Inventories at 23,000+ in August 2007?

I got offered a bet by a real estate agent friend regarding housing inventory levels for this year. He wanted to be me $100 that OC real estate inventory won't pass 20,000 units at ANY point this year. I didn't know if that was a good bet or not, so I built a crude (yes, crude) forecasting model to see whether or not I ought to take the bet.

The results were a little surprising:

The bars are color-coded in two ways, first by season (red is Summer; Orange is Fall, Blue, Winter and Green, Spring), and then the darker bars are actual figures where the lighter are the results of the forecast (also with an E at the front of their label.) Cool, huh?

I took as an assumption that because sales are so strongly seasonal, that one might be able to take the January inventory levels and use it as a "baseline" inventory number for the year. So I compared to inventory levels in Jan 2006 to each of the respective months, giving me a relative inventory increase from January by month. I then used a tampening factor to adjust the forecast downward so that it would come out to where we will be at the end of this month. I know, not very fancy or inclusive, but it's better than a Scotch-inspired guess.

Well, the model came up with some pretty interesting numbers, namely in the inventory levels for the months in and around summer:

E Jun-07 21,724
E Jul-07 22,721
E Aug-07 23,111
E Sep-07 22,743
E Oct-07 21,669

So my model, such as it is, predicted that we'd probably have five months with inventories over 20,000; so, this morning I gave my agent friend a call and accepted his bet. With that said, if inventories actually do break through 20,000 units, I think the $100 bet is that last thing either of us will be worrying about....