No, Partisan Gerrymandering Did Not Cost Democrats Seats in the House of Representatives: A State-by-State Analysis

Many commentators have claimed that the current Republican majority in the U.S. House of Representatives, or much of it, is the result of gerrymandering.  Recently, for example, the Associated Press prepared an “analysis” that purported to show that Republicans won 22 additional House seats due to gerrymandering. The key piece of evidence cited for this claim is that Democrats won have won a higher share of the popular vote than seats in the House. For example, in 2016, the Republicans won 49.1% of House votes (to 48% for Democrats), but won 55.4% of House seats (to 45.6% for Democrats). Because Republicans controlled redistricting in more states than Democrats (the argument goes), the redistricting process is responsible for most of the Republican’s majority in Congress.

The competing argument is that “clustering” is responsible for these anomalies. That is, Democrats tend to be clustered in urban areas that typically vote over 70% for Democrats in contested elections.  Republicans, by contrast, tend to live in suburban and rural/small city areas that favor Republican candidates reliably, but not as overwhelmingly as urban areas favor Democrats.  Thus, under single-member district maps that are drawn to maximize compactness (i.e., that keep urban areas within a single district and do not extend from urban areas out into the countryside), Republicans will have an advantage in U.S. House elections.

In addition, Section 2 of the Voting Rights Act guarantees that minorities will have the opportunity to elect candidates of their choice, where minority voters are sufficiently large and geographically compact to constitute a majority in a single-member district.  Under U.S. Supreme Court precedent, this will typically mean that a state must draw districts according to the “minority-majority” concept—a single minority group that engages in bloc voting must account for over 50% of the voting age population within a given single-member district. Because minorities tend to vote overwhelmingly for Democrats, maps that comply with Section 2 of the Voting Rights Act will group minorities together, typically making surrounding districts safe for a Republican candidate.

I have decided to test these competing hypotheses by going state-by-state and drawing districts that comply with traditional redistricting criteria, and judging whether the changes would have made any difference in the outcome in House elections.  I have found that the net changes, when adjusting for departures from neutral criteria, would have elected more Republicans. That is, gerrymandering probably cost Republicans 5 or so seats in the House.

How is this so? Democrats had two of the most successful gerrymanders of this cycle (Illinois and Maryland). Those two states netted the Democrats about eight seats, when compared to neutral districts. Further, supposedly non-partisan commissions in Arizona and California produced Democrat-friendly maps that shifted five seats to the Democrats. Meanwhile, in the two largest states were Republicans were in complete control of the process (Florida and Texas), Republicans faced hurdles in gerrymandering to the extent of their desires. Florida’s Fair Districts Amendment, enforced via state court review, meant that Democrats gained four seats from the Florida redistricting process, with Republicans losing two seats. In Texas, a Voting Rights Act lawsuit stopped Texas’s attempt to allocate all four of Texas’s additional seats to the Republicans, and the two additional seats were split 2-2. In Virginia, a federal court lawsuit also undid part of a Republican gerrymander, handing an additional seat to the Democrats. Republicans certainly did gerrymander in Wisconsin, Michigan, Ohio and Pennsylvania, but these states were trending Republican anyway (they were the key parts of the Trump coalition). Gerrymandering in these states probably netted Republicans two seats or so.  Add that to Republican gerrymanders in Texas, Georgia, Virginia, Utah and North Carolina (8 seats total shifting partisan control), and you have a net of 10 seats or so for Republicans from gerrymandering alone. That is only two seats more than Democrats secured via gerrymandering in Maryland and Illinois. Clean up jagged lines in Connecticut, Nevada, New Hampshire, New Jersey and New York, and you end up with a +5 Democratic advantage from redistricting.

My analysis confirms the “clustering” hypothesis. Democrats are concentrated disproportionately in urban areas. Any allocation of seats that respects “neutral redistricting criteria” (i.e., communities of interest, not breaking county and city boundaries, obeying the Voting Rights Act) was going to yield substantial Republican majorities in the House in the years 2012-2018.

Under the lines below, Republicans would have won 241 seats in 2012 (a gain of 9), 245 seats in 2014 (even), and 248 seats in 2016 (a gain of 7).

“Neutral Districting Criteria”

It is notoriously difficult to define what is “neutral redistricting criteria.”  It is not impossible, however.  In descending order of importance, I have attempted to: maintain equal population in each district; comply with the Voting Rights Act; group together “communities of interest”; respect county lines and other municipal boundaries; ensure that each district is compact and has regular-looking boundaries; and preserve the cores of existing districts.  With the exception of adherence to the equal population and compliance with the VRA, I have not mechanistically favored one criteria over the other.  Nor do I think it makes sense to do so: there are too many factors at play in the process.  Courts and commissions that have attempted to draw non-partisan maps have similarly refused to follow a strict order-of-operations.  I maintain, however, that all of the resulting districts could have been drawn by a neutral body (such as a court or an independent commission) that was forced to explain its reasoning for choosing various boundaries in non-political terms.

Alabama

My Alabama districts are below. I have mostly cleaned up the lines prepared by the Alabama legislature. Montgomery is no longer split between three districts, for example. It is theoretically possible to draw a second African American majority district, but this would require creating a splitting Mobile– that is, creating an additional gerrymander. The 7th district had to add about 70,000 people, and the most logical place to go, for community-of-interest purposes, was Montgomery. The result is a 6-1 Republican split, and no change.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 682838 63% 34.6% 61.3% 37.9% 68.5% 25.7% 2.5% 1.3% 1.1%
2 682636 64.9% 33.0% 62.9% 36.4% 66.5% 27.7% 3.0% 1.3% 0.5%
3 682997 65.8% 31.8% 62.8% 36.1% 72.2% 23.4% 2.2% 1.0% 0.3%
4 682756 81.4% 16.4% 75.8% 23.0% 87.1% 6.0% 4.7% 0.4% 0.8%
5 683019 64.7% 31.3% 63.9% 34.9% 76.1% 16.3% 3.9% 1.6% 0.7%
6 682492 67.5% 29.4% 71% 28% 77.2% 15.7% 4.3% 1.8% 0.3%
7 682998 30.9% 67.5% 29.4% 70.1% 38% 58.5% 2.1% 0.5% 0.2%

Arizona

Arizona gained a seat in redistricting, and the process was controlled by an independent commission.  But Arizona’s redistricting commission, in taking “competitiveness” into account in drawing districts, essentially produced a Democratic gerrymander, with its ostensibly non-partisan chairperson siding with commission Democrats in drawing the new maps.

Removing politics from the process would have resulted in a very different map.  For example, in the enacted map, Yavapai, Gila, and Yuma counties are unnecessarily split between districts; in my map, each county of those counties is kept intact.  In the enacted map, Tucson’s Pima County is split between three different districts, as is Pinal County; in my map, both counties are split between only two districts (and Pinal is split only so that an Indian reservations would not be split between districts).  The Commission’s Ninth District (drawn to be competitive between the parties) awkwardly combines Tempe, western Mesa and Chandler with precincts in Scottsdale and Phoenix via an arm-like appendage; in my map, the Maricopa County districts have much more regular shapes. This district has been held by Democrat Kyrsten Sinema since 2012.

In drawing my plan, I preserved the core of each of the eight previously-existing districts, which were drawn by the previous independent redistricting commission using good-government principles (the previous commission does not appear to have taken “competitiveness” into account to nearly the same extent that the 2012 commission did). In order to keep Pinal in one district, I split Flagstaff from the rest of Coconino County. I maintained two Hispanic-majority districts. I made the Seventh more compact by removing Yuma and extending it further into Phoenix. I also maintained a substantial Native American population in my First District.

The net result would be two additional districts for Republicans, for a split of 7-2. Adding the rest of Pinal County and removing Flagstaff from the First District means it would have been won by a Republican, if not in 2012 then in 2014. Keeping the Phoenix districts in their previous configuration means a Republican would have been elected in place of Kyrsten Sinema in 2012.

NET CHANGE: R +2 

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 710071 50.7% 43.6% 53.4% 44.9% 55.5% 2.6% 18.8% 1.5% 20.5%
2 710476 57.1% 38.0% 60.7% 37.9% 74.0% 3.8% 17.1% 3.0% 0.7%
3 710373 50.4% 44.4% 57.5% 40.8% 76.7% 2.6% 14.6% 3.5% 1.1%
4 710413 23.1% 71.3% 27.0% 71.2% 30.3% 9.0% 54.2% 2.8% 2.4%
5 710069 44.7% 49.3% 52.9% 45.7% 71.5% 3.6% 16.7% 4.5% 2.2%
6 710064 56.6% 37.5% 62.8% 35.6% 75.6% 3.0% 14.7% 4.6% 0.8%
7 709954 28.5% 66.4% 32.9% 65.4% 46.3% 4.4% 52.3% 2.5% 3.1%
8 709904 47.7% 46.6% 52.9% 45.4% 72.9% 3.2% 18.7% 2.9% 0.7%
9 710693 62.2% 33.0% 61.7% 36.5% 71.6% 1.1% 22.6% 1.2% 2.2%

Arkansas

I have simply cleaned up the Arkansas’s legislatures lines below. This does not yield a change in partisan makeup.

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 729463 65.0% 30.2% 61.0% 36.3% 80.4% 15.6% 2.2% 0.5% 0.4%
2 729192 52.4% 41.7% 54.7% 42.9% 70.3% 19.6% 4.0% 1.4% 0.4%
3 728805 64.2% 31.3% 65.5% 31.6% 82.1% 2.2% 9.9% 2.9% 1.3%
4 728458 63.7% 31.8% 61.3% 36.4% 74.4% 19.3% 4.1% 0.6% 0.6%

California

California did not gain or lose any of its seats in 2012, and, for the first time in its history, its lines were drawn by a non-partisan Citizens’ Redistricting Commission.  As documented by ProPublica, Democratic operatives were able to covertly influence the workings of the Citizens’ Commission by creating sham “citizen’s groups” to push for districts favored by Democrats.  Despite this, however, I do not think that the Commission’s plan unduly favored Democrats.  Although it did not gain or lose any seats, there were significant population shifts in California that required extensive redrawing of the districts. Moreover, Republicans have been losing favor in California for many years among voters in urban areas.  Coastal Northern California (including San Francisco and the Silicon Valley area) has not elected a Republican to Congress since the 1990s, for example. The Central Valley has trended Republican, but the Commission added a new Central Valley district that elected a Republican (David Valdao) in 2012, and preserved Republican incumbents in the other Central Valley districts.

In metro Los Angeles, where Republicans are not competitive in urban areas but retain some strength in the suburbs, the Commission eliminated one Republican seat (the Diamond Bar-based seat formerly held by Gary Miller) and severely weakened Republicans in two others (the Ventura County-based district of Elton Gallegly and the district held by David Dreier, which stretched from the San Gabriel Valley to Rancho Cucamonga).   But Gallegly had benefitted from egregious gerrymandering in the neighboring 23d, which had siphoned off the most Democratic parts of Ventura County. And the Commission sought, quite reasonably, to preserve significant Hispanic majorities in the six Los Angeles-area districts held by Hispanics  (most of which lost population relative to other parts of the state), while also creating new Hispanic-majority seats in Riverside County and the San Fernando Valley, and an Asian-plurality district in the San Gabriel Valley.  There was no way that the Commission could have accomplished all of these goals while preserving Miller and Dreier’s seats, both of which had substantial Hispanic and Asian populations.

It would have been possible, in theory, to draw compact districts that would have favored incumbent Republicans Dan Lungren (from the Sacramento area) and Brian Bilbray (from the San Diego area) to a greater extent than those drawn by the Commission (not so for fellow Republican Mary Bono Mack, whose Riverside County seat could not have been measurably strengthened for her without sacrificing its compactness).  But the changes would have been slight, and the fact is that Lungren and Bilbray, like Mack, lost close elections for reasons having nothing to do with redistricting: the Democrats recruited strong candidates; congressional Republicans were unpopular; and Obama thrashed Romney in California, creating a headwind at the top of the ticket.

NET CHANGE: NONE

Colorado

Colorado did not gain or lose any seats in 2012.  In Colorado, control of the legislature was split between the parties, and no agreement was reached on a map.  That meant the matter went to a federal district judge, who chose between a Republican-proposed map that made minimal changes in the districts and a Democratic map that made very significant changes to the districts with the goal of making the Sixth District, held by Republican Mike Coffman, more Democratic and hence more winnable.  Inexplicably, the judge chose the Democratic plan, based in part on the fact that it would make the Sixth District more competitive.  As in Arizona, then, consideration of “partisan balance” injected partisan politics into the redistricting process in Colorado.

The Democratic plan, however, would not have been produced by a commission or a court tasked with drawing the lines in the first instance on a non-political basis.  Colorado’s previous map was itself drawn by a federal court, and the resulting districts were compact, maintained the cores of existing districts, and preserved communities of interest. The previous map created an urban Denver district; two suburban Denver districts; a district combining Boulder with Denver suburbs and some rural mountain counties; a Colorado Springs district; a Western Slope district, and an Eastern Slope district.  Population changes during the 2000s did not require that the districts be significantly redrawn.

My map preserves the cores of the pre-2012 districts, while cleaning up some of the unnecessary county splits. So, Boulder County is entirely in the Second District; Weld County entirely in the Third District; and Adams County entirely in the Seventh District. The Seventh District loses its portion of Jefferson County and includes nearly all of Aurora, increasing the district’s Hispanic percentage significantly.

There would be no net partisan change, although the Sixth District would get much safer for Republican Mike Coffman.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 718493 21.2% 71.0% 26.8% 71.0% 60.0% 8.2% 26.0% 3.4% 0.6%
2 718638 32.9% 58.2% 37.5% 59.9% 82.5% 0.8% 11.9% 3.0% 0.4%
3 719022 52.0% 40.0% 51.8% 45.8% 75.0% 0.7% 21.1% 0.7% 1.3%
4 718545 50.4% 40.3% 51.5% 45.2% 79.6% 1.2% 15.7% 1.7% 0.5%
5 717872 57.2% 33.2% 59.1% 38.3% 77.0% 5.2% 12.0% 2.8% 0.7%
6 718360 50.3% 41.2% 54.5% 42.6% 86.1% 1.8% 6.7% 3.8% 0.3%
7 718266 36.4% 54.4% 38.3% 59.1% 56.8% 7.8% 28.7% 4.4% 0.6%

 

Connecticut

Connecticut did not gain or lose any seats in 2012.  In Connecticut, a bipartisan commission that included politcos from both parties was not able to reach an agreement on redrawing the lines.  With the commission deadlocked, the Connecticut Supreme Court appointed a Special Master to draw new boundaries.  But in a reversal of the outcome in Colorado, the Supreme Court required the Special Master to draw a least-change map, as had been favored by Democrats, rejecting Republican calls to make significant changes.

