Politicians Choosing Their Voters
The Consent of the Governed Part 1 // Listen to this as a podcast.
Part I: The Gerry-mander, REDMAP, and How to Win an Election Without Really Trying
Gerrymandering is the process of redrawing a Congressional or state legislative map in order to advantage a political party or protect an incumbent politician from being removed from power. The Constitution mandates that there be a reallocation of seats in the House of Representatives every ten years, which coincides with the Census (Article 1, Section 2). The Census records the population statistics for each state, and as state populations change, so too must the number of seats each state receives in the House of Representatives. In the upcoming 2020 reallocation, for example, states like New York, Alabama, and Pennsylvania are projected to lose 1 or more seats, while states such as Texas, Florida, and Colorado are projected to gain seats.
In addition to affecting the balance of power between states, the reallocation demands that all states redistrict with now-updated Census data; as districts within a state must have (roughly) equal populations, as stipulated by a series of Supreme Court cases in the 1960s, the new Census data forces all states to redraw their lines, even if the total number of districts in a state stays the same. Thus, every decade new maps must be produced across the nation, providing an opportunity for partisan interests to determine the political trajectory for (at least) the next decade, until another round of reapportionment forces the production of new maps.
In 31 states, the state legislature draws the Congressional district lines; additionally (and ironically), in 30 states, the state legislature draws their own district lines. Thus, in the majority of states, the redistricting process is controlled by partisan interests, and this truth is reflected in the Congressional delegations from states in which the districts are controlled by state legislatures. Comparing a study from the Brennan Center for Justice by Laura Royden and Michael Li entitled Extreme Maps to my own data, in states where the Congressional redistricting process was completely controlled by Republicans in 2011, in the last election, in 2018, there were 28 more Republican Representatives than would be proportional to their share of the vote. In states where redistricting was under sole Democratic control in 2011, in 2018 Democrats netted 5 extra seats relative to their share of the vote. One reason for this is that in 2011, the most recent redistricting year, Republicans had sole control over the redistricting process in 16 states while Democrats enjoyed that power in only 6 states.
Partisan gerrymandering has had an incredible affect on the partisan composition of Congress in the past decade, since the 2011 redistricting. In the past four election cycles — 2012, 2014, 2016, and 2018 — Republicans, who had sole control over redistricting in many more states than Democrats following their wave election in 2010, have netted a total of 70 more seats than would be proportional to their vote share, adjusting for equal voter turnout across districts. Without adjusting for voter turnout, and just taking the nationwide House of Representatives popular vote, Republicans have received 46 more seats than their vote share would translate to under a proportional system. The distinction between adjusting for voter turnout and using the national House popular vote is the difference between a single-member district system, like the current American institution, mandating proportionality and a nationwide proportional system, which is much less feasible in the context of our 250-year old republic; thus, the net 70 extra Republican seats is the more representative measure of the impact gerrymandering has had on the composition of the House in the last decade.
Gerrymandering has been an American issue since this nation’s inception. Even before Elbridge Gerry’s eponymous Gerrymander, in 1788 Patrick Henry redistricted the Virginia Congressional map to pit James Madison against James Monroe (David Daley, RatF**ked, xviii). Henry, an anti-Federalist, sought to undermine Madison’s accelerating political career by forcing him to run against the popular Monroe. Despite the gerrymander, Madison won all the same. Thus, the first gerrymander, rather than being a product of Elbridge Gerry, seems, as William Rives wrote in the first major biography of James Madison in 1859, “to have been first put in practice, though ineffectually, by the great Virginia orator and tribune [Patrick Henry], against Mr. Madison in the first election of representatives under the Constitution,” (William Rives, History of the Life and Times of James Madison).
The legacy of gerrymandering extends, therefore, from the first American election to our most recent, and continues to become an institution, rather than an exception.
