By Michael Olenick, founder and CEO of Legalprise, and creator of FindtheFraud, a crowd sourced foreclosure document review system (still in alpha)
The National Association of Realtors (NAR) has announced that their estimates for home sales have been materially incorrect since 2007, and that they plan to restate the number of homes sales downward. Apparently the NAR derives their homes sales information from the Multiple Listing Services, the proprietary “want-ads” real-estate agents use to list houses for sale.
Their error stems from several causes, but one of the biggest is overestimating the number of people who sell their homes without a real-estate agent. During the height of the housing boom many people skipped real estate agents, and their 6% commissions, opting to sell houses on their own.
I can see how some people would have decided to skip an agent back then. When I moved from CA to FL in 2005, I went on a weekend vacation with the family, left the house with a real estate agent for an open-house, and came back to nine bids, all well above asking price. I was amazed, but also wasn’t sure what the real estate agent did since my former house obviously sold itself.
NAR took the number of people who came to this same realization, and projected many people were selling homes without real estate agents. They ignored one small factor — the national housing collapse — and the apparently difficult to infer fact that when houses became harder to sell more people hired real estate agents. They also used old Census data to project population trends and they didn’t factor in changes based on consolidation in the MLS market. In other words, they massively blew it.
NAR data is the sole benchmark for existing home sales in the monthly scorecard released by the Dept. of Housing and Urban Development (HUD), under supervision of the White House. As of Nov., 2011, the NAR statistics are the exclusive measure used to measure existing home sales, a vital component used by government agencies, banks, economists, and others to gauge the health of the housing market. Additionally, the data has been cited by the Dept. of the Treasury and various branches of the Federal Reserve.
My own data, derived from sales and mortgage information filed at the courthouse, showed that the NAR figures were wrong. So did data from CoreLogic, a title insurance spin-off based in CA that works on real estate data, who also takes their data from court records. Court records are the best source of real estate data, because any definitive move in real property requires a filing at the courthouse.
I asked colleagues at other companies in my industry and every one of them — from my tiny FL-based one to giant behemoths — uses courthouse data. We all knew that the NAR data was incorrect. Banks must have known, Fannie Mae and Freddie Mac, with their large portfolios of real-estate, surely must have noticed that their properties weren’t selling. It seems impossible that the small army of economists employed by the Federal Reserve could have overlooked the error. Some of us complained but nobody listened because the NAR data was presumed to be definitive. If ours was different, then ours must be incorrect, even if all of ours showed the same trend.
I’m regularly asked to analyze, comment on, or write about trends in housing and foreclosure data. But this week’s non-surprise from the NAR brings up a deeper issue; the methods used to compile that data. Besides home purchases, which isn’t really something I focus on, I’ve found some reports about foreclosure filing estimates to be wildly off.
Last week I wrote that Bank of America had transitioned from filing Countrywide foreclosures as “BAC Home Loans” to “Bank of America.” My analysis about their reason for doing so was the crux of the article. But it’s also noteworthy that there were a number of materially incorrect articles published about an enormous “surge” of filings by BOA when, in fact, there was an increase but most of the “surge” came from changing the name under which they file.
That is, people saw “Bank of America” filings spike, but didn’t notice “BAC Home Loans” filings declined. I spoke to a couple reporters that wrote those stories. They verified the corresponding decrease with county officials, but nobody ran a correction.
It’s not only backwards-looking data collection that’s questionable. My systems wade through a morass of garbage generated by poorly written bank and law firm computer systems chock-full of obvious errors.
One of my colleagues is Prof. Linda Allen, Chair in Banking & Finance at Baruch College of business, at CUNY. Prof. Allen is a long-time academic who understands finance, and especially real estate finance. Her long list of publications speak for themselves; she’s a genuine expert in the field. She predicted the real estate bust in academic papers long before it came along, and I’m fortunate to have her using my data for some of her research.
I ask Prof. Allen when I’m stumped about finance questions. Most of her answers are along the lines of “Michael, you know that language is a swap to hedge against daily LIBOR fluctuations” (um .. yep .. I knew that). But she was more straightforward when asked why the data collection systems at banks suck.
“Someone high on the food chain in one of the biggest banks admitted to me that they were not being paid to do the back office stuff,” Prof. Allen wrote. “You have to understand the disdain on the street for the operational end of the business. I wrote an article entitled ‘Revenge of the Back Office Nerd,’ that states that this inattention to the back office brought down the entire system.”
I know that banks hire great software engineers, but I also know that they’re more likely to end up creating quant-trading systems than writing operational software. But you’d think that with amounts of money swooshing around the system one of them would stop downstairs to take a peek, if for no other reason than to ensure that the data their systems use for trading is accurate.
One of President Obama’s planks surrounded data transparency; making sure that regular people had access to the information that affects their lives. Part of the meltdown in the derivative markets was a lack of transparency, which exists to this day. We still don’t have a full list of the loans Fannie Mae owns, despite that Freddie Mac has been publishing this data, albeit in a difficult to aggregate manner, for years.
Finance is difficult, and hitting an iceberg is bad. But hitting one because somebody thought a lowly looking should wear blinders, because they weren’t worth of enjoying the view, is worse. The NAR fiasco is our big and already injured ship scraping needlessly and perilously close to disaster. There’s a possibility that when their revisions are released we’ll know that, in hindsight, we should have made different decisions than we did.
Even if we manage to escape disaster we should take this mess as a warning. Let’s demand open, honest, full disclosure about data, data collections, and collection methodologies from any organization generating information used to build public policy.