Statistics published by the National Association of Realtors appear to overstate sales of existing homes by 15 to 20 percent, mortgage and property data aggregator CoreLogic says in a new report that concludes home sales fell more sharply last year than previously thought.
A NAR spokesman said the CoreLogic claim "is premature at best," and NAR will be making some benchmark revisions to its historic sales data later this year.
NAR’s figures — based on data collected from multiple listing services and large brokerages — show sales of existing homes fell 5 percent in 2010, to 4.9 million. But CoreLogic, which collects public sales records from county recorders and courts, estimates that home sales actually fell 12 percent, to 3.6 million.
The implications are not trivial: A slower rate of sales means that it will take longer to burn through unsold inventory, and a glut of homes for sale in a given market can undermine prices. CoreLogic says the unsold inventory on the market in November represented 16 months of supply, compared with NAR’s estimate of 9.5 months.
Weak sales following the expiration of the federal homebuyer tax credits, an excess supply of unsold homes, and the impact of sales of distressed homes is driving home prices down, CoreLogic said. A national, repeat-sales home-price index compiled by the company was down 5.1 percent in November from a year ago.
If that trend continues, national home prices will probably be down 10 percent year-over-year by spring, CoreLogic said.
In its latest forecast, NAR projects that the median existing-home price will be down 0.6 percent from a year ago during the first quarter of 2011, but post year-over year gains for the next five consecutive quarters.
CoreLogic says one reason NAR’s existing-home sales data may be inflated is because the benchmark multiplier NAR analysts use to adjust for MLSs that they aren’t getting data from hasn’t been calibrated since 2004.
But there’s been consolidation among MLSs since then, CoreLogic noted, and a decline in the number of for-sale-by-owner sales outside the MLS and brokerage process. That means NAR is now capturing a greater percentage of existing-home sales and doesn’t need to make so large an adjustment when extrapolating its results.
CoreLogic said that historically it’s been able to account for only 85 to 90 percent of the existing-home sales tallied by NAR.
Beginning in 2006, NAR’s sales numbers began to look even more inflated relative to data collected by CoreLogic, the Mortgage Bankers Association, and the U.S. Census Bureau, a trend that has "continued and become more pronounced through 2010," CoreLogic said in the February edition of its monthly report, "U.S. Housing and Mortgage Trends."
2010 Existing home sales: NAR vs. CoreLogic
While NAR numbers show home sales bottomed in 2008 and then rebounded in 2009, CoreLogic data shows no such rebound in 2009.
CoreLogic Senior Economist Sam Khater said the analysis "is less about NAR’s data than a critique of data in general."
"Anytime you’ve got fundamental changes in the market like that, it’s going to cause market data to go haywire," Khater said. It’s important to have data from a wide range of sources, Khater said, in order to "see where the truth lies in between them."
NAR spokesman Walt Molony said that while NAR will be making benchmark revisions later this year to its historic sales data, "data drift" issues are expected to be "relatively minor."
"Under the circumstances, the CoreLogic claim is premature at best, especially given the process that is currently under way," Molony said in an e-mail.
The last benchmark revisions of the existing-home sales series was based on 2000 Census data, Molony said, and NAR will soon be rebenchmarking using independent sources. NAR will be consulting with outside housing economists on the methodology to determine if there is any drift in the data, and by how much, he said.
"There’s been a notable increase in nontraditional sales outside MLSs, so a major function in consulting with outside housing economists and government agencies is to determine methodology and obtain consensus on the benchmarking," Molony said.
He said NAR will also be looking for a new way to rebenchmark existing-home sales on a more frequent basis instead of waiting for Census data to be updated every 10 years. NAR already updates sales rates and months’ supply benchmarks on an annual basis, Molony said.
Molony said the rebenchmarking of existing-home sales will result in "no notable changes" to NAR’s previous characterizations of monthly sales changes, and no impact on price data.
Khater said CoreLogic’s public records data captures all sales, whether they involve a mortgage or are all-cash purchases, and regardless of whether a home was listed in an MLS or not.
One drawback with public records data is the lag time before sales are reported and data are collected. Khater said CoreLogic estimated December 2010 existing-home sales in the February report using preliminary data.