AP's Kravitz Ignored Available Data in Holding Out False Hope For Improved Full-Year New-Home Sales

January 4th, 2012 11:19 PM

A few readers asked me for my reaction to Derek Kravitz's December 23 report at the Associated Press on new-home sales. I thought that it was reasonably good, but felt that his leaving open in readers' minds the idea that this year's sales could conceivably top last year's was in bad form.

I was too kind. Based on data available elsewhere, Kravitz should have known (and maybe did) that instead of holding out the possibility that "December would have to produce its best monthly sales total in four years for 2011 to finish ahead of last year's total," he should have written something along the lines of: "There is virtually no chance that 2011 will be better than 2010."


Here are the new-home sales results for the past three full years, and for 2008 - 2011 through November:

HomeSales2010andThruNov11

To top last year, new-home sales in December (unless there's a tiny bit of help from upward revisions to November) will have to be 43,000. Yes, as Kravitz wrote, it's true that the last time single-month sales were that high was four years ago, in December 2007, but he never gave readers that 43,000 number and didn't tell them November's actual number was only 22,000. Thus, readers had no idea of how deep the Kravitz was shoveling it.

But even worse, it turns out that mortgage application reports from the Mortgage Bankers Association during November 2011 were already telling us (and Mr. Kravitz, whose job it is to be aware of these things) that the chance of December's sales significantly topping last December's 23,000, let alone hitting 43,000, is in the neighborhood of zero. Since the time from mortgage application to loan and property closing is typically about four weeks or 30 days, November's mortgage app reports are good indicators of how strong homes sales (new and existing) will be in December.

Mortgage application data for weeks containing days in November came in as follows:

  • Week ended November 4 -- "The unadjusted (mortgage application) Purchase Index increased 2.7 percent compared with the previous week and was 2.5 percent lower than the same week one year ago."
  • Week ended November 11 -- "The unadjusted Purchase Index decreased 14.8 percent compared with the previous week and was 9.5 percent lower than the same week one year ago."
  • Week ended November 18 -- "The unadjusted Purchase Index increased 15.2 percent compared with the previous week and was 4.8 percent lower than the same week one year ago."
  • Week ended November 25 -- "The unadjusted Purchase Index decreased 33.7 percent compared with the previous week and was 18.2 percent lower than the same week one year ago."
  • Week ended December 2 -- "The unadjusted Purchase Index increased 47.2 percent compared with the previous week and was 0.8 percent lower than the same week one year ago."

Note that all five weekly numbers are lower than they were a year ago. Last year's higher mortgage-app numbers led to 23,000 new-home sales in December. This is not the stuff of which a doubling of new-home sales from 22,000 in November 2011 to 43,000 in December will be made.

The only conceivable bits of wiggle room Kravitz might concoct would to argue that the mix of new mortgage loan apps somehow radically swung towards new homes in November and early December (on what basis?), or that tens of thousands of cash buyers suddenly came out of the woodwork and bought new homes in December (on what planet?), or that there was a sudden rush to tighten closing time frames during the last few weeks of the year (during the Christmas season?). I don't think so.

Bottom line: Kravitz "cleverly" gave readers reason to hold out hope that 2011 will end up being a better year for new-home sales, while ignoring data which would clearly have told him that, barring a miracle, it's not going to happen. You and the AP should be better than that, Derek.

Cross-posted at BizzyBlog.com.