The headlines say new home sales fell off a cliff. Median sales prices were little changed - but the backlog of unsold homes grew with over 7 months of supply.

Analyst Opinion of New Home Sales

This month the backward revisions were significantly downward. Because of weather and other factors, the rolling averages are the way to view this series. The rolling averages also significantly fell helped by the downward revision of the last two months of data.

This data series is suffering from methodology issues which manifest as significant backward revision. Home sales move in spurts and jumps - so this is why we view this series using a three-month rolling average.

Econintersect analysis:

• unadjusted sales growth decelerated 20.2 % month-over-month.
• unadjusted year-over-year sales was NEGATIVE 18.0 % year-over-year. Year-over-year growth rate this month was the lowest growth since 2011.
• three-month unadjusted trend rate of growth decelerated 6.2 % month-over-month - is down 2.8 % year-over-year.

US Census Headlines:

• seasonally adjusted sales down 5.5 % month-over-month
• seasonally adjusted year-over-year sales down 13.2 %
• market expected (from Econoday) seasonally adjusted annualized sales of 610 K to 640 K (consensus 625 K) versus the actual at 553 K.

The quantity of new single family homes for sale remains well below historical levels.

Seasonally Adjusted New Homes for Sale

The headlines of the data release:

Sales of new single‐family houses in September 2018 were at a seasonally adjusted annual rate of 553,000, according to estimates released jointly today by the U.S. Census Bureau and the Department of Housing and Urban Development. This is 5.5 percent (±12.1 percent)* below the revised August rate of 585,000 and is 13.2 percent (±13.6 percent)* below the September 2017 estimate of 637,000.

Unadjusted New Home Sales Monthly Volumes In Thousands

The median sales price of new houses sold in September 2018 was $320,000. The average sales price was $377,200.

Unadjusted Median New Home Sales Price

The seasonally‐adjusted estimate of new houses for sale at the end of September was 327,000. This represents a supply of 7.1 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. This data is based on contracts signed - not actual properties conveyed.

To provide nationwide coverage of building activity, a multi-stage stratified random sample procedure was used to select approximately 900 building permit-issuing offices, and a sample of more than 70 land areas not covered by building permits.

Each month, for permit-issuing places, a sample of residential building permits is selected from each of the sampled permit offices. The probability of selecting a permit is proportional to the number of units authorized by the permit. Permits for one-to-four-unit buildings are sampled at an overall rate of 1 in 50. All permits authorizing buildings with 5 or more housing units in the sampled permit offices are selected.

Each month, for areas that do not require building permits, field representatives conduct a road canvass in each of the sampled non-permit land areas to identify the start of new buildings. All new residential buildings found are selected for the survey.

Once a permit or building is selected, a field representative contacts the owner or builder, by telephone or in person, to conduct the interview each month as necessary. Contact continues until the project is either completed or abandoned. If a single-family home is not sold by the time of completion, the project will continue to be followed until the sale occurs. Each month, interviews are required for about half of the buildings currently being followed up.

Each month, housing starts, completions, and sales estimates derived from this survey are adjusted by the total numbers of authorized housing units (obtained from the Building Permits Survey) to develop national and regional estimates. Estimates are adjusted to reflect variations by region and type of construction, and to account for late reports and houses started or sold before a permit has been issued. Reported data are seasonally adjusted. The Construction Methodology (PDF) document contains further information.

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).

It is more informative to look at these changes over the nearly fifty-year history. The following graph shows new home sales normalized to the population from St, Louis Fed:

Seasonally Adjusted New Home Sales Ratio to Population