Once again, the process in Connecticut resulted in a map that never would have been drawn by a court or commission.  The reason is that Connecticut’s previous congressional map, which the new map preserved almost entirely, was a blatant gerrymander, drawn as part of a bipartisan compromise in 2001.  With the state losing a congressional district, the Commission drew a map that attempted to create a “fair fight” Fifth District between a Democratic and Republican incumbent, while protecting the remaining incumbents.  The map extended the Fifth District, based in Northwestern Connecticut, into New Britain, outside of Hartford, because the Republican incumbent lived in New Britain.  So that the Fifth District could take in a significant part of the Democratic incumbent’s hometown of Waterbury, several small towns in Litchfield County were added to the Hartford-based First District.  The result was to give both the First and Fifth grotesque shapes and to split communities of interest.

By 2010, all of Connecticut’s representatives were Democrats, and the inclusion of heavily Democratic New Britain in the Fifth District benefitted Democrats.  With the Fifth District coming open due to incumbent Democrat Chris Murphy running for the U.S. Senate, Commission Republicans pushed to smooth out the lines, while Democrats sought, successfully, to preserve the 2001 gerrymandering of the First and Fifth Districts.

The map I have drawn better maintains communities of interest, avoids splitting towns, and results in more normally-shaped districts.  It keeps all of Litchfield County intact within the Fifth District.  Unlike the court-approved map, it does not split Waterbury between districts, and it loses its arm into New Britain (I also swapped in Bristol and swapped out Meriden to make the boundaries more compact).  The First District is now entirely within Hartford County.

As a result of the changes, the Fifth District becomes 3-4% more Republican, and would have elected Republican Andrew Roraback over Democrat Elizabeth Esty (Esty won by 51%-49% in 2012, with her entire margin of victory coming from New Britain.) The other districts would not have changed partisan hands.

NET CHANGE: R +1

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 714925 35.3% 60.4% 34.6% 64.3% 65.4% 13.7% 14.9% 4.4% 0.1%
2 714504 45.8% 48.7% 42.6% 55.9% 85.8% 3.7% 5.7% 3.0% 0.4%
3 714745 39.4% 56.9% 35.3% 63.6% 72.1% 11.5% 11.1% 3.8% 0.2%
4 714857 34.6% 61.6% 42.0% 57.1% 63.3% 11.4% 18.4% 5.1% 0.1%
5 715066 49.8% 45.9% 49.3% 49.5% 84.2% 4.4% 8.0% 2.1% 0.1%

Florida

Florida passed the Fair Districts Amendment in 2010, which mandates that its congressional districts be compact, preserve minority voting strength, respect city and county boundaries where possible, and not favor or disfavor incumbents or political parties.  The Fair Districts Amendment severely limited what the Republicans could do to help themselves in redistricting Florida’s congressional seats (which had been subject to a Republican gerrymander in 2002). A subsequent ruling prior to the 2016 elections limited any gains further. It forced Republicans to draw the black-majority Fifth District as a Tallahassee-t0-Jacksonville district, as opposed to the previous Jacksonville-to-Orlando configuration. It also forced Republicans to redraw the St. Petersburg-based 13th District and Tampa-based 14th District. The changes resulted in the loss of three seats for the Republicans.

The closest race was the Seventh District. It became just a couple points more Democratic. But the changes led the Democrats to recruit a quality candidate and heavily target the race. Although the Seventh trended away from Donald Trump in 2016, it is questionable whether the Democrats would have gained this seat if it had the same boundaries since 2012. It would have been more difficult to recruit a candidate to face the entrenched incumbent Mica, and Mica would have been familiar to voters throughout the district. Mica won by enough in 2012 and 2014 to suggest that he would not have been in danger in those years, even with the modified lines.

The current Florida map has been heavily vetted by the courts to ensure that it is not a gerrymander. Therefore, I will not attempt to improve upon it.

NET CHANGE: NONE

Georgia

 Georgia redistricted mid-decade during the 2000s, after an extreme Democratic attempt at gerrymandering the state’s legislative and congressional districts was dismantled under court order.  In a good-government mood, Republicans redrew the congressional lines in a very fair manner, splitting few counties, preserving communities of interest, and drawing three African American majority districts, as well as two districts where African Americans comprised over 40% of the population.

In 2012, Georgia gained a seat, and the Republicans in control of redistricting implemented a far more aggressive map. The Republicans targeted John Barrow’s 12th District, lowering the district’s African American percentage by about 8 points, from 41% to 33% and replacing its Democratic Savannah portions with Republican Augusta suburbs.  This plan resulted in Barrow’s loss in 2014 to Republican Rick Allen.

I re-did the Georgia map to make it adhere more closely to the pre-2012 map.  I restored the 12th and 1st districts to their configuration prior to redistricting, boosting the African American percentage in the 12th back up to 41% and restoring rural and small-city southeast Georgia to the 1st District.  I placed northern Augusta and its suburbs back in the 10th, and pushed the 10th District northward so that it no longer splits the Athens area.

The layout in Atlanta is more-or-less dictated by the need to maintain three African-American majority seats.

NET CHANGE: D +1

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 692010 63.9% 33.4% 63.4% 35.6% 68.8% 23.2% 5.2% 1.4% 0.3%
2 691666 42.8% 55.5% 40.3% 59.1% 45.0% 48.9% 3.9% 1.0% 0.3%
3 692225 63.3% 33.8% 64.9% 34.0% 69.7% 22.8% 4.1% 2.1% 0.2%
4 691586 21.2% 76.3% 24.6% 74.6% 26.1% 52.7% 14.4% 5.1% 0.2%
5 692204 12.4% 84.5% 16.3% 82.6% 34.7% 52.6% 6.9% 4.0% 0.2%
6 691481 49.3% 45.8% 61.8% 36.5% 70.4% 10.9% 7.7% 9.3% 0.1%
7 692193 52.1% 43.8% 61.2% 37.3% 56.7% 17.5% 13.5% 10.6% 0.2%
8 692087 63.3% 34.4% 61.6% 37.5% 65.1% 28.2% 4.3% 1.3% 0.2%
9 691843 77.8% 19.3% 78.1% 20.5% 82.1% 4.2% 11.6% 1.0% 0.3%
10 692143 64.3% 32.8% 65.5% 33.3% 72.7% 18.7% 4.8% 2.5% 0.2%
11 691760 60.3% 35.3% 66.9% 31.5% 71.1% 14.5% 9.7% 2.9% 20.0%
12 692119 45.9% 51.7% 44.4% 54.6% 52.0% 40.8% 4.3% 1.5% 0.2%
13 692323 21.1% 76.5% 24.5% 74.7% 28.3% 56.8% 10.0% 3.3% 20.0%
14 692013 74.0% 23.1% 72.2% 26.3% 82.1% 10.8% 5.0% 0.9% 0.3%

Illinois

Illinois lost a seat in the 2012 reapportionment, and Democrats were in complete control of the redistricting process.  Democrats executed the most successful gerrymander of the 2012 congressional redistricting cycle in Illinois, meticulously packing areas of Republican strength into five of the state’s 18 districts, turning an 11-7 disadvantage into a 12-6 majority in congressional seats.

A group of affected Republican incumbents brought suit, alleging that the Illinois congressional map diluted the influence of Hispanic voters in the Chicago area by packing them into an earmuff-shaped Fourth District, instead of creating two seats with substantial Hispanic populations (a South Side district with a majority-Hispanic VAP, and North Side district where Hispanics would constitute a plurality of the VAP).  The suit also alleged partisan gerrymandering.

The claims were ultimately dismissed by a three-judge panel of a federal district court.  While the judges acknowledged that the Democrats’ map was a blatant partisan gerrymander, the Court held, like nearly every court to hear such claims after Vieth v. Jubelirer, that the plaintiffs had not proposed a judicially manageable standard for evaluating partisan gerrymandering. The Court also held that, while the legislature could have created a much more compact Hispanic-majority Fourth District, the legislature’s interest in incumbent protection was a sufficient reason for it not to have done so.

In sharp contrast to the Democrats’ gerrymander, the districts I have drawn are quite normally shaped, preserve the cores of the previous districts and group together communities of interest. In particular, I have broken apart the “earmuff” Fourth District, one of the most notorious gerrymanders in existence, creating a new Hispanic-plurality VAP Fourth District in the North Side of Chicago and a Hispanic-majority VAP Third District on the South Side. Breaking apart the Fourth District led to changes in other seats, with Chicago-area Republicans Dan Lipinski and Mike Quigley having their districts effectively eliminated. I have created a new, somewhat Democratic-leaning suburban district (the 5th) in Lake County.

The map I have produced would have resulted in a 10-8 Republican advantage in the delegation in 2012, with 10 of 11 incumbent Republicans winning reelection (I eliminated the district of Downstate Republican Tim Johnson, splitting it between Adam Kinzinger’s 11th District and John Shimkus’s [renumbered] 15th District).   It would have resulted in a 11-7 advantage in 2014 and 2016, when Republicans took the 12th District for reasons having nothing to do with redistricting and everything to do with the trend towards Republicans downstate. While suburban Republicans Bob Dold (10th District) and Peter Roskam (6th District) would have been targeted, they probably would have survived under this map. Dold’s district is 5 points more Republican than the district in which he barely lost to Brad Schneider in 2012 and 2016. Roskam’s district trended strongly Democratic, but he’s consistently run well ahead of his party in the 6th District. The remaining Republican districts are more-or-less completely safe.

While a 10-8 split may seem unfair in a state that is reliably Democratic at the national level, Democratic strength in Illinois is quite limited outside of Cook County. While the collar counties are closely split at the presidential level, these areas still typically vote Republican for other offices.   Under my map, Democrats would hold a 6-1 advantage in districts based in Cook, but lose 10 of the 11 seats outside of Cook (the previous congressional map, which reflected a bipartisan compromise, yielded similar results throughout the 2000s).    Again, the fact that Democratic strength is disproportionately concentrated in urban areas means that fairly-drawn congressional maps will often not yield a majority of congressman for the party that wins a majority of the votes.

NET CHANGE: R +4

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 712692 18.9% 77.9% 17.7% 81.5% 37.5% 52.0% 7.5% 1.7% 0.1%
2 712576 18.6% 78.5% 18.0% 81.2% 32.2% 51.5% 14.3% 0.8% 0.2%
3 712665 20.7% 74.6% 24.6% 73.0% 32.5% 5.6% 57.6% 3.6% 0.1%
4 712633 16.9% 77.5% 24.8% 73.0% 45.0% 3.6% 45.1% 5.0% 0.2%
5 712191 36.2% 56.7% 45.2% 53.7% 69.3% 6.3% 16.8% 6.2% 0.2%
6 712392 37.1% 54.5% 47.3% 51.1% 71.3% 3.4% 13.0% 11.1% 0.1%
7 712517 9.2% 87.4% 11.8% 87.2% 30.6% 51.9% 9.4% 6.7% 0.1%
8 713093 46.7% 46.5% 52.1% 46.5% 77.6% 2.2% 14.7% 4.6% 0.1%
9 713391 15.4% 79.1% 23.8% 74.5% 65.7% 9.3% 11.3% 11.6% 0.2%
10 713007 37.6% 57.0% 46.1% 52.5% 76.8% 1.5% 9.6% 11.0% 0.1%
11 712727 50.4% 44.3% 52.8% 45.5% 81.6% 8.4% 4.6% 4.1% 0.2%
12 713047 55.5% 39.7% 48.7% 49.2% 79.8% 15.5% 2.4% 1.1% 0.2%
13 713213 50.9% 44.0% 54.6% 43.9% 79.7% 5.1% 8.0% 6.3% 0.1%
14 713053 43.4% 50.1% 48.9% 49.5% 70.3% 5.8% 17.5% 5.4% 0.1%
15 713040 69.7% 25.5% 62.9% 35.1% 93.8% 3.1% 1.6% 60.0% 0.2%
16 713070 52.5% 42.7% 47.6% 50.7% 83.3% 6.4% 7.7% 1.6% 0.2%
17 712946 55.4% 38.7% 48.6% 49.6% 90.0% 4.3% 3.7% 0.9% 0.2%
18 712379 55.1% 38.4% 55.2% 42.6% 86.6% 9.0% 1.9% 1.4% 0.2%

NET CHANGE: NONE

Hawaii, Idaho, Indiana, Iowa, Kansas, Maine, Minnesota, Mississippi, Nebraska, Oklahoma, Rhode Island

 Each of these states produced a non-gerrymandered map for the 2012 that I could not have improved upon.  In Indiana, Republicans drew a map with aesthetically pleasing boundaries that yielded a 7-2 split in the delegation in their favor.  Iowa lost a seat in 2012, and its non-partisan commission, as usual, produced a map with compact districts that did not split any counties. In Minnesota, a federal court drew a least-change map that preserved most of the pre-existing, court-drawn map. In Mississippi, a federal court drew lines that preserved the state’s one African American-majority district while splitting only four counties.  In Oklahoma, Republicans held a solid 5-0 advantage in the delegation, and there was no need for any gerrymandering.

