Don’t ‘Stress Test’ The Giant Banks—— Break Them Up!

It’s that time of year again. Not for spring cleaning – but for bank stress tests.

Too bad stress tests are nothing more than a sweeping-dirt-under-the-rug exercise, despite the rug-beater supposedly hanging them on the line.

The whole show, which first started in April 2009, is just a public performance.

After all, is any bank ever going to fail and be expelled from the club?


The stress tests, in fact, are nothing more than Dances with Wolves…

Domino Theory

Back in April 2009, I wrote, “The key problem with the whole stress-test exercise is that it does nothing to improve financial-system transparency.”

bank stress testHow much of a farce are the tests? We don’t really know because they’re not transparent. The public has no way of knowing what’s being looked at through what prisms.

What we do know is that the tests are a joke.

Last year Bank of America Corp. (NYSE:BAC) breezed through Round 1 (there are two rounds), where quantitative number crunching determines how much capital and reserves banks really have. After that, Bank of America submitted its “we have so much excess capital we might want to raise our dividend or buyback shares plans” to Federal Reserve carpet-bangers as part of Round 2.

After the quantitative dance is over, Round 2 is a qualitative dance known as Comprehensive Capital Analysis and Review, or CCAR. That’s where the banks are looked at through some other prisms to determine if their capital plans make sense, and if rewarding shareholders is okay, or if they need to keep more of the capital they say they have to be safer institutions.

The joke is that last year, after Bank of America passed the quantitative round and the CCAR round where it got approval to reward shareholders for its good capital management, it had to notify the Fed that… oops, it miscalculated its capital by a few billion dollars.

Here’s how The Wall Street Journal reported the kerfuffle:

The second-largest U.S. bank by assets said it discovered a mistake in certain figures submitted to the Federal Reserve for the regulator’s annual ‘stress tests’ of major U.S. financial institutions. The error leaves the Charlotte, N.C., bank with $4 billion less in capital than it thought it had. The bank had been making the same calculation error since 2009, according to a person close to the bank. Bank of America brought the mistake to the Fed’s attention. The regulator revoked its prior approval of a dividend increase and gave the bank until May 27 to submit a new plan.

That’s scary. The Federal Reserve, which conducts the tests, supposedly, had no idea that the numbers Bank of America submitted since 2009 were calculated incorrectly. Does that mean the Fed doesn’t do its own calculations? Is this an open-book test? Are there teachers’ pets?

How’s that for confidence in the stress tests?

It gets worse…

According to a working paper from the U.S. Department of the Treasury’s Office of Financial Research (OFR), published March 3, all the top stress-tested banks are remarkably alike in how they come out of the tests. According to the paper’s authors – Columbia University researchers Paul Glasserman, who used to be with the OFR, and Gowtham Tangirala – that’s because all those banks go in remarkably alike:

The results are striking. The patterns appear to be an artifact of the bank stress testingprocess rather than an accurate reflection of potential bank losses… The main concern with a routinized stress test is the danger that it will lead banks to optimize their choices for a particular supervisory hurdle and implicitly create new, harder-to-detect risks in doing so. This concern applies to any fixed supervisory scheme, including one based on risk-weighted assets.

For one, the banks all use the same consultants to guide them through the testing processes. These consultants are often former federal regulators – from the very agencies that conduct the tests – who have joined the lucrative consulting game.

What’s not funny about the stress tests is that they don’t recognize that the correlation (specifically and empirically measured in the OFR working paper) between the banks is staggering. Stress testing each bank individually says nothing about testing them collectively.

The banks are connected dominos.

The only thing the stress tests tell us is if capital requirements, reserve requirements, and other buffers aren’t enough to protect the financial system from its own greed.

Stress testing the banks isn’t the answer.

The answer is to break up the monster banks before they break the economy again.