AIG Bailout: A Goldman Rescue in Drag?

 |  Includes: AIG, GS, KBE
by: Craig Pirrong

Krugman is bashing Timmy! Geithner for his role in the AIG bailout. This poses something of a dilemma for yours truly. My sentiments parallel those of Henry Kissinger during the Iran-Iraq War: too bad they both can’t lose.

All snark aside, the SIGTARP report that has put Timmy! (oh, that was snarky–sorry) on the hotseat raises some questions that have been totally ignored over the debate over whether the Fed should have sent Goldman et al to the barber shop to get a haircut on the valuations of their swaps with AIG.

The indisputable fact is that billions of dollars in cash went out the door to Goldman et al as a result of the Fed’s actions. The Fed took ownership of the CDOs underlying the swaps that AIG had entered with the banks, and effectively paid the banks 100 cents on the dollar. That is, they ensured that the CDO hedges were perfect (belying the old trader adage that the only perfect hedge is in a Japanese garden).

The question is, therefore, what were the alternatives? The alternative that has garnered all the attention is that the Fed should have paid less than 100 cents on the dollar.

But that’s not the only alternative. Hank Greenburg has suggested that the Fed should have simply guaranteed the swaps, thereby vitiating the need to provide any collateral payments. (I made a similar suggestion in an earlier post on AIG).

The SIGTARP report states clearly (p. 14) that this alternative was considered, but dismissed. The ostensible reasons for the rejection seem very dubious, indeed.

First, “FRBNY told SIGTARP that a perceived downside of this structure from FRBNY’s perspective was that it could involve FRBNY in long-term credit relationships with supervised institutions.” Please. The Fed has gone hog wild in extending credit (through repos, for instance) with supervised institutions. It has taken all kinds of dodgy collateral at all kinds of dodgy valuations. That certainly involves taking a long term credit exposure. (Spare me any protests that there is no credit risk here because these repos are collateralized. Given the quality of the collateral, and the counterparties, there is an appreciable probability that the Fed will suffer a credit loss on these deals.) And if the Fed’s actions were a response to an existential event, which is the gravamen of its defense of its actions, such prissiness over protocol appears decidedly inappropriate–making this explanation exceedingly implausible.

Further thought (added at around 1900 CT): Given that the CDS were so far underwater to AIG, if the government had guaranteed them, the likelihood that the Fed would have become a creditor to the banks on the other sides of the deals was exceedingly remote. That is, it was highly unlikely that the Fed would have been exposed to default losses on these deals, meaning that the “credit relationship” was a fiction. (Besides, at the time, were most of the counterparties even under Fed supervision? Most were foreign banks, and even the US counterparties, with the exception of Wachovia, were investment banks that I do not believe were under direct Fed supervision, except perhaps as Treasury primary dealers, rather than as banks.)

Second, “there was a lack of statutory authority of the Federal Reserve to provide such a guarantee.” Please, again. There are a variety of structures that effectively create guarantees. For instance, if the Fed could see its way clear to setting up and capitalizing a special purpose vehicle (SPV) to buy the CDOs, it could have set up and capitalized an SPV, and then novated the deals to the SPV. If it was concerns about counterparty risk that made the banks so insistent on receiving collateral payments, this structure would have allayed their concerns–and required no cash to go out the door.

In this structure, the government’s risk exposure would have been the same as under Maiden Lane and its purchase of the underlying CDOs: it would have been long the CDOs.

In sum, the rationales given for not providing some sort of guarantee are completely unpersuasive. Completely. A guarantee would also not have required agreement on valuation with the counterparties. They would have been assured of receiving their contractual payments, and that should have been that.

The transparently implausible rationale for eschewing the guarantee alternative tells me that the Fed’s–and Geithner’s–injured and adamant denial that “the financial condition in the counterparties was not a relevant factor” (p. 15) in deciding to pay 100 cents on the dollar is dishonest. Geithner has said many other things that do not pass the honesty smell test, so it wouldn’t surprise me that if this was the case here as well.

