Monday, November 24, 2008

What will it take to get to the other side?

For the equities markets, the news has been steadily bad without a single positive trend.  Except gasoline and commodity prices providing a quasi stimulus.

Here is something positive.  AIG is selling its Aircraft Leasing business, ILFC for $10 billion.  Deal to close sometime after the end of the year.  
Troubled insurer, the American International Group Inc., has reached an agreement for sale of its plane-leasing arm International Lease Finance Corp. to a group of investors and the unit's management. The sale is likely to be concluded by early next year, according to its chief executive officer, Steven Udvar-Hazy.

"We're in the process of selling ILFC to a group of investors including management that will take back the company from AIG," he said. Udvar-Hazy is not only the chief executive, but also the founder of the leasing arm. He was speaking at an aviation conference in Cancun, Mexico.

"Early next year, we will consummate the closing," he said in a subsequent interview.

The ILFC is one of the biggest buyers of commercial jets from both Boeing Co. and Airbus SAS.

So, the government is getting $10 billion back. Even though ILFC is one of their better asset, this should be a huge boost to the Government/AIG bailout agreements. It indicates that AIG's businesses both have value and can stand on their own. The "buzz" is that AIG is a black hole and that it is part of one of the worst corporate give aways in history. Or something like that.

Industry experts responded positively to the news pointing out that ILFC's strong position was being dragged down by the negativity surrounding AIG.

ILFC is expected to report revenue this year of about $5 billion, with net income for the first nine months of the current fiscal at $913 million. This will be up 39 per cent from the same period last year.

ILFC is clearly quit profitable. $10 billion is a good deal for buyers, but in this environment, capital is scarce. They are also trying to sell their life insurers and their direct personal auto insurance business. These assets could net another big chunk of cash when financing is possible.

The mere fact that the government will get its first real chunk of cash back should ease some of the black hole concerns. The $10 can also be recycled to other treasury programs. This is huge, psychological boost.

What I am really waiting for is the first material write up. It has to happen and some of the CDO tranches have shorter durations. This will be another shock headline.

Another Day, Another Theory

Felix blog has another entry with even more conjecture regarding the mysterious Berkshire credit default swaps.  The assumption that can't die is this: 

 The quoted price on Berkshire credit defaults last week is meaningful.  Exactly how many swaps were sold and at what prices?  We don't know, because the market is an unregulated, over the counter market that seems to be used for speculation more then hedging.  How else can you account for the $50 trillion in notional swaps -- which is a staggering figure?  Much higher then the value of the referenced debt.  

There has been a lot of discussion lately about how this market may be involved in manipulation of stock prices.  See my previous posts for references to the WSJ article about the investigation regarding manipulation of Morgan Stanley's market price using a combination of swaps and shorting. 

The quoted cds prices are meaningless without volume data.  

Anyway, today's version is that the counter party itself is trying to hedge its exposure, after the fact, since the swaps are now more valuable.  

Let's say you're the investor who bought the equity puts from Berkshire -- and let's say that although Goldman might have brokered the deal, your contract is with Berkshire directly. As a large institutional investor, you mark your positions to market daily. The puts are part of your portfolio, and whatever value you put on them, that value has surely been rising substantially in recent months.

As the value of those puts rises, so your counterparty risk with Berkshire Hathaway rises as well.... [and] your counterparty risk is much higher than it was. After all, there's no point in paying billions of dollars for puts, if the counterparty you buy the puts from isn't going to be around to pay out on them at maturity.
Let's examine the assumptions.  A large institutional investor bought the puts.  The mark to market daily, and even though they weren't worried about collateral when they bought the puts, they are now.

First, the puts were bought for $4 billion.  A large sum of money.  But the firm(s) didn't see the need for hedging its $4 billion exposure.  However, now that the puts are worth maybe $8 billion, it is a huge issue for them.  $8 billion is bigger then $4 billion, but if you didn't ask for collateral at $4, why is is crucial at $8?

You just paid $4 billion for "insurance" from a AAA insurer.  You are now going to pay another 4% per year to insure your insurer?  And this is going to be bought from whom?  Goldman?  Citi?  A hedge fund?

