There's an uproar about whether the government should let AIG fail, a debate re-energized by the latest revelation of bonus payments going to AIG's executives. In fact, there's a good case to be made that AIG should fail, and it has nothing to do with bonuses.
The rescue of AIG is warping the banking system and unnecessarily extending the credit crisis. This misguided effort stems from a lack of transparency and some basic misconceptions about AIG's business. (Read "Obama's Challenge: Containing the AIG Bonus Outrage.")
Let's take a moment to review how AIG made money and why it's now losing so much. Most of AIG's losses have been attributed to its failing positions in credit-default swaps (CDSs). Essentially, AIG is swapping cash flows with other institutions: those banks pay AIG a small sum on a regular basis, and then under certain conditions like mass foreclosure or corporations' defaulting on their loans AIG pays out a large sum. In other words, AIG sold insurance; its problem is that it is paying out too much.
Of course, it's a bit more complicated than that because of the funky nature of AIG's insurance. Other types of insurance do not carry the same risks because when one claim pays out, it does not snowball. AIG's insurance on foreclosures and other defaults is not like insurance for accidents and disasters; while earthquakes in California are not correlated with earthquakes in New York, foreclosures are spiraling out of control together, fueled by a widespread recession.
Moreover, investors and banks that hold credit-default swaps do not necessarily own the tranches of mortgages and bonds that the CDSs insure. AIG may have even written multiple swaps over the same mortgages and bonds. It would be as if an insurance company had sold earthquake insurance on one house to multiple investors. When the house falls, so does AIG.
But there's a true insight into this mess if you just step back and consider the bigger picture, not just AIG. Regardless of the details of the various swap contracts, they all represent potential transfers of wealth between financial institutions. If we consolidated the entire financial sector, all these debts would effectively vanish.
Think again of the insured house. Many institutions hold insurance on the house; on the other side are insurance companies and the like making an opposite bet. If the house is destroyed, one group of institutions wins and the other group loses. Considering all institutions together, no money was truly lost it's what economists call a zero-sum game. In good times, risk-hungry banks loved this game, but now they have become risk-averse, and the game seems to have changed. So how can many of the banks simultaneously claim enormous swap losses without a single bank claiming significant profit?
Here are two possibilities: either the vast majority of all swaps not just AIG's are held by investment banks, or a significant portion is held by other financial institutions like hedge funds. Suppose all swaps are held by banks. Since swaps are a zero-sum game, the banking industry as a whole cannot lose money on swaps. Then there is no need for a bailout. (See 25 people to blame for the financial crisis.)
Alternatively, if hedge funds hold significant positions, then it is possible for the banking industry as a whole to net a loss on swaps. That loss would be the hedge funds' gain. This means the bailout is ultimately saving the hedge funds.
Whichever it is, if the number of institutions involved in swap-trading were limited to those trading with AIG, then AIG is probably not too big to fail. We have to worry about chains of claims. Just because AIG dealt only with banks does not mean those banks did not rewrite similar contracts with hedge funds.
The most direct solution for the swap problem is to settle all such agreements and eliminate their uncertainty from the equation. If the payments are reversed, or the payments are stopped, or they are settled once and for all, the uncertainty will vanish.
The problem with keeping the swaps on the banks' books is that their potential payoff or loss is random, depending on the particular details of the contract and various outcomes in the world. Moreover, banks today are risk-averse and often factor worst-case scenarios into current pricing. Thus, they are far more likely to claim losses than profits on such instruments.
At the very least, there should be full transparency. Any institution receiving money from the government and ultimately from American taxpayers should reveal its holdings. Even institutions that do not require a bailout should be more closely tracked by regulators. The government can and should monitor all transactions, even those over-the-counter.
We have focused too much on each individual bank and its possibility for failure. The economy does not need every bank to survive; it needs most. Right now, we need to know which ones. By propping up financial institutions that are subject to unknown potential losses, the government is prolonging the uncertainty about whether they will fail. This perpetuates the crisis of confidence in which banks do not trust one another enough to loan money.