Risk Guarantees Will Convert Automated Evaluation Skeptics
Banton, H. Stan, American Banker
The automated valuation model was first introduced to academia in 1977 at the University of Tennessee by two young neophytes: Joseph Albert and me.
An academic toy? Perhaps. At the time it certainly was not accepted as a useful or meaningful tool by the banking and mortgage industries.
Twenty years later the concept began to see limited acceptance in the banking community. In more recent years that has grown to use by most financial institutions. AVMs are here to stay, if for no other reason than that they are faster and cheaper -- and they work.
End of story? Not quite. With a very low fault rate, the AVM has a low but real cost (in the case of loan default). Just like appraisers, automated valuation will fail from time to time.
Spread over a large portfolio, the losses are relatively insignificant compared to the very real cost saving. This risk of AVMs -- the default losses -- is therefore more comfortably accepted by large banks with high volumes of loan origination.
The real acceptance question lies with the smaller portfolio, the lower-credit-quality portfolio, and -- realistically -- the faint-of-heart lender. It's this segment of mortgage lending that chokes at the very idea of automated valuation.
Among the many technological innovations in the recent past, automated valuation is perhaps one of the most notable. The industry is truly in a paradigm shift. As with all technological changes, innovation breeds innovation -- and resistance.
Many lenders are hesitant to use automated valuation. In order to continue the industry's progress, what is needed for the technology to be universally accepted?
First, by recognizing that the basic principle of risk management requires that risk be spread over large numbers, innovative AVM providers are appealing directly to the insurance industry. Combining the servicing of numerous clients makes the risk reasonable and manageable. …