$400,000 for $10,000,000 is another $460,000,000 per year for the next 5 years.  That's a lot of money for a remote risk.  And that's only buying cds's on the additional $4 billion.  

How about a single entity bought a single contract for $400,000 or $500,000 after shorting Berkshire stock?  

As far as the purchaser, they would be marking to market daily?????  There is no market -- this is a level 3 asset.  But they would remark these daily?  Maybe, but why?  Why did they buy them in the first place?  So they could worry about them?  

Which brings me to the only really interesting thing to me..... who bought them and why?  Here is a hypothesis.  Maybe a large insurer is selling variable annuities or some other product that is tied to equity indices.  And maybe the product guarantees that, at a minimum, the the payout won't be less then the initial investment.  That is a straight up hedge.

Another possibility is an entity (say a pension fund) that can't buy (more) equities for regulatory reasons but can buy securities that offer protection against loss of principal.  In other words, they should be buying bonds, but with the puts, stocks qualify.  Kind of crazy, but there are a lot of legacy regulations out there.

But here is the interesting part.  In a mark to market world, if the buyer owns the indices.  $39 billion of them, then the "delta" on that asset is 1.0.  For every dollar the index goes down -- the market value is one dollar less.  Thats mark to market, no?

If they bought American options, they would be hedged.  They could exercise the option at any time and have $0 losses.

With the European options, they wouldn't be hedged.  The delta on the European options is less then 25%.  They would still have to book $0.75 for every dollar their long position lost.  Something to think about.  

Felix Salmon's Can't Let Go

I suppose it is bad form in the blogosphere to pick on a fellow blogger, but Felix Salmon can't seem to let go of the idea that something, anything must be wrong with Berkshire's accounting for long dated index puts.  A prior post asserted that really bad things could happen if Berkshire were downgraded because of collateral calls associated with index puts.  This didn't hold up to the facts, so now he is back with a "helpful" short seller who is trying to dig up more reasons that it just can't be right.

This little romp begins with the idea that Berkshire used a volatility of "about 23" in the second quarter and a volatility of "about 22" in the third.[volatility as an input to Black Scholes]  Something must be horribly wrong.  A little primary research (as opposed to relying on a helpful friend with a short position) would show that "about 23" was, in fact, 22.8.  It would have also shown that these are 4 large world indices and that foreign currency adjustments have an impact on the results.  Easily enough to account for the difference between about 22 and about 23.  You don't need a conspiracy to explain the change.

Then some conjecture about how volatility should really be 50, and all of a sudden, you have an additional $15 billion in losses.  Non cash, of course.
The VIX is now at 80, but the VIX is much more short-dated than Berkshire's equity puts. But let's say that a reasonable volatility number for Berkshire would be somewhere around 50: that would mean the value of those equity puts going up by about $7 billion, before taking into account that the S&P has fallen by a good 35% since September 30. [50??? not that the VIX is the right number, but look at the history and 50 is MUCH higher then prior periods]

Lets just revisit the facts.  The nominal value of the contracts is $39 billion.  Berkshire has booked $6 billion through 3Q 2008.  They have an average duration of about 13.5 years.

At 6%, the present value of $39 billion in 13.5 years is about $18.  The idea that the losses could easily be $15 billion would put Berkshire's liability at $21 billion.  This is $3 billion more then the present value of 4 global indices going to ZERO in 13.5 years.  They would have to be, well, NEGATIVE, to justify booking $15 billion.

In its insurance business, Berkshire tries to limit its loss exposure to any single event or deal to $5 billion.  The maximum possible loss on this deal is about $14 billion pre tax.  But the maximum probable loss is much closer to the $5 billion figure, and the most likely outcome is no loss. [in my opinion]  

What is the correct number to use for volatility?  I would say it is the expected future volatility over the next 13.5 years.  How much weight should one give to unusually high volatility during a liquidity crisis?  I'm not sure, but not much.  But since this is an estimate of future volatility -- it should be a relatively stable number.   

