% fortune -ae paul murphy

Fun with business inteligence

The financial industries, particularly those concerned with securities trading, employ a lot of genuine rocket scientists - physics grads who aren't intimidated by complexity and consider the transformational models used in finance about as unnatural as breathing.

Unfortunately most of their work is not understood by the executives, traders and market makers driving the industry - most of whom, at least in my experience, combine high IQs with short attention spans, a gift for buzztalk, and a frenetic commitment to money-as-tribal-token that leaves them both emotionally unsupported and intellectually incapable outside their work environments.

The resulting disconnect between methods and the decisions being made about the application of those methods is largely responsible for the current sub-prime mess and represents a classic failure in business inteligence.

The term "sub-prime" refers to the credit status of the borrowers to whom the mortages were issued. not the mortage terms. In fact most sub-prime mortgages were structured as adjustable rate contracts with low initial rates followed by a rapid rise to the above market rates needed to justify the loan in light of the "sub-prime" credit risk posed by borrower incomes and track records.

What happened when the higher rates started to kick in was that a lot of the risk materialized as defaults - and that reduced the expected value of the loan portfolios involved below issue value and thus triggered a classic ERP style whip effect reducing the book value of all of the related assets across the industry.

In the past the industry has had both legal and financial firewalls in place to protect itself from effects like this - the big players on Wall Street, for example, were hardly threatened by the REIT mess in the 90s. What's different about it this time is that the executives involved at every level from issuer to reserve bank completely failed to understand, or at least to act on any understanding they might have had, that the risk spreading processes developed by the rocket scientists effectively bypassed many of the financial system's internal firedoors to leave everybody exposed when the predicted percentage of the high risk borrowers ultimately responsible for the sustaining cash flow defaulted.

The second act is currently underway -and it's another consequence of the decision makers not understanding what their own automated tools do. In this case, credit card holders achieving home owner status automatically had their credit limits adjusted upwards - and a lot of the sub-prime lending candidates have now justified their status as sub-prime by using those credit cards to further overspend their incomes.

From an IT perspective the sub-prime ranking is itself largely an automated process and thus represents a giant "I told you so" to the bosses - but that's so obvious it still isn't widely understood.

Equally importantly, some of those most directly involved in developing and valuing securitization methods also warned that transfering much of the risk away from the issuer and then spreading it across the industry really did create a moral hazard for issuers while breaking with long established risk containment policies - but they weren't understood either.

Yesterday I talked about two incidents about twenty years apart in which I was unable to get senior people to understand numerical analysis going beyond simple crosstabs - and pointed out that almost all supposed business inteligence processing today really consists of more or less cool ways of using PCs to calculate and present very simple tabulations. What the sub-prime mess shows, I think, is the effect of that same communications failure on a much larger stage.

So what's the bottom line? - besides proving that us IT grunts is smarter than them bosses? I think the bottom line is that the gap between our tools, skills, and access to data and their comprehension creates a moral hazard of its own - because at some point we have to accept that our obligations to the organizations we work for exceed our obligations to the people we report to.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specializing in Unix and Unix-related management issues.