Housing Numbers -- New Home Sales Statistics
(and how the numbers are often misinterpreted)
Last Update -- March 6, 2008
Is the U.S. Department of Commerce significantly over estimating new home sales? In the latest report, they indicated that annualized, seasonalized number of new homes sales in January was 588,000. The New Home Sales Report is always an annualized, seasonalized 12 month projection. The Commerce Department used their voodoo to determine that the 43,000 homes sold in January 2008 projects out to an annualized, seasonalized 588,000 home sales.However, using simple math, it seems that 506,000 new home sales might be a more accurate projection. Of course this would have resulted in reported new home sales being down 17% instead of just down 3%.
As shown in following chart, estimated new home sales were typically about 5% above the 2004 projections as new home sales increased during 2005. However, the Commerce Department consistantly overestimated sales by 20% during 2006 as the market declined. If the market continues to decline in 2008, it seems likely that the Commerce Department's January projection is 15% overstated.
In 2007, January home sales totaled 66,000. The Commerce Department projected 937,000 new home sales. Over the course of the next three months they revised this number down to 890,00 and the total ended up being 776,000. [See the chart at the bottom of this page for details and follow the previous link - however beware, the chart at the bottom of this page provides the revised annualized, seasonalized number (890,000) and the chart available via the above link provides the originally published number by the Department of Commerce (937,000). Yes indeed, it can be very confusing]. The Department of Commerce over projected by 21% in January 2007. Based on the final result for 2007, January accounted for 0.085% of the year's total new home sales. In January of 2008, new homes sales were reported to be 43,000. By simple math, if it is assumed that January will account for 0.085% of the total (as in 2007), it would lead to projected new home sales of 506,000.
Details are provided in the chart below:
Results for 2004 - 2007 (in thousands)
2004 2005 2006 2007 Actual January 88 92 89 66 Projected 12 month 1,106 1,106 1,233 937 Actual 12 Month Sum 1,203 1,283 1,052 776 January as a %
of actual 12 month0.073% 0.072% 0.084% 0.085% Thus, if 2008 is like 2005, then new home sales for the next 12 months will be 599,000, just about what the Commerce Department reported. However if the 2007 number is used (43,000 divided by 0.085%), the result is 506,000 new home sales. If January, 2008 turns out to be like 2006 and 2007, then the U.S. Commerce Department will have dramatically overstated projected new home sales. If January sets a pattern for the year that behaves more like 2005 and 2006, then the projection will turn out to be fairly accurate.
----------------------------------------------------------------------
The best commentary on the January New Home Sales number was by Dirk Van Dijk of Zacks.
More Bad News on New Home SalesWhile the stock market has become conditioned to the builders reporting weak fourth quarter 2007 results, sales during the first quarter of 2008 may be well below the modest expectations. Typically, the turning of the page on the calendar to a new year results in a boost in new home sales. Conventional wisdom in the real estate field indicates that the start of the spring selling season starts after the Super Bowl. However, with new home sales the pick up usually starts a few weeks earlier, in large part due to the lag time between the writing of a sales contract and setting move in dates prior to the start of the new school year.
This year is off to a very slow start. A weekday visit to a builder sales office/model has become a lonely experience, as there is a painful paucity of potential buyers touring the models. Also, staffing and hours the sales offices are open have been drastically curtailed. This weakness was confirmed by the Pulte Homes report on 1/30/2008 that "Given the ongoing weakness in the homebuilding industry, for the first quarter of 2008 we are projecting a net loss from continuing operations in the range from $0.15 to $0.30 per share". Further, Standard Pacific Homes indicated during their 2/5/08 conference call that January sales were down 39% versus January '07. As shown in the Alexa chart below, visitors to the home builders' websites have plunged, and the January effect has been very muted this year. During the boom, January and February website visitor traffic surpassed the levels from the previous fall. This year, website traffic to the leading builders websites has only just barely returned to the levels from this year's weak fall season.
When the December new home sales report came out on 1/28/2008, the financial headlines were misleading as usual. The 41% decline in year over year sales was widely reported as a 4.7% decline (see below for a detailed discussion of the source of the misreporting). An aspect of the December report that received little attention was the precipitious decline in sales of $500,000+ homes. Sales of $500,000 homes dropped by 67% on a year over year basis to just 3,000 in December, 2008 from 9,000 the previous year. Obviously, the higher rates and tighter credit standards for jumbo mortgages is starting to really take a bite out of sales of the high end of the market.The Monthly New Home Sales Report
The headline numbers in published reports about new home sales and construction activity often seem to be voodoo economics. This is because the headline number is usually an annualized, seasonalized estimate (aka highly manipulated number) that is compared to the previous period. For reference, the term headline number refers to the subject line in a news story. In articles reporting on the release of new home sales data, which is a monthly joint release by the U.S Census Bureau and the Department of Housing and Urban Development, the headline number is a comparison of the current annual estimate versus the annual estimate the previous month. A comparison to the previous month does not provide the same flavor that a comparison to the year ago period would. And while a comparison to the year ago period is usually included in most articles, it is typically buried in the body of the text, and seemingly goes virtually unnoticed by the financial markets. Thus, a month in which sales are down by 14% versus the same year ago month can be reported as a 5% gain in sales in headlines, if the previous month was down 21% and the full year estimate is raised. This is what happened when the September 2,006 new home sales were reported on October 26. Although new home sales were down significantly versus the year ago result, they were not as bad as the August result. Thus, the estimate for the full year was increased by 5%, and that was the headline number.The stock market reacted very positively to the news, in part because investors and speculators did not understand that the headline, "sales up 5%", disguised the fact the sales were actually down 14% versus the year ago period. The S&P 500 average was up almost 7 points that day and the Dow Jones Industrial Average was up almost 94 points. While many factors fueled the jump in stock prices on October 26, misinterpretaion of the housing numbers was probably a strong contributor. (The headline number was reported as up 5%, but the results were later revised down to 3%, as shown below. Here is the press release).
