AIG...

I still remember the morning I woke up to find that the Federal Reserve had bought an insurance company.

Now Felix Salmon has thoughts on the restructuring of the deal:

The Curious Capitalist: The government's dilemma is that the people at AIG responsible for this mess really do deserve to be tarred and feathered, or maybe drawn and quartered. There's ample reason to be punitive. But once they made the decision that AIG was too systematically important to fail--and I'm not saying that was the wrong decision--the Fed and Treasury needed to structure their bailout in a way that allowed AIG to survive, at least until the financial situation improved dramatically. They didn't do that the first time around, which is why they had to restructure the deal on Sunday...

I think this is a significant mistake: the Federal Reserve is handing a present to AIG's shareholders and bondholders that they do not deserve.

I say full nationalization, and then privatization of the insurance company parts of it. It is not clear to me that any private firm has any business offering insurance against systematic mortgage financial risk anyway.

Comments

[Ken Houghton]

Uh, your link is to Justin Fox.

I might argue you had to not have the firm go into bankruptcy (the counterparty-default uncertainty argument is strong), but "the Fed and Treasury needed to structure their bailout in a way that allowed AIG to survive" is at best wishful thinking, except under his alternate scenario ("full nationalization, and then privatization of the insurance company parts of it").

[camp]

The point was not for AIG to survive; the point was for Goldman Sachs to survive. GS invented the CDS market to borrow AIG's balance sheet. There were only 4 guys in the meeting at the NY FED when Bernanke bought AIG..Bernanke, Paulson, Geithner and Lloyd Bankfein. Its all about GS.

[steve]

The point was that if they owned over 79.9% of AIG, they would have to put it on the government books, which they didn't want to sully. Recall, this was before the big bailout. Still, with they could have forced management changes, but that would have been in violated Paulson's principle of cronyism.

[Neal]

The answer is the same as for GM. Break the company apart into its functional components. The parts with a viable future get seed capital to make the transition, a pat on the ass, and out the door. The non-viable parts go into receivership and dissolution---why should they survive? GM can function as a car manufacturer, but why should it continue to function as a financial firm? AIG can insure buildings, businesses and people against risk, but why should they be allowed to continue as a backstop for gamblers?

[crack]

The difference between GM and AIG is that there would be immediate beneficial effects if AIG were nationalized. It would once again have a AAA credit rating and could take back all the collateral AIG has had to put up for CDSs. The Gov would still payout as needed on those, but they wouldn't have to have all those billions tied up until then.

[SvN]

"It is not clear to me that any private firm has any business offering insurance against systematic mortgage financial risk anyway."

1.) In what way is having firms sell insurance against systematic mortgage risk worse than, say, selling private insurance against earthquake, flood or hurricane risk?

2.) I think the point is that AIG (and perhaps many of its customers) did not *think* they were selling insurance against systematic risk; they thought they were just insuring a bunch of mortgage pools with idiosyncratic risk.

3.) You're right that the idea of the Fed selling systemic insurance directly is attractive.

[low-tech cyclist]

Salmon: "once they made the decision that AIG was too systematically important to fail--and I'm not saying that was the wrong decision--"

For once, I'd at least like to see the *argument* that it was the right decision. Two months later, I still have no idea why Bernanke & Co. decided to sink megabillions into AIG.

[Robert Waldmann]

I agree that no private firm ought to offer systematic mortgage financial risk. I'd add that it would be even worse for a public agency to offer the kind of insurance AIG was offering. Publicly insuring mortgage backed securities and mortgage backed security backed securities etc etc is a way to guarantee that people will find early 21st century subprime mortgage contracts quaint. It would make the bailout automatic.

Naaah. I'd say, just thinking out loud, there ought to be a government agency which insures mortgages only if they meet certain standard conditions (call them prime mortgages). Hmmmm might give that agency a funny name and get back to where we were when I was born -- naaaaahh an impossible dream.

[theDAWG]

I thought GM was a healthcare company.

How much more competitive do they get if they pay the state to take that component off their books?

[abuzer]

I think some of you guys are missing how CDSs were used. They had almost nothing to do with buying/selling of the crappy tranche CDO risk. They were used to multiplier for the crappy CDOs (and their interest payments) in the most greedy , unethical manner. Here is a great article from Michael Lewis: http://www.portfolio.com/news-markets/national-news/portfolio/2008/11/11/The-End-of-Wall-Streets-Boom . For a layman like me , especially page seven is the most shocking thing that I have ever seen/read. Two girls 0ne cup doesn't even come close.

[christofay]

This is a sign of policy buyer's remorse. It was over-enthusiasm to cheer the early Bernanke policy moves in this de-leveraging recession, and you as a recognized liberal economist provided intellectual flight cover. One step after another on the slippery-slope. More of the attitude of Thucydides rather than the weird, "today Bernanke earned his pay."

