gives you an edge in real estate. Find homes »

S&P Cautions About the Use of the Seasonally Adjusted Case-Shiller Index

Date:April 27, 2010 | Author:Stan Humphries | Category:Market Trends

Standard and Poor’s issued a statement Friday detailing recent research into the differences between the seasonally adjusted and non-seasonally adjusted Case-Shiller Home Price Indices (HPI). The essence of the statement is that recent market dislocations, primarily in the form of foreclosures (but probably also including the impact of homebuyer tax credits), have made the seasonal adjustments less than reliable. Because of this fact, S&P urges the use of year-over-year comparisons of the index or, if month-over-month comparisons are desired, using the non-seasonally adjusted index.

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.

While the seasonally adjusted and unadjusted index time series seem pretty similar (see Figure 1), the month-over-month changes for the two indices tell quite different stories in recent months. Figure 2 shows the month-over-month change in the two flavors of Case-Shiller HPI for the Composite-20 set of markets (along with the composite Zillow Home Value Index for the same set of markets). The seasonally adjusted HPI has shown positive monthly appreciation for the last nine months (since May 2009) whereas the unadjusted HPI has shown negative month-over-month changes for the past four months (since October 2009).

I agree that the trend of the unadjusted HPI represents what’s happening in the market right now better than the seasonally adjusted HPI. The month-over-month change in the unadjusted HPI for January was -0.4% (versus +0.3% for the seasonally adjusted change) which matches the monthly ZHVI change for January of -0.4% (the monthly ZHVI change remained the same in February).

I do have difficulty, however, with the volatility of the monthly change in both the seasonally adjusted and unadjusted HPI. It seems incredible that monthly home price appreciation in the unadjusted HPI shifted from a low of -2.8% in January 2009 to a high of 1.7% in July 2009.

Clearly, a lot of this change in appreciation is related to seasonality, which was presumably the impetus for their effort to seasonally adjust the index in the first place. I suspect what is confounding the effort to remove this seasonality is the changing mix of foreclosure re-sales in the monthly sales data.

In the winter months, when sale volumes are lower, foreclosure re-sales make up a larger proportion of overall transactions and, because the Case-Shiller HPI includes foreclosure re-sales, the steep depreciation seen in these types of sales push appreciation rates down. Conversely, in the summer months, when sale volumes pick up, foreclosure re-sales make up a small proportion of overall transactions and appreciation rates increase.

Whatever the cause, I do believe that the ZHVI gives a better, more consistent indication of what is happening to home values for homeowners who sell their homes via traditional, arms-length transactions (i.e., not via a foreclosure process). I don’t believe that home values were appreciating monthly last July at a rate that equates to more than 20% annualized, as the Case-Shiller HPI would have us believe. We believe home values have indeed stabilized over the course of 2009 but never got positive on a month-over-month basis and have deteriorated a bit during the winter months. We do, however, expect monthly depreciation to start to decline again on their way to getting flat by mid-year.

Kudos to S&P to being as transparent as possible about issues in the construction and interpretation of their various indices.

Dr. Stan Humphries is a real estate economist and real estate expert for Zillow. Stan is in charge of the data and analytics team at Zillow, which develops housing market data for most major metropolitan statistical areas in the U.S., and provides economic research for current real estate market conditions. He helped create the algorithms for the popular Zestimate® home value and the Zillow Home Value Index (ZHVI).

« Previous Post How Soft is the Housing Market? Next Post » Renovation Financing: Homepath Renovation

May. 25 162 Responses

Home Values Rise Again in April

Negative Equity Rates: How Does Your Neighborhood Fare?

May. 24 80 Responses

Zillow Negative Equity Report Shows 1 in 3 Mortgage Holders Underwater

Responses (9)

0 tweets 1 likes 8 comments

DebtFree 11:16 AM - Apr 27,2010

It’s a useful exercise to compare wages against housing prices, to see that it’s very unlikely that housing prices will cease their decline anytime soon, and in fact, have MUCH further to fall:

Wages:
uspolitics.about.com/od/economy/ig/Consumer-Income–Insurance/Real-Median-Income–2007.htm

Home Prices:
graphics.nytimes.com/images/2006/08/26/weekinreview/27leon_graph2.large.gif

When 3X median income can purchase a median priced home, the housing market will have “stabilized.” Until then…
DebtFree 11:23 AM - Apr 27,2010

Fixed links below…

Wages:

tinyurl.com/HouseholdIncome1967-2006

Home Prices:

graphics.nytimes.com/images/2006/08/26/weekinreview/27leon_graph2.large.gif
Stan Humphries 11:50 AM - Apr 27,2010

Hi DebtFree. You’re nothing if not consistent! I agree with your larger point that home prices are still not in line with either median household incomes or rental equivalents. Home prices can return to this equilibrium in either of two ways: a) they drop sharply to equilibrium and then return to a more normal annual appreciation rate sooner; or b) they stay flat for a *much* longer period of time (perhaps negative after adjusting for inflation) and allow income and rents to catch up over the long-term.

In both scenarios, we end up in the same position 5-10 years from now, just via different trajectories. I favor the latter scenario because I think household formation, depressed new home construction, and decreased home turnover rates will combine to stabilize prices. I believe that they will be flat for a *long* time because of the countervailing pressures of negative equity, foreclosures, persistent high unemployment, and pent-up supply of homeowners who want to sell (and a decent portion of these sellers who won’t turn around and buy another home).

Would I be right in assuming that you believe that these countervailing pressures will dominate the ones on the upside? Are there other considerations that lead you to favor the first scenario (rapid drop in short term)?
DebtFree 12:30 PM - Apr 27,2010

The thing about housing prices being “flat for a long time,” is that there’s no precedent in modern times where a housing price boom occurs, then prices remain at a plateau, then prices increase again. Rather, for housing price boom/bust cycles, each bust is equal in both magnitude and duration to the preceding boom:

graphics.nytimes.com/images/2006/08/26/weekinreview/27leon_graph2.large.gif

In the Shiller chart, look at the late-1970s boom, the late-1980s boom, and of course, the late-1990s boom; each bust cycle returned prices back to the 110 level. One can print out that chart, literally fold it at each apex, and the downside will overlay the upside almost exactly.

The countervailing pressures you list (foreclosures, unemployment, seller oversupply) are why home prices will continue to decline, until they reach their equilibrium determined by incomes. This bust cycle shouldn’t be any more rapid than previous bust cycles, in that the bust cycle will equal in both magnitude and duration the preceding boom cycle.

Now, this could all be “Doom & Gloom,” and even a broken clock is right twice a day, but the fundamentals (incomes) seem quite clear as to where the housing market is headed. No tea leaves needed.
Arizona Homes For Sale 8:18 AM - Apr 28,2010

I agree that the trend of the unadjusted HPI represents what’s happening in the market right now better than the seasonally adjusted HPI. i agree with you also!
Stan Humphries 5:40 PM - Apr 28,2010

@ DebtFree: The NYT graphic that you reference (showing the Case-Shiller time series) is inflation adjusted which is why it shows sharp declines following the boom cycles. Most of these bust cycles were actually long periods of flat home prices (in nominal terms) that only look like declines once you look at them in real terms (adjusting for inflation). See this chart for a comparison: http://bit.ly/aSshkd

My expectation of flat prices for a sustained period of time is in nominal terms (i.e., the prices we see every day) and these may look like declines in real terms. Perhaps we’re not that far apart in expectations after all.