Dear Norm: 
 
You should probably take a look at the Souleyrette et al article in TRR 1768, 2001.
"Applications of State Employment Data to Transportation Planning"
 
Also, Wende Mix and Phil Fulton prepared a poster at the recent TRB Census Conference (May 2005) where both ES-202 and Census data were used.  She has a full paper available and I believe she has submitted it for review to TRB for Jan 2006.   
 
The biggest problems have been:
 
ES-202 files do not require multiple site businesses to report specific workplace locations.  Some states are better than others.  For example, MN is considered excellent (LEHD project by Census Bureau).  One estimate is that 40% of employment is by businesses with multiple sites. 
 
ES-202 file may lack a physical workplace location (e.g. PO Box, or location of office responsible for employment records)
 
"Personnel Supply"  companies have become increasingly larger shares of total employment, and employees of these companies will not have their specific work location captured, only the office of where they were hired. 
 
On the Census side:
 
Approximately 25 percent of workplace locations were imputed. 
The Census Long Form went to 1:6 households nationwide, so it is a sample survey, not a universe.
 
 
Elaine Murakami
FHWA Office of Planning
206-220-4460
 
-----Original Message-----
From: ctpp-news-bounces@chrispy.net [mailto:ctpp-news-bounces@chrispy.net]On Behalf Of Norm Marshall
Sent: Tuesday, August 30, 2005 11:41 AM
To: ctpp-news@chrispy.net
Subject: [CTPP] county-to-county flows vs. ES202, BEA county control totals

I am comparing CTPP 2000 county-to-county flows vs. travel demand model for a mid-sized region. There are a number of complicating issues, including:

 

Census work trip may not be made every day

Census work tour may be coded as other than home-based-work (HBW) trips in trip-based model

Census only records one commute for workers with more than one job

 

In trying to massage the data to be as consistent as possible, I also reviewed ES-202 and BEA county employment data. The BEA totals including proprietors are very high, but the BEA wage and salary totals are reasonably consistent with ES-202 data, i.e. consistently higher as they include non-covered employment including government.

 

However, the CTPP flow data are not particularly consistent with ES-202 and BEA, i.e. some counties appear to attract more commuters than indicated by ES-202 and BEA, and other counties attract less than ES-202 and BEA (after adjusting so the totals are the same).

 

Have others observed similar discrepancies? Does anyone have advice for reconciling these datasets?

 

Norm Marshall