Caution: Use the Case-Shiller Home Price Index with care

The S&P/Case-Shiller Home Price Indexis one of the most commonly referenced indices available to benchmark and track the changes in U.S. housing prices. Although it is a valuable resource, we recommend a clear understanding of how the methodology works before relying heavily on the Case-Shiller results for critical decision making.

Case-Shiller's 10-City and 20-City Composite indices continue to report record declines, falling 19 percent and 18 percent, respectively, over the 12 months through November 2008 (the most recently available data). As of November, the 10-city index is down 27 percent from its mid-2006 peak, and the 20-city is down 25 percent. The two indices have fallen every month since August 2006, which makes 28 straight months. Home prices have returned to their March 2004 levels, according to this index.

Click to view larger image. *Peak activity since 1985. **Annual Mortgage Costs + 1/7th of the down payment divided by income. ***Proprietary affordability scale with 0 meaning most affordable time since 1983, 5 meaning median affordability, and 10 meaning least affordable time. Data on all markets available at www.realestateconsulting.com. Sources: John Burns Real Estate Consulting, Bureau of Labor Statistics and U.S. Census Bureau through the month ending November 2008.

Today, distressed sales are one distinct type of home sale. Although foreclosures are putting significant downward pressure on the index, foreclosure sales can't be fairly evaluated in isolation, because distressed pricing in a market also affects pricing on non-distressed sales (a second distinct type of sale). New construction is the third distinctive type of sale, and it is not figured into the index.

Below are some of the methodology variables that influence the results of the Case-Shiller indices.

The Case-Shiller indices use a paired sales analysis so that sales of the same home are tracked over time. Once sales pairs are identified, they are then weighted according to:

n Price anomalies — Smaller weights are applied to homes that “appear to have changed in quality, or sales that are not representative of market price trends.”

• High turnover frequency — Sales for the same home must be at least six months apart.
• Time interval adjustments — Pairs with longer intervals are given less weight.
• Initial home value — The weight for the pair is determined by the first sale price.
• Case-Shiller eliminates the transaction if “the order of magnitude in values appears unrealistic.”
• An allocation to one of three “tiers” is made for each paired sale. Allocation to a tier is made according to first sale price (i.e., the lowest tier is the bottom third of sales prices). Lower tiers receive a lower weighting.

Many of these nuances help prevent data errors in a stable market, but in today's environment of distressed sales, they could affect the reporting as values are dropping. The median home is now a foreclosure, which tends to be in a worse location and does not represent what you might consider to be the median home in many markets.

There are other considerations we feel could affect the Case-Shiller results:

In 2006, the sales mix shifted to subprime-type buyers who were, for the most part, purchasing less expensive homes. This resulted in:

• Understated median prices, because the volume shifted more to cheaper homes.
• Overstated price appreciation on paired sales, as the cheap homes in inner cities and outlying areas experienced high appreciation rates.