Fake News Alert: Existing Home Sales Report - Dave Kranzler

December 21, 2016

I’m going to have to throw a flag on the existing home sales report for November published today by the National Association of Realtors. The NAR would have us believe that home sales occurred in November at 5.6mm annualized rate for the month, up 15.4% from November 2015 and up .7% from October. I will point out that, of course, the orignal report for October was revised lower. But who pays attention to those details?

Take a look at this graphic sourced from Zerohedge which shows existing home sales plotted vs mortgage applications back to 2013:

I hate to be cynical, or accuse anyone of presenting “fake news,” but the mortgage application data completely contradicts the NAR’s “seasonally adjusted, annualized rate” interpretation of the data it collected. Existing home sales are based on closings (escrow clears), which means the sales report for November is based on contracts signed primarily from October and some in late Sept/early November. But mortgage applications began dropping off a cliff in late August. Clearly the NAR’s seasonal adjustment interpretation of the data is highly suspect.

Looking at the data itself – LINK – you’ll note that the NAR’s data sampling shows that home sales dropped 6.7% from October. Yet, it’s “seasonal adjustments” suggest that home sales increased from October to November, despite a massive plunge in mortgage applications during the period in which contracts would have been signed for November closings.

I’ve emailed the NAR several times over the years to have them explain their seasonal adjustments calculus. Every time I am politely declined. I will note that they use the same regression analysis model used by the Census Bureau.

The annual rate for a particular month represents what the total number of sales for a year would be if the relative pace for that month were maintained for 12 consecutive months. Seasonal adjustments, which are determined by using the X-12 Variant created by the Census Bureau, are then used to factor out seasonal variances in resale activity.

They do at least disclose that, although, if anything, that fact detracts from the credibility of their calculations. Of course, if I were looking for credibility, I would not advertise that I use the statistical guesstimate package created by the Government…

I would suggest that a year from now, anyone who looks back at the data produced today by the NAR will discover that the number was revised lower by a significant amount. But who looks at revisions? The Government and industry promotion organizations know this. It doesn’t matter how far off the rails their initial “seasonally adjusted” data strays, as long as they revise them at some point in the future, when no one is looking, it gives them plausible deniability if they are ever held accountable.