There seems to be lots of confusion about today’s Commerce Department release on new home sales. The report showed that new home sales were at a seasonally adjusted annual rate of 467,000. Part of the confusions stems from the fact that that is a completely bogus, made up, fictitious number in the first place. Seasonal adjustments are arbitrary attempts to create a smooth curve and they aren’t finalized until years after the fact. In addition, annualizing a single month is insane to begin with. Analysts are trying to make sense out of fiction. In the meantime, the headline number missed the Wall Street conomist guesspectation of 475,000.
The new home sales data has the benefit of being contract data reported just a few weeks after the survey date. It’s the closest thing we have to an official real time measure of the state of the housing market. All of the widely followed existing home sales data suffer from severe lag and the error is compounded by extreme smoothing in worthless measures like Case Shiller. By the time of the Case Shiller release, it represents an idealized market from deals that went under contract 5 months ago, not the market as it is today. The Commerce Department, to its credit, at least measures current contracts, and it does make available the actual, as reported, unsmoothed, unmanipulated data. The mainstream media, to its discredit, completely ignores the actual monthly data and only reports the made up, seasonally adjusted crap.
That being said, the actual data is also severely deficient, primarily because the survey sample is an infinitesimal, and therefore unreliable, slice of the market. As more data comes in during subsequent months, even the unmanipulated raw data is subject to large revisions. This month, the number for August was revised down by a whopping 10%. The July number suffered the same fate.
The data also suffers from the fact that the survey does not take into account contract cancellations. When the market is falling apart, cancellations can be epic. But they’re not subtracted from previous month’s data. We don’t see the impact until sales actually go off the cliff.
All that aside, the number of sales reported for September was 38,000. That’s the actual number of sales extrapolated from the builder survey sample taken early in the month. That number was 22.5% above September 2013. Any way you slice it, it’s a strong rebound after a series of down months on a year to year basis throughout 2014. On the other hand, 22.5% gain or not, compared with historically typical sales levels, single family housing construction remains in a historic depression. There’s just no way around that fact.
