by Calculated Risk on 2/19/2012 12:44:00 PM

The Census Bureau will report new home sales on Friday, and the consensus is for sales of 315 thousand on a seasonally adjusted annual rate (SAAR) basis. This is up less than 2% from the 310 thousand SAAR sales reported in January 2011.

That seems a little puzzling. Consider the following ...

First, look at the NAHB builder Housing Market Index. More builders still view sales as "poor" as opposed to "fair" or "good", but the HMI - and all of the components - are up sharply from a year ago (the most recent report was for February, but compare January 2012 to January 2011):

	Housing Market Index	Traffic of Prospective Buyers	Current Sales
Jan-11	16	12	15
Jan-12	25	21	25
Feb-12	29	22	30

This would seem to suggest more than a 1% or 2% increase in sales.

Second, look at the recent builder reports (from Tom Lawler):

	Settlements			Net Orders			Backlog
	End 2011	End 2010	End 2009	End 2011	End 2010	End 2009	End 2011	End 2010	End 2009
D.R. Horton	4,118	3,637	5,529	3,794	3,363	4,037	4,530	3,854	4,136
PulteGroup	4,303	4,405	6,200	3,084	3,044	3,748	3,924	3,984	5,931
NVR	2,391	2,639	2,550	2,158	1,765	2,000	3,676	2,916	3,531
The Ryland Group	1,040	909	1,666	915	775	969	1,514	1,187	1,732
Meritage Homes	894	837	1,202	749	713	621	915	778	1,095
Beazer Homes	882	549	961	724	553	728	1,309	800	950
MDC Holdings	950	865	1,109	523	519	637	1,043	842	826
Standard Pacific	782	619	943	615	428	547	681	414	599
M/I Homes	667	650	858	505	460	448	676	532	650
Total	16,027	15,110	21,018	13,067	11,620	13,735	18,268	15,307	19,450
YoY % Change	6.1%	-28.1%		12.5%	-15.4%		19.3%	-21.3%

From economist Tom Lawler on February 7th:

The latest Census report on new SF sales showed a YOY increase in Q4/2011 sales of just 3%, and a YOY decline in Q4/2010 sales of 20.5%.

The nine-builder group’s order backlog at the end of 2011 was up 19.3% from the end of 2010.

As I’ve noted many times, Census’ methodology for measured new SF sales is not directly comparable to reports from builders. I’m guessing that part of the “stronger than Census” builder reports reflect gains in market share, but I’m also guessing that overall new home sales were a bit better than preliminary Census data suggested.

The combination of higher order backlogs, stronger sales, and unusually mild weather in much of the country is likely to result in single-family starts numbers in the first few months of 2012 that are significantly higher than “consensus.”

Click on graph for larger image.

Sure enough. Single family housing starts were revised up sharply for December and were above 500 thousand SAAR in January. As Lawler notes, some of this was probably weather related, but some of the pick up was evident in the builder reports.

So if the builders are reporting a “stronger than Census” increase in sales (even accounting for market share gains), confidence is up (actually less pessimism), and single family starts are up sharply from a year ago, it seems surprising that new home sales were essentially unchanged in January.

Goldman Sachs is forecasting sales of 310 thousand SAAR in January 2012 (no change year-over-year), and Merrill Lynch is forecasting 315 thousand. I think I'll take the over.