Lucian Bebchuk says AIG is not too big to fail:
AIG Still Isn't Too Big to Fail, by Lucian Bebchuk, Commentary, WSJ: The AIG bailout -- at $170 billion and rising -- may end up as the costliest rescue of a single firm in history. There is much debate about bonuses paid to AIG's executives. But there is far too little debate on the government's willingness to back all of AIG's obligations.
The company claims any failure by the government to do so would have catastrophic consequences. This claim is exaggerated. Serious consideration should be given to forcing AIG's partners in derivative transactions -- which are mainly buyers of credit default swaps from the company -- to take a substantial haircut. ...
While AIG has thus far been able to cover derivative losses using government funds, the possibility of large additional losses must be recognized. AIG recently stated that it still has about $1.6 trillion in "notional derivatives exposure." Suppose, for example, that AIG ends up with losses equal to, say, 20% of this exposure -- that is, $320 billion. Suppose also that the value of AIG's current assets ... is $160 billion. In this scenario, the government's fully backing AIG's obligations would produce an additional loss of $160 billion for taxpayers. Should the government be prepared to do so?
The alternative would be to put AIG into Chapter 11. In this case, AIG's creditors, including its derivative counterparties, would obtain the company's assets. They would end up with a 50% recovery on their claims, bearing those $160 billion of losses themselves.
AIG recently stated that failure to meet all of the company's obligations could lead to a "run on the bank" by customers seeking to surrender insurance policies and "would have sweeping impacts across the economy." But insurance policyholders wouldn't be at risk if AIG failed to meet its obligations. The insurance subsidiaries are not responsible for the debts of their parent AIG, and insurance policy claims are backed both by the subsidiaries' required reserves and state insurance funds.
Still, what about the concern that losses to derivative counterparties -- which are now known to include major U.S. and foreign banks -- would substantially deplete the capital of some of them? That concern would be best addressed by the U.S. government (or foreign governments in the case of their banks) infusing capital directly -- in return for shares -- into the banks that need it. There is no reason to back AIG's obligations as an instrument for infusing capital (with taxpayers getting nothing in return) into, say, Goldman Sachs or Spain's Banco Santander.
It is true that the collapse of Lehman Brothers last September led to a crisis of confidence among depositors..., which had a dramatic effect on markets. Letting AIG's derivative counterparties take a significant haircut, however, should not lead to such a crisis. AIG's obligations are to derivative counterparties, not to depositors. Moreover, governments world-wide are now committed to backing fully the claims of depositors in financial institutions.
It is important to understand that the government can also employ intermediate approaches between fully backing AIG's derivative obligations and no backing. ... At a minimum, the government should conduct "stress tests," estimating potential losses in alternative scenarios, and formulate a policy on the magnitude and fraction of derivative losses it would be willing to cover. A policy that doesn't fully back AIG's obligations should be seriously considered.
Monetary policy makers often run potential policies through a wide variety of models under differing assumptions about the state of the economy, and they avoid policies that have a very bad downside even if the upside is potentially very attractive. The proposal above can be interpreted within this type of risk management approach to policy that tries to insure against really bad policy outcomes. It may very well be that financial markets can withstand the failure of AIG, but the potential downside if that assumption is wrong is very large, larger than any policymaker wants to take the blame for. Thus, in such cases, you often observe policymakers pursuing what they feel is the safest policy that moves the ball forward rather than policies that might have a better upside, but also come with the possibility of, say, market meltdown. I'm sympathetic to the argument that we should put AIG through a carefully managed failure, but saying, as above, that letting "AIG's derivative counterparties take a significant haircut ... should not lead to ... a crisis" still leaves the door open, even if only a crack, to an outcome that no policymaker wants to be responsible for.
Update: James Kwak:
Now, the government has not explicitly guaranteed AIG’s liabilities. But the main reason for bailing out AIG in the first place was the fear that an uncontrolled failure would have ripple effects that would take down many other financial institutions who were dependent in some way on AIG; most commonly, they had bought insurance, in the form of credit default swaps... And a major usage of bailout money has been to make whole AIG’s counterparties...
I still think it was a mistake to let Lehman fail, because of the sudden panic it created. But we are in a very different situation today. Many people now believe that the government may decide to let bank creditors lose some of their money. As Bebchuk says, instead of continually giving AIG taxpayer money that is effectively used to bail out other banks (many of which are in Europe, allowing European governments to free ride on the U.S.), the government could let AIG fail and bail out those other banks directly, thereby at least getting increased ownership stakes in return. Bebchuck also explains that AIG’s insurance subsidiaries would not become insolvent if the AIG holding company went bankrupt, because they have their own reserves. ... Furthermore, he argues, failure is not an all-or-nothing proposition...
I think that the government could let AIG fail, if - and this is a big if - it can first identify which creditors and counterparties would be hurt, determine which of those cannot be allowed to fail (which should not be all of them), design a program to provide them enough capital directly, and announce everything on the same day. The net cost to the taxpayer cannot be higher than under the Too Big To Fail strategy, which implies a 100% guarantee for all counterparties and creditors...
There was clearly no time to do this between September 15 and September 16. But the government by now has had six months to study the books of AIG and its major domestic counterparties. People are no longer willing to take it on faith that the future of the free world depends on an implicit blanket guarantee for AIG. At least we want to see some evidence.
How sure are we that if we let AIG fail all of the necessary pieces - all the big and little ifs above - will fall into place, how sure are we of any evidence based upon asset valuations when there's no market price for them, and how sure are we that we've thought of all of the things that could go wrong?