Readers may recall that the Federal Reserve created three vehicles to hold dodgy assets it obtained via the Bear and AIG bailouts, namely Maiden Lane (for Bear), Maiden Lane II (for AIG residential mortgage backed securities) and Maiden Lane III (for CDOs the Fed bought as part of taking out AIG credit default swap counterparties at 100% of notional value).
The Fed yesterday reported gains on its exposures in these entities due to improved market conditions. Per the Financial Times:
The US public’s hope of getting repaid for the bail-outs of Bear Stearns and AIG in the financial crisis increased on Thursday after the Federal Reserve reported a paper profit for the first time on all the holdings of securities bought from the companies.
A rise in the value of the mortgage-related securities that caused Bear’s demise and AIG’s near-collapse enabled the Fed to report unrealised gains on all three vehicles it set up to hold assets from the two stricken financial groups.
Yves here. The question is how seriously do we take this report. The authorities havea a funny way of touting mark to market gains, even in bubbly markets, as a sign that All Is Well, then deriding the same MTM values as irrationally depressed when they don’t like the outcome.
Our Tom Adams, who has done extensive valuation work on Maiden Lane III, weights in on the latest report. Not surprisingly, the central bank bank does not provide enough information on its website to allow for quantitative analysis. Tom’s bottom line is that while he finds the change in value reported this quarter to be not entirely implausible, he finds the earlier valuation to be exaggerated. In other words, the percentage gains shown may be defensible, but they were applied to a base number that looks inflated.
From Tom Adams:
The latest report on the Fed’s website for the Maiden Lane III is as of 3/31/10. Overview information:
The original Fed loan was $24.339 billion, AIG’s equity contribution was $5 billion.
As of 12/31/09, the Fed loan had been amortized down to $18.5 billion. In the first quarter, there was another $1.229 billion of amortization, bringing the amortized Fed loan balance to $17.325 billion.
Against this, the Fed claims to have CDOs with a fair market value of $23.699 billion, as of 3/31/10, up 4% from 12/31 of $22.794 billion.
96.6% of the deals in the portfolio are rated BB+ or below, and we know that many of these are in the CCC category.
The Fed’s report continues to break out CDOs by year of origination, even though we know now that companies like TCW traded much of the older collateral out of those earlier CDOs and into new, much worse 06 and 07 MBS. So this categorization by year of CDO closing is meaningless as a credit quality indicator.
By fair market value, 65.1% of the CDOs are “high grade”, 8.9% are mezzanine, 23.3% are CMBS CDOs. Interestingly, $269 million of the portfolio is RMBS – presumably, this is from deals that were liquidated. Another $354 million is cash, which may also have come from liquidations which together with the RMBS equals about 2.6% of the portfolio.
We also know now that the high grade designation is meaningless because so much of the collateral in these deals was other CDOs (an inner CDO squared).
I believe it is not improbable that during the first quarter of this year, the trading value of some CDOs appreciated, given the optimism about the markets recovery. I will expect that the recoveries would be concentrated in the higher rated portions of the collateral underlying high grade deals and CMBS CDOs.
In fact, the change in fair market value for the high grades was 0.24% – basically flat.
The mezzanine CDO change in FMV was 5.48% – this seems improbable.
The CMBS change in FMV was 17.53%. This doesn’t seem too outrageous.
In total, the portfolio’s change in FMV was $905 million, of which $823 million was from the CMBS CDOs.
In general, I thought that the CMBS CDOs were not in as bad a shape as originally treated. In fact, they had much of their posted collateral backed out via the Maiden Lane exchange. These were the late additions from Deutsche Bank. As a result, in concept, I don’t seem it as unreasonable that these have recovered a bit.
However, the problem I have is with the valuation of the high grade deals, prior to the 12/31/09. while they didn’t show much change during the quarter, they appear to be seriously inflated prior to this point, especially given their current rating status.
I would argue that this is the area where Blackrock and the Fed are taking the most liberties by continuing to maintain the illusion that they are high grade and that their year of origination matters, when we know that they are packed with worthless CDO bonds and newer vintage RMBS collateral.
So why is the Fed still holding the assets? Why not dump at cost or small profit, and be done with it? There isn’t even the need to avoid the negative appearance of actual losses.
The Fed provided bridge financing in a liquidity crisis (so they say), and it seems like the bridge is no longer necessary if private market participants think that those assets are worth something near what the Fed paid for them.
