By ANDREW JEFFERY

In contrast to the silver lining of higher-than-expected home sales tallied in yesterday’s release of Existing Home Sales, this morning’s Case-Shiller November Home Price Index registered the worst year-over-year performance on record.

Property values in November 2008 tumbled 18.2% from the prior year and 2.2% from the previous month. As measured by the Case-Shiller’s 20-city index, home prices have retreated 25% from their mid-2006 peak and now sit close to levels not seen since 2004.

Like a CD stuck on repeat in the guy down the hall’s dorm room while he is away at class (you try listening to Chumbawumba’s Tubthumping for 3-hours straight and not being driven to drink at 3pm), annual price declines were the worst in Phoenix and Las Vegas, followed by San Francisco, Miami, Los Angeles and San Diego.

News that home prices are falling, however, isn’t exactly news. So let’s look at what this Case-Shiller Home Price Index is anyway, and how it differs from the less murky Median Home Price data released every month by our friends at the National Association of Realtors (the NAR).

As Cirios statistics’ wizard Austin Nelson explained last week, the median home price is simply the middle price in a list of sales. So, the NAR grabs data on all home sales in a given month and picks out the one with the same number of sales on either side. The logic behind using this measure, rather than the more commonly understood average, is that it reduces the effect of outliers, or sale prices that are either much higher or much lower than the prevailing trend.

Case-Shiller, on the other hand, uses what are known as “paired sales” to evaluate home prices. By examining sales in a given month, then looking back to find the most recent sale price and date for each house, statisticians are able to back into an annual rate of decline.

For example, if house A sold in November ‘07 for $250,000 and again in November ‘08 for $200,000, Case-Shiller would say the annual home price decline was 20%. By limiting the search only to sales that have a relevant pair (ie, previous sale), the methodology limits sample size, potentially leaving margin for sampling errors. Sample size is something I’ll let Professor Nelson explain in further detail at a later date.

So, which method is better, paired sales or median price? As is typical with such economic questions, the answer is a resounding neither.

Median Home Price measurements provide good information about the distribution of sales, and in the current environment reflect that since Jumbo mortgages are nigh impossible to get and foreclosure sales are driving most markets, cheaper houses are selling far more often than more expensive ones.

Case-Shiller, on the other hand, represents a more accurate level of home price declines for specific homes — far more important to the average homeowners. Of course, a number slapped on a metro area is fairly meaningless when it comes to an individual property, but it’s still more useful than a bucket of houses being sold for different reasons by different sellers in different neighborhoods.

With Case-Shiller a month behind the NAR data, keep an eye out next month to see if the silver lining found in December’s median home sales data is reflected in the paired sales analysis. If it is, you can be sure calls for a bottom in housing will once again abound.

Tags: case shiller, Housing, NAR, prices, property values

This entry was posted on Tuesday, January 27th, 2009 at 10:12 am and is filed under Housing Perspective, Property Valuations. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.