Sunday, October 5, 2008

Notes on valuation of AIG

This post is to collect various notes on valuing AIG.

This story attempts to estimate a break-up value for AIG.

"CreditSights estimates that the company's general insurance business is worth $95.2 billion, its life insurance and retirement services business is worth $147.7 billion, its financial services business, including its aircraft leasing unit, is worth $9 billion and its asset management operations are worth $3.6 billion, said Haines. That valuation, which assumes limited stress in the financing markets, comes to a total of $255.5 billion."

That works out to $19/share after the 80% dilution. That seems much higher than any other estimate. I am also not sure of they are including the mark-to-market losses. Probably not. If we take out another say $30B for CDS losses, that would still be $16/share. The book value as of last 10-Q was $5.80/share including the dilution.

Recently (after the bailout) Bill Ackman, the hedge fund manager of the Pershing Square, bought AIG. I think he paid around $3.50. He
estimates
book value to be around $6. This estimate takes into account the mark-to-market losses and the 80% government dilution. He also thinks there is a chance that some of the dilution can be reversed if AIG pays the loan back very soon. I agree.

The real valuation of AIG would have to include the following:

1) Figure out what they are going to sell and what they will get for it and how long it will take them to pay off the loan. Also consider the interest on this toxic loan.

2) Figure out what the real losses will be on the CDSs. Also consider the fact that the mark-to-market losses could get worse. They could also reverse if the TARP bailout moves these MBS priced higher. This will have some effect on liquidity and might effect how much of the company they have to sell.

3) Then figure out the earnings and earnings growth of whatever is left over and model it in the usual way with a DCF.

4) Estimate the actual dilution that the government will demand. Will they take the full 80% or something less.

5) Think of anything else that could go wrong.

################# UPDATE ###############

They are reworking the deal. This info comes from the Wall Street Journal.

Lets go over the details. They are scrapping the original deal. Now the are giving AIG a 5-year loan for $60B at LIBOR plus 3% which comes to about 6% given that LIBOR is about 3%. That is much better than the original 14% loan. They are also giving them $40B in preferred stock with 10% dividend yield. Not sure if it is cumulative. In addition they are injecting $50B in common capital. The government still will keep a 79.9% equity stake. I assume that means common equity but I am not sure.

First of all this is pretty good so far. Instead of paying 14% on $85B or $11.9B pretax per year, they will pay $3.6B pretax in loan interest and $4B aftertax for preferred dividend. At 40% tax rates that saves them about $1B per year after tax. Not a big difference but it helps. The real advantage is that this increases total capital by $90B which will greatly improve the capitalization of AIG and should raise their credit rating. Also it increases common capital by $50B. So instead of giving away 79.9% of the company in exchange for just a loan, common shareholders get $50B common capital in return. This seems a lot more fair to me. At least it is better than before. Now they will have to sell roughly $40B worth of assets over the next 5 years to pay back the loan. That is only half of the $85B in asset sales that was required by the first deal. Plus, they get 3 extra years in which to get it done.

Then the story gets a bit more complicated. The government is setting up one "vehicle" to handle AIG's CDSs on CDOs and another to handle the securities lending business. The first one will be capitalized with $30B from the government and $5B from AIG. That would take on the $70B CDS porfolio on the CDOs. As of last credit presentation, they has $57.8B net exposure to multi-sector CDOs including subprime and $22.5B in multi-sector CDOs without subprime for a total of $80.3B. This may have shrank down the $70B quoted in the journal. It is not clear if this capital is in addition to the other capital injected into AIG. I think almost certainly not. So this would be setting aside $30B out of the total $90B in new capital and $5B of current AIG capital. In addition it says that AIG (or presumably this new vehicle) will attempt to buy the insured CDOs from their counterparties for 50 cents on the dollar so that it can tear up the CDS contracts. The notional exposure (before AIG's subordination) was $112.6B as of last presentation. So 50 cents on the dollar would be $56.3B. If it buys the CDOs back, it gets the collateral back that it had to post.

I can't say I understand this bit. There are 112 different transactions. What makes them think that the counter parties want to take a 50% hit on these if they can make AIG pay anything over the subordination level. AIG's average subordination on the CDSs including subprime is 23.9%. The subordination on CDSs without subprime is 16.3%. So if we just take the ones with subprime, those counterparties can only take a 23.9% hit before AIG covers them. So why sell it back to AIG for a 50% hit? Is it because they don't think AIG could actually pay them off? Perhaps it has more to do with asset reduction. This is confusing.

The other vehicle would be set up to handle the problems in the securities lending facility. AIG lent out securities to short sellers and took in collateral. It used this collateral to buy mortgage backed securities collateralized by subprime loans. Pretty dumb move. When the short sellers closed their shorts (presumably because of the short selling ban and/or deleveraging), they wanted their money back and AIG couldn't sell the RMBSs. So they borrowed money from the Fed to pay back the shorts. The new vehicle would take $20B from the government and $1B from AIG and buy the illiquid RMBSs from AIG at 50 cents on the dollar. So does that mean that AIG will realize a $21B loss on those or is some of that money simply for liquidity?

So the first part of this is good news. The second part concerning these two vehicles is confusing. It sounds like this will improve AIG's liquidity, solve the securities lending problem and protect them from future CDS losses. But what is the cost to AIG? It isn't clear to me what this unwind will cost them. They have already written down about $30B to cover these losses. Will this result in losses above that?

Lets do a rough calculation. Lets say that this unwind results in a $50B total loss. That is $20B on top of what was already written down. Last 10-Q showed $78B in equity. So after the unwind that would be $20B pretax and maybe $12B after tax so that would make equity $66B before the equity infusion. They added $90B but appear to putting $35B on this into the first vehicle. So that leaves $55B of equity left in AIG. Total equity is now $121B including the $40B preferred stock leaving $81B in common equity. Original shareholders own only 20% of that which is $16.2B. There are 2.7 billion original shares so book value per share would be $6/share.

If this is correct (a big if), the stock should be worth at least book value and so should go roughly to $6/share but probably not right away. If the break up value by CreditSights is correct, then it should be worth even more.

A worst case scenario might look like this. Assume the $70B in CDSs are completely worthless and are written off entirely. This is $40B in additional write-downs pretax and $24B after tax. Even this leaves book value per share at $5.11/share. Seem like the stock is a good investment at less than $3/share. I expect it will trade above $3 tomorrow morning.

There are other stories out now saying that if the securities purchased at 50 cents on the dollar result in profits, the government will share them with AIG but keep 2/3 of them. This adds a little more value to AIG.