I really would like to see you use your knowledge to address how to lower our risky exposure to the Himalayas of derivatives and CDS that is out there, switching metaphors the Buffetian finance WMD. This mountain of liabilities is what the American public woke up to one morning to find out the window facing our future. This is the crap that sunk AIG yet provided the oomph for the mega-bonuses to management at that company.

[christofay]

From the Nov 14, 2008, Grant's Interest Rate Observer:

Divergent rates of productivity growth are the problem. The private, or quasi-private, sector is making bad debts faster than the Treasury is guaranteeing them. It may be making bad debts faster, even, than the Fed is printing dollars. The roll call of the government's wards is fast lengthening, but the lineup of prospective wards is building faster.

[Twofish]

Low-tech cyclist: For once, I'd at least like to see the *argument* that it was the right decision. Two months later, I still have no idea why Bernanke & Co. decided to sink megabillions into AIG.

The problem is that a lot of people have exposure to AIG that don't realize that they have exposure to AIG. For example, a lot of colleges, universities, and hospitals use AIG as a plan administrator for 403(b) retirement savings. If AIG goes under, then all those plans are at risk.

The deeper problem is that people want to cancel "evil fat cat" contracts and save "widow and orphan" contracts, however there isn't a label that says "evil fat cat" or 'widow and orphan" on the contracts that AIG signs. For example, if you cancel CDS contracts, then a pension fund that had invested in General Motors bonds ten years ago and then bought bond insurance from AIG is suddenly going to find itself without money.

One thing that I find scary about this whole experience is that most people do not have a good idea what happens to their money.

[Twofish]

SvN: 1.) In what way is having firms sell insurance against systematic mortgage risk worse than, say, selling private insurance against earthquake, flood or hurricane risk?

Nothing. As long as the firm has some clue what it is doing, and AIG didn't. The problem isn't that AIG insured bond defaults, the problem is that AIG *incompetently* insured bond defaults.

If you give me $100, I can give you a policy that gives you 100% coverage for all medical bills if you are in a car accident. And that's a great deal, until you actually get into a car accident, and it turns out that I spent the $100. This is basically what AIG did.

One problem here is since I am a corrupt idiot, my insurance is going to be much, much cheaper than that guy across the street, that actually knows what he is doing. He is worrying about actually paying out money if something bad happens. Also, I can hire a lot more people than he can, because he is doing something like saving money in case he has to pay out a claim. If you work for my insurance company, then every cent that we get from clients goes to pay for commissions and salaries. Great deal if you sell insurance.

SvN: 2.) I think the point is that AIG (and perhaps many of its customers) did not *think* they were selling insurance against systematic risk; they thought they were just insuring a bunch of mortgage pools with idiosyncratic risk.

Well then. They were idiots. It's would be very satisfying to bash them, but that doesn't fix the big problem which is that now you have thousands of people in the hospital who signed up for bad insurance, and you have to figure out what to do with them.

abuzer: I think some of you guys are missing how CDSs were used.

CDS's can be used for good things and for bad things. The trouble is that you can't tell from the contract how they are being used, so you really can't say "let's cancel all of the contracts in which CDS's were used for things we don't think are good, and keep the one's that protect widows and orphans."

[low-tech cyclist]

Twofish: thanks for the quick summary. A couple of comments:

1) Not being familiar with the legal status of a plan administrator, I'd have figured that such a creature would hold and manage the plan's assets, but in trust - that it wouldn't own any of the assets, and they'd be beyond the reach of the administrator's creditors.

I gather I've been wrong on this?

2) I'd have thought that, rather than buy bond insurance from AIG, the primary way pension funds would manage the sort of risk you describe would be that of maintaining a diversified portfolio. If they did so, and bought insurance from AIG on top of that, then even the double calamity of having AIG go bust, and having the assets tank that AIG was insuring for them, should only cause a mild dent in their portfolio.

How widespread was it that pension fund managers were buying insurance from AIG *in lieu of* maintaining a diversified portfolio? And how often and by what mechanism, in normal times, do we bail out pension funds whose managers handle the funds poorly? I'd have to know a lot more before I could agree that it made sense to bail AIG rather than aiding pension funds on a case-by-case basis.

3) "One thing that I find scary about this whole experience is that most people do not have a good idea what happens to their money."

Damned if I know how people *could* have a good idea about that. Trying to unravel this world of collateralized debt obligations and credit default swaps and chopping up pools of mortgages into tranches and whatnot is the most complicated thing I've looked into since I wrapped up my Ph.D. dissertation in math. Every time I think I've got a basic picture of what this crisis was about, I read about something new that makes me have to erase half the picture and start over.

