Last month I started my note on August new home sales by saying:  "Yesterday's new home sales report, on a nationwide basis anyway, can be thrown into the trash basket."

This was due to the failure of about 30% of counties in Florida and Texas constituting 14% of the entire national market being unable to report.  

 

This month they obviously did report. I frequently caution that this report is extremely volatile, and any given month is subject to huge revisions, so at very least it would be a good idea to average August and September together for a number of 614,000. 

In particular, of the 116,000 total increase this month, a grand total of 23,000 were contributed by the Northeast, Midwest, and West Census regions combined.  The remaining 83,000 came from the South..  At 405,000 annualized, this month's reading for the South was 18% higher than the next highest number this of 343,000, and 4% higher than its prior peak for this entire expansion of 389,000.

To consider the hurricane impacts, here are new home sales for the South census region (blue) vs. all three other census regions combined (red):

For the record, though, sales of new single family homes did make a new expansion high (blue in the graph below), in contrast to the flat performance of single family housing permits (red):

Interestingly, new home sales continue to be most consistent with purchase mortgage applications, which most recently (this morning) were up 10% YoY (h/t Bill McBride):

Because of the volatility of new home sales, I like to track the 3 month nationwide moving average. Here's what that looks like so far this year:

Jan 575k 
Feb 584k
Mar 617k
Apr 614k
May 615k
Jun 603k
Jul 600k
Aug 585k*
Sep 603k*

Notice there was a persistent decline from the peak in March, through August. 

*Had sales in the South census region not been delayed due to the hurricanes, but had occurred proportionately as per prior years,  the three month averages would be as follows:

Aug 596k

Sep 601k

Further, the seasonal adjustment in September is roughly +7% vs. August at +3%, so shifting some sales to August brings the September adjustment down slightly.

Thus the most reasonable optimistic spin is that the decline in the three month moving average this year has likely stopped. But even with this month's big gain in sales, it has nevertheless failed to make a new high since March.

Let's also briefly look at prices.   These are even more volatile than sales.  Here's what the YoY change in those look like, monthly (blue) vs. quarterly (red):

g

These continue a slow, persistent rise, generally slightly in excess of wage increases, but likely in line with the increases in median household incomes over the last several years.

We have one more month before we get more favorable YoY comparisons in interest rates, which should lead to less pressure on the new home market. I would not be surprised to see this month's big increase  revised significantly downward, but on the other hand, I suspect the three month average is going to improve as interest rate comparisons ebb. But I do not see a great deal of upside to the housing sales unless affordability improves.