The more widely recognized origin story of gerrymandering is Massachusetts Governor Elbridge Gerry’s infamous 1812 state senate map, which was designed to preserve the Democratic-Republican — Gerry’s party’s — legislative majority. The new state senate map looked so wild that one district, the Essex South District, included parts of 12 counties. Federalists seized on the odd shape of the district, and a cartoonist drew wings on the district, called it the Gerry-mander (as it somewhat resembled a salamander), and published it in the Boston Gazette. Despite the media attention, Gerry was much more successful at packing and cracking the Federalist opposition than Patrick Henry had been. According to David Daley’s 2017 book Rat F**ked, “the Federalists won 51,766 votes that year and elected 11 senators, while Gerry’s party won 50,164 votes but 29 senators,” (Daley, xviii).
The history of Patrick Henry’s and Elbridge Gerry’s gerrymanders goes to show that gerrymandering is by no means a new problem facing our democracy — and it really is an obstacle for American democratic expression. However, in the past 30 years — and especially in the last decade — partisan gerrymandering has become an especially pertinent obstacle, not just for Democrats, for whom the gerrymandering bells have primarily tolled, but for all citizens who seek representation in “the people’s house.” Gerrymandering is, ultimately, as many editorial boards at large publications have remarked, a way for politicians to pick their voters, rather than voters choosing their representatives.
Additionally, gerrymandering is by no means — or should not be — a partisan issue for voters; though the vast majority of gerrymandered states today were redistricted by Republicans, this is primarily a product of the fact that the vast majority of states were controlled by Republicans in 2011, the most recent redistricting year. In 2011, 26 state legislatures were under sole Republican control, while only 17 state legislatures were under sole Democratic control. It should also be noted that in a much higher portion of those Democratically-controlled states (relative to Republican-controlled states) the state legislature did not have power over the redistricting process, as this was delegated to independent commissions, political commissions, or the courts. Regardless, there are still some states that have Democratic gerrymanders, the prime examples being Massachusetts and Maryland.
In the United States today we have the greatest and most severe gerrymandering problem in modern history. Though the problem did not begin in 2008 (it began, perhaps, in 1788), 2008 is a good starting point for understanding our current predicament. 2008 was an incredible Democratic wave election: Barack Obama was elected the first African-American president of the United States, and, frankly, it wasn’t even close. Obama won 365 Electoral Votes to John McCain’s 173, with Obama flipping North Carolina, Florida, Virginia, Colorado, New Mexico, Nevada, Ohio, Indiana, and Iowa. Obama came within a hair of also flipping Missouri and Montana. Democrats picked up 7 governorships and 8 senate seats, and held onto senate seats in deep red states such as Montana, South Dakota, Arkansas, Louisiana, and West Virginia. Finally, Democrats rounded out the landslide with 21 pickups in the House of Representatives, giving them 257 seats compared with Republicans’ 178.
The 2008 election stunned Republicans: the editor of National Review, Rich Lowry, said of the 2008 election, “It’s a bad thing for the Republicans when you drill down into these states. It’s like, where did all the Republicans go? Did they all move to Utah?” (Daley, xii).
Some of the most important elections, however, were generally overlooked, as they still are: the elections for state legislatures and the gubernatorial races. As I mentioned earlier, in the majority of states, the redistricting process is controlled by the state legislature, often with gubernatorial oversight. In 2008, the Democrats had majorities in most state legislatures, controlling 28 state senates and 32 state houses. This continued a trend from the 2006 election, as Democrats gained majorities in 4 state legislatures in 2006 and gained another 4 in 2008. Despite the importance of state legislative elections — they are important for many reasons, among them being control over the Congressional redistricting process — both voters and the political parties paid little attention to these unglamorous races. The 2008 legislative races fell where they may, and little party money was spent on legislative elections, campaign funds being directed instead to the Presidential, U.S. Senate, and House of Representatives races.
After the stunning Republican defeat in 2008, the Republican Party regrouped and came up with a strategy to take back power in 2010 and beyond. The plan was, at the same time, both ingenious and terrible for American democracy. It was called REDMAP, which stood for Redistricting Majority Project. The idea was to invest heavily in competitive state legislative races in the upcoming 2010 election, to take control of the redistricting process in as many states as possible, and to gerrymander Congressional maps across the United States. The goal was not just to take back the House in 2010, but to maintain a majority in the legislative body for the next decade at least.
