Ok Census Bureau: enough is enough.
In the government's endless desire to show just how blistering, nay, stupendously amazingful the gargantuantest housing recovery has been, we just got news that New Home Sales in January soared, SOARED, to 437K from an upward revised 378K, slamming expectations of a 380K print (chart). "This is the best New Home Sales print since August 2008!" the mainstream media roared with adoration and approval (hoping for an avalanche of ad RFPs from Trulia, Zillow and of course, the NAR). Alas, as so often happens, there was more than meets the eye.
Much, much more.
For one: the actual, unadjusted number of homes sold in January was a meager 31K (of which 1,000 houses sold in the $750K+ range): a tiny 4K increase from December, the same as August, and lower than all months from March to July 2012 (chart); the houses for sale rose to the highest since December 2011; the Median Price plunged to $226,400, the lowest since January 2012 and down $23k from December's $249,800. Finally of the 31K houses sold in January, just 12K were actually completed, with 10K under construction and 10K not even started. So who cares: seasonal adjustments happen all the time, and January just happens to be an important inflection point right? Yes.
Which is precisely why we took the December-to-January change in New Home Sales as reported since the peak of the housing bubble, to get an accurate sense of how this inversion has behaved in history. The result is plotted below: the blue bar shows the sequential change in actual data. The red one shows the seasonally adjusted one.
Please highlight the full retard seasonal adjustment.