Targeting Housing Choice Vouchers

In 1999, the Welfare to Work Voucher Program was introduced to give 50,000 additional housing vouchers through local Public Housing Authorities across the country. During the years 2000 and 2001, the Department of Housing and Urban Development ran a randomized controlled trial where applicants on waiting lists were randomly selected to immediately receive a voucher offer. This was a departure from the priority rules often used by Public Housing Authorities to allocate voucher offers. Ruby (2025) evaluates the welfare implications of three priority rules – random assignment, vouchers offered to households with positive wage earnings, vouchers offered to households with zero wage earnings.

The paper uses data from the Welfare to Work Voucher Experiment which includes a baseline survey of demographic information and baseline rent conducted prior to randomization, as well as linked administrative databases which document quarterly labor earnings and welfare income. However, the paper does not observe rents at any point after randomization. For non-voucher recipients, rent is imputed from census-tract level rent data for households with income below $50,000 using the U.S. Census Bureau, 2000 Census of Population and Housing, Summary File 3. For voucher recipients, rent is imputed from census-tract level household and federal government spending on housing from the Picture of Subsidized Housing 2000.

The paper finds that the MVPF does not vary whether one distributes housing vouchers equally across eligible participants, targets those with no wage earnings, or targets those with positive wage earnings. Below is a discussion of the estimated MVPF of offering an additional housing voucher with equal probability to a family with zero or positive wage earnings.