TY  - JOUR
AU  - Finlay, Keith
TI  - Effect of Employer Access to Criminal History Data on the Labor Market Outcomes of Ex-Offenders and Non-Offenders
JF  - National Bureau of Economic Research Working Paper Series
VL  - No. 13935
PY  - 2008
Y2  - April 2008
DO  - 10.3386/w13935
UR  - http://www.nber.org/papers/w13935
L1  - http://www.nber.org/papers/w13935.pdf
N1  - Author contact info:
Keith Finlay
U.S. Census Bureau
4600 Silver Hill Road
Washington, DC 20233
Tel: 301-763-6056
E-Mail: kfinlay@gmail.com
M1  - published as Keith Finlay. "Effect of Employer Access to Criminal History Data on the Labor Market Outcomes of Ex-Offenders and Non-Offenders," in David H. Autor, editor, "Studies of Labor Market Intermediation " University of Chicago Press (2009)

M3  - presented at "Labor Market Intermediation Conf.", May 17-18, 2007
AB  - Since 1997, states have begun to make criminal history records publicly available over the Internet. This paper exploits this previously unexamined variation to identify the effect of expanded employer access to criminal history data on the labor market outcomes of ex-offenders and non-offenders. Employers express a strong aversion to hiring ex-offenders, but there is likely asymmetric information about criminal records. Wider availability of criminal history records should adversely affect the labor market outcomes of ex-offenders. A model of statistical discrimination also predicts that non-offenders from groups with high rates of criminal offense should have improved labor market outcomes when criminal history records become more accessible. This paper tests these hypotheses with criminal and labor market histories from the 1997 cohort of the National Longitudinal Survey of Youth. I find evidence that labor market outcomes are worse for ex-offenders once state criminal history records become available over the Internet. Non-offenders from highly offending groups do not appear, however, to have significantly better labor market outcomes. The sign of the non-offenders estimates are consistent with the predictions of the statistical discrimination model, but the estimates are not significantly different from zero. These estimates may be confounded by a short sample period and ongoing human capital investments, but the research design provides a unique setting for testing theories of statistical discrimination.
ER  -