Unreliable housing statistic of the day

December 21, 2011

How many existing homes (as opposed to new homes) were sold between 2007 and 2010? The job of counting such things is outsourced to the National Association of Realtors, which up until yesterday said that the number was 20,629,000. Today, however, it released revised figures, saying that the true figure is 17,680,000 — a difference of 3 million homes. At an average of say $250,000 apiece, that means the economy saw $750 billion less in economic activity, over those four years, than the NAR had given us to believe. That’s real money.

Here’s the NAR’s official chart of the old and new numbers:


In one sense, this shows that the housing slump was much worse than we were told. But in another sense, what we’re seeing here is fewer people selling their homes at a loss. And what that says to me is that it’s going to take a very long time yet before we get a healthy, clearing housing market.

There are three factors making today’s housing market highly artificial. The first is historically unprecedented interest rates: the average 30-year fixed rate mortgage in November was taken out at just 3.99%. That creates a temporary speculative lift for the housing market, and raises serious questions about whether today’s housing prices can withstand a return to normality in the mortgage market. The connection between mortgage rates and house prices is by no means simple or predictable. But insofar as houses are being bought — and house prices are being supported — by investors looking to make money by renting them out, we could well see another downward lurch if and when today’s insanely low mortgage rates go away.

The second factor, related to the first, is that no sensible banker will lend money for 30 years at 3.99% — you just can’t make money that way. Which means that the US government has essentially become the sole lender to Americans looking to buy houses. At the same time, Democrats and Republicans are agreed that the current situation can’t be allowed to continue indefinitely. But the private sector has no interest whatsoever in stepping in where it was so badly burned in the past — neither banks nor bond investors want to buy mortgages these days, and it’s hard to think of what would make them change their mind. Except for a tiny sliver of jumbo mortgages, the private mortgage market in the US is dead, and showing no signs of being resuscitated. How much would you pay for a house today if you had no assurance that, when you come to sell it, most potential buyers of your home will be able to get a mortgage? It’s a real worry, and it’s going to become increasingly salient.

Finally, there’s the well-known phenomenon that house prices are very sticky on the way down. If you can’t sell your house without bringing a check to the closing, you’re likely to delay selling your house for as long as possible. And there are millions of houses on the market which have just been sitting there at unrealistic prices for well over a year — the owner won’t take less, but no buyer would dream of paying that much. That has created an unhappy overhang of unsold houses — and an even greater number of houses which people would love to sell, if only there were a decent market to sell them into. It’s the very definition of a non-clearing market, and today’s revisions only go to show just how few houses have been clearing the market for years now.

How all these factors are likely to play out over the next few years is impossible to predict; Neil Irwin, for one, reckons that housing “finally seems to have found its bottom — and may even be starting to bounce back.” Today’s news from the NAR, if nothing else, serves as an important reminder that housing data is messy, and prone to very large errors. Which means that anybody crunching numbers to come to a considered conclusion has to build in a large amount of GIGO risk.


Comments are closed.