Working in the GOP’s favor going into 2010 was, first, the fact that the president’s party generally loses seats in the House of Representatives in midterm elections. Second, was the Citizens United ruling, which essentially allowed for unlimited “dark money” to pour into party coffers from corporations and Super PACs (Daley, xv). Third, the Democratic canary in the coal mine was the surprise election of Republican Scott Brown, who succeeded Ted Kennedy, the “liberal lion of the Senate,” as Senator of Massachusetts. Brown’s election broke the Democratic supermajority in the Senate, severely hinduring President Obama’s push for universal health care (Daley, xvi). And fourth, Democratic strategy did not focus on state legislatures like the Republican REDMAP did, and the Democrats did not fight for many of the critical state-level seats that Republicans targeted.
Karl Rove, the Former White House Deputy Chief of Staff for President George W. Bush, described REDMAP bluntly in the Wall Street Journal:
“There are 18 state legislative chambers that have four or fewer seats separating the two parties that are important for redistricting. Seven of these are controlled by Republicans and the other 11 are controlled by Democrats, including the lower houses in Ohio, Wisconsin, Indiana and Pennsylvania. Republican strategists are focused on 107 seats in 16 states. Winning these seats would give them control of drawing district lines for nearly 190 congressional seats.”
David Daley put it well in RatF**ked:
“The assertion is so bold, yet so sensical, that one does not know whether to stand back and admire the audaciousness, indict the Democrats for gross negligence and lack of imagination, or simply howl over the undemocraticness of it all,” (Daley, xvii).
REDMAP worked. Republicans flipped 3 state senates and 6 state houses in 7 states by spending about $8.2 million (Daley, xix). That’s about as much money as was spent on a single U.S. House election, the race for the Michigan 7th district, in 2010. Republican control over redistricting in those same 7 states has netted Republicans 50 seats over the past four elections relative to their portion of the vote. REDMAP has proven to be incredibly consequential and effective, maintaining a Republican majority in the House of Representatives through the 2012, 2014, and 2016 elections, despite Democrats receiving a majority of the popular vote in House elections in 2012.
REDMAP had its limits, however, and a series of gerrymandering court cases throughout the 2010s slightly eroded Republican gerrymanders. By 2016, three states — Florida, Texas, and Virginia — had had their Republican-gerrymandered maps modified by courts, losing Republicans 2 House seats between 2014 and 2016, when the maps where redrawn.
The most significant victory for anti-gerrymandering advocates in recent years came on January 22, 2018, when the Pennsylvania Supreme Court struck down the 2011 Republican legislature-drawn Congressional map, deeming it an unconstitutional partisan gerrymander under the Pennsylvania Constitution. The case, League of Women Voters of Pennsylvania v. Commonwealth of Pennsylvania, was unlike other cases brought by anti-gerrymandering groups in that it attempted to prove partisan gerrymandering unconstitutional under state constitution rather than the U.S. Constitution — I’ll discuss Supreme Court gerrymandering cases more later. The Pennsylvania Supreme Court found partisan gerrymandering unconstitutional under Article I, Section 5 of the Pennsylvania Constitution, which reads: “Elections shall be free and equal; and no power, civil or military, shall at any time interfere to prevent the free exercise of the right of suffrage.”
The court ordered that the legislature draw new maps before the 2018 midterm election, but the new maps were vetoed by Governor Tom Wolf, a Democrat, and the courts decided to draw the maps themselves. The new, court-drawn maps eliminated the Republican gerrymander, and Democrats, who had previously held only 5 of Pennsylvania’s 18 Congressional seats (despite receiving about 48.1% of the House popular vote in 2016), now control 9 of Pennsylvania’s 18 Congressional seats (having received about 53.9% of the House popular vote in 2018).
