Disability Analysis File (DAF) Annual Public Use File (PUF)
Intro and Overview
The Disability Analysis File (DAF) Annual Public Use File (PUF) is a 10 percent sample of the population of Supplemental Security Income (SSI) or Social Security Disability Income (SSDI) beneficiaries from 1994-2015, and child SSI beneficiaries from 2005 forward. Demographic data include dates of birth and death, gender, full retirement age, monthly benefits, participation in Ticket to Work, and earnings information. Notably, persons do not appear in the dataset unless they receive benefits. Unsuccessful applications for benefits are not included.
The data come in two types of files:
- Demographic
- Annuals
The Demographic file is beneficiary level and includes one-time information like date of birth and information from the time of application for benefits. Researchers can use this file to draw a subsample of records that meet study inclusion criteria before working with the Annuals. Currently, there are over 3 million observations in the Demographic file.
The Annuals consist of one file per year and contain monthly data for each beneficiary, including information about benefits received and due, and participation in Ticket to Work. These files can be linked at the person level across years. Not all beneficiaries received benefits in all years, so there are more records in the Demographic file than in any Annual file.
Tips for Getting Started
SSA has helpful documentation on the webpage Links to an external site. and bulk download files as SAS, Stata and csv.
Mathematic researchers have provided a helpful methods report Links to an external site. for new users.
This short flier explains the highlights of the DAF. Links to an external site.
Key Advantages
- Administrative information about benefit receipt will be more accurate than reported benefits in survey data.
- Longitudinal linkage across records is feasible.
Key Disadvantages
- Most recent DAF PUF ends in 2015.
- No state or more detailed geo identifiers are available on the PUF and other masking to protect confidentiality may affect analysis (e.g. diagnosis codes are collapsed to 8 categories, dollar values are topcoded, and only birth month and year are observable, not day). Read the user guides carefully.
Papers Using these Data
O’Leary, Paul, Elisa Walker, and Emily Roessel. "Social Security Disability Insurance at Age 60: Does It Still Reflect Congress’ Original Intent?." Issue paper 2015-01 (2015).
https://www.ssa.gov/policy/docs/issuepapers/ip2015-01.html Links to an external site.
Bound, J., Burkhauser, R. and Nichols, A. "TRACKING THE HOUSEHOLD INCOME OF SSDI AND SSI APPLICANTS", Polachek, S. (Ed.) Worker Well-Being and Public Policy (Research in Labor Economics, Vol. 22), Emerald Group Publishing Limited, Bingley, pp. 113-158 (2003).
Benıtez-Silva, Hugo, Moshe Buchinsky, and John Rust. "Induced entry effects of a $1 for $2 offset in SSDI benefits." manuscript, SUNY-Stony Brook, UCLA, and University of Maryland (2006).
https://www.sunysb.edu/commcms/economics/research/papers/2005/jpam05.pdf Links to an external site.
or