An update to this post:

Reader AdamJ writes:

Paulson paid Goldman 15 million, not the other way around. Which is why Goldman was trying to help keep Paulson involved in portfolio selection. And the SEC doesn't claim that Goldman wasn't betting on Abacus to fail or to succeed. I suspect Goldman only lost that 90 million because they couldn't find a buyer for the 45-50 tranche in time before Abacus imploded- Goldman was making bets against the housing market at the time (they were the only bank which knew which way the wind was blowing)

It merely is saying that Goldman said ACA, an independent company with experience picking CDOs and omitted to mention that Paulson helped pick the bonds that went in the CDO. Obviously, an investor (particularly a sophisticated one) would want to know that someone helping to pick the bonds is shorting the CDO and therefore wants to pick the worst possible bonds. Paulson wasn't sued because Paulson didn't have any disclosure requirements.

I reply:

I'm pretty sure my source, to which I failed to link, and cannot now relocate, had it that way. Figures. I maintain a policy of once I put it up it's there and I have to live with the occasional embarrassment which comes from blogging in too much haste. This is, of course, one of the reasons why bloggers should always leave open the possibility of comments. If one can't always be right, one should at least be open to correction. I stand corrected, and chastened. If that is the way the source had it, I should have done a better job of double checking the precise relationship between Paulson and Goldman Sachs. I was attempting simply to translate my source. Clearly, I should have appraised and then translated.

On the other hand, it must just be that in "translating" I mis-wrote, thus rotating the roles played by Paulson and Goldman Sachs. Rather than writing Paulson paid Goldman, I wrote it the other way round.

Larry Kudlow provides a time line of the securities-selection process that was made by ACA management, the portfolio selector, from the actual SEC complaint:

ACA/PAULSON PORTFOLIO

January 9, 2007
Goldman sends email to ACA, titled "Paulson Portfolio," containing list of 123 RMBS selected by Paulson for the Abacus 2007-AC1 reference portfolio

January 22, 2007
ACA sends email to Fabrice Tourre & others at Goldman containing list of 86 RMBS, including 55 of the 123 selected by Paulson; 68 were rejected. This is very important. Goldman maintains that ACA was in fact the portfolio selector. ACA rejected 68 of Paulson’s recommendations. They accepted 55.

February 2, 2007
After meetings with Paulson & Tourre, ACA emails Paulson, Tourre & others at Goldman a list of 82 RMBS on which Paulson & ACA concurred, plus 21 others. So at this point, they are in agreement on 82, but they insert 21 others.

February 5, 2007
Paulson sends email to ACA & Tourre deleting 8 of the RMBS recommended by ACA and leaves the rest alone.

February 26, 2007
After further discussion, Paulson & ACA agree on a reference portfolio of 90 RMBS for Abacus 2007-AC1.

Clearly, as AdamJ, writes, ACA management was the portfolio selector. Confusion is rooted in the fact that some (but only some) of the RMBS were selected by Paulson, and of the 123 he selected, ACA accepted 55. The one doing the accepting, is pretty clearly the portfolio selector.

I didn't double check my source's recitation of the facts because I knew that Paulson had a role in the selection. I trusted my source's understanding of the nature of that role. I've never relied on that source before. Won't do it again, at least not uncritically.