Wednesday, May 7, 2008

Who Cares About the Employment Birth/Death Model?

The debate over the reliability of BLS employment figures, and specifically the impact of the birth/death model, rages on. The latest posts I've seen (both very helpful) are from A Dash of Insight and Across the Curve; older posts from The Big Picture and Econbrowser demonstrate the debate has not moved forward much. My belief is the concern is overstated if you look at the employment figures in the right way (i.e., year over year change) over time.

My point is illustrated by comparing the trend in Sacramento:


with what's going on in Seattle:



Clearly, something is deeply amiss in Sacramento, while Seattle is holding it's own. If the birth/death model is understating job losses (and that's not a given, there are smart people on both sides of that issue), the Sacramento curve should be even steeper and Seattle might have started trending down. But if you're trying to use the published numbers to make investment decisions, this way of looking at the data gives you a pretty clear idea of what's going on.

This example also demonstrates the interesting stories happening in individual markets that get averaged out when you roll up to the national figures.