Page 300 - untitled
P. 300

LATE  TO EARLY : BILLIONS IN SUBPRIME LOSSES                   


         ‘guesstimating’ pricing.”   Cassano agreed. “No one seems to know how to discern a
         market valuation price from the current opaque market environment,” Cassano
         wrote to a colleague. “This information is limited due to the lack of participants [will-
         ing] to even give indications on these obligations.” 
            One week later, Cassano called Sherwood in Goldman’s London office and de-
         manded reimbursement of . billion. He told both AIG and Goldman executives
         that independent third-party pricing for  of the , securities underlying the
         CDOs on which AIG FP had written CDS and AIG’s own valuation for the other 
         indicated that Goldman’s demand was unsupported—therefore Goldman should re-
         turn the money.   Goldman refused, and instead demanded more. 
            By late November, there was relative agreement within AIG and with its auditor
         that the Moody’s model incorporated into AIG’s valuation system was inadequate for
         valuing the super-senior book.   But there was no consensus on how that book
         should be valued. Inputting generic CDO collateral data into the Moody’s model
         would result in a . billion valuation loss; using Goldman’s marks would result in a
          billion valuation loss, which would wipe out the quarter’s profits.   On November
         , PwC auditors met with senior executives from AIG and the Financial Products
         subsidiary to discuss the whole situation. According to PwC meeting notes, AIG re-
         ported that disagreements with Goldman continued, and AIG did not have data to
         dispute Goldman’s marks. Forster recalled that Sullivan said that he was going to have
         a heart attack when he learned that using Goldman’s marks would eliminate the
         quarter’s profits.   Sullivan told FCIC staff that he did not remember this part of the
         meeting. 
            AIG adjusted the number, and in doing so it chose not to rely on dealer quotes.
         James Bridgewater, the Financial Products executive vice president in charge of mod-
         els, came up with a solution. Convinced that there was a calculable difference be-
         tween the value of the underlying bonds and the value of the swap protection AIG
         had written on those bonds, Bridgewater suggested using a “negative basis adjust-
         ment,” which would reduce the unrealized loss estimate from . billion (Goldman’s
         figure) to about . billion. With their auditor’s knowledge, Cassano and others
         agreed that the negative basis adjustment was the way to go.
            Several documents given to the FCIC by PwC, AIG, and Cassano reflect discus-
         sions during and after the November  meeting. During a second meeting at which
         only the auditor and parent company executives were present (Financial Products ex-
         ecutives, including Cassano and Forster, did not attend), PwC expressed significant
         concerns about risk management, specifically related to the valuation of the credit
         default swap portfolio, as well as to the company’s procedures in posting collateral.
         AIG Financial Products had paid out  billion without active involvement from the
         parent company’s Enterprise Risk Management group. Another issue was “the way in
         which AIGFP [had] been ‘managing’ the SS [super senior] valuation process—saying
         PwC will not get any more information until after the investor day presentation.”  
            The auditors laid out their concerns about conflicting strategies pursued by AIG
         subsidiaries. Notably, the securities-lending subsidiary had been purchasing mort-
         gage-backed securities, using cash raised by lending securities that AIG held on
   295   296   297   298   299   300   301   302   303   304   305