Zillow has a few issues with a paper published in the Appraisal Journal (a publication of industry trade group the Appraisal Institute), which claims the estimates of home values on its website are no more accurate than the estimates of value made by homeowners.
The paper is authored by a trio of University of Texas-San Antonio professors and looks at Zillow’s estimates of home values vs. the actual sales prices on 2,045 single-family residential properties sold in Arlington, Texas, in 2006.
Among Zillow’s complaints, according to a statement issued in response by the company, are that the survey uses data over three years old and is restricted to one city, thus it is “out of date and limited in scope.”
In addition, the researchers look at sales in 2006 but use “Zestimate” (what Zillow terms its price estimates) values from January and February 2007, which the company claims is “apples and oranges as it’s two separate periods of time.”According to the paper, the average overestimation of a Zestimate is 11.67% or $15,576, which led the authors—Daniel R. Hallas, Ronald C. Rutherford and Thomas A. Thomson—to conclude, “Zillow’s magnitude of overestimation is marginally higher than the value overestimation by recent homebuyers.”
The paper cites data that state on average homeowners overestimate the value of their homes by 5.1% and that new owners overestimate by 8.4%.
They conclude, “In most cases, Zillow should not be relied on to provide an accurate estimate of value.”
However, Zillow responded, “In the paper’s own comparison, more of Zillow’s estimates are within 10% of what the authors consider 'accurate’ (59%) vs. homeowner estimates within 10% (37%).”
In its statement, Zillow said the professors did not reach out to it to discuss its approach to home valuation, as it “would have been happy to talk to them as we are very open and transparent about both our accuracy and the intended use of Zestimates as a starting point for learning about home values.”
Mr. Rutherford, in an interview, stated the ideal was to get the price information as of the date of sale, but that was not possible. He also rejected Zillow’s “apples and oranges” contention.
Part of the reason for doing just one city was the data collection issue, especially in Texas, which is a “nondisclosure state,” he said.
As for a follow-up study, unless Zillow was willing to provide access to its database, so that the researchers do not have to go through the time consuming data collection process, it is unlikely another study using more recent information would take place, Mr. Rutherford said.The paper does state that Zillow says Zestimates are not appraisals and should be used as one estimate of market value and the company encourages getting a valuation from a professional appraiser. Zillow’s disclaimer notes that these values are a starting point.
The authors said they chose Arlington based on the availability of housing sales, its stability “and Zillow’s indication that the Zestimates for this location are at its highest accuracy level because it has the best data for this area,” a point Mr. Rutherford reiterated in the interview.