NET CHANGE: NONE

Kentucky

 As a last vestige of the state’s loss of a congressional seat in 1992, Kentucky’s 1st District contains an arm linking it to Republican counties along the Tennessee border.  This arm continues to exist because it keeps the 1st District even more heavily Republican than it otherwise would be.  I have produced a map that eliminates this and a few other quirks in Kentucky’s congressional boundaries, making the lines smoother.  The changes would not have altered the partisan split in the state’s House delegation.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 723206 69.4% 26.9% 63.4% 35.1% 89.8% 6.6% 2.0% 0.6% 0.2%
2 723350 70.5% 24.6% 66.3% 32.1% 89.9% 5.2% 2.6% 1.1% 0.3%
3 723441 39.9% 54.9% 42.8% 55.7% 73.5% 19.2% 3.8% 2.1% 0.2%
4 723343 64.7% 29.8% 62.9% 35.3% 93.6% 2.6% 1.9% 0.9% 0.2%
5 722767 78.9% 18.2% 74.3% 23.9% 96.4% 1.5% 0.9% 0.3% 0.2%
6 723255 56.2% 37.9% 57.3% 40.7% 84.9% 8.4% 3.8% 1.7% 0.2%

 

Louisiana

I have smoothed out the lines adopted by the state legislature. I also sought to maintain more of the Cajun Country within the Third District, instead of spreading it to districts based in New Orleans and Baton Rouge. The changes do not affect the state’s partisan split.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 755,684 70.7% 24.7% 72.9% 24.9% 74.2% 14.2% 7.7% 2.3% 0.4%
2 755,553 22.2% 74.6% 22.8% 75.8% 31.5% 59.4% 5.4% 2.3% 0.3%
3 755,748 69.8% 26.7% 68.6% 29.8% 72.6% 20.6% 3.2% 1.3% 1.4%
4 755,352 61.6% 35.6% 60.0% 38.7% 64.9% 29.0% 2.9% 1.2% 0.8%
5 755.601 63.5% 34.1% 61.0% 37.7% 63.5% 32.7% 1.8% 0.7% 0.5%
6 755,434 63.8% 32.0% 65.1% 33.0% 70.6% 23.4% 3.1% 1.8% 0.3%

Maryland

 Maryland did not gain or lose any seats, but it was the site of probably the most grotesque gerrymander of the 2012 cycle.  In the previous redistricting cycle in 2001 and 2002, Democrats in control of the process drew an aggressive gerrymander that succeeded in ousting two Republican incumbents, Bob Ehrlich of suburban Baltimore and Connie Morella of suburban Washington, D.C., for a 6-2 Democratic edge in the delegation.

In 2012, Democrats were in charge of the process again, and sought to build on their 2002 gerrymander to gain a 7-1 advantage.  Their main obstacle was geography.  Both Republicans, Andy Harris and Roscoe Bartlett, represented heavily Republican areas at the edges of Maryland (the Eastern Shore and Western Maryland, respectively) that would be difficult to parcel out to other districts.

Nonetheless, Democrats pushed ahead.  They tacked extreme Western Maryland to heavily Democratic portions of Montgomery County, in suburban Washington, D.C.  They then extended the existing Montgomery County-based 8th district out to heavily Republican parts of Carroll and Frederick counties.  The two African American majority districts, the 4th and the 7th, gained new, heavily Republican territory in Anne Arundel and Baltimore counties, respectively.  The crazy-quilt 3d District, represented by John Sarbanes of suburban Baltimore, got even more bizarre looking, becoming perhaps the most grotesque looking district in the country as it slithered from the northwest Baltimore County suburbs, through bits of Baltimore City and Howard County, and into Montgomery County, with a tentacle going into Anne Arundel to take in Annapolis (which Sarbanes said he wanted to continue to represent).

The enacted Democratic map is drastically unfair.  It continues to give the Baltimore area representation in excess of its population, as a full three districts continue to be based in the Baltimore area despite population losses relative to the Washington, D.C. suburbs.  The districts also brazenly ignore communities of interest, joining rural, impoverished, culturally conservative Western Maryland with suburban, wealthy, culturally liberal Montgomery County suburbs.  Baltimore City is now small enough population-wise to fit into a single district, yet it is split between three districts, with heavily African American areas siphoned off to help white Democrats in surrounding districts.  The Fourth District, which previously joined African American and Hispanic communities of interest in Montgomery and Prince George’s County, lost its portion of Montgomery, which was replaced with far-off parts of Anne Arundel County. The 7th District, based in urban West Baltimore, now extends outward to include exurban parts of Howard and Baltimore counties.

I sought to restore the pre-2002, non-gerrymandered lines to the greatest extent possible.  That meant keeping the 8th District entirely within Montgomery County, and joining the remaining portions of eastern Montgomery County to adjoining areas in Prince George’s in an African American-majority 4th District.  The 4th continues to stretch along the D.C. border to encompass the inside-the-Beltway portions of Prince George’s.  I added College Park to the 4th, along with heavily Hispanic areas around Langley Park that were moved to the 8th District by the 2002 gerrymander.  These changes ensure that much of the sizable Hispanic population in Montgomery and Prince George’s is contained within the 4th District.  In exchange, have dropped some outside-the-Beltway portions of the 4th around Largo and Upper Marlboro.  The district remains majority African American VAP at 54%, with a sizable Hispanic population as well.  Incumbent Rep. Donna Edwards, who decried the 4th’s loss of Montgomery County in redistricting, would have been much happier with this map, as would her constituents in the previous 4th District.

Since 1992 redistricting, the remainder of Prince George’s has been lumped together with the rural and exurban Southern Maryland counties and part of Anne Arundel in order to create a safely Democratic white-majority 5th District for Prince George’s Democrat Steny Hoyer.  But there is a far better alternative to this configuration, in light of heavy African American population growth in outer Prince George’s.  My map creates a third African American-majority district running from outer Prince George’s, up Route 29 through North Laurel and Columbia (both of which have large African American populations), and into the heavily African American suburbs west of Baltimore.  These exurban African American communities are large enough and compact enough that the traditional redistricting goals of preserving communities of interest and not diluting minority voting strength militate in favor of drawing the lines in this way, even if a third African American-majority district is not strictly required by the VRA.  This district also has the virtue of uniting exurban areas in Prince George’s, Howard, and Baltimore counties, instead of combining them with more rural or more urban areas (as under the existing maps). It is 52% African American VAP.

Creating this district also allows for the African American majority 7th District to encompass all of Baltimore City in a single district.  To meet the population requirement, I have included the northwest Baltimore suburbs of Pikesville, Owings Mills, and Reiserstown, each with growing African American populations, in the 7th (this configuration also has the advantage of uniting heavily Jewish neighborhoods in Northwest Baltimore and the suburbs).  The 6th District encompasses all of Western Maryland, taking in rural and exurban parts of Howard County to meet the population requirement (as it did prior to the 2002 gerrymander).  The 2d District is almost identical to its configuration before the 2002 gerrymander, combining Harford County with the Baltimore County suburbs north and east of Baltimore City.

Since the 1992 redistricting, Anne Arundel County has been sliced and diced among an ever-increasing number of districts.  My map changes that, keeping all of Anne Arundel (except for a few precincts near the Bay Bridge) in a single district, the 3d.  The 3d District also reaches into areas of Howard and Baltimore counties (including Elkridge and Catonsville) that are similar to the north Anne Arundel suburbs.  The district also reaches south to take in all of Calvert County, which is similar to the southern part of Anne Arundel.

That leaves the Eastern Shore and the remainder of Southern Maryland.  These areas were joined together prior to the 1992 redistricting, and I have joined them together again.  This configuration ensures that the Eastern Shore is kept whole and is not placed in a district dominated by the Baltimore suburbs (as it is at present).  Southern Maryland is culturally similar to the Eastern Shore, as both are still predominately rural and small-city and populated by the descendants of original settlers (although that is changing).  The configuration also unites the large African American population in Charles County with the large African American population on the Eastern Shore Rather than split Calvert County, I had the 1st take in a few precincts near the Bay Bridge in Anne Arundel to meet the population requirement.

The net result of dismantling the Democrats’ gerrymander would be a 4-4 split in Maryland’s congressional delegation, the same as before to the 2002 redistricting.  Republican Andy Harris would prevail in the new Baltimore County-based 2d, which is reliably Republican.  A Republican likely would have won the new Anne Arundel-based 3d, which voted for McCain in 2008 and Romney in 2012 and lacks a strong Democratic base of support.  (John Sarbanes represented much of this district prior to 2012, but he does not live in the district and had struggled to carry his portion of Anne Arundel against weak challengers.) A Republican also would likely have prevailed in the Eastern Shore-based 1st District, which voted 51.5% for McCain in 2008.

A 4-4 split in Maryland is actually much fairer than the current 7-1 split.   While Democrats dominate Maryland elections, their margins of victory have typically come entirely from just three areas: Montgomery County, Prince George’s County, and Baltimore City.  If you remove those areas, Democrats would have lost most every statewide election of the past few decades.  Sure enough, the four seats held by Democrats under my plan are each based in those areas (one Montgomery-based, one Baltimore City-based, and two Prince George’s-based).  The four seats that would have been won by Republicans are the four that are entirely outside those areas (the Eastern Shore, Baltimore County, Anne Arundel, and Western Maryland districts).  The reason Democrats have needed to gerrymander Maryland so heavily is that, in order to win a supermajority of congressional seats in a diverse state, each district must have similar demographics to the state as a whole.  That, in turn, requires slicing and dicing cities and counties to an extreme extent.

NET CHANGE: R +3

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 721614 53.5% 43.3% 51.5% 46.8% 73.6% 19.5% 3.4% 1.8% 0.3%
2 721857 53.4% 40.2% 55.4% 42.7% 79.2% 12.2% 3.2% 3.8% 0.3%
3 721563 46.2% 47.8% 49.5% 48.5% 74.8% 14.3% 4.8% 4.2% 0.3%
4 722008 8.2% 88.7% 9.5% 89.8% 20.7% 52.9% 17.8% 6.4% 0.2%
5 721573 15.9% 80.6% 16.7% 82.6% 30.9% 52.4% 7.1% 7.3% 0.2%
6 721734 59.4% 36.5% 58.4% 39.2% 85.5% 6.7% 3.5% 3.1% 0.2%
7 721700 14.5% 80.7% 15.1% 83.4% 36.2% 54.9% 4.1% 3.1% 0.3%
8 721503 21.9% 72.2% 29.9% 68.4% 57.0% 10.9% 14.7% 15.1% 0.2%

Massachusetts

Massachusetts lost a district in 2012, and Democrats controlled the redistricting process.  Massachusetts, the birthplace of the gerrymander, has long has some of the most unusual-looking congressional districts in the nation.  With slow population growth in Western Massachusetts, and the retirement of incumbent John Olver, the legislature chose to eliminate Olver’s district, drawing liberal college towns from Olver’s district into Jim McGovern’s Worcester-based 2d District in order to make the 2d District reliably Democratic for when McGovern retires.  The legislature also chose to maintain the unusual configuration of the 4th District, which runs from Brookline and Newton, inner-ring suburbs of Boston, down to Bristol County in southeastern Massachusetts, splitting the old industrial towns of Fall River (which is included in the 4th) and New Bedford (which is not).   To shore up the 9th District, which was won by freshman Bill Keating in a competitive race in 2010, the legislature removed Republican-friendly areas in the northern part of the district and added heavily Democratic New Bedford.  The legislature also changed the lines in the 7th district in the Boston area to make it minority-majority.

For my map, I kept all of Bristol County together in the same district, the 8th.  Bristol County has a distinctive economy and history, and the towns of New Bedford and Fall River, with their large Portuguese-American populations, form a clear community of interest.  Up through the present, Massachusetts congressional maps have split Bristol County, placing much of it in the Boston suburb-based districts, thereby aggrandizing the representation of the Boston suburbs.

Basing the 8th in Bristol County allows the 4th District to become much more compact, both geographically and from a communities of interest perspective. It also includes the inner-ring Boston suburbs south and west of the city, including Brookline, Newton, Wellesley, Sharon and Quincy, as well as Cambridge and heavily white areas of North and West Boston.

I also tried to keep the three Western Massachusetts counties of Franklin, Hampshire, and Berkshire together to the extent possible, combining all but a few towns in those counties with Springfield and its immediate suburbs. With their many liberal arts colleges, they have a distinctive history and form a community of interest. This configuration drops the Democratic percentage in the 2d and increases it in the 1st. I made minor changes to the enacted 6th District to make it more compact and Essex County-focused.

All of these changes likely would have had no impact on the partisan makeup of the state’s congressional delegation, which would remain all-Democratic.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 727119 31.0% 62.7% 28.4% 69.5% 78.7% 5.4% 11.9% 2.4% 0.2%
2 726987 41.3% 50.4% 45.4% 53.3% 85.4% 3.1% 6.5% 3.5% 0.2%
3 727382 36.4% 57.2% 42.4% 55.9% 75.8% 2.7% 12.9% 6.9% 0.1%
4 727941 21.6% 71.3% 30.3% 68.4% 78.3% 4.4% 5.1% 10.0% 0.1%
5 726951 28.2% 66.8% 35.6% 62.7% 78.8% 3.8% 6.3% 8.2% 0.1%
6 727518 37.7% 56.6% 43.4% 55.2% 86.0% 2.3% 7.6% 2.8% 0.1%
7 727793 15.9% 80.1% 19.6% 78.5% 46.3% 23.8% 17.1% 7.3% 0.2%
8 727807 45.5% 48.3% 42.7% 56.4% 89.2% 2.5% 4.1% 1.6% 0.2%
9 728131 44.8% 49.5% 46.1% 52.5% 89.0% 2.2% 1.9% 4.4% 0.3%

Michigan

Michigan was the only state to lose population during the last decade, and it lost one seat in 2012.  Republicans controlled the process completely.  Already holding a 9-6 advantage in the delegation, Republicans set out to eliminate one of the five Democratic incumbents from Metro Detroit while preserving two African American-majority Detroit districts and shoring up vulnerable GOP incumbents Thad McCotter, Tim Walberg, and Dan Benishek. The Republicans  came up with a map that split apart the Oakland County district of incumbent Democrat Gary Peters in suburban Detroit, combining its more Republican portions with the Republican parts of McCotter’s 11th District in western Wayne and western Oakland, and tacking its more Democratic areas (Farmington Hills, West Bloomfield and Pontiac) to the Detroit-based 14th District, which took on an unusual, serpentine shape as it combined some of the poorest and wealthiest parts of the Metro Detroit area.  To shore up Walberg, the legislature dropped Battle Creek, the home of Walberg’s 2010 opponent, Mark Schauer, from the 7th, replacing it with marginal Monroe County.  Benishek’s 1st District picked up Republican-leaning Traverse City from the 4th District, and lost its portion of Democratic-leaning Bay County to the 5th.

