For the record, however, sales of new single family homes have failed to make a new high for 5 months (blue in the graph below), and do accord with the stagnation shown by the much less volatile metric of permits for single family homes (red):
Here's what the 3 month nationwide moving average of new home sales, that smooths out most of the volatility, looks like so far this year:
Jan 575k
Feb 584k
Mar 617k
Apr 614k
May 615k
Jun 603k
Jul 600k
The three month moving average has failed to make a new high since March.
Since the Census Bureau does collect this data on a regional basis, we can avoid the hurricane complications by showing only those sales for the Northeast, Midwest, and Western regions and omitting the South. Here's what that looks like:
Like the national data, this data peaked in March both on a monthly basis as shown, but also on as a 3 month rolling average. Since this is *not* affected by the hurricanes, it absolutely does confirm a slowdown.
Median prices were probably less affected by the hurricanes, but with that caveat, the very slow uptrend since the beginning of last year that was indicated last month may have been broken, as the median price of a new home in August 2017 was just slightly above what it was 12 months before:
The three month moving average of YoY prices also shows this deceleration -- although the overall trend is still positive:
We had a similar stall, for a similar reason, in the first part of 2014. This was due to the interest rate "taper tantrum of mid-2013. The increase in mortgage rates that started last November is probably having a similar effect. YoY interest rates may therefore be lower beginning this November, and that will begin to help the housing market thereafter.
The question is whether a general lack of affordability has reached a point where sales have indeed peaked for this expansion.