Existing home sales (seasonally adjusted) were flat in March, after a jump in February that seemed counter to notion of a popping housing bubble. Rising to flat existing home sales are definitely better news than plummeting values. But, the news is not really as okay as it seems.

Houses differ from many commodities in that sales should be affected by inventory, even at constant prices. Why? Because an inventory of houses is not like an inventory of identical widgets. No two homes are exactly alike. Cautious buyers seek homes that must match very idiosynchratic criteria. Holding "pickiness" constant, the likelihood that any buyer finds a suitable home will corelate positively with the number, and therefore "selection", of homes is target neighborhoods. Therefore, all else being equal, existing home sales should be expected (as an absolute number) with increasing inventory.

Even with February's "jump", existing home sales didn't keep up with rising inventory. Flat sales in the face of rising inventory this month is suggestive of a weakening home market.

Here's a graph of existing home sales in absolute numbers (seasonally adjusted):

And here's existing home sales as a fraction of inventory:

In the context of rising inventories, the graph of sales in absolute terms would be expected to "better" than reality, because sales should rise with inventory. The graph as a fraction of inventory looks "worse" than market reality, because even in a healthy housing market, the increase in sales attributable to higher inventories would not keep up with the increase in inventory when there is already a fairly reasonable selection of homes available.

Hat tip: Calculated Risk

Update: Commenter tryuj notes, correctly, that I'm off by a year on the months graphed. I'll try to redo these graphics soon... Okay, they're fixed now. Thanks to tryuj and qwerty.

Steve Randy Waldman — Tuesday April 25, 2006 at 3:57pm permalink
tryuj (mail):
Your years are off there, sport.
4.25.2006 10:12pm
Steve Randy Waldman (mail) (www):
Oops! And so they are...
4.26.2006 3:17pm
Steve Randy Waldman (mail) (www):
...okay. fixed. thanks again.
4.27.2006 8:28am
Note: You don't have to register to comment! Just use the pull-down menu to comment as "Guest".