It may not seem possible, but the horrific housing market crash of the past few years has actually been worse than we previously thought.
In the past two years alone, more than three million fewer homes may have been sold than was previously reported. On Tuesday, the National Association of Realtors, which releases the most widely watched monthly data on the real estate market, said it has been overstating home sales for as much as five years. The trade organization said it will release revised numbers going back to 2007 next week. The restatements may chop the number of homes sold since the beginning of 2010 by a third.
(LISTS: Homes with Million Dollar Discounts)
For at least a year, housing economists have suspected that the NAR was overstating home sales. We reported on the growing discrepancy between the NAR’s sales numbers and outside data providers back in February. At the time, a NAR spokesperson said the assumption that the trade group’s data were inflated was “premature at best.” They also pointed out that real estate analytics company CoreLogic, one of the firms raising questions NAR’s sales data, was a competitor.
Now NAR appears to be admitting its error. The question what that means for the housing market.
It’s not clear how off NAR was, but it could be a lot. As of last month, NAR said that nearly 4.2 million existing homes (NAR does not track new home sales) had been sold in the U.S so far this year. By CoreLogic’s estimate, though, only 2.4 million homes had changed hands in 2011. Why such a big difference?
(MORE: Bulldoze: The Way to Save the Housing Market)
NAR uses 2000 Census data as its benchmark numbers, and bases its estimates for home sales on a sampling collected from multiple listing services, which track local housing markets, and larger brokerages. The problem is that 2000 is the last year the Census collected the data that NAR has long used as a benchmark. So the farther we got from 2000, the more flawed NAR’s data became. That’s what has led to suspicions about NAR’s sales data. NAR’s last benchmark revision in 2000 cut sales by just over 10%
But the problem could be even larger than some suspect. NAR now says there were also reporting problems with the MLS data as well. For instance, local MLSes sometimes overlap, listing some of the same houses twice. So some sales may have been double counted. And in the past few years, NAR says many MLSes have expanded, so the overlapping problem has probably gotten bigger. What’s more, NAR has always had to guess at the number of for-sale-by-owner transactions each month, since those sales often happen outside of MLSes, and not in NAR’s databases. But in recent years, it appears FSBO sales have been dropping faster than the rest of the market. NAR now says it was probably overestimating those sales as well.
(MORE: Latest Housing Bust Casualty: Babies)
CoreLogic, on the other hand, calculates its figures through public sales records from county recorders and courts. That has led some to believe its data is better at estimating sales than NAR’s. In the recent revise, though, NAR says it looked into using county recorder data, and found that collecting the data that way, because some counties are better at recording sales than others, it was only able to capture about 80% of all sales. That suggests the CoreLogic numbers could be somewhat underestimating the total number of U.S. sales. In the end, NAR decided to stick with MLS numbers. It says it has corrected the reporting and double counting problems it discovered, and it has switched to benchmarking its calculations to data that the Census updates annually, instead of data that used to be updated once a decade, but is now not collected at all. Sam Khater, a senior economist at CoreLogic who has been briefed on NAR’s new methodology says it makes sense. “NAR’s data will be more consistent with ours,” says Khater.
What does this all mean? First of all, the housing bust was worse than we thought. According to current NAR numbers, sales have fallen 31% from their 2005 peak. If CoreLogic’s numbers are closer to the truth, real estate transactions have actually dropped more like 50%. As for the future, Khater says he believes housing prices aren’t going to rise anytime soon, and NAR’s downward sales revision doesn’t significantly change his outlook. But maybe it should. According to NAR’s current data, the number of homes sold this year will be the lowest since 1998. The revision, though, could knock the sales pace back to the mid-1980s. But the U.S. population and number of households had grown dramatically since then. That suggests there might be more pent-up demand in the housing market, and that we could be closer to a rebound, than we thought.