After being the big optimist who was bashing the double-dip gang in the summer and fall, I am now back to being the killjoy who refuses to join in the celebrations over the November data on housing starts reported yesterday. The point that I made in a prior post is that these numbers are highly erratic. This is especially true of the monthly data on starts of multi-family units, which were driving the jump reported for November.

This chart gives the basic picture.

Click for Larger Image

Source: Census Bureau.

This chart shows three important pieces of information. First construction of both single family units and multi-family units has plummeted since the peak of the bubble in 2006. Second, the monthly data on starts are far more volatile for multi-family units than single family units. Third, in the last year, starts of multi-family units have recovered much more than starts for single family units, which are still near their 2009 trough.

Combining points 2 and point 3, we can conclude that the monthly number on starts will be far more volatile now that multifamily units account for around one-third of all starts as compared to the good old days when they accounted for less than one-fifth. So this is why I am not celebrating just yet (except for Hanukkah).