Crosswalks

Crosswalks between PUMAs and CZs

Many researchers are using Census files from IPUMS (Ruggles et al., 2017) to study U.S. local markets defined at the Commuting Zone (CZ) level. In those files, individual respondents' exact locations are suppressed to protect their anonymity. Instead, locations are identified by the group of counties in which respondents reside. In the 1980 sample these are called county-groups. Starting with the 1990 samples they are called PUMAs. County-groups/PUMAs may be either wholly contained within a CZ or overlap CZ boundaries. This page includes crosswalks between PUMAs and CZs. In cases where the county-group/PUMA crosses CZ boundaries, the crosswalks assign respondents to CZs in proportion to the share of the county-group/PUMA population that falls within each CZ. I used county-level population counts from the Census Bureau to generate the shares.

There are four crosswalks (one each for the 1980, 1990, 2000, and 2010 versions of IPUMS county groups). The readme file has instructions at the bottom for merging with an IPUMS data set. That example allocates PUMA-specific counts of respondents to CZs. Calculating CZ means would involve a tweak to the merging code.

Here is a zipped folder that includes the following files:

puma_cz_cross_1980.dta: 1980 Census 5 percent sample

puma_cz_cross_1990.dta: 1990 Census 5 percent sample

puma_cz_cross_2000.dta: 2000 Census 5 percent sample and the ACS/PRCS samples from 2005 to 2011

puma_cz_cross_2010.dta: 2010 Census and 2012-onward ACS/PRCS

readme_forweb.txt: includes instructions for merging with an IPUMS data set

Cited:

Ruggles, S., K. Genadek, R. Goeken, J. Grover, and M. Sobek (2017). Integrated Public Use Mi- crodata Series: Version 7.0. https://usa.ipums.org/usa/index.shtml. Accessed: 18AUG2017.