Issues with 2008 ACS employment variable ESR

USA
1

Hi there, first-time poster. I am having issues recreating the FRED all employees, total nonfarm graph (variable name PAYEMS) for years 2005-2019 using the ACS 1-year estimates. I am getting an overinflated employment number for the year 2008 as shown by the graph below:

image
image1217×1017 45.3 KB

I read that the employment questions on the ACS changed in 2008, and that comparing 2008 ACS estimates for employment is not advised.

census.gov/programs-surveys/acs/guidance/comparing-acs-data/2008

census.gov/library/working-papers/2007/acs/2007_Holder_01

How would I recover the correct employment number for 2008? If this is not possible, how should I treat 2008 going forward?

2

Bumping this post

3

It may not be possible to create a consistent graph of unemployment estimates using the ACS 1-year files for the time period you are interested in (2005 - 2019). As you noted, there are comparability issues between 2007-earlier and 2008-later employment estimates, due to changes to the survey questionnaire implemented by the Census Bureau in 2008 (this is noted on the comparability tab for EMPSTAT (USA)). You might consider using employment information from the Current Population Survey (IPUMS CPS) as an alternative or to adjust pre-2008 estimates; there are no comparability issues for EMPSTAT (CPS) from 2005 to 2019.