The gerrymandered lines worked as intended, producing a 9-5 split in the delegation.  Peters chose to run in the 14th District, despite not living there, and beat freshman Rep. Hansen Clarke of Detroit.  Even though McCotter failed to qualify for the ballot, the Republicans managed to hold his 11th District with a weak candidate. All other incumbents survived comfortably.

I drew a map that undoes the Republican gerrymandering, producing a map that closely follows the lines that a federal court drew for the 1992 redistricting cycle.  Consistent with Michigan law, which mandates that county and township splits be minimized in redistricting, my map splits only five cities and townships in the entire state: Detroit (which would have to be split under any reasonable map), small parts of the Oakland County suburbs of Birmingham and Rochester Hills, the small Downriver city of Trenton, and the city of Holland in Western Michigan.  In addition, my map only splits six counties outside of Metro Detroit.

As compared to the actual map, this map would have resulted in a gain of 1 seat for the Democrats in 2012. Gary Peters likely would have chosen to run in the 9th District, which contains 55% of his old district and is only a couple points more Republican than his previous district. However, it is likely that Republicans would have won it in 2014 when Peters stepped aside to run for Senate. This part of Oakland County elects mostly Republicans at the local level; Peters almost lost in 2010 in a more Democratic version of this district; 2014 was a wave year for Republicans; and this district is only a 3-4 points more Republican than the seat Dave Trott won by 15 points in 2014.

Otherwise, the partisan balance would not have shifted due to the changes. While Walberg might have had a tougher race in 2012, there is no way Democrats would have been able to hold the Seventh District in 2014 and 2016 given the pronounced GOP trend in his district.  Given the pronounced shift towards Republicans of Michigan outside of Metro Detroit, there was no need to gerrymander the Outstate Michigan districts to protect incumbents.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 705895 58.4% 36.1% 54.1% 44.8% 93.4% 1.4% 1.0% 0.6% 2.3%
2 706155 56.3% 37.7% 56.5% 42.6% 87.9% 4.4% 4.9% 1.2% 0.6%
3 705880 51.0% 42.8% 52.6% 46.4% 81.0% 8.4% 6.9% 2.0% 0.4%
4 706022 59.2% 35.0% 53.2% 45.8% 92.9% 2.2% 2.3% 1.0% 0.7%
5 705719 46.0% 49.2% 38.8% 60.2% 78.2% 15.9% 3.5% 0.7% 0.4%
6 705464 51.3% 42.9% 50.2% 48.8% 85.1% 7.6% 4.1% 1.3% 0.5%
7 705833 55.7% 38.7% 51.0% 48.8% 89.9% 4.8% 3.0% 0.8% 0.4%
8 706341 50.6% 43.9% 51.1% 48.0% 85.7% 5.2% 3.7% 3.7% 0.4%
9 706474 45.9% 48.6% 48.3% 51.0% 79.9% 10.2% 3.4% 5.1% 0.2%
10 705551 63.2% 32.2% 54.7% 44.3% 92.4% 3.1% 2.2% 1.1% 0.3%
11 705599 32.6% 62.0% 36.6% 62.4% 76.4% 11.3% 3.2% 7.0% 0.3%
12 706556 43.7% 51.5% 41.9% 57.2% 83.4% 8.0% 1.8% 5.0% 0.3%
13 706196 22.3% 74.4% 18.2% 81.1% 36.1% 53.0% 7.1% 2.0% 0.3%
14 705955 20.1% 76.8% 16.3% 83.2% 41.7% 52.7% 2.4% 1.1% 0.3%

Missouri

Missouri lost a seat in 2012.  Republicans held the legislature; a Democrat, Jay Nixon, held the governorship.  The slow growth in St. Louis, plus the need to preserve the African American-plurality 1st District based in St. Louis, made the Democrat-held 3d District, anchored in the southern part of St. Louis, the obvious choice to be eliminated.  With the support of Missouri’s two African American members of Congress, Lacy Clay and Emmanuel Cleaver, the legislature was able to override Gov. Nixon’s veto and enact a map dismantling the 3d District.  There were some other partisan flourishes: Cleaver’s Kansas City-based 5th District picked up some traditionally Democratic rural counties to help neighboring Republican incumbents, and Jefferson County in suburban St. Louis, which had previously been in the 3d, was split between districts to ensure that the 3d District’s incumbent, Russ Carnahan, would not run against a Republican incumbent.  As it turned out, Carnahan lost in a primary to Clay, and all other incumbents were reelected, for a 6-2 split in favor of Republicans in the delegation.

I produced a version of the map that removes the partisan flourishes and keeps the lines regular, but otherwise preserves the districts in the actual map.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 748564 18.8% 77.0% 18.9% 79.9% 47.6% 45.4% 2.7% 2.6% 0.2%
2 748103 52.6% 42.3% 57.1% 41.4% 89.8% 3.3% 2.0% 3.9% 0.1%
3 749142 66.1% 29.1% 61.0% 37.1% 93.3% 3.0% 1.7% 0.9% 0.3%
4 748952 64.8% 29.8% 60.7% 36.9% 89.7% 4.2% 2.8% 1.5% 0.5%
5 748293 41.1% 53.6% 39.9% 58.4% 68.6% 20.5% 6.9% 1.9% 0.4%
6 749059 63.1% 31.7% 60.0% 37.9% 91.5% 3.6% 2.5% 1.0% 0.3%
7 748403 70.4% 24.7% 67.6% 30.3% 91.6% 1.5% 3.3% 1.2% 0.9%
8 748411 75.4% 21.0% 65.9% 32.0% 92.5% 4.1% 1.3% 0.6% 0.4%

 

Nevada

Redistricting in Nevada fell to a court-appointed commission after the legislature failed to reach agreement with Republican Gov. Brian Sandoval.  Nevada gained a fourth seat after the 2010 census, and Sandoval pushed for a Hispanic VAP-plurality congressional district.  Democrats opposed this, as did many Hispanic groups sympathetic to Democrats.  In the end, the commission did not create a Hispanic-plurality seat, creating an urban Las Vegas 1st District, a Reno-based 2d District, a 3d District consisting of the southern Las Vegas suburbs and points south, and a new 4th District stretching from North Las Vegas to the Cow Counties.

From a communities of interest standpoint, urban North Las Vegas should not be in a district with the Cow Counties. Thus, I have created a map that creates two districts based in Las Vegas’s urban core, and a Third District that combines outlying parts of Clark County such as Henderson with the Cow Counties. The result would be two solidly Democratic districts and two solidly Republican districts.  The Fourth District (my First District) has changed partisan control in each of the past three election cycles, while the Third District has flipped from Republican to Democratic control in 2016. Thus, the result of these changes would be a gain of one seat for the Republicans in 2016; a loss of one in 2014; and no change in 2012.

NET CHANGE: R +1

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 675169 37.6% 56.5% 36.7% 61.4% 48.1% 13.0% 22.6% 13.2% 0.4%
2 674850 52.0% 39.7% 52.9% 44.8% 73.7% 1.7% 16.6% 4.3% 1.8%
3 675528 54.5% 39.5% 53.7% 44.5% 72.8% 5.6% 12.0% 6.7% 0.7%
4 675004 31.6% 62.6% 31.4% 66.6% 40.0% 9.9% 38.9% 8.4% 0.5%


New Hampshire

New Hampshire’s two congressional districts have had roughly the same boundaries since the 1880s, splitting the state’s two largest cities, Manchester and Nashua. However, the Manchester, Nashua and the surrounding towns in southeast New Hampshire form a clear community of interest. Creating a compact First District centered around these towns and a Second District that includes all of rural New Hampshire better serves the values of compactness and preserving communities of interest.

These changes would result in the First District flipping firmly to Republican control.

NET CHANGE: R +1

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 658298 50.2% 44.6% 50.6% 48.2% 91.7% 1.2% 3.3% 2.6% 0.2%
2 658172 44.2% 50.6% 42.5% 56.2% 95.4% 0.6% 1.2% 1.6% 0.2%

New Jersey

After New Jersey’s tumultuous experience redistricting in the 1980s, when a Democratic attempt to gerrymander the congressional districts was struck down by the U.S. Supreme Court in Karcher v. Daggett, New Jersey created a bipartisan commission to handle redistricting.  The commission features an equal number of Democratic and Republican partisans, with a tiebreaker chosen by agreement of the major parties.  The state lost a seat in the 2010 redistricting cycle, dropping it to 12 representatives.  Democrats argued that they should continue to have an edge in the state’s delegation (the state had previously had a 7-6 split in congressional representation in their favor), because the state usually votes for Democrats in presidential contests.  Republicans argued that the state’s slowest growth had been in urban northern New Jersey, an area represented by four Democrats.  Because the Voting Rights Act required that the districts of African American Democrat Donald Payne, Jr. of Newark and Hispanic Democrat Albio Sires of Hudson County must remain intact, Republicans contended that the slow-growing districts of Democrat Steve Rothman of Bergen County and Bill Pascrell of Paterson should be merged.  The tiebreaker, former Republican Attorney General John Farmer, chose the Republican plan, on the (correct) ground that population movement should trump political considerations.

The enacted New Jersey map takes the right approach given population changes within New Jersey in the past decade, but it does contain features designed to protect incumbents that make the lines somewhat jagged in places.  In my map, I have attempted to remove the incumbent-protection features of the map, make the districts more compact, and more closely follow county lines, while retaining the cores of the districts, which have remained relatively constant since the 1980s. Unlike the enacted map, I have kept all of Warren, Sussex, Camden, and Mercer counties whole.

The most notable change is that I have moved Teaneck (with its large African-American population) and Hackensack (with its large Hispanic population) from the Fifth District to the Ninth District. Placing these areas with the wealthy suburban and exurban Fifth District is not consonant with maintaining communities of interest. In exchange, I have added more suburban and exurban territory to the district. The changes mean that Republican Scott Garrett would not have lost the seat in 2016.

NET CHANGE: R +1

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 732851 37.1% 59.5% 34.8% 64.1% 69.6% 14.8% 9.8% 4.4% 0.2%
2 732862 48.6% 48.0% 43.4% 55.5% 69.8% 12.8% 12.3% 3.6% 0.3%
3 732634 52.4% 44.2% 48.2% 50.8% 80.5% 9.6% 5.5% 3.2% 0.1%
4 732403 55.8% 41.0% 54.2% 44.7% 79.7% 6.4% 9.4% 3.4% 0.1%
5 732363 51.3% 45.2% 53.5% 45.4% 80.2% 1.8% 7.8% 9.1% 0.1%
6 732509 40.3% 56.5% 36.7% 62.1% 52.4% 11.0% 18.2% 16.7% 0.1%
7 732445 47.5% 48.6% 52.5% 46.3% 81.1% 3.3% 7.3% 7.3% 0.1%
8 732933 21.5% 75.7% 20.7% 78.3% 31.0% 8.5% 49.7% 8.7% 0.1%
9 732898 30.1% 67.3% 27.8% 71.3% 41.1% 12.0% 33.6% 11.7% 0.1%
10 732669 13.3% 84.7% 12.0% 87.4% 23.3% 50.6% 16.6% 7.1% 0.2%
11 732752 48.8% 47.9% 52.4% 46.6% 76.4% 3.3% 10.2% 8.9% 0.1%
12 732575 32.3% 64.5% 31.9% 66.0% 56.1% 14.8% 13.9% 13.6% 0.1%

New Mexico

I have cleaned up the lines a bit in New Mexico. I also moved Clovis to the Second District from the Third District, as it forms a community of interest with other parts of Little Texas in the Second District. These changes would not have changed the outcome of any elections.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 686096 35.1% 51.6% 39.6% 55.3% 46.9% 2.5% 42.9% 2.4% 3.8%
2 686013 52.6% 37.4% 54.2% 42.4% 48.2% 2.0% 46.3% 0.9% 1.5%
3 687070 35.7% 52.8% 37.7% 58.5% 40.8% 0.9% 37.7% 1.0% 18.4%

New York

New York lost two seats in the 2012 reapportionment cycle. New York’s congressional districts were drawn by a federal court, based mostly on a proposed map prepared by Common Cause. The federal court sought to embody a bipartisan consensus that the Brooklyn/Queens district of freshman Republican Bob Turner and the upstate district of retiring Democrat Maurice Hinchey should be the ones eliminated. As a result, Turner’s old Ninth District was carved up among several neighboring districts (mostly the districts currently held by black Democrats Hakeem Jeffries, Yvette Clarke and Gregory Meeks, which needed to gain substantial population).

The court carefully preserved the two white-majority Manhattan districts held by Jerrold Nadler and Carolyn Maloney, both of which now extend deep into the outer boroughs (Borough Park in Nadler’s case; Astoria, Long Island City and Greenpoint in Maloney’s). It expanded Charlie Rangel’s Northern Manhattan district into the Bronx. The court also left intact the unusually-shaped Hispanic-plurality district held by Nydia Velazquez, now numbered the 7th District, which combines Ridgewood, Bushwick, Sunset Park, and Manhattan’s Lower East Side.

I disagree with the decision to carve up Turner’s district, and not Maloney’s. Turner’s old Ninth District combined distinctive communities of interest: the heavily Orthodox and Russian Jewish neighborhoods of Midwood, Sheepshead Bay and Brighton Beach in Brooklyn, and several older white ethnic neighborhoods in Queens: Breezy Point, Howard Beach, Glendale, Ozone Park, Middle Village and Maspeth. Far from the bright lights and affluence of Manhattan, these downscale parts of the Outer Boroughs are frequently ignored by New York City politicians. Brooklyn has the largest Orthodox Jewish population anywhere outside of Israel. Of all the districts in the U.S., the Orthodox community could effectively elect its candidate of choice only in the former Ninth District. Moreover, the old Ninth District could easily have been expanded to meet the population requirement by adding Borough Park, thereby uniting most of Brooklyn’s vast Orthodox Jewish community in a single district.

Manhattan is only entitled by population to two districts. However, three different districts (the current 10th, 12th and 13th) are based in Manhattan. They only meet the population requirements by extending into adjoining boroughs. Moreover, the 10th District, in combining the secular, progressive Upper West Side with heavily Orthodox Jewish Borough Park in Queens, clearly does not unite communities of interest. Indeed, it is difficult to imagine two parts of New York City that are more dissimilar culturally and politically. The 10th combines these two far-flung regions across a narrow salient of the Brooklyn waterfront, violating principles of compactness.

