Tuesday, February 17, 2009

Geithner's Plan

Geithner unveiled a vague plan to address the crisis, and it's a what they call in Washington, a crap sandwich.


"First, we're going to require banking institutions to go through a carefully designed comprehensive stress test," he announced. "This borrows the medical term."

Sounds reasonable at 30,000 feet, but what kind of stress test do you imagine this group will generate? Anything with enough granularity to be really fair, taking into account the risk of obligors and facilities as banks evaluate them, is beyond the ken of any regulatory body. That they are going to rush it through, just means it will be 100 times worse than what they might have come up with.

Risk management in a large, complex financial institution implies a lot of detail. There is no easy way for banks to demonstrate their viability with a top down, easily verifiable number, for the main reason that such information is difficult to produce. If it was easy produce, one of them (certainly there is one) would have generated such information to ease their liquidity strain.

The only people who know less about portfolio risk than the senior executives at a bank, are the regulators. They will generate numbers that must ignore nuance because they can only apply cookie cutter solutions. They will feel obligated to fail some banks, merely because otherwise this will look toothless. Thus, you have an arbitrary axe hanging over the top 100 financial institutions, and we will only know who it will strike when we see how the stress tests are biased--towards community banks, consumer banks, investment banks, etc. This has weighed heavily on the markets, and I don't see it subsiding until we hear something comforting from the Treasury, like 'Heh. Just kidding! We aren't going to apply stress tests.'

Asking for specifics, a Treasury official made this ponderous statement:

"this stress test is . . . designed to be more forward-looking and to make sure we're taking into account a quite broad range of economic scenarios."

"What could be worse than the current situation?" the reporter asked.

The official smiled. "I'm not going to answer that," he said.

He should have said, "I can't answer that", because that's the truth. He has no clue. The bottom line is, no regulator with any power in Washington understands the specifics needed to assess credit risk. They merely measure asset performance, looking at late payments and the like. Their assessment of credit risk is merely various gradations of overdue accounts (30 days past due, 90 days past due, OAEM). All current assets are not the same, and if they apply the same stress test to all such assets, it is merely a leverage test, which is a very, very crude way to assess solvency. The good stuff, loans that are not past due, they have no experience grouping and classifying. Will they apply there stress test to consumer loans? How will they adjust by vintage? FICO score? Collateral type? Auto vs. home equity loan? Do the stress test multiples based on historic volatility, if so, is it using market value stress or cash flows? Does it apply 2, or 3 times these losses? These are essential question, and to assume they will figure it out on the fly, is like a surgeon just opening up the patient and hoping something obvious asks to be removed.

Further, the idea of making sure banks have capital for an unforeseen event, after the unforeseen event happens, is like asking banks have capital for 2 unforeseen events. It is the regulatory equivalent of mark-to-market, amplifying a bad event.

Too bad so many feel that if Washington is the solution, rather than the problem right now. I would merely demand that banks present all their credit information by product type: home loans, auto loans, indirect and direct lending, leases/lines/outstandings, and then, via their internal grading, the losses experienced by banks in these various buckets. Those firms with greater granularity, better data, will be able to show they are of lower risk. We need more information, and because banks are fearful of revealing secrets, they are making everyone too scared. I personally understand that you should never reveal information as a general rule, because there are enough people out there with bad faith that it is not a good general idea. But this is a crisis, and the public needs it. Then let the market punish the losers, as they will have to be taken over by the winners. The Fed or Treasury should merely help banks that have liquidity problems, by just trying to value their current assets and liabilities, so that you only do this to solvent banks (where solvency is based on current values, not value cum stress test).

3 comments:

AHWest said...

Good idea. More information is better. And lying about the numbers should be a death sentence for a bank. Problem is, I don't think the banks have a single valuation for a lot of their assets and liabilities - they have scenarios that would result in widely different valuations.

Eric Falkenstein said...
This comment has been removed by the author.
Anonymous said...

"I personally understand that you should never reveal information as a general rule..."

Did some lawyers teach you that recently?