms correctly and keep certain books and records. And that when you put the SEC, an organization famous for its forms, in charge of approving ratings agencies, then rating agencies will adapt and become expert at forms and inexpert at, um, evaluating credit.
I went to this conference thingy last week and I was struck by this paper by John Hull (you know him from this book) and Alan White, which makes a point that is not exactly news and that I feel like I had rattling around in some dim corner of my brain before but hadn’t really seen the implications of:
The criterion used by S&P and Fitch [in rating structured products] aims to ensure that the probability of a loss on a structured product with a certain rating is similar to the probability of a loss on a corporate bond with the same rating. The criterion used by Moody’s aims to ensure that the expected loss on a structured product with a certain rating is similar to the expected loss on a corporate bond with the same rating.
So if say a BBB company has a 2.5% chance of defaulting, a BBB MBS tranche should have a 2.5% chance of experiencing a loss, of any size. But as Hull & White point out, that leads to a blindingly obvious arbitrage: if you have a portfolio with a 2.5% chance of experiencing a loss (of uncertain size), just cut it in half, with a junior tranche that takes the first 50% of losses in the portfolio and a senior tranche that loses only if losses are greater than 50%. The senior tranche has to have a lower probability of loss than the portfolio as a whole – since if there are any defaults that lead to less than 50% losses the senior tranche avoids them – whereas the junior tranche has the same probability of loss as the portfolio as a whole. So the senior tranche should be AAA or whatever, and the junior tranche should be BBB. You’ve taken $100 of BBB stuff and turned it into $50 of AAA and $50 of BBB. Repeat as necessary on the BBB tranche, with the result that that BBB tranche will have the same probability of experiencing a loss as a BBB bond (maybe) but a vastly higher loss in the event of default – and therefore won’t be at all the same credit quality. Hull & White go into more detail and make a plausible case – based on disparities with Moody’s ratings (which are relatively non-arbitrate and loss-based) among other things – that this is exactly what happened with credit ratings for mortgage-backed securities and (more worryingly) CDOs and CDO^2s thereupons.
Imagine a world where the SEC inspector general asked questions about that.
To be fair this was all disclosed and if you had a brain and a heart you’d notice that arbitrage and determine that BBB mortgage bonds are terrifying and so they’d trade – and in fact did trade – at a discount to BBB corporate bonds. But it’s hard not to agree with Hull and White that many investors used ratings as proxies for credit quality, not probability of default, and that in any case financial regulators treated mortgage bonds at least as well as corporate bonds of the same rating even though they were mathematically different beasts.
In a sense it doesn’t matter if Egan-Jones is five drunk ex-cons in a room forging their own character references and financials, as long as they have ratings that are better predictors of credit quality than … whatever your baseline is, call it “S&P” or “Moody’s” or “nothing” (right? why not have nothing be your baseline?). What you want is a model where someone is asking of Egan-Jones, “are you right?” and if the answer is yes then they listen to them and if it’s no then they don’t. If that someone is the market then, well, maybe that’s good, or maybe market participants will be terrible at figuring out if Egan-Jones is right and then it’s bad. An overlay of Official Government Approval weakens market participants’ incentives to do that evaluation, which is maybe fine if the government is doing a better job than the market would, and given information disparities and incentives and whatnot maybe that’s plausible. But if the government is (1) evaluating ratings agencies at least in part on the basis of things that look … kind of non-substantive? and (2) also doing that wrong, that does not bode particularly well for the quality of the ratings.
SEC to vote on charges against Egan Jones-sources [Reuters]
Small Credit Rater Raised SEC Concern [WSJ]
Ratings, Mortgage Securitizations, and the Apparent Creation of Value [Russell Sage Foundation]
* Also: resident Dealbreaker classical scholar B. Levin (no relation): “I assumed that that headline would mean ‘keeping up with the Egan-Joneses.’ No?”
** Oh btw do you know my favorite thing about Egan-Jones? That they advertise and evaluate themselves based on Moody’s and S&P ratings. Like, “hi, we’re Egan-Jones, and we are pretty pretty good at predicting what Moody’s and S&P will do.” That is I suppose a valuable service but it’s kind of a funny one for Moody’s and S&P’s competitor to provide.