The Epicurean Dealmaker has a very smart gloss on the CELF transaction I wrote about on Friday:
Investment banks traditionally thought of market making as a client service. An agency business. We put our capital at risk to facilitate the trading of our investing clients. In exchange, we earned a small commission, the occasional chance to put our capital to work in longer-term trades where we thought we had an edge, and—most importantly—priceless insight into the daily operations of particular securities markets, including the appetites, biases, and weaknesses of countless third party market participants. This insight is incredibly valuable, not only in market making itself, but also in making the investment bank possessing it a better informed underwriter for new securities. Securities markets are hotbeds of asymmetric information. The party with the best information has the greatest power. Market making can provide that power.
All of this is well and good, says TED until the market makers’ balance sheets balloon into the trillions of dollars. At that point, market marking becomes indistinguishable from large-scale prop trading.
And in the CELF trade, for all Goldman’s protestations that it was making a market in otherwise-illiquid structured securities, in fact it was simply bidding on those securities, against other bidders:
Goldman’s actions in 2010 bear absolutely no resemblance to behaving like an agent when it purchased the outstanding CELF securities and liquidated them. It did not behave like a normal market maker, buying securities from one investor and selling them to another. It paid an arm’s length price, determined after an auction run by a third party, to the investor it originally sold the AAA rated tranche to. It then triggered the liquidation of the securitization by purchasing a majority stake in its equity. With respect to the seller of the AAA tranche, it acted as a pure trading counterparty. A principal.
I’ve since learned a few more interesting facts about this transaction, none of which make Goldman look particularly client-centric.
For one thing, the investor selling the AAA tranche was exactly the same as the investor from whom Goldman bought enough equity that it could gain control of the CLO and liquidate it. That’s really weird. It makes sense that a Dutch pension fund might want risk-free assets — pension funds are like that. But why on earth would they want a piece of the incredibly risky and volatile equity tranche as well?
Goldman’s trade, here, was to buy the AAA tranche at a discount, and then to also buy up the equity tranche — which, when added to the residual equity which Goldman already held, pushed its total shareholding over 50% and gave it control of the deal. It then found a couple of outside investors (hedge funds, I assume), to come along for the ride and to promise that they would vote as a bloc, and sold them a package of bonds and shares on the understanding that if the shareholders voted to liquidate, the bonds were sure to be paid off in full.
This is a really complicated way of adding value which could have gone much more easily, with much less use of Goldman’s balance sheet. Goldman’s financial advisers, with the best interests of their client at heart, could have looked at the pension fund’s assets and noticed that it had almost enough equity to liquidate the CLO and be paid off in full on its bonds. And looking into their own bottom drawer, they could have noticed that Goldman’s own shareholding would be enough to push the pension fund over the 50% mark.
The most obvious thing to do would be to simply tell their client that Goldman would vote with them to liquidate the deal, and thereby help them get 100 cents on the dollar for their AAA tranche. But since that might involve Goldman taking a loss on its equity tranche, the next most obvious thing to do would be for Goldman to sell its shares to the pension fund at a small premium, and simultaneously help the fund liquidate the CLO, making a substantial mark-to-market profit on the deal.
Both those actions would have been client-centric things to do, and neither would have constituted proprietary trading. But Goldman instead decided to wheel out more than a billion dollars of its own money to buy up the pension fund’s holdings at a discount and take the profit for itself. It was so much money, indeed, that Goldman felt the need to bring in outside partners to share the risk — risk it never even needed to take in the first place.
A clearer example of what TED is talking about could hardly be found. Goldman had asymmetric information: it knew that there was a good enough bid for the loans which made up the CLO that if the structure were unwound, all the debtholders would be paid off in full, and even the equity holders would make decent money. (I’m hearing that the equity holders walked away with 80 cents on the dollar.) Goldman then used its monster balance sheet to profit from that information itself, rather than simply sharing the information with its client and letting the client do the trade instead.
Goldman is adamant that what it did on this trade was client-focused market-making. “This is a good example of helping a client achieve its objective,” the bank told Jesse Eisinger, “and underscores the critical importance banks play in using their capital to facilitate transactions on behalf of clients.”
TED, on the other hand, sees things slightly differently, calling that statement “a patently disingenuous dodge” and “pure, unadulterated horse****”.
I’m tempted to side with TED on this one. Yes, the client’s objective was to get out of its position. But Goldman could have helped it do so more profitably than it did, and what’s more could have done so without using its capital at all. The only reason it didn’t was that it made more money this way round.