After Governor Wolf rejected the Scarnati-Turzai Congressional map last week, the Pennsylvania Supreme Court promised to unveil their new remedial Congressional District Map by February 19th.Figure 1
The Scarnati-Turzai submission explicitly aimed to keep constituents in their current districts as much as possible – with a stated purpose of avoiding confusion, but with a real impact of preserving much of the partisan advantage from the 2011 maps. This new map from the Supreme Court does not follow that same strategy. This map is wildly different from the one we have had since 2011, complicating the comparison of new districts to old. The composition of the newly unveiled districts, relative to their districts of origin, is presented in Figure 2.
On this chart, the new districts from the remedial plan are listed vertically down the left side, and the district of origin for their new constituents is represented by the size of each colored bar. For example, over 90% of the population of the new PA-01 were constituents of the old 8th district. The new PA-09, on the other hand, was created from nearly equal portions of the old 6th, 11, 15th, and 17th districts.
For the comparisons that follow, each new district was paired initially with the old district that contributed the largest share of constituents. However, that resulted in the old 13th and 17th districts being used twice each. To maintain a comparison where each new district is compared to a distinct previous district, the new PA-09 was paired with the old 11th district, and the new PA-02 was paired with the old 1st district, even though those were only the second largest source of constituents for each.
Figure 3 shows the result of these comparisons, in terms of average Democratic candidate margin based on statewide candidates from 2014 and 2016.
The performance of the old district is represented in each case by a hollow outlined bar, and the performance of the new districts from the Court’s map are presented as filled bars. For each pair of districts, the label on the axis shows not only the identity of each district, but also the percent of the constituents within the new district that resided in the old district to which the comparison is being made.
Overall, on performance this is a map that much more closely aligns with the electorate in the Commonwealth. There are eight districts that on average perform Republican for statewide candidates, and ten district that on average perform Democratic, including four where the average margin for Democrats is under four points, making them clearly battleground districts. Under the old map, there were only five districts that on average performed Democratic. This in spite of the fact that four of the six statewide Democratic candidates whose election results go into this performance calculation won their contests.
There is still a lot to unpack about the remaining implications of this map, as it relates to open seats, primary candidates being cast into different districts, and – notably, the 18th Congressional District which will once again be an open seat, regardless of the results of the March 13th Special Election, as Rick Saccone will have to run in a heavily Democratic district against Rep. Doyle, and Conor Lamb will have to take on a well-funded incumbent lawmaker in Rep. Rothfus in the newly drawn PA-17.
The following 18 charts show individual candidate performance in each district, represented by Democratic margin of victory, for statewide candidates between 2010 and 2016. The grey bars represent the districts from the 2011 Congressional map that the new districts are being compared to, and once again, in the legend, the percent of shared constituents is noted.
The Supreme Court released a report on municipal and precinct splits in their map, that included the population allotted to each district for the 18 municipalities that spanned multiple districts. As a shortcut, to account for these splits, I took the municipal-level election result history and divided the vote totals proportionately between the districts based on the share of the municipal population in each district. This could introduce a small degree of error, if the population being split within the municipality is not uniform in their voting habits, but this would amount to no more than a rounding error. Using this shortcut also accounted for split precincts, which otherwise would have entered a different small source of error into the calculations.
Mike Johnson is Data Director for the Pennsylvania AFL-CIO. The analysis reflects his individual views. Twitter: @MikeJohnsonPA