Thus, it is highly likely in my view that this was a backdoor way of providing liquidity to systemically important institutions at a time that their financial condition was in serious question.

In this regard, it could well be that Goldman (NYSE:GS) was the firm that was in greatest need of an injection of cash. The SIGTARP report states that, unlike the other AIG counterparties, “Goldman Sachs did not hold the underlying CDOs but rather had sold equivalent credit protection to its clients who held those positions.” Very interesting. It is likely that these client counterparties were demanding collateral from Goldman. If so, if Goldman didn’t receive cash from AIG–or the government–it would have needed to find additional cash to make these payments.

Yes, Goldman states that it was hedged by its purchase of credit protection on AIG. But, (a) in prevailing conditions, there was considerable credit risk in those CDS, meaning that Goldman may not have been paid out 100 percent of what it was owed, and (b) even if the CDS paid out, there would almost certainly have been a cash flow date mismatch, with Goldman needing the cash to make margin calls to its clients immediately, and receiving any cash payments on CDS at some later date. Given the state of the credit markets at the time, funding this gap would have been an expensive, and dicey, proposition.

Given the supposed First Commandment to Treat All Banks Equal (p. 29), the Fed could not have bought out Goldman and not the other banks.

Against that, if providing liquidity to Goldman alone was the objective, there should have been ways of doing that directly–unless the Fed was concerned that special treatment of Goldman would have commenced a destablizing run on it like the one that cratered Lehman.

I therefore can’t conclude for certain that the AIG bailout was really a Goldman rescue in drag. One can tell that story, but there are alternative explanations. However, given that the Fed’s explanation for not taking actions that would have required no cash payments is so weak, my conclusions are that the AIG bailout was an indirect way of providing liquidity to systemically important institutions, and that one cannot exclude the possibility that this was an indirect way of providing liquidity to one institution in particular–Goldman.

(One question unanswered by the SIGTARP report: if Goldman didn’t own the CDOs that eventually wound up in Maiden Lane, how did they get there? Did Goldman buy them from its clients in a mirror image deal that involved swap tearups, and then sell them to Maiden Lane? It would seem that would be necessary to deal with Goldman’s own sales of protection. )

A couple of other points related to the SIGTARP report. First, as I emphasized in “It’s a Wonderful Life: AIG Edition,” if AIG hadn’t been born and hence not around to sell protection, the owners of the CDOs would have taken a bath. Thus, it is not credit default swaps per se that were the ultimate source of the problem; it was the underlying CDOs. Only to the extent that the existence of AIG contributed to a larger CDO market could CDS have contributed to the financial crisis. Indeed, the crisis–that is, the losses suffered by big banks–could have been worse if AIG hadn’t taken a $50 billion hit.

Second, one of the narratives has been that AIG didn’t have to post collateral, and hence took on bigger positions than it would have if it had been required to do so. It indeed didn’t post any initial margin, but it is clear that the deals contemplated the posting of collateral even absent an AIG credit event. AIG had posted at least $22 billion in collateral prior to its downgrade. Perhaps the necessity of posting initial margin would have reduced AIG’s appetite, but likely not, in my view. First, by not requiring original margin, counterparties were extending AIG credit, and presumably charged for it; only to the extent that it would have been costlier to finance initial margin payments would the posting of such margin have made AIG reduce its positions. Second, given that it lost huge sums on mortgage backed in its security lending program and other operations, it is clear that AIG viewed these as very attractively priced risks. Sure, a slightly higher cost (due to the necessity of posting initial margin) might have induced it to cut back some, but likely not very much.

To conclude: given the availability of another alternative to buying out the banks at 100 percent of par, that would not have required a cash payment, and the weak justifications for avoiding that option, make it highly likely that the AIG bailout was structured in part to provide liquidity to major banks (and perhaps, but not conclusively, one particular bank). Which makes the Fed’s–and Geithner’s–denial that the financial health of these firms was an irrelevance highly dubious, not to say, a lie.