
There were two main schools of thought as to what should be done to improve the ratings agencies. One said that the problem was conflicts of interest and wanted to solve this through random assignment of raters. Another said the problem was cartelization and wanted to solve this through forcing agencies to sell accuracy rather than franchise value. Today, it seems to me that today the Senate passed amendments reflecting both those ideas.
First, Al Franken’s amendment passed:
Franken’s measure would create a central clearinghouse, regulated by the Securities and Exchange Commission, to assign firms a credit rating agency to give an initial rating on a security, eliminating the practice of banks shopping around for the highest marks. Financial firms could seek out subsequent ratings on their own, but any discrepancies between the assigned and unassigned ratings would be made public.
But George LeMieux’s amendment also passed:
“There is a handful of federally-approved rating agencies that gave their top marks to some of the worst investments. Those stellar ratings made bad investments seem sound, up until when they crashed the markets,” said LeMieux. “An investment rating should mean something, but today it doesn’t. Removing their federal endorsement will end the dangerous over-reliance on these ratings and allow sound measures of risk to re-emerge in the marketplace-giving investors confidence an investment’s true risk is known.”
It’s not totally clear to me what LeMieux means by this and I haven’t seen a clearer explanation of the amendment anywhere. But he seems to be saying that he’s going to have the SEC stop designating official Nationally Recognized Statistical Ratings Organizations.
These are both decent ideas, but they’re also contradictory. The Franken process assumes there’s some finite number of authorized raters—NRSROs or some equivalent, but the LeMieux concept is precisely that there shouldn’t be any such set of privileged raters. Sometimes when faced with two ideas about how to solve a problem, doing both is smart. But as best I can tell that doesn’t apply in this case.
Update
Tim Fernholz IMs to say I’m misunderstanding what the LeMieux amendment does. He says it doesn’t get rid of NRSROs, it gets rid of regulatory dependence on NRSROs and enjoins regulators to develop their own independent risk assessments. That’s a good idea that would complement Franken’s.
Note that the actual language has been made available, Tim’s story seems to check out. Would be nice if this had been posted a couple of hours ago when the vote happened.
Update
,Here’s a more detailed account of LeMieux’s amendment from Mike Konczal. It looks good to me.