Skip to main content

Identifying Subsidized Housing Units Within the American Community Survey Through Administrative Record Linkage: A Technical Report


Report Acceptance Date: August 2019 (25 pages)

Posted Date: August 24, 2020

In this report, we describe the development of a multifaceted record linkage approach to identify federally subsidized rental housing units in the U.S. Census Bureau’s American Community Survey (ACS). Administrative data from the U.S. Department of Housing and Urban Development’s (HUD’s) rental housing assistance programs are linked to ACS housing units using a combination of three linkage algorithms. The first linkage algorithm is traditional address matching and is used for HUD rental assistance administrative records with “clean” addresses. The second linkage algorithm, which we call “imputed address” matching, is used when HUD’s rental assistance administrative records have incomplete addresses. This algorithm relies on the Census Bureau’s Master Address File (MAF) as a source for imputing missing apartment numbers in HUD records. The third linkage algorithm is a person-based linkage that relies on the Census Bureau’s Numerical Identification (Numident) file and the Census Bureau’s Person Identification Validation System (PVS), which uses a compilation of administrative records from various federal sources. When we used our multifaceted record linkage approach, our annual linked data sets resulted in an estimated total number of HUD-assisted rental units between 98 and 100 percent of the known number receiving HUD rental assistance.

Publication Categories: Affordable Housing     Housing Market Characteristics     Housing Finance     Other    


All Publications
Search for Publications
Search for Ongoing Research