There is a similarity between bombing Belgrade and banking in Belmont. Our information about what is really there is out of date. We can deliver a product to a very precise geographic location, but have no clear idea that it is appropriate. The cost in lives and foreign policy of bombing errors in Belgrade is clear today. The policy implications of our ignorance of what lies on the ground in the USA is less obvious. A couple of months ago we launched an air offensive in Yugoslavia with the blissful assurance that precision guided munitions were at our disposal. One of the unintended consequences is that we very precisely delivered three missiles to the wrong address. As we consider launching the desktop underwriting of real estate loans, have we given much consideration to the possibility that we might be about to commit lots of federally insured dollars to the wrong properties?
In California the inventory of real property is poorly known. In most states that inventory is kept and updated more or less well by property tax assessors. In California Prop 13, which assesses value by purchase prices, has led to a degradation of that knowledge. If a property changes hands, whatever it is, the price is the value. If it doesn't, then the assessment is effectively fixed anyway. In both cases, who cares what the property is? Although the assessor still has a duty to pick up new construction, assessment of the alteration of existing properties is so arcane that it is not cost-effective. Some effort is made, no doubt, but its efficacy in recording changes to the real estate inventory is necessarily much less than perfect.
But, doesn't desktop underwriting rely on physical inspection of the properties on which the loans are to be placed? Well, yes, but what about the properties fed into the Automated Valuation Models (AVMs) to predict value? If there was any effort to do inspect them, then there would be no saving in time or money over conventional appraisal. So how predictive, in fact, will the AVMs be? I suppose that could, itself be predicted as a function of the original homogeneity and age of neighborhoods. But I haven't seen any proposal to do so, nor would such an effort probably pass muster under civil rights requirements and the Community Reinvestment Act. While it is arguable that the Berkeley Hills and Pacific Heights in San Francisco are intractable problems for AVMs, so will Hunters Point, San Francisco or East Oakland be. It is probably alright in the first two cases to discriminate against wealthy white people if one is so inclined, but the second two examples are problematic.
The least of our problems, however, may be ignorance. Deception and deceit will always have a place when money is involved. If fortunes can now be made at the speed of light, real estate fraud will probably also be expedited.