Chapter 6 - Welfare II

This chapter tests several prediction of the Gerrymandering model simulation with respect to the effect of redistricting institutions on four different measures of voter welfare. Empirical voter utility is estimated by generating ideology scores for respondents to the 2008 CCES survey on a common scale with members of the 110 Congress, and then rescaling these estimates to match DW-NOMINATE. The chapter finds that the predictions of the model with respect to personal and policy median ideology are largely affirmed: nonpartisan gerrymanders perform very well with respect to policy median representation, but poorly with respect to personal representation. Partisan maps perform better with respect to policy median representation when tides are greater and the gerrymander is more aggressive. Predictions of the model with respect to compositional and discursive representation are not consistently supported. The chapter also finds preliminary evidence that increased polarization hurts policy median and personal representation, but that this effect is reduced under strong tides.