Skip to main content

The effect of the SSI program on labor supply: Improved evidence from Social Security administrative files (Neumark & Powers 2005)

Absence of conflict of interest.

Citation

Neumark, D., & Powers, E. T. (2005). The effect of the SSI program on labor supply: Improved evidence from Social Security administrative files. Social Security Bulletin, 65(3), 45-60.

Highlights

  • The study examined the relationship between generous state Supplemental Security Income (SSI) benefits and the employment of older worker nearing SSI eligibility age.
  • The authors used a regression model and data from the Survey of Income and Program Participants (SIPP) and Social Security Administration (SSA) to estimate the effect.
  • The study found that older males who were likely SSI participants in generous states worked significantly fewer hours and had lower employment rates compared with those living in less generous states. The authors, using the SSA administrative records, also found that likely older participants (ages 60 to 64) worked significantly fewer hours than younger ones.
  • The quality of causal evidence presented in this report is low because the authors did not ensure that the groups being compared were similar before the intervention. This means we are not confident that the estimated effects are attributable to generosity of state SSI benefits; other factors are likely to have contributed.

Intervention Examined

State Supplemental Security Income (SSI) benefits

Features of the Study

SSI provides income to low-income elderly and disabled people or couples, with benefit levels varying by state and by income level. This study focused on SSI for elderly people. States can supplement the federal SSI benefits with additional state-level benefits. In some cases, these state-level benefits can be generous. For this study, the authors referred to benefits 20 percent or more than the federal maximum benefit for individuals or couples.

In this study, the authors used data from the SIPP from 1984 to 1993 and Social Security Administration data from 1974 to 1999. They included about 18,000 observations of male heads of household ages 40 to 64. The people in the study were not yet eligible for SSI given their age, but the authors estimated their likelihood of SSI participation using probabilities drawn from an analysis of SSI recipients older than age 65 in the SIPP data. The authors considered those with a probability of participating in SSI at or above the 90th percentile as likely participants.

Drawing on the variation in benefits across states, the authors used a statistical model (difference-in-difference-in-differences) to compare employment outcomes between older and younger likely participants associated with changes in SSI benefits. The authors estimated the model separately using SIPP and SSA data.

Findings

Employment

  • The study found that likely SSI participants ages 60 to 64 who lived in generous states were less likely to be employed compared with those living in states with no state SSI benefits, depending on the data source used for the analysis. These findings were statistically significant.
  • Using administrative records, the study found that likely SSI participants ages 60 to 64 who lived in generous states worked significantly fewer hours than those living in states with no state SSI benefits.
  • The study also found that likely participants ages 60 to 64 worked 10 fewer hours than likely participants ages 40 to 59, which was a significant difference.

Considerations for Interpreting the Findings

Although the authors’ analysis accounts for race, age, and gender, the authors did not establish that there were similar trends in employment before the introduction or eligibility of state SSI benefits or account for these potential existing employment differences between the groups. These existing differences between the groups—and not the generosity of state SSI benefits— could explain the observed differences in outcomes.

Causal Evidence Rating

The quality of causal evidence presented in this report is low because the authors did not ensure that the groups being compared were similar before the intervention. This means we are not confident that the estimated effects are attributable to state SSI benefits; other factors are likely to have contributed.

Reviewed by CLEAR

January 2020

Topic Area