Misleading Housing Numbers

I'm sick of articles like this which report data as follows:

The 20-city slice of the S&P/Case-Shiller Home Price index recorded a drop of 0.6% from March to April, compared with a 2.2% drop in the prior month. The index has declined every month since July 2006.

The problem is, as Seattle Bubble is fond of pointing out, there's normally a bump in housing prices from March to April due to the seasonality of home sales.

A more fair comparison would be comparing the year-over-year price changes from March and April. The Wall Street Journal reports that the 20-city index reported a year-over-year decline of 18.7% in March, and The New York Times reports that the 20-city index reported a year-over-year decline of 18.1% in April.

The numbers CNN reports sound as though we saw a 72% improvement, but once you remove the seasonality and look at the year-over-year numbers, the improvement is only 3.2% better. That's quite a difference.

And in either case we're looking at a second-derivative number here: the change in the rate of decline. We're still talking about a very significant decline, which appears to be ongoing. Even if we see a few months of positive changes, Japan saw several multi-month periods of positive improvement in their housing market before it bottomed out.