It's true: the rating agencies really screwed up before the financial crisis. They got a lot wrong about complex securities; investors trusted them and got burned. Indeed, some of them are still screwing things up, as S&P announced a few weeks ago that it was downgrading billions of dollars in complex securities that it rated in 2009. But considering that we, and especially investors, should broadly understand their ineptitude by now, can we really continue to blame the agencies for any problems that arise after the financial crisis?
Jesse Eisinger from ProPublica thinks so. He continues to complain about the rating agencies in a column this week. He notes the problem with "re-remics" that I just mentioned above. These were products created when banks repackaged a bunch of toxic securities and agencies slapped on high ratings. Then, S&P had a Homer Simpson "Doh!" moment after realizing a mistake in their analysis. This makes Eisinger, and no doubt others, angry. He writes:
During the mortgage securities boom, bankers knew more about their bonds than the ratings agencies did and took advantage. A similar problem occurred with re-remics. "Chances are that if a bond is getting re-remicked, it's a bad bond and the holder wants to forestall the inevitable reckoning," Mr. Kolchinsky said. The ratings agencies somehow missed that.
There also looks to have been "ratings shopping," where issuers seek out the most lenient firms, rather than the best. S.&P., according to Mr. Kolchinsky, was slower to downgrade residential mortgages than Moody's was. Lo and behold, S.&P. nabbed the bigger market share in new offerings of residential securities. Then came the big debacle.
These were common post-crisis complaints about the agencies, reapplied. He goes on to worry that financial reform didn't do enough to reform the rating agencies and to complain that the little the legislation did do has been mostly nullified by the Securities and Exchange Commission or won't be funded by Republicans. All these points could be validity. He's even somewhat right to question the practice of creating these re-remics in the first place, though I would argue that they could theoretically be created to perform well if structured properly. But none of this should really matter.