My map keeps Turner’s old district (renumbered the 10th). The Harlem- based 13th District remains entirely within Northern Manhattan, as it had been prior to redistricting. Although it picks up more of the heavily white Upper West Side, it remains about 40% Hispanic and 25% African American, similar to the population mix pre-2012. In fact, the 13th would be far more likely to elect the candidate of choice of the African-American community without the extension into the Bronx. Rangel would have beaten Dominican-American primary opponent Adriano Espaillat more easily within these borders, and African-American Keith Wright would probably have beaten Espaillat in the 2016 primary within my lines.

The 10th and 12th Districts could then be combined into a single Manhattan district, with the Lower East Side and Chinatown reserved for the 7th District (as they were prior to 2012). The black-majority 9th District would then extend into Sunset Park which (like the 9th) is heavily populated by immigrants. The 8th District would take in Park Slope. The Seventh District would replace Sunset Park with Jackson Heights, which has a similar ethnic mix to Sunset Park, allowing the Seventh District to remain a Hispanic plurality district. The 14th would lose Jackson Heights, but would become even more heavily Hispanic by adding Corona and Hunts Point and losing heavily-white Whitestone and College Point. The 14th would also add Astoria and Long Island City. There would continue to be an Asian-plurality Sixth District in Queens. Thus, my map effectively swaps out Maloney’s district for the old Ninth District without affecting the ethnic mix of any of the other districts.

I have also made changes in Long Island. The court unnecessarily split the western half of Suffolk County between the Second District and Third District. My map keeps two full districts within Suffolk County. The Second District would continue to be based in the South Shore of Nassau County, as it was prior to redistricting. I also preserve the African American majority in the Fifth District by extending it Elmont and Hempstead in Nassau County (the court’s map reduces the black percentage in the Fifth District to under 50%). The remaining district would unite the North Shore of Nassau County, extending into the North Shore Queens neighborhoods of Douglaston, Little Neck, Whitestone, Bay Terrace and College Point, and Throgs Neck in the Bronx.

I did not change the court’s map hardly at all upstate.

The net result from these changes would be a gain of two seats for Republicans by 2016. The 10th District that I have drawn would have voted solidly for Trump in 2016, and would have continued to elect a Republican to Congress. In addition, keeping the Second District entirely within Suffolk makes the district substantially more Republican. Republicans would have been favored to pick up this district after veteran Democrat Steve Israel retired in 2016: it voted narrowly for Trump, and a Republican came close to unseating Israel in a similar district in 2010.

NET CHANGE: R +2

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 717602 54.5% 42.2% 49.1% 49.6% 80.2% 4.2% 11.1% 3.3% 0.2%
2 717681 49.0% 46.4% 47.5% 51.4% 69.6% 7.6% 17.9% 3.6% 0.1%
3 717692 49.6% 48.9% 47.9% 51.3% 74.5% 7.1% 12.4% 4.8% 0.1%
4 717886 44.7% 52.4% 47.4% 51.6% 66.1% 3.9% 13.4% 15.0% 0.1%
5 717886 12.0% 86.4% 8.4% 91.3% 13.4% 54.7% 19.4% 7.3% 0.4%
6 717856 27.6% 69.6% 26.5% 72.3% 30.3% 5.4% 18.8% 41.1% 0.2%
7 717647 10.4% 86.9% 10.3% 88.4% 27.4% 8.1% 43.6% 18.7% 0.2%
8 717726 7.0% 91.1% 3.5% 95.7% 26.7% 51.0% 16.1% 3.9% 0.2%
9 717867 8.0% 89.9% 7.5% 91.6% 17.3% 51.1% 18.2% 11.2% 0.2%
10 717872 51.7% 44.9% 50.8% 48.4% 66.9% 4.0% 14.2% 13.1% 0.1%
11 717847 53.1% 44.3% 46.8% 52.1% 64.9% 6.0% 14.1% 13.5% 0.1%
12 717864 12.9% 83.9% 20.9% 77.5% 73.8% 3.6% 8.4% 12.3% 0.1%
13 717410 6.4% 91.3% 5.6% 93.6% 29.8% 23.3% 39.8% 5.2% 0.2%
14 718020 10.8% 86.7% 9.3% 89.7% 23.6% 17.2% 46.4% 10.3% 0.2%
15 717347 4.6% 94.1% 2.7% 97.0% 5.0% 27.1% 63.6% 2.7% 0.2%
16 717863 23.0% 74.6% 26.0% 73.2% 39.9% 30.2% 23.6% 4.2% 0.2%
17 717921 37.4% 59.6% 40.9% 58.1% 65.8% 9.9% 16.6% 6.4% 0.1%
18 717662 49.4% 46.7% 47.5% 51.0% 74.4% 8.1% 13.0% 2.9% 0.2%
19 717624 50.8% 44.0% 45.9% 52.1% 87.9% 4.0% 5.4% 1.3% 0.2%
20 718100 40.5% 54.0% 38.8% 59.2% 82.6% 7.7% 4.2% 3.5% 0.2%
21 717680 53.9% 40.0% 46.1% 52.2% 92.2% 2.9% 2.4% 0.8% 0.8%
22 717446 54.8% 39.3% 49.2% 48.8% 90.7% 3.3% 2.5% 2.1% 0.3%
23 717176 54.5% 39.7% 49.6% 48.4% 91.1% 2.6% 2.6% 2.1% 0.5%
24 717111 45.3% 48.9% 41.1% 57.0% 86.2% 6.8% 2.9% 2.2% 0.6%
25 717894 39.1% 55.5% 39.4% 58.8% 75.9% 13.2% 6.0% 3.3% 0.2%
26 717833 38.0% 57.6% 34.3% 63.9% 75.2% 15.9% 4.3% 2.9% 0.5%
27 717561 59.7% 35.2% 55.3% 42.9% 93.5% 2.5% 1.8% 0.9% 0.7%

 

North Carolina

North Carolina has had the most tumultuous redistricting process of any state. The state was forced to draw two black-majority districts by the George H.W. Bush Department of Justice in the 1990s. Not wanting to endanger any of the state’s 7 white Democratic incumbents, the Democratic legislature drew the ultra-grotesque 12th District. In its original incarnation, the 12th was a thin line stretching from Gastonia to Durham, impossible to create without computer mapping technology. The state also created an only slightly-less grotesque 1st District in the eastern part of the state. The Supreme Court struck down the 12th District in the case of Shaw v. Reno, creating the Supreme Court’s modern racial gerrymandering jurisprudence. After several additional rounds of remapping and subsequent trips to the Supreme Court, the Supreme Court upheld a somewhat modified version of the 12th District (stretching now from Charlotte to Winston-Salem to Greensboro). This time, the Court found that the 12th was a permissible political gerrymander, as race and voting behavior are closely correlated in the South.

The legislature remained under Democratic control for the next round of redistricting in 2002. The Democratic legislature drew a fairly successful gerrymander that yielded a 7-6 Democratic-Republican split in the state’s delegation by 2012. The Republicans controlled the process in 2012, and drew an aggressive map that yielded a 10-3 split in favor of the Republicans by 2014. This created three districts that were overwhelmingly Democratic, two of which (the 1st and 12th) were black-majority. The Fourth Circuit struck this map down, and the Supreme Court affirmed. The “political gerrymandering” defense accepted by the Court in Hunt v. Cromartie in 2001 had mysteriously vanished in the intervening 16 years, as Justice Alito pointed out in dissent. The legislature responded with a new map that yielded a 10-3 split, albeit with cleaner lines.

I have produced a non-gerrymandered map that follows county lines and protects communities of interest. There are four “crossover” (that is effective majority-minority districts) districts, two more than in the current map: the 1st in Raleigh-Durham and the neighboring black-majority counties; the 7th, which combines Fayetteville, the Sand Hill counties, and Robeson County (with its large Native American population); the 12th, based now entirely in Charlotte; and the 13th, based in Greensboro and Winston-Salem. Two of these districts are new, displacing districts drawn to elect Republicans. There is also a Research Triangle district that combines Chapel Hill, Raleigh and Durham (including Duke University). The remaining districts would be safe for Republicans. Heath Shuler might or might not have held the 12th District for the Democrats in 2012, but he would have been unlikely to have done so in 2014 and 2016 given the heavy trend towards Republicans in Western North Carolina.

NET CHANGE: D +2

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 733944 27.2% 70.8% 27.8% 71.3% 39.3% 49.9% 7.5% 1.5% 0.8%
2 733301 53.2% 43.6% 56.0% 43.0% 70.7% 17.1% 9.0% 1.7% 0.5%
3 733079 58.4% 39% 56.6% 42.8% 68.6% 24.5% 4.5% 1.0% 0.3%
4 733816 32.9% 63.5% 38.8% 60.2% 70.8% 13.4% 7.1% 7.0% 0.2%
5 733123 65.9% 31.3% 65.0% 33.6% 89.1% 4.7% 4.3% 1.0% 0.2%
6 733503 63.1% 34.4% 62.9% 36.1% 77.8% 13.8% 6.3% 0.9% 0.4%
7 732836 57.6% 39.9% 56.4% 43.6% 72.4% 17.6% 6.9% 1.2% 0.6%
8 733421 44.5% 52.5% 42.6% 56.8% 48.2% 32.3% 7.3% 1.6% 8.8%
9 733355 60.9% 36.3% 61.8% 37.2% 17.3% 51.1% 18.2% 11.2% 0.2%
10 733605 65.0% 32.4% 62.6% 36.4% 80.6% 11.6% 5.1% 1.6% 0.3%
11 733785 59.0% 38.2% 55.5% 43.1% 89.0% 3.9% 4.2% 0.7% 1.2%
12 733921 29.9% 66.9% 32.8% 66.6% 47.4% 34.4% 11.9% 4.5% 0.3%
13 733794 36.7% 59.4% 39.3% 59.2% 56.0% 31.5% 7.8% 3.0% 0.4%

 

Ohio

Ohio was the site of one of the nation’s most aggressive Republican gerrymanders. All the intricate maneuvering reflected in that map only netted Republicans two seats, however. The map I have drawn no longer splits Cincinnati down the middle. It also unpacks the Toledo-to-Cleveland district designed to combine Democrats Marcy Kaptur and Dennis Kucinich.  I think the decision to keep the 10th District (numbered 14 on my map) as a black-majority district was a permissible redistricting goal, and that required extending it into Akron. Treating Akron and Cleveland as a single metro area allows a map that combines Akron, West Cleveland, and close-in suburbs into a single district (the 13th); making the 16th district (renumbered the 12th) entirely suburban/exurban; and keeping the 14th district (renumbered the 16th) focused in Northeast Ohio with few changes.  I have created a 15th District that combines Youngstown, Warren and Canton, which share an industrial heritage, along with most of their close-in suburbs.

The net result of this map would be two additional districts for the Democrats: the 13th and the 2d. This yields a 10-6 split in the delegation in favor of Republicans.

NET CHANGE: D+2

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 721121 69.1% 26.6% 65.9% 32.4% 89.8% 5.1% 2.1% 2.0% 0.2%
2 721239 42.2% 53.6% 43.4% 54.3% 68.8% 25.0% 2.4% 2.3% 0.2%
3 720926 53.2% 41.1% 54.0% 44.4% 79.4% 15.4% 1.8% 1.9% 0.2%
4 720988 64.9% 28.4% 61.5% 36.7% 90.4% 5.5% 1.5% 1.2% 0.2%
5 720952 61.7% 32.6% 55.9% 42.1% 92.8% 1.9% 3.7% 0.8% 0.2%
6 721119 69.4% 26.6% 59.8% 38.2% 94.9% 2.4% 0.8% 0.6% 0.3%
7 721335 50.0% 45.1% 54.4% 43.9% 87.3% 4.4% 2.2% 4.7% 0.1%
8 720706 26.3% 69.3% 26.6% 72.0% 62.0% 28.1% 4.8% 2.9% 0.2%
9 721115 39.0% 56.2% 33.4% 65.1% 77.3% 14.1% 5.8% 1.2% 0.2%
10 721243 67.8% 28.2% 54.7% 43.2% 95.2% 2.5% 0.8% 0.4% 0.2%
11 720910 66.5% 28.8% 57.7% 40.2% 93.9% 3.2% 1.3% 0.5% 0.2%
12 721402 57.6% 38.0% 54.9% 43.7% 94.4% 1.8% 1.5% 1.5% 0.1%
13 721009 43.8% 50.7% 42.8% 55.7% 88.6% 5.3% 2.8% 2.0% 0.2%
14 720803 17.0% 80.5% 16.5% 82.8% 39.8% 51.8% 4.4% 2.3% 0.2%
15 720921 45.6% 50.1% 36.4% 61.9% 85.3% 10.5% 2.2% 0.7% 0.2%
16 720715 51.1% 42.6% 50.4% 48.1% 92.1% 3.6% 1.7% 1.8% 0.1%

Oregon

I have redrawn the 5th District so it is more compact and normal-looking. This has the effect of making it much more Republican, to the point that it would have voted solidly Mitt Romney and Donald Trump for president. It also has the effect of flipping the district from Democratic to Republican control.

NET CHANGE: R +1

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 766196 30.7% 59.0% 37.5% 60.8% 78.2% 1.5% 10.0% 7.3% 0.6%
2 766269 55.0% 35.0% 56.8% 40.5% 85.3% 0.5% 9.5% 1.1% 1.8%
3 765905 18.3% 71.7% 23.8% 73.2% 77.1% 4.6% 8.7% 6.3% 0.7%
4 766048 43.4% 45.4% 43.5% 53.2% 88.0% 0.6% 5.3% 2.3% 1.2%
5 766656 48.0% 41.1% 52.1% 45.5% 81.4% 0.7% 12.3% 2.6% 1.1%

 

Pennsylvania

Pennsylvania was also the site of a Republican gerrymander in 2012. It is entirely possible, however, to produce a map with the exact same 13-5 Republican-Democratic breakdown without the intricate gerrymandering. Democrats in Pennsylvania are ultra-concentrated in Philadelphia, Pittsburgh and their close-in suburbs. Take away just three counties (Philadelphia, Montgomery and Allegheny), and the two-party vote in Pennsylvania was 58.7%-41.2% for Donald Trump, comparable to Louisiana, Mississippi and bright-red other parts of the Deep South and Great Plains. In areas outside metro Philadelphia and Pittsburgh where Democrats have historically been strong (Erie, Beaver and Lawrence County, Scranton/Wilkes-Barre, Allentown) their vote share has declined rapidly in the past 10-20 years.

