Over the summer, on remand from the Supreme Court, a district court in Virginia struck down eleven House of Delegates districts on racial gerrymandering grounds. The court also gave the Virginia General Assembly until October 30 to pass a remedial map. Over the last few weeks, the House Democrats offered a remedy (HB 7001), which was rejected in committee. The House Republicans also put forward a pair of proposals (HB 7002 and HB 7003), one of which (HB 7003) was passed by the Committee on Privileges and Elections.
To allow these and other Virginia House plans to be evaluated, PlanScore recently added Virginia to the list of states for which district maps can be uploaded and scored. Here’s a quick summary of both the existing plan (the one partly invalidated by the court) and the Democratic and Republican proposals.
First, the existing plan is significantly skewed in a Republican direction. Using a model based on the 2016 election, it has a pro-Republican efficiency gap of 6.5%, a pro-Republican partisan bias of 3.7%, and a pro-Republican mean-median difference of 3.0%.
Second, the Democratic proposal is highly symmetric. Using the same model, it has a pro-Republican efficiency gap of 1.8%, a pro-Republican partisan bias of 0.2%, and a pro-Republican mean-median difference of 0.4%.
And third, both Republican proposals are almost exactly as asymmetric as the existing plan. The map passed by the committee, for example, has a pro-Republican efficiency gap of 6.6%, a pro-Republican partisan bias of 3.6%, and a pro-Republican mean-median difference of 3.3%.
Of course, the partisan implications of these proposals are not directly relevant to whether they cure the racial gerrymandering violations found by the court. These implications are of great political interest, though, and they highlight one of the important truths of redistricting: that, frequently, plans with very different partisan effects are quite similar in their nonpartisan characteristics. Maps’ nonpartisan features (like their avoidance of racial gerrymandering) are thus often a poor guide to their partisan performance.