The market also may have misread the October report. The headline number was down 3%, which disguided the fact that new home sales were actually down over 25% versus the year ago month. While there was lots of other positive news that day that fueled the stock markets gains of 12 points in the S & P 500 and 92 points in the DJIA on November 29, the new sales number was not a big drag on the market. Remarkably, the DJIA average has been up on nine of the ten days during which the Census Bureau released the 2007 monthly new home sales report.
The headline number on 12/19/06 reporting on the November release of new home starts and building permits was also wildly misleading. The headline on the report from the Associated Press (AP) stated ""Wholesale Prices Surge; Housing Rebounds" due to the housing start estimate for the full year being revised upward. However, comparing the result to the previous year shows that total starts plummeted down 28% and starts of one unit structures (single family homes) were down 31% versus November 2005. The impact on the stock market was probably muted due to the closely watched Produced Price Index (PPI) having been released concurrently. The PPI caught the market by surprise with a larger than expected increase in prices.
There is not any fraud or obvious intent to mislead in the reported numbers, but they are very misleading if one does not understand how they are calculated. This website shows the actual sales for a specific period and comparisons to the same year ago period, which may provide a more accurate indication of the status of the new home market
In the chart below, the left table provides unadjusted monthly results. The middle table includes the seldom published comparison versus year ago. The right hand table shows the results that the headline numbers are drawn from. Notice how small the percentage changes that have been reported in headlines have been the last few months (+5%, -13%, -5%, and -3%, respectively) which has been viewed very favorably by the stock market. However, glance over at the middle and left columns, and it is fairly obvious that new home sales during the last four months have been very soft (-27%, -36%, -39% and -35% versus year ago).
Source - U.S. Department of Commerce
New Homes Sales
(in '000's)
Month
2005
2006
2007
2008YOY %
ChangeJanuary 92 89 66 43 -35% February 109 88 68 -23% March 127 108 80 -26% April 116 100 83 -17% May 120 102 79 -23% June 115 98 73 -26% July 117 83 68 -18% August 110 86 60 -31% September 99 82 53 -36% October 105 77 57 -27% November 86 72 46 -36% December 87 71 43 -39%
Cumulative
New Homes Sales
(in '000's)
2005
2006
2007YOY %
Change92 89 66 -26% 201 177 134 -24% 328 285 214 -25% 444 385 297 -23% 564 487 376 -23% 679 585 449 -23% 796 668 517 -23% 906 755 577 -24% 1,005 840 630 -25% 1,110 916 687 -25% 1,196 988 733 -26% 1,283 1,059 775 -27%
Seasonally Adjusted
Annual Rate
& Change Vs Previous
Month
(in '000's)
2005
2006
2007
2008YOY %
Change1,193 1,173 890 588 -3% 1,252 1,038 840 -6% 1,324 1,121 830 -1% 1,270 1,121 907 9% 1,311 1,101 861 -5% 1,272 1,078 797 -7% 1,367 969 796 0% 1,271 1,009 701 -12% 1,253 1,004 693 -1% 1,346 952 725 5% 1,236 987 630 -13% 1,259 1,019 605 -5%
The links below are badly out of date. We'll get around to updating them one of these days
Interested in digging a little deeper? View a comparison of monthly new home starts to new home completions. Not surprisingly, new home starts are down for the year. However, it does seem surprising new home completions remain up for the year.
Additional information available on this website includes details on the monthly increase in inventory of ready to move in homes that the builders are trying to unload. There is also a table showing the daily movement in stock prices and interest rates on the day of the report release.
Post a Comment | Enter the Housing Numbers Prediction Contest | Subscribe to Our Mailing List Looking for a new home? Find new homes and learn about a 1% cash back rebate program at iNest.com. iNest clients have used the service to purchase over 16,000 new homes and over 10,000 visitors daily search for new homes on iNest.
Find information on deals, incentives, and builder closeouts available on Chicago area new homes at the New Home Deal Finder blog. Award winning author Aricka Flowers provides the scoop where to find great deals on Chicagoland area new constructon homes.
©2008 Housing Numbers Contact Us