This is the third negative Berkshire post for Salmon over the last 5 days.  I think he should just give it up.  However, I doubt it.  

I do find it interesting that the thing that seemed to kick started this series of posts was a quote regarding Berkshire credit default swaps.  Seems like they contain some sort of rock solid predictive information about the real financial strength of a company.  However, today, in the WSJ, there is a case history of Morgan Stanley's short siege from last month.  And one of the culprits was the use and possible manipulation of credit default swaps.  Go figure.
Investigators are attempting to unravel what produced the market mayhem in mid-September, and whether Morgan Stanley swaps or shares were traded improperly. New York Attorney General Andrew Cuomo, the U.S. Attorney's office in Manhattan and the Securities and Exchange Commission are looking into whether traders manipulated markets by intentionally disseminating false rumors in order to profit on their bets. The investigations also are examining whether traders bought swaps at high prices to spark fear about Morgan Stanley's stability in order to profit on other trading positions, and whether trading involved bogus price quotes and sham trades, people familiar with the probes say.

I generally like Felix's blog. And he does qualify his comments and has made at least some effort to present a balanced case. But the blogosphere is an echo chamber, and opinion and fact are hopelessly intertwined. I just expect a little more from

Friday, November 21, 2008

The Case for Stocks and Bonds

Is anyone ever going to want to own an old fashioned portfolio of stocks and bonds again?  I think it is back to the future -- a 1974 Ben Graham sort of future where people think of stocks and bonds as a place to invest one's savings.  

Stocks are little slices of businesses.  Bonds are little slices of loans.  People know what businesses are and understand them.  At least some of them.  They work for them, buy from them, sell to them.  People can also get their mind around a loan.  They have loans - car loans, mortgages, student loans.  It does not take a lot of abstraction to see that these familiar things can be sold in small chunks.  Stocks and Bonds.

In addition, people have an idea about what makes a good business and what makes a bad business.  Same for loans.  They have been around for a long time and there are methods, tools, and processes to evaluate them, buy them, sell them, and deal with the sorts of issues that have arisen over the last 70 years.  

There is an economic reason for their existence.  Firms need to borrow money and bonds are an efficient way for large firms to borrow.  They also need capital, and the stock market is once again, an efficient way for large firms to raise capital.

Although this financial Eden never really existed outside of a textbook, a messier version did exist and work pretty well in the United States from the establishment of the SEC in 1934 and the Securities Act of 1940.  No need to go over the familiar history of publicized failures and the far more frequent plodding successes.  

Against the backdrop of this familiar system, two strands of innovation began slowly developing.  The first being the idea of diversification and the second being that of, for lack of a better word, fungibility.  The benefits of diversification have been obvious since someone first recommended not to put all of one's eggs in a single basket.  I will skip the well known evolution of products based on this concept, and go to its logical extreme.  The Yale Model popularized by David Swensen.  The Yale Daily News reported:
Yale School of Management professor Roger Ibbotson said Yale's endowment does not closely follow the overall equity market, which means that it was better equipped to weather the economic downturn. Goetzmann said other schools have started following Yale's innovative investment strategies, and attributed much of the success to Yale's Chief Investment Officer David Swensen -- an expert in alternative investments.

"A few years ago, other schools saw the success of Yale and started to imitate Yale's strategy," he said. "The larger schools tended to do this extended diversification into smaller investments. The smaller schools didn't do it as quickly."

The Yale Model consists broadly of dividing a portfolio into five or six roughly equal parts and investing each in a different asset class. One should avoid asset classes with low expected returns like bonds and that liquidity is an expensive luxury that simply isn't needed for an endowment.  So far so good.

According to the New York Times:
His target asset allocation for fiscal 2008 is fairly similar to the allocation for 2007. The biggest allocation — 28 percent of Yale’s funds — is in real assets. Last year, those included real estate, oil and gas and timberland, Yale’s report showed.

The second-largest allocation is to what Yale calls absolute return investments, which generally means hedge funds. Yale is slightly reducing that stake to 23 percent, from 25 percent, according to Mr. Swensen.