It seems like the amount is too small for a Fed sale of balance sheet assets to be a monetary concern (i.e. a monetary contraction) that would impact the anticipated QE2, so that probably isn’t it.
If they could they would; or, is that what you’re saying.
I think this report of the Fed doing well is a fiction.
I believe what Pravda tells me.
The five year plan will be finished ahead of schedule.
All the unemployed would do well to prepare for jobs in the growing disinformation sector. The need is growing exponentially until the empire crashes.
I agree with Steve, be done with it, the Fed should sell the MLIII portfolio, gain some credibility (if market confirms the mark) and move on.
That would resolves Yves’ concern about how seriously we should take the Fed’s (and by extension the rest of the market’s) valuation of the drek they’re carrying on their books.
It would also comfirm Tom’s implicit point that the benefit of the previous overestimations is diminishing.
It would also be nice to know to what extent the prices improved (or are distorted) because the FED intends to withhold these assets from the market.
The practical me says, This is as good a time as any to dump ’em and prove the costs.
Since the same argument that the gov’t is going to make money on the GM bailout is also getting lots of MSM coverage, it seems like a good time to shout ‘Bullshit, sell it and prove your case’.
The realist me says, there’s still too big a risk the valuation assumptions (used by everyone) are too fragile to be put to the test just yet.
And the tin hatted conspiracy me thinks the MS trial baloon recommending a blanket GSE portfolio forgiveness program will fly, and lift this boat as well, allowing the Fed to claim a huge ‘profit’.
Re: “he finds the earlier valuation to be exaggerated. In other words, the percentage gains shown may be defensible, but they were applied to a base number that looks inflated.”
==> See: “Since the harmonic mean of a list of numbers tends strongly toward the least elements of the list, it tends (compared to the arithmetic mean) to mitigate the impact of large outliers and aggravate the impact of small ones.”
>>> Also see: In mathematics, Kachurovskii’s theorem is a theorem relating the convexity of a function on a Banach space to the monotonicity of its Fréchet derivative.
The Lehman/AIG Random Phenomena Problem (L/ARP2) is a puzzle that still has missing data, but as with every black hole, there are methods to extrapolate the probability as to where the hidden kitty is most likely to be found in the tree:
Also see: Click on the hidden objects in these puzzles
Then see and hear: http://www.youtube.com/watch?v=vCQDkrFJiCQ
It seems clear that the USG and Fed’s MO is the ends justify the means, whatever it takes, the law be damned.
There has been a discussion of the recent Scott Garrett line of inquiry with Ben Bernanke regarding the Fed’s purchase of GSE securities. (http://www.hussman.net/wmc/wmc100726.htm)
The issue is whether the Fed engaged in fiscal policy without Congressional authorization when it purchased the GSE securities. These securities are not explicitly backed by the full faith and credit of the USG. Therefore, there is a real probability that this trillion dollar investment can lose money if the implicit backing is withdrawn, which Garrett indicates could happen in 2012. The Fed losing money on an investment that it directly holds (unlike Maiden Lane) is equivalent to directly creating money, fiscal policy, which it is not authorized to pursue.
Garrett’s questioning of Bernanke focuses on that future timepoint when the implicit USG guarantee on GSE secutiries is withdrawn. At that point, Garrett states that the Fed will have lost money on the investments, and be in violation of its charter.
“If Congress chooses to restructure that debt after 2012, the Federal Reserve will have created money without an offsetting asset of equal value on its balance sheet. It will have spent money out of thin air to pay off the holders of Fannie and Freddie securities. This would constitute a fiscal policy decision that was not actually voted on by elected representatives in Congress.”
I would argue that the USG guarantee has to be priced as an optional component, and therefore at the point of purchasing the GSE securities, the Fed has engaged in fiscal policy, as there is some probability that the USG backing will be withdrawn.
The Federal Reserve has engaged in ulawful activity.
But it’s OK, I guess.
More evidence – to the extent that your will accept it as evidence – supporting Cassano’s assertions during his testimony to the House Oversight Committee that the CDS’ haven’t breached subordination levels.
Laws only apply to the little people . Just like in the country that produces PRAVDA news . Do we get it yet ?
Well, me too think that five year plan will be finished ahead of schedule. Thanks Yves for sharing your post with us.