Affiliations: [a] Population Research Center, University of Texas at
Austin, Austin, TX 78712, USA
Note: [1] We are grateful to the William and Flora Hewlett Foundation, the Andrew W.
Mellon Foundation, and the Tomas Rivera Center for providing support for
this research. The comments of Michael Fix, Jeffrey Passel and Guillermina
Jasso on portions of the research are gratefully acknowledged, as is the
helpful assistance of Molly Martin, K.T. Park, and Karen Wilkinson.
Abstract: Policy relevant social and economic research that involves comparing rates of
public assistance receipt among various groups must rely extensively on the U.S.
Census Public Use Microdata Samples (PUMS) for reasons of sample size. For
instance, the PUMS is the only data source that contains a sufficient number of
cases to examine patterns of welfare receipt among immigrant groups
disaggregated by national origin. The usage of census data, however, suffers two
drawbacks. One is that the PUMS does not ascertain the program (AFDC, SSI,
General Assistance) from which recipients report having received payments. The
other is that little is known about the accuracy of census measures of public
assistance receipt. This paper develops and presents a procedure to estimate
AFDC and SSI receipt based on the correspondence between the recipient's
characteristics and the eligibility criteria for these programs. The research
also assesses the accuracy of estimates of the number of persons receiving any
kind of cash public assistance (the presently available measure), as well as the
number receiving AFDC and SSI (the new measures presented here). The results
indicate that census data provide very satisfactory assessments of levels of SSI
usage and assessments of levels of AFDC usage that are adequate for some
analytic purposes.