The endowment is also planning a slight reduction in domestic equities to 11 percent, from 12 percent. Fixed-income and foreign stocks will remain at 4 percent and 15 percent respectively.
We don't know how things turned out this year, but we do know it has been rough on virtually all the asset classes.  The fund operates on a fiscal year ending in June, but by October, the benefits of foreign stocks, oil and gas, hedge funds and Real Estate were all losers.  The exceptions being hedge funds that were truly hedged and not highly leveraged.  It is also likely that Yale and the top foundations had the best managers and were starting from a string of highly profitable years.  It is likely that the copy cats did not fare as well.  

It is also possible that Yale's illiquid investments will not be fully marked down to reflect this year's carnage in the markets.  And what Yale did well, others may have done much more poorly.  Prior to the Fall, virtually any of the alternatives to the US equity markets would have significantly improved returns.   There is a saying that in a bear market, the only thing that goes up is correlation.  This was proven in October in a way that will significantly reduce the perceived benefits of diversification.

There is also a shortage of a number of alternative investments.  Pension funds manage well over an order of magnitude more assets then the Universities.  There simply isn't that much Timberland.  Hedge funds performance decreases with size at some point, if for no other reason then their fee structure.  There is simply not enough positive alpha to go around.

The second trend, what I referred to above as fungibility, is better known now as structured finance.  The fundamental idea is that cash is simply cash, and all financial assets can be chopped up as finely as one wants and aggregated in any way imaginable.  Further, they can be chopped by risk (tranches) or time so that an investor can find a product that meets any imaginable cash flow need.  The CDO^2 or CDO squared is this year's example of structured finance run amok.  

A combination of the two trends is seen in the various "portable alpha" strategies.  I think it is safe to say that most bankers would prefer to never hear the word, "CDO."

Which brings us back to individuals.  Residential real estate was the man on the street's hedge fund of this century.  Never have so many people had access to so much leverage.  The man on the street may have never studied finance, but he understood cash flow.  And when you could borrow for 4% on property that was appreciating at 10%, he wasn't interested in financial theory.  He just signed up.

Alternative investments for individuals included emerging market mutual funds, precious metals, REITs, mutual funds specializing in energy, currency funds, and ETF's to bet for or against just about any financial asset.  Not to mention access to various types of derivatives.

What will individuals want in the future?  I believe they will want the features of traditional investment grade stocks and bonds.  In this new paradigm, people will favor tradition over innovation and simplicity to complexity.  Well established, deep, liquid markets.  Relatively low, standard fees.  The transparency and tangibility of a security issued by leading public companies and governmental entities.  And things like dividend checks.  

Finally, given the decline in stock prices, once (or if) we are through this recession, the financial prospects may be quite favorable.  


Thursday, November 20, 2008

Citi-advised SIVs --- It seems like a decade ago

Remember back in the day when Citi had a little problem with off balance sheet assets. Remember theMaster-Liquidity Enhancement Conduit, or MLEC?  And the Citi SIV?  It is now mostly over, with the remaining assets ($17 billion) folded into their balance sheet.  $17 billion isn't much, but in the 15 months that this has been in the process of liquidation, the stock went from a market cap of $200 billion to $25 billion.  Their announcement states: 
In order to complete the wind-down of the Citi-advised Structured Investment Vehicles ("SIVs"), Citi announced today that, in a nearly cashless transaction, it has committed to acquire the remaining assets of the SIVs at their current fair value, estimated to be approximately $17.4 billion, net of cash.

Lessons learned? Stupidity trumped greed. It wasn't the SIV, it was everything else. And finally, they really made virtually nothing on this, prior to the blowup. Maybe that's the definition of greed -- risking everything for nothing. From Barrons
SIVS earn narrow spreads estimated at 0.30 to 0.40 percentage point. This means that Citigroup, which had $100 billion of SIV assets this summer, might have netted $300 million annually, perhaps split 50/50 between Citigroup and the outside equity holders. The equity investors were getting a relatively modest 8% annual return.

