Payroll Employment and Household-Survey Unemployment
The household-survey unemployment rate tends to fall when the monthly establishment-survey payroll employment number is more than... it's now 130,000 jobs gained a month, and tends to rise when the monthly establishment-survey payroll employment number is less than 130,000 jobs gained a month.
Figure updated to cover 1981:1 to 2006:7. Datafiles courtesy of http://www.economagic.com
(Since 2000:1, the estimated intercept has risen to 166,000--but is very imprecisely estimated.)
Wouldn’t it make more sense to use the proportional growth rate of employment (or alternatively, use the change in the number of people unemployed on the y-axis)? As it is, you would expect the intercept to increase over time, since the number of workers represented by a given change in the unemployment rate is increasing.
Posted by: knzn | August 26, 2006 at 08:10 PM
The household survey also tends to lead the payroll data -- especially at bottoms.
I suspect this is because smaller firms react quicker and this show up in the household survery.
Posted by: spencer | August 27, 2006 at 11:23 AM
Actually, spencer, firing people is one area that the large businesses are very fast and "efficient" at.
Posted by: Jim S | August 27, 2006 at 01:19 PM