The headlines for new home sales tell you that they are not good. What is not good also is the seasonal adjustment methods and noisy data which are providing this answer.
New home sales were relatively robust in February causing the March data to look bad in comparison. But overall the trend shows increasing volumes.
Let us start with perspective – new home sales are less than 1/4 of the peak values seen in 2005 – and are running at levels last seen in the 1970’s (non seasonally adjusted data). But still, since mid-2011 new home sales have been growing year-over-year. US Census Headlines for March 2012:
- sales down 7.1% month-over-month
- year-over-year sales up 7.5%
- market expected annualized sales between 318K to 320k (actual was 328K – seasonally adjusted)
- sales down 16.6% month-over-month
- year-over-year sales up 10.7%
- three month rolling average up 2.1% month-over-month, up 15.8% comparatively year-over-year
As the data is noisy (large monthly variations) – it would be best to view new home sales growth using a three month moving average. This graph suggests that new home sales volumes have bottomed (and are now growing) in the last seven months.
The headlines of the data release:
Sales of new single-family houses in March 2012 were at a seasonally adjusted annual rate of 328,000, according to estimates released jointly today by the U.S. Census Bureau and the Department of Housing and Urban Development. This is 7.1 percent (±20.7%)* below the revised February rate of 353,000, but is 7.5 percent (±19.6%)* above the March 2011 estimate of 305,000.
The median sales price of new houses sold in March 2012 was $234,500; the average sales price was $291,200. The seasonally adjusted estimate of new houses for sale at the end of March was 144,000. This represents a supply of 5.3 months at the current sales rate.
Caveats on Use of New Home Sales Data
This data is compiled by sampling, and historically has little revision:
First, an estimate is made monthly of the number of housing units for which building permits are issued in all 19,000 permit-issuing places as described above.
Second, for each permit selected in the 900 permit-issuing places, an inquiry is made of the owner or the builder to determine if the house was already sold or is for sale. If already sold, the month and year the house covered by the permit was sold is obtained. In case the unit authorized by a permit in a particular month is for sale, follow-ups are made in successive months to find out when the unit is actually sold.
Ratios are calculated (by type of structure) of the number of units authorized by permits, based on the Building Permits Survey to the number of units authorized by permits based on estimates generated from the 900 SOC permit offices…..
As in most US Census reports, Econintersect does not agree with the seasonal adjustment methodology used and provides an alternate analysis. The issue is that the exceptionally large recession and subsequent economic roller coaster has caused data distortions that become exaggerated when the seasonal adjustment methodology uses several years of data. Further, Econintersect believes there may be a New Normal seasonality and using data prior to the end of the recession for seasonal analysis could provide the wrong conclusion.
Econintersect determines the month-over-month change by subtracting the current month’s year-over-year change from the previous month’s year-over-year change. This is the best of the bad options available to determine month-over-month trends – as the preferred methodology would be to use multi-year data (but the New Normal effects and the Great Recession distort historical data).
With new home sales at 25% of past rates, whatever your interpretation of the new home sales data is not significant enough to matter. Also the data is distorted by the first home buyer’s stimulus which required contract signing before 30 April 2010 – causing a data bubble and subsequent trough. In spite of Econintersect‘s reservations about the efficacy of seasonal adjustment at the present time, it is interesting to look at the deep history of the seasonally adjusted data. [note graph below updated through February 2012]
The broad bottoming process for new home sales in 2010 may not be confirmed or denied for another year or more. The critical factor will be whether the one-year positive trend can continue as year-over-year comparisons will no longer be against the very low sales after the collapse of the tax credit stimulus micro-bubble.
The seasonally adjusted new home sales rate is the lowest it has been for 50 years and has been at that level for almost two years. At the beginning of 1963 the U.S. population was around 188 million. With annual new home sales averaging around 550,000 per year in 1963, the extreme depression in the new home market is evident. In 1963 the rate of new home sales was about 290,000 per 100 million of population. In 2011 the number is about 100,000 per 100 million.
It is more informative to look at these changes over the nearly fifty-year history. The following graph shows new home sales normalized to population from from St, Louis Fed (graph updated through December 2011):
The same data is plotted below to include the average for the entire period and two moving averages (graph updated through October 2011):
The bottom line is that the new home market is in an extreme depression and the apparent bottoming process has been dragging on for two years, if in fact the bottom has been reached. Recent review of the Fed 2011 stress tests for banks has a new recession scenario that would see home prices decline another 20% from here. It is unlikely that the attempts to complete a bottom here could hold under those conditions. Econintersect analysis of recession indicators is still not seeing the start of new U.S. recession, however. We can only hope that outlook continues.