Citi's factsheet on the SIV's was relatively upbeat.  As December, the SIVs had $62 billion in liabilities, less $13 billion in cash and $2.5 billion in junior notes (which take the first hit) for a net exposure of $57 billion.  I looked through the earnings announcements, and didn't notice that there were significant losses related to these assets.  Very difficult to tell without heavy digging.  I'm sure the $2.5 billion was wiped out (SIV junior notes) and they did mention a few hundred million.  Under a billion unless it was hidden somewhere else.

The assets had a maturity of 3.5 years, so I would think the remaining ones would have a lower duration.  And there may be another haircut or two on the $17 billion.  They obviously can't sell them easily, so they may have to just hold them and see what happens.

In conclusion, they didn't make much on these things.  They may not have lost more then a billion or so on them.  They might get most of the remaining $17 billion back.  In the meanwhile, they managed to lose money on virtually every other part of their balance sheet.  The remaining assets are all in areas that look sensitive to the economy.  

If we assume that the SIV WASN'T THAT BAD, (note that I could easily have missed billions in writedowns - but lets just say it's true), then it is possible for the company to be correct -- it isn't that bad, and the critics wrong about the SIV.  But being right about the SIV (assuming it is true) doesn't mean that they (and you if you believed them) were right about the company.  It was other stuff.  Huge numbers of things.  Other structured finance securities, marks for the failed monolines, general loan losses, etc. etc. 

Follow Up:
Here is where we are now:  City tries to put itself up for sale.

The had a market cap of $200 billion.  The raised $50 billion in private capital.  They got $25 billion from TARP.  

Result, market cap, $25 billion.  

And it still isn't clear how much (if anything) they lost on the stupid SIV.  I remember that every time Citi said ANYTHING about the SIV, their stock dropped by 10%.  I have a possibly demented idea that if they had just moved it to their Balance Sheet in August, 2007, they might have kept under the radar screen, become the 3rd or 4th worst money center bank, and maybe even thrived.  Probably not, but the SIV -- the way they handled it -- killed them as a serious business.  Please let me know if I am missing something regarding SIV specific losses.


One Rumor Down.....

Brief follow up on yesterdays discussion.

Hypothesis on Berkshire:
1. Derivatives lose value ---> ratings pressure ---->collateral calls ------>ratings pressure

Since this happened at AIG, I suppose it is the first thing on everyone's mind.  

Facts regarding Berkshire:
2. Collateral Calls on derivatives -----> Not with these derivatives.

These contracts don't have collateral calls under any circumstances.*

How do you refute a rumor without raising more questions? BRK today announced that it has nominal exposure to collateral calls on derivatives.  Less then 1% of capital, worst case.

*nominal doesn't mean $0, since it seems that if you do enough of anything, there will be exceptions.  The equity puts very likely have NO collateral requirements.  A few of the CDS's seem to contain contract language requiring them -- so it isn't zero.  Let's just say it rounds to zero.

It is hard to drive a stake through a rumor without raising more questions.    

Wednesday, November 19, 2008

Felix Salmon's Musings on Berkshire Hathaway AAA

 Felix Salmon's musings on Berkshire Hathaway reifies the blogosphere regarding this subject, but basically gets it wrong.

This is the crux of Felix's musings:
A downgrade could be very, very bad for Berkshire, depending on how its collateral agreements are worded. At some point, Berkshire's counterparties are going to be able to ask it to put up a lot of collateral against the derivatives contracts it has written -- not only the CDS contracts, mind, but quite possibly also the long-dated put options it's written on broad stock-market indices. Such collateral calls could be extremely harmful to Berkshire's business model -- and that's before taking into account the loss of business at its new monoline subsidiary.

I don't know where to start.  First, a downgrade would have very little effect on its operating businesses.  Secondly, most of its derivative contracts do not require posting collateral under any circumstances.  Thirdly, Berkshire's monoline subsidiary is an opportunistic venture which produces modest profits currently and is not a major business.  It is a Statutory Insurer with a dedicated balance sheet in its own Legal Entity.  Berkshire Hathaway Assurance is regulated by the State of New York, and could, if needed, support a AAA rating independent of Berkshire's corporate rating.  Finally, Berkshire does not rely on external financing for the majority of its business.  Its modest debt is either borrowed on a non recourse basis by its utility subs, or is borrowed to support the Clayton manufactured home finance business, which is neither large nor particularly important (but has been very profitable).  

