Using Wasted Votes to Evaluate Legislative Redistricting Plans: The Inefficiency Gap

Abstract

Courts in the United States have expressed a need for a manageable standard for evaluating redistricting plans speaking to concepts of equal protection and one-person/one-vote. This paper builds upon the concept of “wasted votes” proposed by Stephanopolous and McGhee (2015). What share of wasted votes cast due to gerrymandering, dilutes a group’s voting power sufficiently to shift legislative seats from that group to the group whose members predominantly drew the redistricting plan? Courts can conditionally reject a redistricting plan based on the existence of an alternative plan more responsive to voter preferences and having sufficiently equal population and manageability based on geographic and political boundaries. Although cited recently by expert witnesses and justices, the “efficiency gap (EG)” and the “mean-median” gap, as previously constituted, are inconsistent measures upon which to base a decision to accept or reject redistricting plans. Three corrections to the EG measure results in a reliable inefficiency gap (IG) measure. Using the IG measure, a threshold may be chosen to indicate when redistricting plans shift one or more seats between voter groups due to diluting group members’ voting power via partisan gerrymandering. However, one correction to the EG measure proposed herein – weighting the wasted vote shares by each voter groups’ lost seat share – improves the EG’s reliability in most likely scenarios. This paper tests several standards against hypothetical reapportionment plans having varying amounts of vote shares, legislative seats, and “stacking” and “cracking” or no gerrymander bias. The results show the proposed “inefficiency gap (IG)” measure, by adopting three mathematical improvements to the EG measure, reliably meets the legal needs of state and federal courts for a manageable standard.