The map below would likely have yielded no partisan change. It is unlikely a Democrat would have been able to hold the 16th District, which voted by 17 points for Romney and 20 points for Trump. Democrats might have attempted to target Pat Meehan of the 7th District more aggressively under this map, but his district only gets about 3 points more Democratic, and he is a popular incumbent whi has consistently run well ahead of his party in a district with a long Republican heritage.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 705492 13.2% 84.5% 11.9% 87.3% 35.8% 39.3% 16.5% 6.5% 0.2%
2 705991 8.6% 89.4% 10.0% 89.4% 34.1% 54.2% 3.8% 5.8% 0.3%
3 705609 34.7% 62.3% 35.9% 63.2% 74.6% 10.6% 6.6% 6.8% 0.1%
4 705244 47.2% 49.0% 48.4% 50.3% 88.1% 3.2% 3.4% 4.2% 0.1%
5 706003 49.0% 46.7% 46.7% 51.8% 81.3% 3.9% 11.3% 2.4% 0.1%
6 705797 49.5% 45.6% 50.6% 48.1% 84.6% 3.7% 8.6% 2.1% 0.1%
7 706005 42.9% 53.1% 47.4% 51.5% 86.4% 4.9% 2.5% 5.2% 0.1%
8 705743 53.0% 42.2% 54.4% 44.3% 84.5% 4.7% 7.9% 1.9% 0.1%
9 705285 61.6% 34.1% 60.1% 38.5% 90.0% 3.8% 3.9% 1.2% 0.1%
10 705467 58.4% 37.5% 56.5% 42.1% 85.3% 7.1% 4.4% 2.1% 0.1%
11 706221 66.1% 30.2% 60.1% 38.4% 94.1% 2.5% 2.1% 0.6% 0.2%
12 705197 51.7% 44.5% 41.4% 55.8% 87.3% 4.3% 6.0% 1.4% 0.1%
13 705477 71.2% 25.7% 64.3% 34.4% 94.4% 3.0% 1.4% 0.5% 0.1%
14 705498 63.8% 32.0% 58.6% 40.0% 94.0% 2.2% 1.4% 1.6% 0.1%
15 705369 59.2% 36.9% 53.4% 45.2% 93.1% 3.7% 1.4% 0.8% 0.1%
16 705999 58.7% 37.9% 57.8% 40.9% 94.2% 2.9% 0.9% 1.3% 0.1%
17 706121 30.5% 66.0% 30.6% 68.0% 75.0% 18.7% 1.6% 3.2% 0.2%
18 705861 57.6% 39.0% 57.4% 41.5% 94.8% 2.5% 0.9% 1.1% 0.1%

 

South Carolina

South Carolina’s 2012 redistricting witnessed a battle between Republicans in the State House who wanted to place the state’s new 7th District in the Pee Dee/Myrtle Beach area, and a potential bipartisan compromise map that would have placed the new district in Beaufort and suburban Charleston. The Senate’s 7th District was much less safe for Republicans than the House’s plan, largely because it created a 6th District that had a higher black population. The legislature eventually went with the Pee Dee-based 7th District.

It is undeniable, however, that the bipartisan compromise map had more regular lines, by virtue of not splitting Charleston and allowing the 6th District to become less compact. The proposed 7th District voted for John McCain in 2008, and, like the rest of the Deep South, it trended even more Republican as the decade wore on. It would have likely been a safe seat for the Republicans.

NET CHANGE: NONE

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 661119 53.1% 40.8% 54.5% 44.0% 71.8% 20.9% 4.8% 1.2% 0.3%
2 660264 56.3% 38.0% 59.1% 39.4% 71.2% 20.7% 4.8% 1.8% 0.3%
3 660805 67.5% 28.5% 65.0% 33.4% 77.6% 17.3% 3.4% 0.8% 0.2%
4 661110 59.7% 35.0% 61.7% 36.7% 72.1% 18.5% 6.4% 2.0% 0.2%
5 660894 57.8% 38.3% 55.6% 43.1% 69.3% 25.4% 2.9% 0.8% 0.7%
6 660444 36.3% 60.8% 34.1% 64.9% 43.2% 52.2% 2.3% 1.1% 0.3%
7 660728 52.6% 43.7% 52.4% 46.3% 61.7% 29.5% 5.8% 1.4% 0.4%

Tennessee

I have smoothed out the lines in Tennessee. This yields no partisan change: each of the 7 districts outside of Memphis and Nashville voted for Donald Trump by over 30 percentage points.

NET CHANGE: NONE

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 705083 76.4% 20.0% 72.4% 26.0% 94.1% 2.1% 2.2% 0.6% 0.2%
2 705462 64.4% 30.4% 66.6% 31.6% 88.7% 5.6% 3.0% 1.4% 0.3%
3 704794 66.2% 29.4% 64.1% 34.3% 83.9% 10.5% 3.1% 1.2% 0.3%
4 705021 70.3% 25.7% 67.0% 31.4% 92.6% 3.1% 2.6% 0.5% 0.3%
5 705138 38.2% 56.6% 42.5% 55.9% 65.2% 23.0% 7.5% 2.7% 0.3%
6 705201 70.1% 26.2% 66.6% 32.0% 85.5% 7.6% 4.0% 1.6% 0.3%
7 705256 67.5% 28.2% 65.7% 32.9% 83.6% 9.9% 3.5% 1.6% 0.3%
8 705015 66.3% 30.7% 66.1% 32.8% 76.2% 19.0% 2.3% 1.4% 0.3%
9 705135 20.8% 76.5% 21.9% 77.3% 30.9% 60.5% 5.6% 2.0% 0.2%

Texas

Texas has been the subject of a long-running Voting Rights Act lawsuit regarding its congressional districts since 2011. The legislature was ordered by a federal court to draw a minority-majority 33rd District in the Dallas area back in 2011. Further changes may be coming as a result of the suit.

The lines I have drawn below remove the gerrymandered elements of Texas’s districts. I have created a new, compact Hispanic-majority seat in the Dallas area; a new, compact Hispanic-majority seat in South San Antonio; and a new, compact district based in Austin. This required eliminating three Republican seats. I have also rationalized the boundaries of the remaining Border districts, making them much more compact. I also redrew the lines where necessary to follow county lines and become more compact. This map would have easily satisfied any Voting Rights Act concerns; it creates 8 Hispanic-majority districts (four on the Border, 2 in San Antonio, and 1 each in Dallas and Houston), and 4 heavily African-American districts (2 in Houston, 2 in Dallas). Thus, a full one-third of the state’s delegation would come from black or Hispanic “opportunity to elect” districts. There also would be two white Democrats from Austin. The rest of the districts would be solidly Republican.

NET CHANGE: D +3

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 698425 72.2% 25.3% 71.6% 27.5% 68.2% 17.2% 12.4% 0.9% 0.4%
2 698972 56.4% 39.1% 66.9% 31.6% 54.7% 10.7% 24.4% 8.7% 0.3%
3 698510 54.8% 40.6% 64.3% 34.2% 65.1% 8.4% 13.2% 11.3% 0.4%
4 698730 75.4% 21.8% 74.0% 24.8% 77.0% 10.3% 9.9% 0.9% 0.8%
5 698473 69.7% 27.3% 71.5% 27.4% 70.8% 10.8% 15.6% 1.4% 0.5%
6 698760 26.7% 69.9% 30.1% 69.0% 30.5% 26.2% 36.1% 5.5% 0.3%
7 698775 45.1% 50.5% 57.9% 40.6% 54.8% 7.5% 27.6% 8.5% 0.2%
8 698338 72.7% 23.9% 77.0% 21.7% 71.9% 7.4% 17.1% 2.2% 0.4%
9 698473 18.0% 79.3% 21.1% 78.0% 13.1% 37.7% 35.8% 12.0% 0.2%
10 698492 59.3% 36.2% 66.1% 31.8% 62.9% 9.0% 24.3% 2.5% 0.3%
11 698594 76.8% 20.1% 78.2% 20.6% 62.1% 3.9% 31.9% 0.8% 0.5%
12 698576 64.9% 30.7% 68.8% 29.7% 75.3% 6.5% 13.8% 2.6% 0.5%
13 698925 79.9% 16.9% 80.2% 18.5% 71.4% 5.1% 20.2% 1.6% 0.6%
14 698580 58.2% 38.4% 59.3% 39.5% 57.5% 19.3% 19.1% 2.8% 0.4%
15 698380 27.0% 69.7% 27.8% 71.1% 9.2% 0.5% 89.0% 1.1% 0.1%
16 698487 26.2% 68.9% 33.5% 65.2% 16.8% 3.0% 77.9% 1.3% 0.3%
17 698349 66.3% 29.2% 68.2% 30.3% 65.5% 13.2% 17.8% 2.3% 0.3%
18 698550 18.0% 78.5% 20.8% 78.1% 19.1% 40.6% 34.3% 4.8% 0.2%
19 698419 73.5% 22.5% 74.6% 24.0% 63.5% 5.5% 28.3% 1.4% 0.4%
20 698505 34.3% 61.0% 39.7% 58.9% 27.7% 5.5% 62.3% 3.0% 0.2%
21 698284 61.5% 33.5% 68.8% 28.9% 70.1% 2.8% 23.2% 2.3% 0.3%
22 698386 52.1% 44.2% 62.1% 36.7% 49.3% 11.9% 21.5% 15.7% 0.2%
23 698774 32.4% 62.9% 37.8% 60.8% 26.3% 10.2% 60.7% 1.5% 0.2%
24 698369 53.7% 41.5% 63.4% 35.0% 62.5% 7.5% 18.3% 9.7% 0.4%
25 698391 27.0% 67.6% 31.1% 66.5% 38.5% 11.2% 45.7% 3.1% 0.3%
26 698126 60.9% 34.4% 67.6% 30.7% 70.7% 6.8% 15.4% 5.2% 0.5%
27 698308 33.7% 63.2% 34.3% 64.8% 14.4% 0.5% 84.0% 0.7% 0.1%
28 698289 36.5% 60.3% 36.7% 62.3% 16.6% 0.9% 81.3% 0.5% 0.3%
29 697988 24.4% 72.1% 32.0% 66.9% 14.6% 10.8% 71.9% 2.0% 0.2%
30 698420 19.3% 78.1% 20.6% 78.6% 21.7% 45.7% 29.8% 1.6% 0.3%
31 698734 68.9% 27.2% 69.6% 29.1% 69.0% 10.9% 15.7% 2.3% 0.5%
32 698496 46.6% 48.5% 57.0% 41.5% 58.3% 11.2% 21.0% 7.8% 0.3%
33 698613 32.7% 63.9% 36.1% 63.0% 29.1% 13.2% 52.2% 4.2% 0.4%
34 698327 34.6% 60.5% 43.7% 54.9% 70.6% 3.9% 15.8% 7.7% 0.3%
35 698710 60.1% 36.5% 60.5% 38.2% 46.2% 5.4% 45.9% 1.4% 0.3%
36 698026 72.0% 25.2% 73.2% 25.7% 69.4% 8.9% 18.6% 1.8% 0.5%

Utah

Utah’s congressional districts were drawn in order to eliminate the state’s one Democratic representative, Jim Matheson. This required splitting central Salt Lake City among multiple districts.

My map creates a compact Salt Lake City-based district that would likely have elected a Democrat to Congress throughout this decade. The remaining three districts are solidly Republican.

NET CHANGE: D +1

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 690801 50.7% 21.4% 78.4% 19.4% 85.5% 1.0% 9.8% 2.0% 0.5%
2 691224 29.0% 48% 51.0% 46.2% 73.4% 1.6% 17.3% 5.4% 0.8%
3 690490 42.0% 21% 78.8% 18.2% 85.0% 0.5% 9.6% 3.1% 0.4%
4 691363 58.1% 20.1% 82.2% 15.2% 87.4% 0.3% 8.0% 1.1% 2.2%

 

Virginia

Virginia was another state that was required to redraw its districts mid-decade to produce another effective minority-majority district. My map includes two districts that would have elected African-American candidates of choice. It cleans up the lines considerably by creating a new district in the Shenandoah Valley, a region which is now split among several other districts (this practice goes back to 1992, when the then-Democratic legislature wanted to eliminate the seat of then-congressman George Allen, and has been continued since then in order to benefit incumbents).

I felt it was necessary to keep the 10th District entirely within the Northern Virginia suburbs. The 10th District I have drawn would have voted for Republicans Frank Wolf in 2012 and Barbara Comstock in 2014, but would have elected Democrat LuAnn Bennett in 2016.

NET CHANGE: D +1

 

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 727032 52.8% 41.7% 57.4% 41.6% 77.3% 14.0% 3.4% 3.7% 0.3%
2 727734 50.8% 42.9% 52.1% 47.1% 69.1% 18.5% 5.0% 5.2% 0.3%
3 727567 28.9% 66.6% 28.8% 70.6% 46.8% 42.6% 5.3% 2.8% 0.4%
4 727372 31.1% 64.7% 32.8% 66.6% 45.7% 46.1% 4.7% 1.8% 0.4%
5 726979 54.4% 41.3% 54.3% 44.7% 72.7% 21.5% 2.7% 1.9% 0.2%
6 727395 59.7% 34.9% 59.6% 39.4% 83.4% 10.7% 3.2% 1.4% 0.2%
7 727189 57.4% 36.7% 56.0% 41.9% 79.4% 10.9% 6.1% 1.9% 0.3%
8 727244 20.4% 74.0% 31.4% 67.8% 54.1% 17.5% 17.2% 8.7% 0.2%
9 727793 69.0% 27.1% 63.9% 34.7% 90.7% 5.6% 1.7% 1.1% 0.2%
10 727296 40.0% 54.4% 48.0% 51.1% 62.9% 7.8% 12.9% 14.1% 0.2%
11 727423 28.5% 65.9% 37.9% 61.3% 56.7% 7.2% 15.0% 18.7% 0.2%

 

Washington

As bipartisan redistricting commissions are wont to do, Washington’s bipartisan redistricting commission strove for partisan balance in the state’s congressional lines. Thus, it made changes to the state’s district lines that were designed to make the 1st District more competitive politically. It thus messed with the lines in service of achieving partisan competitiveness. However, the 1st District has continued to elect a Democrat to Congress.

