Lloyd Blankfein’s op-ed in the FT today doesn’t break much in the way of new ground, but it’s still worth reading closely. That’s the only way you can notice where his logic falls apart:
An institution’s assets must also be valued at their fair market value – the price at which willing buyers and sellers transact – not at the (frequently irrelevant) historic value. Some argue that fair value accounting exacerbated the credit crisis. I see it differently. If institutions had been required to recognise their exposures promptly and value them appropriately, they would have been likely to curtail the worst risks. Instead, positions were not monitored, so changes in value were often ignored until losses grew to a point when solvency became an issue.
At Goldman Sachs, we calculate the fair value of our positions every day, because we would not know how to assess or manage risk if market prices were not reflected on our books. This approach provides an essential early warning system that is critical for risk managers and regulators.
The problem here is that Goldman’s mark-to-market system only works because other banks don’t follow suit. Think about what Blankfein is and isn’t saying here. He’s not saying that mark-to-market can stop credit bubbles from happening, because they can’t. In fact, if you mark your risk assets to market during a credit bubble, then the asset side of your balance sheet rises, and you look safer than if you marked to par. As a result, you’re likely to take more risk, not less.
So what happens when the bubble bursts and markets turn south? There are basically three options:
- The banks all mark to par. This is the brave-it-out approach, which actually works very well in mild recessions.
- The banks all mark to par, except Goldman Sachs, which marks to market. When there’s a major disruption this works well for Goldman Sachs, which can start selling as soon as the markets start turning, even as the rest of the market tries to brave things out.
- The banks all mark to market. This risks being disastrously procyclical, with all the banks running to the exits at the same time.
Blankfein would have us all believe that option 3 makes sense, and that somehow all the banks could sell all their assets at the same time and before they declined so far that solvency was imperiled. But who does he think they’d sell their assets to?
Blankfein also thinks it’s easy to get investment banks to hand over to regulators extremely sensitive and valuable market intelligence:
We have to build a culture whereby firms are required to share concerns about systemic risks with regulators…
Regulators could establish a multi-firm business practices committee to examine issues such as underwriting standards. If practices slip, regulators would be among the first to know. They should ask questions such as: “Where are policies being stretched and pressures building? Where are you seeing concentrations in risk, crowded trades or one-way bets?”
The insight here makes sense, and I agree with it: regulators should ask questions, and banks should volunteer information when they see systemic risks. But a large part the reason that the banks have this information in the first place is that they are trusted by their counterparties to keep secrets: that’s one of the reasons that all banks bang on so incessantly about putting clients first. What happens when you get confidential information from a client which would be of great interest to your regulator?
The fact is that there are a lot of hard questions here, and Blankfein’s blithe assurances that conceptually it’s quite easy to answer them ring false to me. Especially when he never mentions the too-big-to-fail elephant in the room. Because he knows that any action on that front would harm, rather than help, Goldman Sachs.