Moral hazard and AIG

We are now suffering the consequences of one of the most spectacular financial miscalculations in history, after investors around the world discovered that trillions of dollars invested in securities derived from U.S. home mortgages were far riskier than they had originally believed.

Part of this miscalculation can be attributed to misguided quantitative models that were used to assess those risks. The key inputs for those models were assumptions about underlying default rates and their correlations across different borrowers. Default rates and correlations were quite low up until 2005, because rising home prices made default a decidedly inferior option to refinancing for even the least credit-worthy borrower. But the rising home prices were themselves caused by the huge flow of capital for lending to this market, sucked in by the illusion of safety. When the flow of credit stopped and house prices began to fall, the same forces operated impressively in reverse, now leading otherwise credit-worthy borrowers to default in increasing numbers.

But I would argue that another factor contributing to the illusion of safety was severe distortions in the markets for derivative contracts based on those underlying mortgage-backed securities. The seller of a credit default swap promises to make a payment to the buyer in the event that the security against which the contract is written goes into default. From the perspective of the buyer, a CDS is like an insurance policy against default. Some institutions might be interested in buying such contracts even if they did not have long positions in the securities against which the CDS was written, as a hedge against risks of other related investments.

And who would want to be on the sell side of these contracts? Insurance giant AIG was one big player. At first blush, you might think this could be a reasonable role for such an institution, since from the point of view of the buyer a CDS could function much like an insurance policy. But from the point of view of the seller, this is a very different product from conventional insurance. Selling more fire insurance policies helps the insurer to diversify, because fires across different communities have little correlation. But there is a common risk factor at the core of recent housing market developments. By selling a bigger volume of CDS, AIG was simply taking a bigger lopsided position on a single one-sided bet. And AIG lacked the financial resources to make good on those contracts in the event that the housing downturn became as severe as it has now proved to be.

But that raises a separate question. Could it make any sense for AIG to sell and someone else to buy a promise on which AIG in fact could not deliver? From the point of view of AIG– at least the specific players within AIG running these operations– one could argue that the answer is yes. In selling the CDS, they were receiving huge payments. As Forbes reported last September:

A big part of the reason was most likely that AIG’s financial products unit, run by Cassano since 1988, was a veritable money machine, pouring $6 billion of riches into AIG’s coffers from 1988 until 2005…. According to The Times, compensation ranged from $423 million to $616 million for Cassano’s group. That would be about 20% of the unit’s revenue, meaning Cassano was being paid like a hedge fund manager.

This understanding of AIG’s incentives may have been what prompted Federal Reserve Chair Ben Bernanke to sound off last week:

if there’s a single episode in this entire 18 months that has made me more angry, I can’t think of one, than AIG…. AIG exploited a huge gap in the regulatory system…. There was no oversight of the Financial Products division. This was a hedge fund, basically, that was attached to a large and stable insurance company, made huge numbers of irresponsible bets– took huge losses. There was no regulatory oversight because there was a gap in the system.

So let’s grant that Cassano and his cohorts may have had ample incentive to sell the product. But who would buy? Though he may not have been thinking of AIG and its counterparties in particular, Bernanke aptly described one potential answer in his remarks today on financial reform to address systemic risk:

In a crisis, the authorities have strong incentives to prevent the failure of a large, highly interconnected financial firm, because of the risks such a failure would pose to the financial system and the broader economy. However, the belief of market participants that a particular firm is considered too big to fail has many undesirable effects. For instance, it reduces market discipline and encourages excessive risk-taking by the firm.

Could AIG’s counterparties have been thinking that their payments would come not from AIG but from the Federal Reserve and taxpayers? If that’s what they thought, some of them at least would appear to have been correct. Gretchen Morgenson reported this weekend:

When A.I.G. couldn’t meet the wave of obligations it owed on the swaps last fall as Wall Street went into a tailspin, the Federal Reserve stepped in with an $85 billion loan to keep the hobbled insurer from going bankrupt; over all, the government has pledged a total of $160 billion to A.I.G. to help it meet its obligations and restructure operations….

Edward M. Liddy, the chief executive of A.I.G., explained to investors last week that “the vast majority” of taxpayer funds “have passed through A.I.G. to other financial institutions” as the company unwound deals with its customers….

On Wall Street, those customers are known as “counterparties,” and Mr. Liddy wouldn’t provide details on who the counterparties were or how much they received. But a person briefed on the deals said A.I.G.’s former customers include Goldman Sachs, Merrill Lynch and two large French banks, Societe Generale and Calyon….

How much money has gone to counterparties since the company’s collapse? The person briefed on the deals put the figure at around $50 billion.

To the extent that the buyer and seller of the CDS were making a bet for which the taxpayers were implicitly picking up the downside, the CDS market, rather than helping institutions effectively manage risk, would have been a factor directly aggravating systemic risk.

I raise this issue not to be yet another voice clucking that we need to get tougher on AIG or others in order to prevent moral hazard, though that is one reasonable inference to draw from the discussion above. But the issue for me has always been not to exact retribution or instill market discipline, but instead the very pragmatic question of how to use available resources to minimize collateral damage. I accept the argument that a complete failure of AIG would have unacceptable consequences. The relevant question then is, what combination of parties is going to absorb the loss?

The concern I wish to raise is that any reasonable answer to that question would include Goldman Sachs, Merrill Lynch, Societe Generale, and Calyon, to pick a few names at random, as major contributors to this particular collateral-damage-minimization relief fund. But if they are to contribute, the plan must be something other than doling out another $100 billion every few months to try to keep the operation going a little longer, but instead requires seizing this bull by the horns. Split AIG into a core business we want to protect– with enough equity to be a viable operation, and a hefty fraction of the existing management team fired– and a derivatives business that’s going to be systematically liquidated in large part by abrogation of outstanding contracts.

Then there’s the domino effect to consider. What do we do when this brings down the next player who can’t continue operations without those payments AIG (or the taxpayers) were supposedly going to deliver? I say, we implement the parallel operation there.

That’s my proposal for how to dismantle the derivatives house of cards. One trillion at a time.

Originally published at Econbrowser and reproduced here with the author’s permission.

One Response to "Moral hazard and AIG"

  1. BlackyBlack   March 11, 2009 at 1:24 pm

    Can the banks show that the derivatives market is essential for the financial system to function? If not, this looks to me like the tail wagging the dog. We’re supporting a sidebetting (gambling) scheme that is not essential to our banking system. In other words, the bailout is largely a wealth redistribution scheme, and, as you say, will not by itself do anything to mitigate future risks. The risks are merely being rolled over to the future through the further use of credit derivatives. So we are paying now and we will again pay later, not to save the banking system from failing, but to save a gambling scheme from being closed up and shut down. The banks are relying on our cumulative ignorance about credit derivatives to spin their requests for continuing the same game. So many people have said the same thing I am saying and you’re sayiing, but the people in charge are not listening. I think they are stalling while the game continues. They will not abrogate those contracts. It has made them too much money to give up.