Ratings want to be free
Embedded video is unfortunately not a very efficient way for me to absorb information, particularly when this video does not show up in my RSS feeds. But I not only clicked through, but watched the whole damn thing when Barry Ritholz' Big Picture pointed me to a video of two gentlemen proposing an alternative to the big ratings agencies.
The two gentlemen - computer scientists Jesper Andersen and Toby Segaran - have formed an open-source venture called Freerisk, which describes itself as "project with the goal of making freely available the data, algorithms and tools necessary to perform risk modeling". More bluntly, the founders, as you can see from the above-linked video, want to loosen the stranglehold of the agencies on credit analysis.
This is a timely, worthy, necessary and maybe even feasible goal. The founders acknowledge that access to the data necessary for debt analysis needs to be improved, and that this data should be provided, presumably to the SEC in a machine-readable format. The algorithms necessary to perform this analysis would be open source.
I say timely because the Obama administration's instincts have been to ask issuers and agencies to provide a greater amount of information, and I say worthy because, as the Freerisk founders demonstrate persuasively, investors need to have access to better means of measuring credit risk. The current system of semi-regulated, issuer-paid, nationally-recognised statistical rating organisations is simply not stable.
I say maybe feasible, because if you're going to start trying to chip away at the agencies' dominance, you'll start, much as newer outfits like Gimme Credit and Egan-Jones have, by looking at corporate debt. The models are simpler, and the inputs and assumptions are much more transparent.
But let me look now at some of the potential pitfalls. I have no idea, for starters, whether publicly-registered bond issuers provide the right information in a homogenous fashion, in the same way that registered equity issuers do. Issuers of bonds that can only be bought by accredited large investors (Rule 144A buyers) do not have to disclose any information at all. You might say that 144A buyers are presumably less in need of this sort of protection, but then you wouldn't have been reading the papers much recently.
Then there's the fact that agencies have not had a particularly horrible time with their corporate ratings of late. Now this is all relative, and their work on financial institutions has been, at the very least, far from timely, but the agencies' biggest failure has been in the rating of structured products.
The Freerisk principals say quite clearly that having proprietary and closed algorithms was one of the factors that got us into this mess in the first place, but if Egan-Jones can't work out a way to break into the rating of structured products, you can assume that the barriers to entry in that business are fairly high. It could be that it's easier than the big boys make it out to be, but he way the industry developed suggests not. Which is not, of course, to say, that issuers can or should be concentrating on structuring fiendishly complicated instruments these days.
They do do things differently in debt markets, this peculiar mixture of wild west and gentleman's club. In equity markets the assumption is that any idiot should be allowed to go out and own any stock they like. In debt markets there's an assumption that some products are definitely left to the professionals. which doesn't make a massive amount of sense because debt products - in general - provide a much more stable income stream and better recovery prospects, though they also require a bit more supervision of issuers, whether directly or by a trustee.
Regulators have traditionally enforced this distinction by leaving the decisions to the agencies, and there's no sign that they will either let investors buy whatever debt securities they feel like or staff up in a sufficient fashion to make these distinctions themselves. That decisions still confronts them.
But chipping away at the edifice of ratings agency dominance will be a gradual process (I see challenges to their use of First Amendment protections to be another means of doing so). I see Freerisk being complimentary to efforts by competing agencies to check their rivals' work by offering unsolicited ratings. Regulators have traditionally frowned on unsolicited ratings, seeing them as a form of blackmail (hire us or we'll put out an unsolicited rating that makes you look bad). But they're probably the quickest way to establish a means of checking ratings shopping among the big agencies.
It would also get investors used to the idea that even if a competitor, whether paid for or free, doesn't have the best analytical tools, then if it has a better record with the way its using assumptions and inputs it could still provide an alternative to the big agencies. I'll be curious how it works out.