Why Goldman Might Sell its Loans at a Discount

The Breaking Views column in today's WSJ features a column by Lauren Silva originally published last Thursday. "Can Goldman Win on Debt?" is the headline on the piece, which asks whether Goldman Sachs is "about to turn the credit crunch to its advantage".

The problem is that the mechanism Silva's talking about doesn't seem to be any more profitable for Goldman than if the bank simply kept all its unsellable debt on its balance sheet. Basically, the bank buys debt from itself at 85 cents on the dollar, recoups "a third of its original loss" if the debt rises to 90 cents, and "would eliminate its loss" if the debt goes all the way back to par. Well, yes: if you issue debt at par and it's worth par, then there's no loss.

It seems to me that the Goldman plan, if Silva's description of it is accurate, is an attempt not so much to turn the credit crunch to its advantage, so much as to limit its own downside if things get a lot worse. Goldman's upside, in Silva's plan, is essentially identical to its upside if it kept all its debt on its balance sheet. But its downside is much smaller: just 17% of face value, rather than 85%.

Goldman's share price seems to be stabilizing around a price-to-book ratio of 2, which is not bad for an investment bank in the midst of a credit crunch. My guess is that Goldman reckons that its losses on the loans it's underwritten are already baked in to its share price, and that taking an up-front charge to move those loans off balance sheet will actually go down quite well among its equity investors. The scheme does seem to be a way of moving risk from Goldman's shareholders to an as-yet-unspecified group of lenders. Which raises an obvious question: who are these lenders?