Reviewing these slides, there are two big messages:
A) The default trend slides (5, 6, 8, 9, 12, etc.) show an unfolding train wreck; and B) Reserves are still way too low.
Just look at slide 2: Ambac is reserving $636 million on $5 billion of exposure to CES, $432 million on $11.4 billion of HELOCs, $200 million on $6.5 billion of mid-prime RMBSs and $16 million on $8.1 billion of subprime RMBSs. I think the HELOC number is the most absurd: 3.8% losses?!
Keep in mind that these are exposures to RMBSs themselves, not CDOs and CDO-squareds comprised of tranches of RMBSs. In our analysis (which was similar to the Open Source Model Pershing Sq. released -- see the last page of our presentation on the mortgage crisis), we had assumed minimal losses on the RMBSs -- and then Ambac confesssed last week to $1.3 billion of expected losses (note that they're not claiming these are mark-to-market losses that will be reversed).
To see how wrong Ambac's models were, check out pages 19 and 20, in which Ambac insured a second-lien RMBS deal with Bear Stearns in 4/07 and projected 10-12% losses, and now they're projecting collateral loss of 81.8%!!!
Pages 23-28 of the attached cover Ambac's $32 billion of exposure to CDOs. On page 28, they've finally admitted that they're going to get clobbered on the CDO-squareds (they've written them down 70% -- in reality, it will likely be 100%), but they've taken only 1.2% impairment on $26.5 billion of exposure on regular CDOs. This is a complete joke -- according to Ambac's internal ratings, $7 billion is already rated below investment grade (page 24)!
The Goldman analyst is starting to get in the right ballpark, estimating $11.8 billion in losses. How this leads to a need to only raise an additional $3.4 billion to maintain Ambac's (increasingly ludicrous) AAA rating is beyond me, but even this amount ain't gonna happen for a company with a $1.1 billion market cap today. He also continues to value MBIA and Ambac as ongoing entities, saying "We do not believe that a run-off valuation scenario is appropriate for either Ambac or MBIA, as both companies have recently raised capital to avert a ratings downgrade from triple-A (from Moody’s and S&P), and both companies continue to write new business (albeit at depressed levels relative to history). Still, we present our “run-off” valuation model below for reference (see Exhibit 5)."
But the clincher is this line: "We lower our estimates for Ambac to -$18.05 in 2008 (from +$0.95) and -$6.40 in 2009 (from +$0.90), but increase our 2010 estimate to +$2.10 (from $1.00). We lower our estimates for MBIA to -$21.75 in 2008 (from -$1.81) and -$8.50 in 2009 (from -$1.33), but increase our 2010 estimate to +$3.25 (from -$1.14)."
Wow, it's trading at 1x 2010 earnings -- what a buy! Ya just can't make this stuff up!
Disclosure: Author's fund is short ABK and MBIA