Social security personal-account participation with government matching

Gary V. Engelhardt, Anil Kumar

Research output: Contribution to journalArticlepeer-review

2 Scopus citations


This paper examines the potential impact of government matching contributions on personal-account participation in the President’s Commission on Strengthening Social Security's Model 3 for Social Security reform. Given the government’s choice of four plan-design parameters, the magnitude of the match is determined solely by the differential return personal-account assets receive above the notional return, referred to as the ‘personal-account premium’, akin to the equity premium. The impact of matching on personal-account participation is simulated for older workers (ages 40 to 65) in the first wave of the Health and Retirement Study (HRS) using empirical estimates from a structural model of the impact of employer matching on participation in corporate 401(k) plans. For a personal-account premium of three percentage points, which implies a match rate of 7.5 % for middle- to lower-income workers, the simulations imply that 72 % of mid-career and older workers would participate in voluntary personal accounts. The response of participation to matching is very inelastic; it seems not unlikely that participation by mid-career and older workers would achieve the mid-range assumption by the Commission of 67 %. There is substantial heterogeneity in participation across subsets of older workers: participation would be the lowest for low-educated, minority, female, and unmarried mid-career and older workers.

Original languageEnglish (US)
Pages (from-to)155-179
Number of pages25
JournalJournal of Pension Economics and Finance
Issue number2
StatePublished - 2005

ASJC Scopus subject areas

  • Finance
  • Economics and Econometrics
  • Organizational Behavior and Human Resource Management


Dive into the research topics of 'Social security personal-account participation with government matching'. Together they form a unique fingerprint.

Cite this