NAR Admits Publishing Bad Existing Sales Data

Real estate professionals and economists who paid close attention to the monthly existing-home sales data provided by the National Association of Realtors were given a sharp shock earlier this week. Last Tuesday it was revealed in the Chicago Tribune that the association had been consistently reporting bad figures each and every month since January 2007, when the housing crisis really took hold.

The NAR has now advised that it will start revising four-years worth of sales data downward, beginning next week. The bad data has been attributed to a statistical error in its software that resulted in some home sales being counted twice.

Lawrence Yun, chief economist at the NAR, explained that the association used the Multiple Listing Service (MLS) to track existing home sales. However, he pointed out that the MLS database is limited to data from sales listed by realtors only. Sales of homes listed by owners are excluded from the database, and so the MLS only provides a narrow view of real estate markets. He also pointed out that because the majority of homeowners use realtors, sales figures for realtor listed homes became artificially inflated.

Also, the NAR often made a number of assumptions each month, based on data from the 2000 Census, which is clearly outdated.

The NAR has yet to reveal how big the revision to its existing-home sales data will be, but Mr. Yun did say that there is likely to be a “meaningful” decrease in the numbers, and that the new revision would show that the housing market downturn was even worse than previously thought.

The bad data is particularly concerning because the US economy is so closely linked to the fortunes of its real estate markets, and real estate policies followed by both Congress and the Federal Reserve have been based in part of the NAR’s existing sales data.