[Ken Houghton]

ltc - Short answer to (2) -- It not a purchase a lieu de; the purchase is =part of= maintaining a diversified portfolio.

Longer answer: you bail out an AIG or a BSC because it's cheaper and cleaner to kill the root than it is to deal with every offshoot. In the BSC case, it worked. In the AIG case, it didn't.

(I'm personally of the opinion that it was inevitable that it didn't--for reasons mentioned or alluded to above, among others--and that the first thing that should have happened was spinning off the Very Profitable, Well-Managed PP&C Insurance/Reinsurance Company from the Financial Insurance sub. Would give a better impression of the area that needed work, and contain the contagion. But renegotiating and making it cheaper for a failed firm to continue mismanaging itself is not a solution.)

[Perry Mehrling]

I agree that AIG had no business selling insurance against systemic risk, and neither does any other private firm. They didn't reserve against these policies, but even if they had it wouldn't have helped. According to the UBS Shareholders Report, they were selling insurance on the supersenior tranches for 15 bp, and at that price UBS saw a nice little arb that could gain them about 15 bp. To me that says the price of insurance should have been at least 30-40 bp.

With Larry Kotlikoff and Alistair Milne, I've been pushing an insurance approach to the crisis, but also an insurance approach to fundamental financial reform: http://cedar.barnard.columbia.edu/faculty/mehrling/creditinsureroflastresortfinal09Oct2008.pdf

[low-tech cyclist]

Ken - with all respect, I don't see that. Insurance itself doesn't make your portfolio any more diverse; it just further reduces your risk. But in the event that the insurer fails, you've still got a diverse portfolio that will only crash if the whole market crashes. And nobody can insure against that.

So where insurance from AIG was an improvement to a sufficiently diverse portfolio, I still don't see where the need for rescue comes in. Suppose those GM bonds that Twofish uses as an example are only $2M of my $100M pension fund. Then if GM goes bust, I'm out $2M, but one can take a 2% hit anytime for any reason, and nobody needs rescuing from it.

I may be missing something here, but I still don't see what it is.

[Twofish]

1) If the plan contains an annuity or fund issued by AIG, then having a trust won't protect that from the creditors.

2) Diversification won't protect you from a systemic credit downturn. The problem is that if one AAA company goes bust, then the credit risk on all of the bonds will go up and so you end up with the bonds being worth less. CDS for most companies is extremely cheap, so that you can buy protection against all of your bonds without too much cost.

low-tech cyclist) Suppose those GM bonds that Twofish uses as an example are only $2M of my $100M pension fund. Then if GM goes bust, I'm out $2M, but one can take a 2% hit anytime for any reason, and nobody needs rescuing from it.

If GM goes bust, then Chrysler, Ford, GE, and every other AAA bond becomes more risky, and you could lose 10% off of a long dated bonds from increased credit risk on all bonds. A 10% loss doesn't sound like much, but remember you are dealing with bonds and not stocks. If you were wanting to deal with 10% losses, you'd be buying stock, and while your bonds are going down 10%, your stocks are going down 40% at the same time.

Part of the reason that you don't want bond losses is that you know that stocks are going to move up and down, so you if are in a down market and you need cash, you limit your losses by selling your bonds. If both stocks and bonds are going down, then you are hosed.

[Twofish]

Also a 2% loss isn't a killer if you are a pension fund, it could be if you are a bank. If you are a bank that is required to keep 8% of its book in securities, then this is very bad. If you lose money, you are required to sell loans to raise cash, at the very moment everyone else is doing the same thing.

[Twofish]

(I'm personally of the opinion that it was inevitable that it didn't--for reasons mentioned or alluded to above, among others--and that the first thing that should have happened was spinning off the Very Profitable, Well-Managed PP&C Insurance/Reinsurance Company from the Financial Insurance sub)

Maybe but that does run into a problem in that as long as you keep the very profitable well-managed company, around you have the possibility that you won't lose too much money. If you spin off the good stuff, you immediately and obviously leave the government with nothing except crap. It's not clear that this is a good idea.

[Twofish]

Quote: Trying to unravel this world of collateralized debt obligations and credit default swaps and chopping up pools of mortgages into tranches and whatnot is the most complicated thing I've looked into since I wrapped up my Ph.D. dissertation in math. Every time I think I've got a basic picture of what this crisis was about, I read about something new that makes me have to erase half the picture and start over.

It might help to step back a bit, and look at the big picture. When times are good people like to borrow lots of money because borrowed money turns a 2% gain into a 20% or 200% gain. People keep borrowing money until times turn bad, and then all those huge gains become huge losses and things start falling apart.

It also helps to read economic history. The details of the boom/bust are different each time, but the psychology is the same.