Let me go through the points:

"depending on how its collateral agreements are worded"       
From the 3Q 10K:  
" However, Berkshire is not required to post collateral with respect to most of its long-dated credit default and equity index option contract liabilities."
 You could debate what "most" means, but I believe it means some of its credit default swaps and none of its equity index options.  The idea, as Felix put it, that "Berkshire's counterparties are going to be able to ask it to put up a lot of collateral" is pure conjecture.  It is highly likely that the potential collateral, under the most highly stressed scenario, would be modest in the context of  Berkshire's finances.  

From the 10Q:
 "As of September 30, 2008, BHAC had approximately $1 billion in capital and has received the highest rating available from two credit rating agencies. BHRG is pursuing opportunities to write financial guarantee insurance on municipal bonds. In its first nine months of operations, BHAC produced $315 million of written premiums. The impact of this new business on underwriting results was nominal."
 Berkshire could easily maintain the entity as AAA regardless of the ratings of other units.  Further, the premiums are earned over a number of years and the impact on current profits is "nominal."  Therefore, this has to be considered an opportunistic rather then a strategic business.  Buffett has expressed the intention to quit writing new business if/when it becomes less profitable.

Most Property Casualty Insurance and Reinsurance is written by companies with ratings significantly lower then AAA.  For example, Munich Re has an S&P rating of AA- .  The Chubb Corporation, a highly regarded primary insurer, has Senior debt is designated A by Standard and Poor’s and A2 by Moody’s.  Chubb's insurance subsidiaries are rated highly for claims paying ability, but the notion that a AAA rating is critical for successfully writing property casualty insurance is incorrect.  The ONLY insurance business that requires a AAA rating is bond insurance, which is a small potatoes for Berkshire.

Berkshire's capital of $120 billion at September 2008 supports a balance sheet of $160 billion in liabilities.  Of course, it is apples and oranges, but banks typically have capital ratios in the neighborhood of 10% of assets.  (note that traditionally, balance sheet uncertainty in banks is on the asset side, and property casualty insurers have their uncertainty on the liability side).  Nevertheless, Berkshire is very well capitalized, does not significantly rely on borrowing (except for the non recourse loans associated with utility assets), and does not rely on holding company credit ratings for its core businesses.

And all this conjecture is based on what?  An quoted price on over the counter market for CDS's that isn't regulated and does not disclose transactions in meaningful detail.  We have seen cases where companies have been in a situation where ratings downgrades triggered collateral calls which further stressed the company and resulted in lower ratings.  Perhaps that is on everyone's mind, but Berkshire's situation is similar only if one focuses on the superficial analogies and ignores facts of substance published in the companies SEC financial statements.

However, an attack by shorts could damage the company.  The only plausible ratings action would be a credit watch or negative implications, but I am not seeing any basis for a downgrade.  Anything that significantly weakens the stock price reduces the flexibility of Berkshire to do deals with stock or raise equity capital if he so desires (unlikely).  This kind of negative attention isn't good, regardless of the facts.  Facts are always complex and subtle -- conjecture can fit a simple story that is hard to shake.  

Nov 20 (Reuters) - Berkshire Hathaway Inc:

* says has 'nominal' collateral requirements that would take effect were credit rating agencies to reconsider triple-A rating - spokeswoman
* collateral requirements would total 'far below 1 percent of assets' - spokeswoman

That settles it. Most have no collateral requirements means there are only nominal collateral requirements.

Monday, November 10, 2008

AIG CDS Solution: Financing Entity #2

What is commonly referred to as AIG bailout #2 has been ripped by bloggers.  One of the more articulate rants by Yves Smith at naked capitalism, is titled AIG, the Looting Continues.

