In an article entitled Veros Succeeds with Forecasting Tool written by Rick Grant on June 27, 2003, Mr. Grant writes that AVM developers are being pressed to provide more than just a snapshot of the value of a residential property. Some are responding to this pressure by attempting to include in their products an estimate of what the future value of the property will be. While this is proving very difficult to do with any degree of accuracy, Veros Software here has found success with its model.

Article published by Thomson Media Visit the AVM Technology web site for more AVM news Mortgage Technology website

Mr. Grant then continues to say that as the company, developer of the VeroVALUE automated property valuation tool, introduced its expanded VeroFORECAST earlier this year at the MBA Technology Show in Orlando, Fla. But the company has been using the tool internally for longer, at least as long as the MBA's annual convention last October. At that conference, the company puts its new forecasting engine to the test and made a prediction specific to home values in Orange County, Calif.

Mr. Grant continues by quoting Darius Bozorgi, CEO of Veros Software, "In the release we made in October, we said we were not seeing a decrease in home prices at that time. What we were seeing was a decrease in the rate of appreciation."

Mr. Grant then continues with his article saying that in fact, Veros said at the time that it expected home values in Southern California to appreciate an additional 10% over the next 12 months.

The forecast was received with a great deal of skepticism, especially from experts in Orange County, who were quoted in the local media projecting property values would fall between 5% and 10% over the next year.

"The result was an actual 7% increase for the six-month period ending April 2003," Mr. Bozorgi told National Mortgage News. "It turned out pretty well."

Mr. Bozorgi added that the Veros estimate was a 12-month projection. "So at the rate it's going, it will probably be higher than the 10% that we projected in October. But it is certainly not a decrease, as many people had projected."

The reason the new VeroFORECAST is so accurate, Mr. Bozorgi said, is that is doesn't rely on the same analytics that other forecasting models use, at least not anymore.

"When we started looking at the forecasting issues, we did what many people attempted to do in the past. Our early models depended very heavily on past home prices and home price trends. That was a major factor, along with economic data," he said. "We found in our early attempts that our forecasts were good for a short window of time, say one quarter to six months, but then their accuracy would fall off."

Mr. Grant than wrote that in order to get more longevity into their estimates, Veros had to find another approach.

"We wiped the slate clean, so to speak. We said, 'Let's develop forecasting models without looking at past home prices, only focusing on economic variables.' We played around with that for some time until we found a mix that works," Mr. Bozorgi said.

The new engine can predict out 12 months into the future, with what the company says is a very high degree of accuracy.

According to Eric P. Fox, vice president of technology at Veros, "Historic price models that rely on past price fluctuations to accurately predict future price changes are not very good at predicting several quarters into the future. Rather, by understanding and automating the complex relationships between price changes and numerous key economic indicators, we are able to understand how declining interest rates, increasing unemployment, a falling affordability index, a steady number of active listings, and numerous other indicators will influence prices into the foreseeable future."

Mr. Grant then writes that Veros plans to combine its AVM offering and its forecasting engine in order to produce a collateral score it hopes to have ready for the market by the end of the year.

Based on current forecasts for the next year, VeroFORECAST continues to report increasing prices in most of the major metropolitan areas across the country.

"In many of these areas, the rate of increase is beginning to or already has started to slip," said Mr. Bozorgi. "As an example, we anticipate that Orlando prices will see an approximate 7% increase over the year. On the West Coast, most of Southern California should see increases of 13%-15%, while San Francisco prices will remain relatively flat in contrast to previous gains. Phoenix, Denver, and Oklahoma City should experience more modest 4%-5% gains compared to Cincinnati or Nashville that will see little if any gain. On the East Coast, Buffalo, N.Y., and Boston should expect increases of 4% and 5%, respectively, while Baltimore should see prices rise approximately 10%."

So what are the company's predictions for the hottest and worst residential real estate markets in the nation?

Veros says the top five real estate markets over the next 12 months will be Miami, Boston, Newark, N.J., Ventura County, Calif., and St. Petersburg, Fla.

The worst real estate markets will be Akron and Columbus, Ohio, Cincinnati, Las Vegas and San Francisco.

Email the author at: rick.grant@thomsonmedia.com