- The Commerce Department said Wednesday that sales fell 3.6 percent to a seasonally adjusted annual rate of 402,000 from a downwardly revised 417,000 in August. Economists surveyed by Thomson Reuters had expected a pace of 440,000.
- It was the first decline since March. Sales in September were down 7.8 percent from a year ago.
- The median sales price of $204,800 was off 9.1 percent from $225,200 a year earlier, but up 2.5 percent from August's level of $199,900.
- There were 251,000 new homes for sale at the end of September, down 3.8 percent from August and the lowest inventory in nearly 17 years. At the current sales pace, that represents 7.5 months of supply.
- ***The data reflect contracts to buy homes, not completed sales. Many new homes are sold while they are still under construction, and buyers may be worried that they won't be able to complete the deal before the Nov. 30 deadline to take advantage of a tax credit of up to $8,000 for first-time buyers.
Of course, there are signs of economic improvement as Y/Y comparisons enter the easy phase (comparing against an ugly 4Q08) and some sectors might even start growing again (aside from secular growers like Amazon and PetMed Express - see prior posts here and here). However, we don't think stabilization and gradual improvement translates into immediate improvement on the consumer front, whether shopping for homes or for discretionary items. Therefore, we would expect normal seasonality to dominate M/M residential housing moves.
We understand that seasonally adjusted data are meant to adjust for seasonality, but let's look at the raw, not-seasonally-adjusted (NSA) data for new home sales. Thanks to CalculatedRiskBlog.com we present the following chart:
Looking at the chart, 2009 new home sales are represented by the red bars. We see that the 2009 pattern is shifted to the middle right (with peak sales in August) relative to typical peaks in the spring. Coming off of depressed sales and the economic tundra that was early 2009, this shift -- with gradual M/M improvement prior to September -- arguably makes sense and provided some rationale for economists to expect, yet again, higher M/M sales in September.
Still, looking closely at the graph, we see that -- for every single year, from 2003 to present, -- sales decline M/M from August to September. This looks like normal seasonality to us. Hence, why expect anything different given a still weak consumer environment, not to mention a real estate market where prices are still falling Y/Y and inventories remain extremely high? Or, put another way, why should the Market be so surprised by a modest M/M decline? Finally, we could even go one step further: forecasting M/M movements for anything is no easy task and subject to regular error.
Briefly, for a point of reference, let's also look at the much larger unit figures for existing home sales (again, thanks to CalculatedRiskBlog):
The raw, NSA data show normal seasonality and a slight M/M decline in September. However, we note that the National Association of Realtors reported a M/M increase for the seasonally adjusted annual rate:
- Existing-home sales – including single-family, townhomes, condominiums and co-ops – jumped 9.4 percent to a seasonally adjusted annual rate1 of 5.57 million units in September from a level of 5.09 million in August, and are 9.2 percent higher than the 5.10 million-unit pace in September 2008. Sales activity is at the highest level in over two years, since it hit 5.73 million in July 2007.
© 2009 Jeffrey Walkenhorst
Please see important Risk Factors & Disclaimer