Dr. Julian Zelizer of Princeton just wrote an OpEd on CNN.com calling for congressional hearings into the current financial meltdown. I subsequently sent him this email:

______ 

www.bgladd.com

______

Dr. Black, author of the book depicted above, wrote this (below), in some comments to a congressional aide:

From: Black, William

Sent: Thu 2/19/2009 5:12 PM

Subject: Why you need to sample loan files to evaluate fraud/credit risk of nonprime loans and paper (and CDS counterparty risk) 

You asked me to explain why it was critical to sample loan files to evaluate fraud/credit risk of nonprime loans and paper and its value.

The old joke in banking about loans is that it's far more important to secure the return of capital than a return on capital. An example helps. If you make a $1 million loan at 10% interest rate that quickly defaults and never repays you don't simply lose the interest you expected to earn (roughly $100,000), but also the amount of the principal that is never repaid (hypothetically, and realistically for a fraudulent mortgage, 50%, $50,000). You can see that one bad default swamps the income from a whole lot of good loans. Historically, prime mortgage loans had default rates around 1% and loss rates well below that (because the loans are collateralized by the borrower's home). Mortgage lending is highly competitive, so the profit margins on prime loans are low. Prime lenders, therefore, have to keep defaults and losses extremely low if they are to earn a profit. Default risk is the primary risk that mortgage lenders face, and mortgage fraud poses the most virulent risk of default and exceptionally large losses...

______ 

 OK, let me run with that. Some simple arithmetic readily reveals that, if you loan out a million dollars at 10% and the loan goes completely bad, you are out $1.1 million and will have to make 11 more such loans that perform perfectly just get to get back even -- a zero % ROI from a total of $12 million placed at risk. A total of 20 such loans, then, -- 19 performing perfectly after your initial loss -- gets you only to a 4% aggregate ROI. If your going-in "ROI Hurdle" was 10%, you can now never get there -- absent raising your subsequent interest rates (as CitiGroup recently, insultingly did to one of my credit cards, notwithstanding my 800-range FICO and small debt balances).

Now, while some loans indeed have eventually (<100%)  recoverable tangible asset collateral (though I would perhaps take issue with Bill Black's expectation example of "50%" during a time of mass foreclosures), you can put all of these data in a spreadsheet and vary the loss frequency ("incident loss rate") and net amount scenarios  (the latter "dollar loss rate" comprising the far more salient concern) and quickly see the highly adverse effect of even a relatively small number of defaults. Particularly given that so much of lending in recent times has itself been done using recursively borrowed money (multiple layers of "leverage"). Leverage cuts sharply both ways, as we are now seeing to our dismay. Again, see "Tranche Warfare" (in particular, scroll down to my take on "The OPIUM Economy").

 We really at this point have no idea as to the extent of utterly insolvent "zombie banks" and kindred institutions out there. And, I  have to wonder about the extent of federal political appetite for fully and publicly investigating it.

TIM GEITHNER'S "STRESS TESTING" 

I worked for a number of years in subprime credit card risk management. We did "stress-tested" ROI analyses for everything. Our CEO had set a "stress-tested 10% ROI hurdle" policy for every marketing and internal ops initiative (basically to assess the "opportunity cost" of the money). We had to use his Excel Sheet-From-Hell "profit model" (he'd formerly been our CFO), which looked out monthly for five years. If you couldn't project -- worst-case -- surmounting the "hurdle" within a specified period (typically a year or less) your project simply wouldn't get funded, period (which led to all manner of bullshit data gaming -- mostly by Marketing --, but that's another story).

Beyond that, even for approved projects and products, execs would be sweatin' basis-point bullets when monthly "actuals" came in that diverged ever so slightly from the "projecteds," such was the anxious awareness of the ever-present random-walk adverse volatility potential (I always found the 5-year look-ahead thing amusingly lipstick-on-a-pig stupid anyway).

 Geithner's post hoc "stress-testing," done openly and forthrightly, might reveal a level of actual insolvency that the government would prefer we not know about.

UPDATE: Also, Dr. Black makes a wonderfully salient point regarding the Quixotic nature of "stress testing" absent analytical access to original loan data:

_____ 

 "...a professional credit rater has told his superiors that he needs to examine the mortgage loan files to evaluate the risk of a complex financial derivative whose risk and market value depend on the credit quality of the nonprime mortgages "underlying" the derivative. A senior manager sends a blistering reply with this forceful punctuation:

UPDATE:

 "The Dukes of Moral Hazard" is now underway. Probably gonna be another long one.

UPDATE:

February 24, 2009 (LPAC)--A senior regulator for the Resolution Trust Corp. during the 1980s savings-and-loan crisis demanded a "new Pecora investigation" and prompt bankruptcy receivership for U.S. banks, in a column on Huffingtonpost.com Feb. 23.