Andrew Ross Sorkin today notes that the Fed and the SEC didn’t do anything about Lehman Brothers, despite the fact that they knew full well that there were problems.
Where was the government while all this “materially misleading” accounting was going on? In the vernacular of teenage instant messaging, let’s just say they had a vantage point as good as POS (parent over shoulder)…
Indeed, it now appears that the federal government itself either didn’t appreciate the significance of what it saw (we’ve seen that movie before with regulators waving off tips about Bernard L. Madoff). Or perhaps they did appreciate the significance and blessed the now-suspect accounting anyway.
Liquidity was an important factor in the stress testing that Lehman was required to run under the CSE Program. After March 2008 when the SEC and FRBNY began on‐ site daily monitoring of Lehman, the SEC deferred to the FRBNY to devise more rigorous stress‐testing scenarios to test Lehman’s ability to withstand a run or potential run on the bank. The FRBNY developed two new stress scenarios: “Bear Stearns” and “Bear Stearns Light.” Lehman failed both tests. The FRBNY then developed a new set of assumptions for an additional round of stress tests, which Lehman also failed. However, Lehman ran stress tests of its own, modeled on similar assumptions, and passed. It does not appear that any agency required any action of Lehman in response to the results of the stress testing.
The CSE program, which stands for Consolidated Supervised Entities, was the SEC’s way of trying to supervise too-big-to-fail banks from the inside. But even the SEC didn’t think it was qualified to actually do that, which is why the SEC brought in Geithner’s New York Fed as a partner — note the bit above about the SEC “deferring” to the Fed to put together stress tests.
Clearly the CSE program was an abject failure: it could put together stress tests, but then the SEC and the NY Fed ignored the results. There’s obvious bad news here: that the Fed is such an incompetent regulator of systemically-important institutions that it can’t even get alarmed when one of those institutions fails its own stress tests. But there’s a possible glimmer of good news too: that the Fed had people capable of putting together a decent stress test, and that the SEC sensibly deferred to those people in terms of stress test design. In other words, the Fed has the ability to regulate; all that’s needed now (and was missing in 2008) is the willingness to do so and to bare teeth once in a while.
A good way to institutionalize that is to implement what David Merkel calls “dumb regulation” — once you put simple rules in place, it becomes much more difficult (although never, of course, impossible) to override those rules or to ignore them. The problem with Lehman was that there were no simple rules, and that no one at the Fed or the SEC felt comfortable making up new ones on the spot, like “you’ve got to be able to pass the stress test which we invented five minutes ago”. I, for one, wouldn’t want to be the regulator who had to receive the phone call from Dick Fuld after implementing a rule like that, using dubious legal authority.
One of the problems with giving lots of supervisory authority to the Fed is that the Fed is run by economists who care primarily about setting monetary policy, as opposed to being run by bankers who care primarily about bank regulation and systemic risk. The base-case scenario is that unless and until we start staffing the Fed with a bunch of poachers-turned-gamekeepers, the biggest banks are likely to be able to smooth-talk their way past the Fed’s regulators. The Fed is still the least bad institution to do this: Any other alternative would be even worse. But that doesn’t mean that I have any confidence in it.