REDMAP was limited by the courts to some extent, but it would still require a Democratic wave election at near-2008 levels in order for control of the House of Representatives to flip — in a kind of miracle for the Democratic Party, 2018 was just that. 2018 was a blue wave, though the Democratic resurgence was limited to a very great extent by REDMAP. In 2018, Democrats received about 54.4% of the national House popular vote and 235 seats in the House. In contrast, in 2008, before REDMAP, Democrats received roughly 55.5% of the national House popular vote and 257 seats. In other words, though Democrats received only 1.1% less of the 2018 vote share compared to 2008, they won 22 fewer seats in 2018 than they won in 2008, which represents about 5.1% of the seats in the House. Had Democrats translated their vote share to Representatives at the same rate they did in 2008, Democrats would have won 17 more seats, which would have put them at 252 seats in the House. Even in the bluest of elections, gerrymandering has limited Democratic representation in the House of Representatives.
For Democrats, ending gerrymandering should be the priority — it should be a centerpiece of national strategy and party platform. The Democratic Party got lucky in 2018, and taking the House proves that the party still has not just energy in it, but also the support of a majority of voters. But the state-level races that enabled REDMAP in the first place should be even more concerning for the Democratic Party; once the 2020 elections are through, state legislatures in a majority of states will be drawing the lines for the next decade, and retaking state governments is going to be an uphill battle for Democrats. The reason for this is, as you may have already guessed, gerrymandering: in 2011, when the Congressional lines were redrawn, so were state legislative lines, in many cases by the same state legislators who would run for reelection in the new map. In 2018, one consequence of extreme Republican gerrymandering at the state house level was that in three states — Pennsylvania, North Carolina, and Michigan — Democrats received a majority of votes (54%, 53%, and 51%, respectively) but Republicans won control of the legislature.
In RatF**ked, by David Daley, Norman Ornstein, political scientist and scholar at the conservative American Enterprise Institute, said, “[REDMAP] means basically that the whole constitutional notion of the House as a mirror of popular views comes into jeopardy… Now, I don’t believe the idea that a majority of the nationwide popular vote should automatically translate to a majority of the seats. But the idea that almost nothing happens when you have a broad public expressing its disfavor with the party in power and it doesn’t do anything? That’s not good,” (Daley, xxii).
Part II: The Judicial Problem
In 1986, the U.S. Supreme Court ruled in Davis v. Bandemer that partisan gerrymandering cases were justiciable, meaning that they could be brought before courts. However, no functional definition of gerrymandering was established: in fact, the criterion required for a map to be ruled a gerrymander were so rigorous that no map was deemed a partisan gerrymander in federal court between 1986 and 2004 (Nicholas Stephanopoulos and Eric Mcghee, Partisan Gerrymandering and the Efficiency Gap, 832–833). Additionally, Bandemer did not, by any means, determine partisan gerrymandering to be unconstitutional. The Indiana General Assembly map, which was challenged in Bandemer, had awarded 57 of 100 seats to Republicans despite Republicans receiving only about 48.1% of statewide votes. Though the Supreme Court ruled that the case was justiciable, they did not strike down the map as an unconstitutional partisan gerrymander.
Racial gerrymandering, in contrast, is unconstitutional, as established by the Voting Rights Act of 1965 and backed up by a series of Supreme Court cases in the past thirty years. It should be noted, however, that the Voting Rights Act has since been essentially gutted by the Supreme Court; nevertheless, the language on racial gerrymandering and minority-majority districts remains somewhat strong, though the same cannot be said, of course, for federal oversight of state-run elections.
Regardless, in 2004, the issue of partisan gerrymandering and its justiciability was again brought before the Supreme Court in Vieth v. Jubelirer. In Vieth, the plaintiffs were three registered Democrats in Pennsylvania, who charged that the Pennsylvania General Assembly, which at the time was controlled by Republicans, had committed an unconstitutionally partisan gerrymander against Democrats in the 2001 Congressional redistricting.
Though the plurality of the court — Scalia, Rehnquist, O’Connor, and Thomas — found partisan gerrymandering to be nonjusticiable, they did not garner a majority; Justice Kennedy joined the plurality only in ruling against the specific anti-gerrymandering plaintiffs in Vieth and did not join the plurality in overturning Davis v. Bandemer.
