The longer-term supply/demand picture in housing
Between January 2001, when the Fed started cutting interest rates and got this whole bubble going, and July 2006 (the worst month of the bust so far), the supply of existing homes for sale shot up 117%. The numbers are 1.777 million units in 1/01 and 3.861 million units in 7/06.
During that same time, the Seasonally Adjusted Annual Rate of sales climbed from 5.1 million units to 6.33 million. That's a 24% increase. Even if you measure from 1/01 to the peak demand month (6/05), it's still an increase of just 43% in the sales rate.
In other words, the rate at which supply has increased has far exceeded the rate at which demand has increased. Since July, we have only whittled that supply mountain down by 1.1% (to 3.82 million units in November).
On the new home side of things, from 1/01 through 7/06, the seasonally adjusted annual rate of sales climbed just 4.6% (936,000 in 1/01 vs. 979,000 7/06). The gain as measured through the peak sales month (7/05 -- 1.367 million SAAR) was 46%.
From 1/01 to 7/06 (the high point to date), for sale supply shot up 94% (296,000 in 1/01 vs. 573,000 in 7/06). It has come down just 4.9% since then.
As I see it, the downturn will likely be a lengthy, drawn out affair, with relatively weak building activity, sales, and pricing lasting until at least 2008. There is simply too much inventory out there, and a lot of it is held by "weak" hands -- homeowners/speculators who bought as second homes or investment purchases. They were 40% of the purchase market in 2005, by the National Association of Realtors' own figures, the highest non-owner-occupied market share percentage in history. Unlike primary homeowners, these owners are more likely to try to sell when the going gets tough, as it has.