The crux of this rather long post is that the "...treasury is going to buy crap assets at amazing (high) prices..."
A price of 50 cents on the dollar for CDOs across all tranches, particularly when the objective is to buy the dreckiest dreck (the ones where AIG's losses on its CDS guarantees would be greatest) is simply breathtaking. It's a wet dream for anyone who owns them.

Remember, this would be the price across ALL tranches. Recall that in Merrill's not-all-that-long-ago sale of its super-senior CDOs (the very best tranches) it got a nominal price of 22 cents on the dollar, but that did not accurately represent the economics of the transaction. The hedge fund Lone Star paid only 25% of that amount (or 5.5 cents) in cash, the rest was contingent on performance. So Merrill might have sold the CDOs for as little as 5.5%.
However passionate the argument, the facts simply don't support this criticism.  The problem is that AIG wrote credit default swaps to insure them at par.  They also plan on buying them at par, putting them into a SIV and commuting the credit default swaps.  

Here is what the press release says about this "vehicle."
Reduction of Exposure to Multi-Sector Credit Default Swaps: AIG and FRBNY will create a second financing entity that will purchase up to approximately $70 billion of Multi-Sector CDO exposure on which AIG has written CDS contracts. Approximately 95% of the write-downs AIG Financial Products has taken to date in its CDS portfolio were related to Multi-Sector CDOs.

In connection with this transaction, CDS contracts on purchased Multi-Sector CDOs will be terminated. AIG will provide up to $5 billion in subordinated funding and FRBNY will provide up to $30 billion in senior funding to the financing entity. As a result of this transaction, AIG’s remaining exposure to losses on the Multi-Sector CDOs underlying the terminated CDS’s will be limited to declines in market valueprior to closing and its up to $5 billion funding to the financing entity. As with the securities lending program, FRBNY and AIG will share in any recoveries in the market prices of assets.
They also have a little graphic on the slides they used at their earnings announcement.  Unfortunately I can't seem to upload pictures, so just go to page 5.

Basically they have a notional value of $70.  They are insured for $70.  AIG will buy them for $70, then put the collateral of $35 billion in the SIV, throw in another $5 billion, and borrow up to $30 billion from the FRBNY.  

The interesting thing will be those situations where the purchaser of the CDS doesn't own the underlying CDO.  They will either have to find them in the market, buy them and book the profit.  Or work out a value with AIG.

Then the CDS's will be in a SIV.  They will then sit until they amortize (not much chance of a sale)  and the Fed gets their cash back first, having $40 billion subordination.  As I recall, that was the original idea of the bailout -- buying bad assets and getting them out of the system.

The owners of the CDO's will no longer have any counter party risk.  They also will no longer have to pay premiums to AIG.  AIG won't be getting collateral calls.  The actual cash losses on these CDO's has been minimal to date.  There will no longer be any argument regarding what they are worth, since they will just amortize.

Like it or not, I don't see how this can be construed as looting.  But whatever, the idea that the price that the holders of the CDS's get isn't much of an issue.  That was settled by contract when the CDS was written (no doubt for a hundred basis points or some pathetically small premium).     

I read the conference call transcript.  Although they talked a lot about how this would be done, they never came out and said they would have to buy them at par.  I suspect the reason is that every deal is unique and they will try to drive the best bargain possible.  Particularly when the CDS holder doesn't own the underlying.

Anyway, it was clear that they would have AIG directly settle/commute with the CDS holder at "market value" and then the new SIV would buy the CDS at "market value".  Not saying that the two "market values" would be identical.  

The one thing that was clear is that the SIV will just buy assets (multi sector CDO's).  They won't assume the liability for the CDS's.  AIG already has written down the CDS's by $30 billion of the $72 billion.  

Anyone that really wants to know what these things are can look at AIG's financial supplement.  The CDO's in question still have about 20% subordination.  They consist of 137 "transactions" which I assume means CDS contracts.  The transactions would have to be on specific tranches of CDO's if they are quoting subordination percentages.

They also are publishing the results of their "roll rate model" for estimating ultimate credit losses and say that it is $12 billion compared to the "accounting" estimate based on CDO market prices.