by Calculated Risk on 1/05/2007 01:12:00 PM

With regards to the differences between the BLS and ADP job reports, Professor Hamilton notes the

"reaction of many people to such disparate indicators is simply to disregard several of the measures."

That describes me; I've just ignored the ADP report.

Dr. Hamilton suggests blending the three measures of employment (BLS Payroll, BLS Survey, ADP):

"I've been using a weighted average based on the assumption that the NFP is considerably more reliable than either of the other two, giving NFP a weight of 80% and each of the others 10%."

That made me wonder if there is a lag, or some other consistent relationship, between the ADP and BLS reports of employment. I found this graph from Northern Trust's Asha Bangalore:

Click on graph for larger image.

Note that this graph does not include today' BLS numbers. There appears to be reasonably strong correlation over time, but the monthly differences between the reports can be significant.

It appears that the two series don't diverge for long, so I'd expect better correlation next month (something to remember). This also suggests Dr. Hamilton's blending approach might be helpful - although I'd like to see a graph with the survey data added.