I cleaned up the politically-oriented changes the Commission made to the 1st and 2nd Districts. I also redrew the 6th and 10th Districts so they would have more regular lines.  These changes do not affect the state’s partisan congressional delegation.

NET CHANGE: NONE

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 672483 33.8% 58.1% 39.3% 58.1% 75.9% 2.0% 6.6% 12.4% 0.6%
2 672332 38.8% 52.9% 42.0% 55.2% 82.7% 1.4% 7.6% 4.2% 1.8%
3 672301 49.9% 42.5% 49.6% 47.9% 86.2% 1.3% 5.9% 3.4% 1.0%
4 672786 57.9% 35.1% 59.7% 37.9% 64.1% 0.9% 29.5% 1.6% 2.3%
5 672569 52.2% 39.1% 53.5% 43.7% 87.9% 1.4% 4.4% 2.5% 1.6%
6 672365 39.5% 51.8% 41.2% 56.1% 77.1% 5.1% 6.3% 6.7% 1.3%
7 672234 12.2% 82.1% 18.1% 79.2% 75.0% 4.0% 6.3% 10.9% 0.7%
8 672395 44.7% 47.7% 48.1% 49.7% 79.9% 2.1% 7.7% 7.0% 0.9%
9 672331 23.3% 70.5% 29.6% 68.3% 53.1% 10.4% 10.0% 22.8% 0.6%
10 672744 39.9% 51.3% 41.1% 56.3% 79.1% 3.8% 6.3% 5.5% 1.9%

 

West Virginia

I cleaned up West Virginia’s lines a bit, mostly so the Second District would be more compact and the panhandle would not be split. The partisan impact is nil– the Second District, which saw the only close West Virginia congressional race this decade (in 2014), becomes slightly more Republican.

NET CHANGE: NONE

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 620082 66.0% 28.4% 60.2% 37.5% 93.1% 3.3% 1.5% 1.0% 0.2%
2 614690 67.8% 27.4% 62.0% 36.0% 94.6% 2.9% 0.8% 0.6% 0.2%
3 618222 72.5% 23.3% 65.0% 32.8% 94.1% 3.7% 0.7% 0.5% 0.2%

Wisconsin

Wisconsin’s congressional lines were drawn with the intent of shoring up the state’s 5 incumbent Republicans. However, the protection wasn’t really needed, given Wisconsin’s Republican trend through the decade. I have drawn a map that follows county lines and generally follows the boundaries of the pre-2012 districts. Each of the state’s 5 Republican incumbents would have a safe district that voted for Trump and Romney.

NET CHANGE: NONE

 

Population Trump Clinton Romney Obama White VAP Black VAP Hisp VAP Asian VAP Nat Amer VAP
1 711073 51.6% 43.3% 50.6% 48.4% 85.6% 4.8% 7.0% 1.5% 0.3%
2 710487 29.0% 65.8% 30.5% 68.3% 86.8% 3.8% 4.6% 3.4% 0.3%
3 711538 48.8% 45.3% 43.3% 55.3% 94.7% 0.9% 1.8% 1.5% 0.6%
4 710923 21.8% 74.0% 23.8% 75.3% 51.5% 30.4% 12.9% 3.2% 0.6%
5 710871 59.4% 35.3% 63.3% 35.7% 92.0% 1.5% 3.4% 2.2% 0.3%
6 710529 55.2% 39.3% 52.6% 46.3% 93.0% 1.4% 3.1% 1.5% 0.4%
7 710631 57.3% 37.8% 50.4% 48.3% 94.6% 0.5% 1.2% 1.3% 1.6%
8 710934 56.2% 38.6% 51.3% 47.6% 91.2% 0.9% 3.2% 1.6% 2.3%

 

 

 

Previous Post Next Post

51 Comments

  • rdelbov July 23, 2017 at 10:13 pm

    A lot of great work here-I will need more time to study some maps that you did but great stuff. I will even not argue about your views on CA but more about that later.

    I love your Texas map!!

  • krazen1211 July 24, 2017 at 10:20 am

    Really amazing work.

  • w920us July 24, 2017 at 10:56 am

    States like Maryland are easily ignored by Democrats when it comes to this faux fair redistricting push. They would definitely lose two if not three seats in Maryland.


    R, South Philly, 47, Gay, WFU Alum
    #TrumpVoter #NeverHillary

    • shamlet July 24, 2017 at 11:02 am

      Unfortunately this MD map is actually a pretty strong GOP gerrymander. You can make a second black-majority seat without going all the way to Randallstown (which is not at all a COI with PG county) by going from Columbia to Charles County. Then the 3rd becomes AA-Calvert-St. Marys and the 2nd drops Harford tothe 1st and picks up the little remnants of Howard. The natural configuration of Maryland is 4-2-2 or something thereabouts.


      R, MD-7. Process is more important than outcome.

      • FreedomJim July 24, 2017 at 7:57 pm

        That makes more sense to me too. Publius’s MD-1 is only contiguous using unbridged water.

      • Publius July 27, 2017 at 1:09 am

        This is how Maryland’s congressional districts were drawn in 1972-1992 (i.e., pre-gerrymandering): Eastern Shore and Western Shore in one district; one Western Maryland district, one Baltimore County district, 2 Baltimore City districts (reduced to 1 now because of heavy population loss), 1 PG County, and 1 Montgomery County district. These districts had almost no partisan aspect to them.

        http://planning.maryland.gov/images/OurProducts/Redistricting/1972CongressionalDistrictmap.gif

        • shamlet July 27, 2017 at 7:22 am

          No doubt that’s a fair impulse. But what I’m saying is that connecting Baltimore and DC like your 5th does is a big no from a COI standpoint.


          R, MD-7. Process is more important than outcome.

          • rdelbov July 27, 2017 at 10:27 am

            A very fair point about CD5–Let me suggest a compromise. Maryland is a state that is like WA if you perserve COI for one district it puts the other one out of wack. Sometimes you just have to do it.

            Push CD5 out of Baltimore county and give it the rest of Howard County. Then the Western District moves into NW Baltimore county. Then you swap parts of Baltimore county between CD7 and CD2 (I think I have the numbers right).

            This does not change the partisan complexion of any of the seats. It just smooths out the lines. CD6 is just as attached to Baltimore county as Howard county. You can’t get perfect COI in Maryland.

    • Jon July 24, 2017 at 6:39 pm

      To me, the way he’s drawn MD-3 looks exactly like districts strongly trending D; and one that even with an R incumbent would be on the Lean D projection list for 2018.
      Even MD-2 on that map would have problems if open.
      So what I’m actually counting this as is R+ 1.5


      45, M, MO-02

      • Publius July 27, 2017 at 1:15 am

        The 2d District and 3d are pretty solidly Republican, in my view. Baltimore (outside of the AA and Jewish areas), Harford, Anne Arundel, and Calvert are all Republican areas, with many blue-collar whites. White voters in Maryland (outside Montgomery and Howard) vote pretty much the same as white voters in Pennsylvania.

  • shamlet July 24, 2017 at 11:03 am

    Really wonderful effort in many respects – any chance you could split this up into a couple diaries so it’s a little easier to take in?


    R, MD-7. Process is more important than outcome.

    • rdelbov July 24, 2017 at 5:18 pm

      The big question with any fair Maryland Map is Anne Arundel county! AA plus Calvert & St Marys is a solid R seat. The Eastern Shore and Harford is another solid R seat. If you keep the R five western counties (keeping Frederick and Carroll together) that’s another safe R seat. Now the 4th seat could be Baltimore county but that’s harder because the Western and Eastern seats need some population

  • rdelbov July 24, 2017 at 5:35 pm

    I might add that I disagree with your conclusions on MA–IMO the GOP would likely have a great chance at 2 or maybe 3 congressional seats under your map. The Ds would absolute sue in court before your map was enacted.

    • BostonPatriot July 24, 2017 at 6:16 pm

      Republicans would certainly have a chance at 9, 8, and 2 (in that order) but Publius’ conclusion that Dems would likely have held all of their seats is probably right. Keating would have held onto 9 in 2012, and likely 2014 and 2016 although there’s a chance he could have gone down either year if he drew a strong competitor (which in real life he didn’t either year).

      8 would have been an open seat won by a Democrat in 2012. If it had come open in 2014 or 2016, or if the incumbent were weak, it certainly would have been more competitive in those years.

      2 would not be competitive while McGovern is still in office outside of an R wave year with a great candidate (and even in 2014 I think McGovern would have won). It would be competitive as an open seat, especially in a midterm year, since Worcester County is more Republican in state elections than presidentially.

      • rdelbov July 24, 2017 at 6:21 pm

        I would love to see the Scott Brown/Baker numbers in 2-8-9. Might be too much for him? Trump ran +20% behind of the areas in those seats. A 45% Trump seat could be a 60% Brown seat? Just saying. They are certainly less D then current maps in those seats.

        • BostonPatriot July 24, 2017 at 6:25 pm

          That’s true–although by that math, Tisei would have won MA-06 easily in 2012 and 2014, and MA-10 would have flipped in 2010. Presidential numbers are much more informative.

          Then again, if this map were in place, there’s a good chance Brown would have run for and won MA-08 in 2014 instead of moving to New Hampshire, and that he would be serving in the House today.

          • rdelbov July 24, 2017 at 9:04 pm

            Great points!! MA6 with the right candidate in the right year could flip. Big Brown and Baker area. There is not a great farm team in MA because state senate and house is so gerrymandered. This map along with a decent house/senate map could change the state over time.

      • BostonPatriot July 24, 2017 at 6:23 pm

        Also, even if it wouldn’t have changed the math, kudos to you for a map that would have pissed off 7 of the 10 Democrats who were in the Massachusetts congressional delegation in 2012!

    • Publius July 27, 2017 at 1:21 am

      Yeah, the 8th and 2d would be prime targets for Rs. Republicans just tend to do abysmally in Massachusetts congressional races, and those districts have strong Democratic heritages.

  • Jon July 24, 2017 at 6:30 pm

    With regards to IL: I think the most natural way to draw #12 would be for it to have rural SE IL instead of St Clair County, but it wouldn’t significantly change partisan impact.


    45, M, MO-02

    • FreedomJim July 24, 2017 at 7:25 pm

      I have been curious about that. Your idea would be more compact, but maybe there are cultural reasons for separating SW IL and SE IL.

  • Jon July 24, 2017 at 6:45 pm

    MO: I’m afraid I really don’t like how MO-01 is drawn on that version; because it contains my own house.
    Also from a COI perspective, there’s no need to split both West County and Mid County between #1 & #2; all of West County (the boundary is I-270 between I-44 & I-70) should be in #2.
    No partisan impact.


    45, M, MO-02

  • Jon July 24, 2017 at 6:53 pm

    NC #1 : That version looks like something that also would have been struck down by the recent court decision; it’s not nearly as bad as what the legislature drew, but they were pretty much mandated to not have any double crossing of two or more counties involving #1 and the same other district.


    45, M, MO-02

    • Jon July 24, 2017 at 6:59 pm

      As to Ohio, I’m thinking that Cleveland to Akron district is on very shaky legal ground; only allowed because no one objected in court to it. (The Ohio NCAAP wanted it unlike the NC districts.)


      45, M, MO-02

  • roguemapper July 24, 2017 at 8:13 pm

    I’m mildly curious to know why splitting Mobile AL for VRA purposes is off-limits when it’s otherwise fine to split municipalities in precisely the way most beneficial to Republicans.


    Dem NC-11

    • Publius July 27, 2017 at 2:17 am

      I don’t think it’s off-limits. I just don’t think it’s gerrymandering to *not* do that, and it does make the lines less clean. I also am not entirely convinced that a non-partisan commission or special master would create such a district. If you look at my Georgia, North Carolina and Texas maps, I do create at least 5 additional “crossover” districts. I recognize you could squeeze out 2-3 more, I just don’t think those additional ones are compact enough that it’s gerrymandering to not create them.

      • roguemapper July 27, 2017 at 12:10 pm

        It’s certainly not gerrymandering to *not* draw VRA seats but it’s inconsistent with the guidelines that you set out in your introduction. It should go without saying that you put a lot of work into these maps so kudos for that. I especially applaud your calculation of 2012 & 2016 presidential results. That said, while your maps are interesting and I looked over all of them, it’s nonetheless quite clear that you apply your criteria arbitrarily, subjectively, and haphazardly. That would be fine in my view if you weren’t claiming otherwise and if you weren’t doing this in the service of a very bold claim in your title.

        In short, whether or not it’s deliberate, you repeatedly make decisions that benefit Republicans, often on very flimsy pretext, while avoiding identical decisions elsewhere if they will benefit Democrats. Your Alabama map versus your Maryland map is simply the most egregious and expressly stated example. In your Maryland map you’re hellbent to create a third majority-AA VRA seat which is of course the only way to justify your blatant GOP gerrymander. Relatedly, in your North Carolina map you have arranged NC-01 in a manner that is not only optimal for the GOP in the Triangle area but that is expressly counter to the recent court rulings that struck down that district. By contrast, you reject creating a new VRA seat in Alabama based on that flimsy Mobile rationale and you also ignore the prospect in other states (e.g., LA, MS) where it would help Democrats.

        You did not claim that you were drawing less gerrymandered maps. You claimed that you were drawing ungerrymandered maps to prove that partisan gerrymandering did not cost Dems seats in the House. Yet you repeatedly grandfather in elements from old gerrymanders where it suits that partisan claim. You don’t define any clear COI standard or standard of compactness and you certainly aren’t applying them consistently. That’s why I didn’t bother to systematically comment on your maps. You did however specify that the VRA took priority over any other criteria but then fail to apply it when it would create additional Democratic seats as in Alabama. So, I’ll just say that your overall claim would be slightly more credible if you had at least conceded that one Alabama seat. If nothing else it would’ve made for a more interesting map drawing exercise for that state.