Vieth did, however, reject essentially every conceived standard to measure gerrymandering developed before 2004. To demonstrate how substantial that rejection was, here’s a list of methods to measure gerrymandering that Vieth basically eliminated, as described in Partisan Gerrymandering and the Efficiency Gap by Nicholas Stephanopoulos and Eric McGhee:
“Both the Bandemer plurality’s approach and that of Justice Powell were judicially unmanageable, in the Vieth plurality’s view. So too was the appellant’s proposal of (1) predominant partisan intent, (2) systematic packing and cracking of a party’s voters, and (3) a party’s inability to translate a majority of votes into a majority of seats. And so too were Justice Stevens’s intent-based tests, Justice Souter’s elaborate five-part framework focused on disregard for traditional districting principles, and Justice Breyer’s minority entrenchment standard,” (Stephanopoulos & McGhee, 841–842).
However, Justice Kennedy did leave the door open for a possible standard for gerrymandering to emerge in the future. Two years later, in 2006, League of United Latin American Citizens (LULAC) v. Perry once again introduced the questions of partisan gerrymandering justiciability and workable standards for gerrymandering. LULAC dealt with the aptly named issue of “re-redistricting”: in 2003, the Texas legislature re-drew the maps despite it not being a decennial redistricting year. The Texas congressional maps had previously been court-drawn, and had favored Republicans slightly. Following the 2003 re-redistricting, the maps greatly advantaged Republicans, and they picked up five congressional seats in the 2004 election.
A similar re-redistricting had occurred in Colorado the same year, and Colorado Attorney General Ken Salazar had filed suit against Colorado Secretary of State Donetta Davidson. The Colorado Supreme Court ruled in Salazar’s favor, striking down the new maps as a violation of the Colorado Constitution. However, as Salazar v. Davidson struck down the re-redistricting under state constitution, the precedent did not directly translate to LULAC, as the plaintiffs argued that the new Texas maps violated the U.S. Constitution.
The plaintiffs in LULAC utilized partisan symmetry measures to argue that the new Texas maps were gerrymandered. Partisan bias, a kind of partisan symmetry measurement method, compares a party’s seat share to their statewide vote share, estimating the outcome (in terms of seats) if the party had received 50% of the statewide vote. The difference between the percentage of seats a party would receive in a hypothetical election in which the aggregate vote is split between two parties and 50% is the partisan bias of a map. For example, if there are two parties in a state — Party A and Party B — and Party A receives only 45% of the seats while capturing 50% of the statewide vote, the map has a 5% partisan bias in favor of Party B.
In LULAC, Justice Stevens was the most adamant advocate of partisan symmetry, but the other three liberal justices — Souter, Ginsberg, and Breyer — and Justice Kennedy did express some interest (Stephanopoulos & McGhee, 844). However, Justice Kennedy did have some problems with partisan symmetry, problems that he hinted might be addressed in future methods to measure gerrymandering (Stephanopoulos & McGhee, 845–846).
Justice Kennedy had four main concerns with partisan symmetry. First, he was concerned with the “uniform partisan swing” assumption. Uniform partisan swing is the assumption that, for calculating a hypothetical election (in the case of partisan symmetry, a hypothetical election with a 50/50 statewide vote split), the vote share would change by the same amount in every district. In other words, it is the idea that between elections, there is the same “swing” in every district. For example, as Stephanopoulos and McGhee described in Partisan Gerrymandering and the Efficiency Gap, “if Democrats received 45 percent of the vote in a state, and a researcher wanted to know how many seats they would have won if they had received 50 percent, the researcher would simply add 5 percentage points to the actual Democratic vote share in each district,” (Stephanopoulos & McGhee, 845 note 83).
Kennedy’s problem with uniform partisan swing was that it makes assumptions that are often inaccurate and unfounded. A district’s swing may be influenced by any number of factors uniform partisan swing doesn’t take into account, such as the candidates themselves — their incumbency, their politics, their record, etc. — or the political geography of the state — the number of swing and independent voters varies greatly between urban, suburban, and rural districts (Stephanopoulos & McGhee, 845, 859–860).
Kennedy’s second objection to partisan symmetry, which directly relates to the first, was: “Even assuming a court could choose reliably among different models of shifting voter preferences, we are wary of adopting a constitutional standard that invalidates a map based on unfair results that would occur in a hypothetical state of affairs.” These hypothetical elections result from the use of the assumption of uniform partisan swing.
