Seasonally vs. Non-Seasonally-Adjusted Housing Market Data

April 23, 2010

by Scott Sambucci


Thanks to CalculatedRisk, I learned about this week’s S&P’s statement on seasonal adjustments to Home Price Index data.   While seasonal adjustments are made to all sorts of economic data – unemployment, retail sales,  manufacturing –  it seems that Standard & Poors has a clear opinion when it comes to housing market adjustments.   From the  S&P statement:

After reviewing the data, the S&P/Case-Shiller Home Price Index Committee believes that, for the present, the unadjusted series is a more reliable indicator and, thus, reports should focus on the year-over-year changes where seasonal shifts are not a factor. Additionally, if monthly changes are considered, the unadjusted series should be used.

Since August 2009, the Case-Shiller HPI has been showing negative movements in their non-adjusted HPI values:

158.16 in Aug ’09
158.91 in Sept ’09
158.81 in Oct ’09
158.42 in Nov ’09
158.16 in Dec ’09
157.89 in Jan ’10

But, the seasonally adjusted Case-Shiller HPI values moved positively during this same period, which of course is a trend much ballyhooed and reported by the general media.  No one seemed to be talking about the consistently negative direction of the non-seasonally adjusted values.  OK, except for us over here at Altos Research because our housing market indicators have always been non-seasonally adjusted.  Take a look at the Altos 10-City Composite values since January 2008 (on a 90-day rolling average basis to stay consistent with the Case-Shiller HPI methodology):

Price and Inventory - Altos-10

Our Price trend peaked in August 2009, then moved negatively each month through early 2010.  A trough only emerges in 2010 in non-seasonally adjusted prices in our April 16 data set.  (Ahem – we also reported the trough of home prices back in January 2009 before anyone else got around to it…)  Looking at the data, there are some other causes for concern. Notice that housing supply is on the rise for the first time in 18 months. The picture gets worse when examining the weekly inventory numbers and you fold in Days-on-market, Price of New Listings, and Active Seller Price Reductions (yes, a shameless plug for our real-time data analytics and statistics). Put these together with Bob Shiller’s comments in this week’s NY Times article:

On March 31, the Federal Reserve ended its program of buying more than $1 trillion of mortgage-backed securities, and the homebuyer tax credit expires on April 30.

Recent polls show that economic forecasters are largely bullish about the housing market for the next year or two. But one wonders about the basis for such a positive forecast.

Momentum may be on the forecasts’ side. But until there is evidence that the fundamental thinking about housing has shifted in an optimistic direction, we cannot trust that momentum to continue.

The Case-Shiller HPI is a very reliable index in our opinion and it’s methodology has been proven over time to accurately reflect transaction prices and housing market trends.  Its challenge is the reporting time lag that exists by using closed transactions.  For example, index values for February 2010 will be released next week, so there’s always that 60-day reporting lag with initially-reported values subject to revision.  Yes, it’s vital to know where prices settled, but if you need to know what’s happening right now as we head into May, looking at the active market trends is what you need to do.

As luck would have it, we publish our Altos Research Real-Time Housing Report around the 10th of each month.  (Hey – it’s free. Monthly is pretty good for free.  If you want the really good stuff, you gotta pay for it…)

Previous post:

Next post: