This morning (9/27) we're presenting our "Map of the Month" to our policy
board (Metropolitan Transportation Commission, MPO for the San Francisco Bay Area). This
month's map of the month is "Change in Total Resident Workers, 2000 to 2005"
based on data from Census 2000 and the 2005 American Community Survey. Data is mapped for
the 54 PUMAs in our region, and the PUMAs are "clipped" to the urban/developed
"footprint" to focus attention on the urban areas of the Bay Area.
This is actually a very interesting map, showing increases in resident workers at the
"edge" of our region (eastern Alameda & Contra Costa Counties; Vacaville in
Solano County; Napa County), and significant decreases in the "information industry
corridors" (northern San Mateo County; northwestern San Francisco - Richmond
District); and "Silicon Valley." Very interesting patterns are emerging. The
comparisons also show some unexpected increases in labor force: the south of Market /
Potrero Hill area of the City of San Francisco; and our East Oakland PUMA.
The problem is that the only year 2000 data we have, at the PUMA level, is from the
decennial census, which significantly undercounts our labor force (compared to ACS, and
compared to Bureau of Labor Statistics). So, our map, and this comparison, actually
downplays the recession in our region. As I told our Executive Director, it's actually
worse than it appears.
In order to compare apples-to-apples-to-apples, we're examining the "civilian
employed residents" from three data sources: Census 2000, ACS 2000 to 2005; and
BLS's LAUS for 2000 to 2005; for our region; for California; and for the United
States. Here's what we found for our region:
Comparing Census 2000 to ACS 2005, our region lost 3.3 percent of our civilian employed
residents, decreasing from 3.366 million to 3.254 million resident workers.
Comparing ACS 2000 to ACS 2005, our region lost 5.7 percent of our civilian employed
residents, decreasing from 3.451 million to 3.254 million resident workers.
Comparing the BLS 2000 to BLS 2005, our region lost 7.0 percent of our civilian employed
residents: 3.614 decreasing to 3.360. This is from the BLS's "Local Area
Unemployment Statistics" (LAUS) program, and includes the non-institutional group
quarters civilian employed. (ACS is strictly household workers, and we removed the GQ
workers from Census 2000 to be comparable.)
The issue -- at least in our region * is that the ACS data is closer to the independent
estimate of employed labor force (Bureau of Labor Statistic's LAUS) than the decennial
census. The concern with the decennial census is that we're getting very precise, yet
inaccurate (biased) results. On the other hand, the decennial census is giving us fairly
credible estimates of transit commuters, but we think it's undercounting the other
non-transit commuters by a non-trivial amount.
An interesting (annoying) discovery is that sometimes the data is suppressed for
place-level AND PUMA-level geography, even for "collapsed" tables. For example,
we like to use the Table C08006 (instead of it's base table: B08006) to obtain workers
by means of transportation to work.So, there is no "collapsed" table data for
one of my PUMAs, and one of my places, for this table. (We can use other tables to get
"total workers" so at least our map-of-the-month has data for my too small
Right now we're busy assembling table shells with historical census and ACS data, and
getting ready to "drop in" the ACS 2005 data once it's released to the hoi
polloi next Tuesday. My guess is that the "area of work" tabulations, expected
next Tuesday, will exclude summary level data for PUMAs and Congressional Districts, and
will only include places and counties with residential populations of 65,000+. (Please
correct me if my guess is wrong. We would love to have data by PUMA-of-work, but I
don't think that's coming this year....)
Magic Number = 0.
Chuck Purvis, MTC
>> "Murakami, Elaine"
<Elaine.Murakami(a)fhwa.dot.gov> 09/25/06 2:09 PM >>>
With the release of 2005 ACS, the Census Bureau has included PUMAs as
a tabulation geography. You can use American FactFinder to access the
PUMA tabulations. http://factfinder.census.gov
<http://factfinder.census.gov/> The population threshold for
tabulation from 1-year data accumulation from ACS is 65,000, so a PUMA,
with a population threshold of 100,000 meets this requirement. This is
mostly a benefit to very large counties (to be able to get sub-county
tabulations), or areas with many small jurisdictions which fall under
the 65,000 population threshold.
Below is the link to the 5% PUMA maps.
Has anyone looked at the PUMA tabulations for their area yet? If so,
please share your results with this listserv. Thanks in advance!
Since PUMAs are being used as tabulation geography, there is more
incentive for transportation planners to work with their State Data
Centers when PUMAs are re-defined for the 2010 Census.
FHWA Office of Planning