Automated real estate valuations gain acceptance

August 2, 2004

AVMs are very reliable in most cases, panel says

By Samantha Peterson
Inman News

Christopher Cagan, First American Real Estate Solutions
Automated valuation models have become very reliable in most cases and the lending industry has accepted them more as an alternative to traditional appraisals, according to a Friday panel at Inman News' Real Estate Connect 2004.

AVMs are programs that generate appraised values of properties by using information from various databases, combined with statistics. They were first developed to accurately value homes for property tax assessment purposes, said Vicki Cassens, managing director of real estate and lending practice for strategic development worldwide.

Over the years, Cassens said, there have been various types of AVMs, from models that value homes based on an indexed price appreciation to those that take into consideration all the different features of a house. Most models these days blend different technologies and use a variety of methods to come up with a value.

The types of data used to come up with values also can vary according to where the property is located, Cassens said. That's because different states make different home sale data public. Texas, for example, doesn't disclose how much a property sold for, but does make public details about the house, she said. New Jersey, on the other hand, makes public the sale price, but doesn't disclose the property details.

Still, today's AVMs have become so sophisticated that they save time by doing appraisals in seconds and minutes as opposed to days and weeks, said Christopher Cagan, direct of research and analytics for First American Real Estate Solutions. They also save money, costing only a few dollars instead of hundreds for a traditional appraisal. They are excellent at appraising the value of standard types of housing in densely populated areas, and were great to use in the recent refinance boom.

Their disadvantages, however, include missing things in the database it uses to determine the value, such as property details, property conditions or special circumstances of the location. They have difficulty with unusual properties such as high-end real estate that is difficult to find comparables for, Cagan said. Properties on the other end of the scale can also be difficult for AVMs as can remote properties that are difficult to comp.

But they're excellent in determining the value of 90 to 95 percent of residential properties, he said.

"AVMs are about as good as they can be or appraisals can be," Cagan said.

That's because price and value are not fixed numbers, he said. Instead, they are a range of numbers that can vary even in traditional appraisals. He's seen, for example, identical or near identical properties in the same neighborhood vary in price by as much as 5 to 10 percent.

AVMs will continue to evolve by companies adding different bells and whistles to their basic appraisal capabilities, Cagan said. They'll add such things as giving the person who requests the appraisal the standard deviation, and the ability to appraise niche properties such as those on the beach. The industry also will likely consolidate to a few major AVMs and will connect AVMs to other products.

Mike Sklarz, chief valuation officer for Fidelity National Financial, said it's fairly well documented that AVMs are very accurate. Also, he said, AVMs are obviously becoming more accepted in the lending industry.

"I think that trend is only going to grow," Sklarz said.

He said Fidelity is working on developing a product that will predict home buyers' lifestyle – and potential house need – changing events, such as getting married or divorced or having a child. That product, in development now, uses publicly available data and combines it with AVM technology, he said. It could be used to help generate leads for real estate agents, lenders or anyone else interested in such consumers.

AVMs also could be used for mass appraisals needed to property tax valuations, helping developers determine the correct presale price of new construction, identifying overvalued property listings, identifying undervalued properties for sale and helping real estate agents correctly price property listings.

Copyright: Inman News Features


Contact ALTA at 202-296-3671 or communications@alta.org.