Punishment and participation

This dissertation explores the expansion of the criminal justice system and its effects on behavior and representation at the individual, community, state, and national levels, arguing that convictions influence how citizens participate in politics. The direct and indirect effects of convictions on individual political behavior suppress voting both as a consequence of disenfranchisement and as a result of the material deprivation, social isolation, and mental distress that such convictions cause. Evidence supporting this point is provided using an original data set that contains information on the demographic background and voting behavior of millions of offenders. The results show that partly as a consequence of the conviction, only 10 percent of current offenders and 20 percent of ex-offenders voted in the 2000 and 2004 presidential elections. Although voter turnout is suppressed entirely for offenders subjected to incarceration or disenfranchisement, a conviction decreases turnout even among offenders who do not experience these policies. Because of this low turnout and the candidate preferences of offenders, convictions have not affected past election outcomes, including the 2000 presidential election. However, examining the implications of the growth of the criminal justice system for the representation of local communities elucidates the true effects of convictions on politics. Because convictions are concentrated among particular racial groups and geographic areas, the increasing incidence of convictions influences the quality of representation experienced by these groups or localities as a whole. Having a neighborhood resident sent to prison decreases voter registration in that neighborhood an average of three percentage points.