Third, Kennedy questioned where the threshold of partisan bias should be, above which a plan would be considered unconstitutional. Neither the plaintiffs nor any of the amicus briefs addressed the threshold, which, in retrospect, seems to be something of a gross oversight on the part of anti-gerrymandering advocates. Kennedy sought a well-established threshold which would provide the court with a clear definition of an unconstitutional partisan gerrymander — the plaintiffs in LULAC did not provide this (Stephanopoulos & McGhee, 845–846).
Fourth, and finally, Justice Kennedy wrote that partisan symmetry alone could not persuade a court to strike down a map. Instead, asymmetry could provide one perspective on a map, among other measures and considerations. Thus he wrote,
“Without altogether discounting its utility in redistricting planning and litigation, we conclude asymmetry alone is not a reliable measure of unconstitutional partisanship.”
In 2018, a trio of high-profile gerrymandering cases were argued before the Supreme Court. The first, Lamone v. Benisek, dealt with the Maryland congressional map drawn in 2011 by the Democratic Maryland General Assembly. The second, Rucho v. Common Cause, regarded the remedial 2016 North Carolina congressional map, which was introduced by the Republican-controlled state legislature after the previous map was struck down as an unconstitutional racial gerrymander. The third case, Gill v. Whitford, was about the Wisconsin state house plan that was enacted in 2011 by the Republican-controlled state legislature.
In Lamone v. Benisek, the conservative wing of the Supreme Court — Roberts, Thomas, Alito, Gorsuch, and Kavanaugh — formed the majority opinion and ruled that, contrary to Davis v. Bandemer, partisan gerrymandering is nonjusticiable, a serious blow to anti-gerrymandering advocates hoping for a strong nationwide ruling on the constitutionality of gerrymandering. The reason for the ruling, as described in an article from the Brennan Center for Justice, was:
“In the Supreme Court’s 5–4 opinion in the cases out of North Carolina and Maryland, Chief Justice John Roberts explained that the federal courts cannot hear partisan gerrymandering cases because there are no “discernable and manageable standards” with which to identify when these gerrymanders are unconstitutional.”
Following the rulings on Lamone v. Benisek and Rucho v. Common Cause, Gill v. Whitford was dismissed by the Supreme Court on July 2nd, 2019.
In the trio of cases, the plaintiffs used the efficiency gap, a method to measure gerrymandering which I will focus the next iteration of The Consent of the Governed on, as evidence of partisan gerrymandering, among other mathematical tools. However, the efficiency gap was dismissed by the majority in Lamone, Rucho, and Gill and by Chief Justice Roberts, who said of the efficiency gap, “I can only describe [it] as sociological gobbledygook.”
The efficiency gap is not sociological gobbledygook. In League of Women Voters of Pennsylvania v. Commonwealth of Pennsylvania, the recent anti-gerrymandering success in the Pennsylvania Supreme Court that struck down and replaced the state’s Republican-gerrymandered Congressional map, George Washington University political scientist Christopher Warshaw used the efficiency gap as a gerrymandering measure before the court. The efficiency gap was used in conjunction with other measures to persuade the court, possibly producing a model to fight gerrymanders under other state constitutions.
This model was used again in Common Cause v. Lewis last year. That case concerned, yet again, the Republican-drawn North Carolina congressional map. However, this time the maps were successfully struck down and redrawn by the North Carolina Supreme Court, which ruled in the plaintiff’s favor on state constitution grounds. The new map is now in place and will be used in the 2020 House elections. Again, the efficiency gap was used by the plaintiff as evidence of partisan gerrymandering, furthering the continued importance of the efficiency gap in state and federal-level court rulings on gerrymandering.
The efficiency gap has been touted by the anti-gerrymandering movement as a solution to judicial calls for a better method to calculate gerrymandering, if not at the federal level, then at least at the state level. But the efficiency gap has some serious problems, and other, more effective solutions are available. Join me in the next iteration of The Consent of the Governed for a deep-dive into the efficiency gap as we continue our journey through the world of partisan gerrymandering.