A bank has to prove to its regulators that it will withstand a crisis, so nothing better than having products that are designed, by definition, to protect against crisis. The problem is that the assets that the banks are piling up in their balances are designated to assuage regulators, not to stabilize the situation during crises. It is management to appease the central bank, not to optimize the firm’s management.
The Federal Reserve’ answer to that and other strategies has been to increase the ‘qualitative’ side of its ‘stress tests’. Last week, the Fed vetoed Citigroup’s plans to raise its dividend and share buyback after deciding it had not passed the ‘stress test’. Spain’s Santander, and British HSBC and RBS are also in the same situation, since they did not pass the test either. For the European banks, this must have been a shock. EU’s stress tests tend to be a ‘free lunch’, and it is still fresh in the memory the case of Franco-Belgian behemoth Dexia, a bank that passed with flying colors the 2011 European ‘stress test’ just to collapse a few months later.
Setting aside those specific cases, the Federal Reserve’s ‘stress test’, even if they are the most reliable among the big central banks, raise important questions. The most critical is the simplest: how can the solvency of a bank be measured in a hypothetical crisis? In Citigroup’s case, for instance, the regulator clearly states that its criticism “in part reflects significantly heightened supervisory expectations for the largest and most complex.” In other words: the Federal Reserve is very serious about the ‘too big to fail’ risk.
The problem with that approach, obviously, is its discretionality. Citigroup’s executives and its investors are enraged, and complain of “lack of transparency” on the central bank’s side. They are right. But, after the 2007-2012 crisis, the regulators have learned that not everything can be reduced to numbers. Now, let’s hope that this ‘qualitative’ judgment is sounder than the ‘quantitative’ one that brought us into this mess.