Sunday, March 01, 2009

The Bank "Not Too Much" Stress Test Plan

The following is an email I thumbed on my Blackberry while on an airplane to Asia (very light editing to clean up some Blackberry typos):

Reading the NY Times cover story on Citi and govenment mandated bank stress testing from today (Monday the 22nd) and in it the author talks about how banks' balance sheets will be stress tested by the Fed and Treasury under Geithner's plan. The testing will include scenarios that assess how banks would perform under a variety of "Depression-like [literally with a capital D] conditions, with unemployment surging to 10 or 12 percent [from 7.6 at last report], for example, or home prices dropping 20 percent further [officials said]".

At first blush, this seems like a good idea to me. We should want to know the answers to those implied questions.

However, the article goes on to say, "Fed officials emphasized that these hypothetical events were 'highly unlikely' to occur." The article actually goes on to call these "nightmarish economic conditions." Does anyone [reading this blog] think those conditions aren't reasonably likely?

Let me state unequivocally that 10%+ unemployment is not "highly unlikely". While it certainly may not occur, at best it is a "reasonably likely" scenario at this point. Also, home prices down another 20% also seems reasonably likely given that's about what it takes to return back to long run affordability averages (see Clay's email from Sunday). I'm not saying better than 50% odds but probably better than 25%, so certainly not "highly unlikely." Do policy makers really believe this and if so...

Nothing is certain, but these are anything but "highly unlikely" scenarios. They may not be "highly likely" either but frankly they should be in people's middle to slightly-worse-than-hoped-for case at this point.

Further, the article states the stresses will be an "or" scenario not an "and" scenario. For example, 10% unemployment OR housing prices declining a further 20%. However, if one of these happens I'd say it's "highly likely" both will happen. These stresses need to be applied under "and" scenarios. Obviously the outcome will be much worse under "and" scenarios, but that's the only sensible way to apply the stresses. They almost certainly will happen together so we need to assess their cumulative effects.

My guess is the answer results in bad outcomes. I think it was BB&T that said a month or so ago (when measured unemployment was closer to 7.0%) that their modeling got ugly at unemployment of 8.0-8.5%. The next unemployment report will have us on the doorstep of 8% (if not over the threshold). Further, the shadow unemployment of reduced pay/shortened work weeks won't be reflected in the stats but is also impactful.

Food for thought.

The culprit in all this, I believe, is in the government's past "mandate" of high leverage. By having the Fed's support of low-reserve fractional reserve banking, it basically ensures that asset spreads (eg, lending margins) on bank owned assets would decline and leverage would increase to compensate. In order to maintain a reasonable ROE, you absolutely had to partake in the leverage orgy. This basically means that the government mandated both high leverage and low spread (risk both ways - high leverage and higher prices for assets funded with the leverage) that caused this crisis. While everyone from borrowers to lenders is to blame, the government (esp the Fed) deserve a double or triple dollop.

As an aside, everything I stated above is compounded by a foolhardy belief or semi-belief in the efficient market hypothesis - how else could someone possibly justify or dispassionately observe (as the Fed did) the massive leverage coursing through and ultimately building up in the system? You have to believe asset prices are fairly valued at all times to operate at 15x-20x levered (or more for investment banks). If a 5-6% general overestimation of asset values can lead to complete wipeout, how can you conceivably not think this is possible (much less likely) without a core belief in market value efficiency?

Don't be fooled by statements that this is a breakdown in free market behavior. Free markets are by definition imperfect and they generally structure and price to account for imperfection. Banks and insurers do not and did not operate in a free market (frankly nobody does given the government controls the printing press and thus manipulates demand signals by changing the pace and volume of printing as well as the accessibility of freshly minted dollars all the time). Banks and insurers are the most highly regulated and government manipulated private market in the U.S. outside of utilities. Free markets would NEVER allow the broad market of banks to lever like this because they'd be at constant risk of bankruptcy (much less lever more and more to fund increasingly risky assets). Rather than interbank lending supported by the Fed, banks would force settlement of assets received backed by other banks (eg, customer checks) which would immediately limit the pyramiding of leverage on bank equity.

Anyway, regardless of who deserves the aim of our damningly directed finger, we should at least demand that the "stress" test actually hypothesize a stressful AND unlikely set of scenarios. I only want to have my money on loan to an institution prepared to weather the highly unlikely not merely the somewhat unlikely.

If anyone identifies said institution, let me know. Your feedback is always welcome.


No comments: