Stressed Again

I don't mean to beat this issue to death, but I have a few more comments on the stress test.

First, now that all the various news and editorial sources have had time to digest the release of the test, its becoming fairly apparent that the evaluation process used by the Fed and overseen by the Treasury were at least in part influenced by the lobbying efforts of the banking industry.

I don't think anyone would really benefit from me repeating myself, but needless to say, the fact that the stress test was less rigorous than was probably necessary and more forgiving towards trading desk activity--points I outlined in my previous unnecessarily lengthy post--is suggestive of some bias.

But apparently, there is still some more explicit evidence.

US banks have been given government assurances they will be allowed to raise less than the $74.6bn in equity mandated by stress tests if earnings over the next six months outstrip regulators’ forecasts, bankers said.

The agreement, which was not mentioned when the government revealed the results on Thursday, means some banks may not have to raise as much equity through share issues and asset sales as the market is expecting. It could also increase the incentive for banks to book profits in the next two quarters. (source: Financial Times)

Which, except for the whole secrecy thing, sounds like perhaps not such an atrocious idea. But some are more incredulous than I:

And in case you missed it, the phrase in the FT, "increase the incentive for banks to book profits in the next two quarters" is code for "fabricate earnings". Per below, there were quite a few instances of permissive accounting this quarter. The powers that be are inviting more of the same. And this is all in the name of boosting confidence. (source: Naked Capitalism)

This past quarter, mark-to-market regulations were suspended (or at least, eased up) for certain classes of bank assets, which some claim account for a bit (or maybe more than a bit) of the the banking sector's surprise recovery since January. So Smith certainly has a point.

And then there's this:

“The Federal Reserve significantly scaled back the size of the capital hole facing some of the nation’s biggest banks shortly before concluding its stress tests, following two weeks of intense bargaining.

In addition, according to bank and government officials, the Fed used a different measurement of bank-capital levels than analysts and investors had been expecting, resulting in much smaller capital deficits. . .

The Fed ultimately accepted some of the banks’ pleas, but rejected others. Shortly before the test results were unveiled Thursday, the capital shortfalls at some banks shrank, in some cases dramatically, according to people familiar with the matter.” (source: WSJ)

Over at Big Picture, there is a corresponding graphic that shows just how "dramatic" these reductions were. To cherry-pick the most outrageous, CitiGroup's capital requirement was reduced from $35 billion to $5.5 billion.

Lastly, focusing on that second WSJ paragraph above, apparently half-way through the process, the type of capital evaluated by the Fed was changed. Instead of looking at "tangible common equity," the Fed started measuring "Tier 1 Common Capital." This may seem a bit of a technical detail, or at the very least, prosaic, but the implication is an important one.

The point of such measures is to gauge leverage. Does a bank have sufficient capital to offset potential losses on assets? The measurement used in the stress test—”risk-weighted assets” / “Tier 1 Common Capital”—is more favorable because it understates assets and overstates capital. (source: Option ARMageddon)

Basically, Tier 1 CC is a broader type of capital than TCE. OA posted a helpful
little table to show the difference, but, like with the Big Picture graphic, let me cherry-pick the most extreme example. The Bank of NY Mellon had a TCE-Asset ratio of 2.1%, but a Tier 1-Asset ratio of 9.5%.

I also bring up this particular accounting detail to point out a mistake a made in my last post on the stress test. Towards the bottom of the post, I mentioned that

Much of the TARP money came to these banks in the form of "preferred shares"--less risky and easier to get out of for the investor than common stock. But all those preferred shares can be converted into common equity and--POOF!--, in a cloud of financial wizardry, the banks are capitalized.

That isn't true anymore. I was working with old information and only a very vague understanding of what I was talking about.

When the Fed was using TCE as a measure of bank health, this particular maneauver could have been used to get better stress-test results. Why? Because TCE, stricter than Tier 1 CC, only looks at common stock and not preferred share capital. If TCE had been the benchmark, the administrative could have swapped preferred for common and thus fiddled with the stress-test results. But not so with the Tier 1 Common Capital, since that actually takes into account preferred shares.

Anyway, my B.

And if you're lost in all of this and are interested enough to read about it, here's a good basic explanation of the various types of equity.