Today’s Financial Times article: Report to highlight alleged conflicts of interest in Goldman’s dealings (Jan 12, 2011), Goldman’s pieties insult our intelligence (Jan 13, 2011)
In the second article, Sebastian Mallaby reminds us: “After one vain attempt to explain market making at a belligerent Senate hearing, Goldman’s boss, Lloyd Blankfein, gave up.” Let’s try to do better than Blankfein was able.
The famous Abacus deal, in which John Paulson and the German bank IKB were on opposite sides of a bet on subprime mortgage securitization tranches, could hypothetically have been arranged in a number of different ways, and the role of Goldman Sachs in the deal could hypothetically have taken a variety of forms.
Most simply, GS could have served as a broker, simply bringing the two parties together, John Paulson as the buyer of a credit default swap and IKB as the seller, with GS taking a one-time fee. From then on, Paulson would make small periodic payments to IKB so long as the referenced mortgages were not in default; and if ever they did default then IKB would make a large one-time payment to Paulson equal to the difference between the face value of the referenced bonds and their liquidation value.
Alternatively, GS could have served as a dealer, selling CDS to Paulson and buying CDS from IKB. So long as the referenced mortgages were not in default, Paulson would make his periodic payments to Goldman, and Goldman would make similar payments to IKB. In the event of default, IKB would make a big payment to Goldman, and Goldman would make a similar payment to Paulson. I say “similar payment” because as a dealer Goldman would be selling and buying at different prices, and that difference is the profit incentive for serving as a dealer in the first place.
If Goldman had in fact been acting as a dealer, Blankfein could have sustained his argument that Goldman was simply making the market. “In the absence of a central exchange acting as counterparty for all trades,” he could have argued, “Goldman was serving that function.” But that is not how the deal was structured; Goldman was not a market-maker, and did not serve as central counterparty for the deal.
The way the deal was actually structured was by creation of an off-balance sheet entity called Abacus that stood between Paulson and IKB. Paulson bought CDS from Abacus, but Abacus did not buy CDS from IKB. Instead, Abacus sold risky bonds to IKB, and used the proceeds to buy Treasury bills. In the event of default, Abacus would use these Treasury bills to make its contracted payment to Paulson, so depleting its own assets and leaving its creditor, IKB, holding an empty shell.
From a standard economic point of view, the difference between these three ways of organizing the deal is all about counterparty risk. In each case, the underlying risk exposure to subprime is the same. From Paulson’s point of view, the only difference is who is on the hook for the big payment: IKB, Goldman Sachs, or Abacus.
From a money view standpoint, however, an additional dimension comes clear.
To see this, suppose that Goldman had in fact been serving as a market-making dealer between Paulson and IKB. Just like an exchange, Goldman would have been concerned about the performance of its two counterparties, and so would have required some posting of collateral to ensure performance. Probably it would also have insisted on marking positions to market, so requiring IKB to post additional collateral when default came to seem more likely. But even with these provisions, there is still the problem that IKB might one day be unwilling or unable to meet a collateral call. In that event, Goldman would of course immediately terminate the IKB contract, but it would be left with the problem of finding an alternative counterparty in a possibly disordered market. Rollover risk, or liquidity risk, remains.
From this perspective, we see Abacus in a different light. By selling bonds to IKB, Abacus was essentially buying CDS from IKB but insisting on 100% collateral to ensure performance. In this way, Abacus completely eliminated counterparty risk from IKB, and also completely eliminated liquidity risk. In effect, all potential future collateral calls were made up front, at the very inception of the contract.
There is a lesson here for those trying to implement the Dodd-Frank call to move derivative trading off the balance sheet of private dealers and onto a public exchange. Some, who resist the change, say that Dodd-Frank is asking for the impossible because the collateral requirements to ensure performance would make the market uneconomic. The Abacus deal suggests otherwise. Note first that the entire deal was already off the balance sheet of Goldman Sachs. Even more, note that the seller of CDS quite willingly posted 100% collateral.
Maybe the SEC suit against Goldman for misrepresentation to IKB was meritorious, and maybe it wasn’t. Either way, Abacus shows one way forward to meeting the call of Dodd-Frank.