        Dem NC-11

        • rdelbov July 27, 2017 at 1:49 pm

          You can’t get a cleaner map for Maryland then the one Publius did. CD6 might favor the GOP but how can you draw a cleaner seat? With the eastern shore does it really make any more sense to connect to Baltimore suburbs or to cross the bridge near Annapolis? You can give Charles county to CD5 and St Marys to CD3 then swap area in Howard county around and it makes little or no political difference. CD6 can take western Montgomery county instead of Howard county and you can move population around but not change the map much. RM or other Ds can criticize how this map of MD is done but you still easily get 4 GOP leaning seats with a clean map.

          You could certainly draw a second D seat in Alabama. Jefferson county plus some sort of snake like appendage to downtown Tuscaloosa. Plus a Black belt set across the state from Sumter to Russell County. Of course half of CD2 will be south of it and half will be north of it with some sort of entry through Russell or Montgomery county.

          As a compromise I would do a map of Alabama and not split Montgomery county. Put it in CD2 –Do an AA seat from Jefferson to the Black Belt. IMO you can’t get a second D seat in Alabama unless you chop and split up counties and that should be done in a non-partisan map. You cannot get to 50% AA with chopping up counties as Jefferson county is 53% white.

          • roguemapper July 27, 2017 at 3:03 pm

            Of course you can get a cleaner map for Maryland than the one above. There’s not the slightest doubt about that. That aside, I’m not much interested in arguing about GOP gerrymanders of Maryland. Maryland isn’t going to draw a GOP gerrymander so it’s pointless to argue over what one should look like.


            Dem NC-11

            • rdelbov July 27, 2017 at 3:28 pm

              I disagree with your idea that a 4-4 map in Maryland is a GOP gerrymander. The point of this diary is that a non-partisan drawing of the states of the USA will not produce a very different outcome for the US house then what we have right now.

              I contend that if you asked 10 non-partisan experts to draw a congressional map for Maryland with their only limitation being to avoid un-necessary county splits you see at least one map that would be 4-4 maybe two maps–several maps that would be 5D-3R and several that would be 4D-3R-1 tossup or 4D-2R- two tossup. I easily see 4 possible R or lean R seats in Maryland. Arundel county plus St Mary(and anything else) Harford and Eastern Shore—Baltimore County minus population that goes to Baltimore city CD plus the Western fringe-Frederick-Carroll(plus something).

              Unlikely to be like the map associated with this diary but a fair non-partisan map that respects county lines will produce 3-4 GOP possible seats. I am in my fair mode here. Is his map perfect no–would I agree with some of the maps of these ten experts–no but a couple of them would be favorable to the GOP. The only way to get a 6D map in Maryland using non-partisan standards is to slice Frederick county in half or slice Anne Arundel in half.

              • roguemapper July 27, 2017 at 3:36 pm

                I disagree with your idea that this isn’t a GOP gerrymander. I know the point of this diary. It failed. I contend that you’re wrong about what “10 non-partisan experts” would do with Maryland if their only limitation was to avoid unnecessary county splits. I also have no idea why that’s relevant since the above map doesn’t follow that guideline (none of them do except AR). I’m also uninterested in imaginary maps that exist only in your head. I thought we established that years ago. That said, you may be right that a fair nonpartisan map of MD might produce 3-4 GOP possible seats. If I see a fair nonpartisan map of Maryland that does so then I’ll certainly comment on it.


                Dem NC-11

                • rdelbov July 27, 2017 at 4:05 pm

                  We all see the world through partisan glasses. I continue to applaud and appreciate the work done in this diary.

                  • roguemapper July 27, 2017 at 4:12 pm

                    It’s a shame that you never took up DRA yourself. I actually would’ve been quite interested to see how your fair nonpartisan maps maximized GOP seats everywhere. 🙂


                    Dem NC-11

                    • rdelbov July 27, 2017 at 4:33 pm

                      If one had to know how to do DRA to comment on redistricting or criticize how a map looks then put me in the same camp as Anthony Kennedy. Probably every federal judge as well. I can’t DRA and can barely spell DRA. I can spell fair and the above maps on this tread all fit within the fair category. They may or may favor the GOP in every case but they close to what a reasonable non-partisan might look like.

              • fzw July 27, 2017 at 4:11 pm

                By your standards, one could make a clean 50D-3R map in California, but I’ll gander you’d call that a gerrymander


                Currently MO-5. From MO-3.
                R-leaning Indy.

            • Publius August 6, 2017 at 7:16 pm

              I think the Maryland map I have drawn satisfies the basic requirements of a non-partisan map: 1) Split as few counties as possible; 2) create districts centered around communities of interest (Eastern Shore, Western Maryland, Baltimore City, Baltimore County; Montgomery County suburbs, Anne Arundel County, a compact black-majority in the Prince George’s suburbs inside the Beltway, and a compact black-majority district in outer Prince George’s). Also, I would point out that this configuration is the same one that Maryland followed for 20 years prior to mild gerrymandering in 1992 (to protect Steny Hoyer) and heavy gerrymandering thereafter. (See: http://planning.maryland.gov/images/OurProducts/Redistricting/1972CongressionalDistrictmap.gif)

              The CD-01 in my North Carolina map is very different from the one struck down by the Court. Centering the district on Raleigh-Durham maintains the district as black-plurality while necessitating fewer county splits.

              • roguemapper August 6, 2017 at 7:51 pm

                Your decisions aren’t especially indefensible for any given state taken on its own, except for North Carolina where that CD-01 tendril into eastern Raleigh is a blatant violation of the relevant court rulings. As I stated above, my objection is that your “non-partisan” decisions are inconsistent from one state to the next. Moreover, the “non-partisan” criteria that you chose to use for any given state as well as the degree to which you chose to rely on past configurations for any given state are routinely those that are most advantageous to Republicans for any given state. Since you don’t actually define objective criteria, except for prioritizing the creation of VRA districts (which you then proceed to ignore wherever it would result in additional D seats), there’s really nothing further for me to evaluate about your maps. Sorry. For what it’s worth, I do appreciate all the hard work you put into this!


                Dem NC-11

              • rdelbov August 7, 2017 at 8:51 am

                Publius I for one am absolutely convinced that your Maryland map is certainly one that a non-partisan redistricting could come up with. I note as well that non-partisan maps do not always please both sides.

                RM does not like to talk about theoretical maps . I note that in Maryland and several other maps that you did minor changes could be made to smooth out lines here and there that would not change the likely partisan outcome. I consider your Washington state map as well as maps for MA /Texas to be fair but they IMO a lean D edge to them. Moving a line here and there IMO would help the GOP win additional seats but the nature of non-partisan work is such that you can’t please everyone.

                I note as well that complaints about VRA implications are a problem to be as non-partisan maps in CA-AZ-FL certainly took race into account in drawing lines. Will that be the case in 2021? Who knows but IMO your maps clearly show that a reasonable case can be made that the GOP gain much of an overall edge due to redistricting. A few seats here and there perhaps but not a lot.

                • rdelbov August 7, 2017 at 8:58 am

                  I might add I 100% agree with your basic outline for Maryland as far a non-partisan map could be drawn. Clearly MD is such that some hard decisions have to take place. Does it make any more sense to attach the Eastern Shore to Baltimore suburbs or Anne Arundel or other counties across the Bay. In WA the commission decided to criss cross the Cascade mountains twice. That got my goat but if don’t then you divide out some of the Yakima area between two CDs and does that sound fair to them. No rather decisions just have to made and benefit the GOP as in your Maryland map but then again others as in your Texas map benefits the Ds

  • Mike1965 July 24, 2017 at 8:34 pm

    I applaud the effort. I barely have the time to skim something this big, I can’t imagine how long it took to write.

    One thing that did catch my eye, why isn’t OH-01 entirely in Hamilton county?


    Corker 2020

    • BostonPatriot July 24, 2017 at 9:03 pm

      OH-02 (the urban Cincinnati district here) probably should be entirely in Hamilton. But making that change here wouldn’t affect any of the races–1, 4, and 6 would just each shift clockwise around 3. All of those districts would remain Safe R, and 2 would remain Likely D.

    • Publius July 27, 2017 at 2:01 am

      I drew another map with the 1st entirely in Hamilton and Miami County back in the 1st, but I loved the 4th district in my map above too much. 9 counties arranged symmetrically in Central Ohio north of Dayton, neatly combining the old 4th and 7th. And I really don’t like the current 2d District from a COI perspective, combining wealthy Cincinnati suburbs with dirt-poor Appalachia. I wanted to avoid that configuration as much as possible, and keep the 6th as an entirely small-city and rural district.

  • rdelbov July 25, 2017 at 12:40 pm

    I looked hard at the Washington state map–Something just strikes me as wrong. I know redistricting is all about, at times, hard or unfortunate decisions. Not sure I am content with CD8 crossing the Cascades. The alternative is having CD3 cross over into Yakima county. Chopping up Pierce county three ways is also a travesty–Likewise Snohomish county is jumped up.

    The partisan commission did a better job of giving the GOP a chance at CD10 or CD1 but this map divides out the R strength. Ironically I think the GOP has a better chance at a 5th under the current map then this one. Not a criticism as your Washington Map would likely get the vote of thanks from incumbent but I think there are too many county splits to be truly non-partisan.

    • Jon July 25, 2017 at 6:11 pm

      Basically on this one at the very least WA 01 & 07 should swap territory so that WA-7 doesn’t cross the county line.
      (Except in places where county governments have been dissolved, Michigan style no recrossing rules should apply; perhaps in those areas it should apply to townships as those boundaries tend to be more rectangular than municipal boundaries.)


      45, M, MO-02

      • rdelbov July 25, 2017 at 10:53 pm

        Washington state is just a mess geographically -Ironically splitting some counties and going over the Cascades in CD8 makes other seats more compact. Still think chopping Pierce up 4 ways (not three like I thought it was) is a travesty. Snohomish is chopped up three times in the current map. The above Map for WA is quite good but I think CD1 wastes too many R rural and outer suburban voters.

  • Merrimackman July 25, 2017 at 9:18 pm

    Nice work. Your thesis is right on and most of your maps are great. Couple quibbles with some of your districts.

    Georgia- Not sure why Savannah needs to be in two districts. The first district should take the Savannah area (Effingham, Chatham) counties and the Georgia coast. Makes it about 55% for McCain. The 12th should take in Columbia and Richmond Counties and the surrounding areas of eastern south-central Georgia, and likewise is about 54% for McCain. Reunite Macon in the 2nd District, and contract the 8th district entirely South of Macon. My 2nd District is 50%+ African American with only Muskogee split.

    Ohio- Not sure the 13th district makes any sense. 14th should be condescend into Cuyahoga County. There should be a suburban west district outside Cleveland which would be in the worst years swingy for the GOP. Mappers have a choice on whether to put Akron or Canton with Youngstown, but could easily put Canton back into the 12th to make a solid GOP seat. Also questionable if Cincinnati needs to be split as you did. Place Western Hamilton County with Cincinnati, putting north non-Cincinnati Hamilton with either Butler or Warren/Clermont. That Cincinnati district would be competitive with a D edge.

    Maryland- Makes more sense I feel like to keep Steny’s district pretty much as is. Keep the 2nd as you drew it, and give Greater Annapolis to the Eastern Shore. Then make a Southern Suburban Baltimore seat with Eastern Howard and Northern Anne Arundel. The 3rd would lean Dem, and the GOP would still have 3 solid seats.

    Washington- Put Kitsap back in the 10th, and make your 6th district entirely a Tacoma seat. Mappers have a choice North of Seattle to either put Everett in 2nd or 1st districts. Then I’d make a Metro North district straddling the Snohomish and King County line, and put the 7th back fully into King.

    Virginia- Probably could rotate the 1st, 2nd and 4th districts clockwise around the 3rd district. Not a big deal

    Texas- Austin area isn’t very clean. No reason to split Williamson County. Could easily put all of eastern Travis in one seat and give the West to district 21. The 25th has no VRA standing. I’d clean up the remainder of the 25th, 10th seats. Also district 36 should have Jefferson County, rotating people around Houston accordingly.


    R, RI-2

    • Publius July 27, 2017 at 1:44 am

      Fair points. I wanted to leave the Georgia districts as they were from 2006-2012, which I thought was a very fair map. And it does dilute the AA percentage in the 12th to do that, which a non-partisan redistricting panel would probably not do.

      I played around with different lines for Ohio, but ultimately I liked that configuration the best. Other scenarios changed the districts around too much, and I didn’t like putting Warren in the 16th or Akron in the 13th or Canton in the 12th.

      Steny’s district was specially gerrymandered for him by combining his old home area in North PG with Southern Maryland. I think combining those areas is the same kind of COI mismatch as the current gerrymandered 8th and 6th combining Western Maryland with Montgomery County.

      For Virginia, I thought doing that would simply dilute the AA percentage in the Fourth District.

      • Left Coast Libertarian July 27, 2017 at 10:38 am

        I haven’t looked at everything here, but I applaud you for doing such thorough extensive work. When we try to do non-partisan maps you shouldn’t ignore that you might let some partisan considerations slip in or that you might create the maps with a geographic preference that might not be the best. Please seriously consider making changes based on the input here. You can get more expertise and objectivity from a group of people who have knowledge.

        • rdelbov July 27, 2017 at 10:48 am

          I love the detail on some maps–I encourage people to look at the CO map. You can see how Denver really seemingly flow into the three counties surrounding it. You can make a good case for filling out CD1 with any of the surrounding counties. Saying that this map is so much better with smoother lines then the current CD map.

  • davybaby July 28, 2017 at 10:47 am

    You have put in a stunning amount of work on this. Congrats and thanks!

    • rdelbov July 28, 2017 at 11:02 am

      One reason I like this diary because it is such a superb rebuttal to the worthless report from the Brennan center on GOP redistricting advantage. The Brennan report suggests that the Ds derived no advantage from the Illinois congressional map. Of course their assumption would be that Illinois CDs 1-3-5-7-8-11 are a model of fairness as the Ds are not deriving any advantage, seat wise, in this CD map. Total Trash from Brennan Center.

      As others have noted you can quibble a bit here and there about particular state maps but IMO all of these maps are very close to what would have been written by non-partisan commissions in 2011. I note some have complained about VRA seats but look at the judges in Florida and their map–the NJ and WA state commission maps. Folks they draw VRA seats in non-partisan maps.

      One plus item from Brennan report they note that Rs do get the shaft in CA and MA.

    Leave a Reply

    Do NOT follow this link or you will be banned from the site!