It is a sad and ironic state-of-affairs that gerrymandering stubbornly remains the de-facto remedy for gerrymandering.
The root-cause of gerrymandering is district voting. If voters’ preferences are uniformly distributed within a state, EVERY district, no matter how compact or how torturous, contains a statistically identical amalgam of voters. If one party controlled just 51% of the electorate, they would hold the majority in every district and win 100% of the elections. Independent of how the districts were drawn. Effectively denying the other 49% any representative say in their government.
This perverse result is somewhat mitigated by the fortuitous propensity of like-minded individuals to cluster geographically. Gerrymandering these “communities of interest” into districts where they hold the majority, tends to bring the values of the elected legislators back into alignment with the voters. And assures the legislature is representative of its citizenry.
Unfortunately, gerrymandering can just as easily manipulate an election district’s voter composition to such an extent, that a well-organized minority party grabs a temporary majority. Once in power, the winners may further tilt the voting laws and districts in their favor, locking in a permanent advantage, undermining any positive benefits of districting.
Separating “fair” from “unfair” districts remains a conundrum. Discouragingly, prescribing gerrymandering as a remedy for the “tyranny of the majority” places the government in the tenuous, and dangerous position of blessing certain communities of interest above all others. Seeking equal protection through unequal treatment. Certainly, allocating a fixed number of seats along religious or ethnic lines, as mandated by other nation’s constitutions1 1
Among others, Afghanistan reserves 64 seats for women, India allocates 18% of parliament by caste and tribe, and the UK holds 26 seats in the House of Lords for Church of England Bishops. , is indefensible. But choosing race, sex, political party, et alia as the standard for evaluating a “fair” district boundary, risks a similar anti-democratic outcome.
Among the myriad of issues with district-centric voting:
- Redistricting is a blunt instrument. Gerrymandering can only rebalance districts against a single community of interest, albeit one that may span multiple, associated values. But who should anoint the “one” salient objective? For example, a district drawn to lower voting participation by Christians would be illegal and reprehensible. But redrawing the district by packing it with Southern Baptists rather than Episcopalians, could dramatically swing the district from liberal to conservative. And might simultaneously unbalance the white/black ratio. Clever computer algorithms and detailed profiling will always be able to undermine even the most neutral-appearing or well-intended election guidelines.
- By living and working together in a community of shared values, citizens enjoy numerous tangible and emotional benefits. Arguably, strong, stable local institutions advantage society as a whole, and should be encouraged. But when that community is in an economically depressed region, families face the Hobson's choice between accepting a lower family income to remain in a supportive social environment where they can elect a representative sympathetic to their concerns, or moving to a less economically depressed area where they may feel isolated and out of place. Where their vote may not matter.
- “All politics are local”. Except when they are not. While in some cases the desires of city voters may differ significantly from agricultural regions, (e.g. stereotypically between liberal and conservative values) at other points in time, “litmus test” issues of same-sex marriage or oil pipeline routes may decide an election, cutting across traditional party and geographic boundaries. In a mobile and networked society, communities are increasingly virtual rather than a day’s horseback ride away. It should not be the role of the judiciary, or even the legislature, to anoint and ossify a pre-ordained list of geographically protected groups, who may then dominate the legislature. Creating a vicious circle of protectionism.
- In a large state with, say 10 Representatives, a community of interest that is distributed widely across the region but makes up only 10% of the population will never be able to elect a Congressman. Yet they have a reasonable expectation of electing one representative as their fair share of the “people’s house”. In addition, new political parties grow from small beginnings. The current district-based system locks-in the existing two-party structure by discouraging third and fourth party electoral success. Such lock-in was not anticipated by the Founding Fathers or enshrined in the Constitution, but has become the new-normal.
Gerrymandered districts cannot be drawn to assure “equal protection under the law” for all possible communities of interest. No do they enable multiple communities of interest to arise spontaneously in exercise of their “right to assemble freely”. At least if those communities expect to wield effective political power.
While independent redistricting committees or other arrangements may address the most egregious misuses of gerrymandering, gerrymandering as a remedy places the judiciary in an untenable bind. It is inflexible, backwards looking and too easily manipulated by those currently in power. 2Currently, only 13/50 states split senatorial party affiliations (11 if independents are not included), and most of these splits are in “battleground states” where there is less than a 5% party affiliation difference. In a two-seat cumulative election, the split would be nearer to 50/50, which more closely echoes party divisions in our county. For example, imagine two candidates from the majority party, and one from the minority competing for two open senatorial seats. If the RED party affiliation was 60%, and the BLUE 40% (a 20% difference between parties), the BLUE backers would dedicate ALL 40% of their votes towards their one candidate. The RED could split their vote between their two candidates, and if one received more than 40%, the other would receive less than 20%. Or focus all 60% on one candidate. In every case, RED takes one seat, and BLUE wins one seat. Providing the percentage difference between party affiliations was no more than 33% and the electorate remained disciplined and did not vote split on independent candidates, the winners would split by parties.
We note the difference in party affiliation in all 50 states is less than 33%.The reliance on geographic districting is fundamentally flawed.
We may observe this flaw even in our senatorial races. Currently, only one senatorial seat is open in any cycle, and the entire state functions as a single “virtual district”. Accordingly, the majority party consistently elects both senators, albeit two years apart. But if both senate seats were open at the same time, in a cumulative voting process, it is more likely a minority party would win alongside the majority party2. Timing, as they say, is everything.
To agree upon an alternative to physical districting, we first must agree on the goals of a fair electoral system. Here we suggest three tests:
- The elected legislature should mimic, as closely as possible, the views and interests of its citizens. This is a foundational principle of a representative government.
- In order to assure broad-based participation in the electoral process, the voting system must be simple and accessible. It must accurately reflect each voter’s intent and instill confidence in the outcome
- Individuals should be free to associate with any and all other citizens in their state to coordinate their votes towards a sympathetic candidate. The Government should not infringe upon their choice as individuals, or limit a self-identified community’s ability to band together in their own interest.
In a Utopian world, a legislator might represent ALL of their constituents on a pro-rata basis, thus making these concerns moot. But in practice legislators increasingly favor a “winner take all” philosophy benefiting only those who backed them in an election. Consequently, a representative democracy is best achieved by enabling a mirrored alignment between the voters and the legislators. Which is a characteristic of at-large elections.
Various at-large voting schemes populate the history of the United States, and many more have been proposed or tested in small trials. However, most fail Test #2. For example, rank-ordering the candidates can be shown –theoretically- to robustly yield fair outcomes. But in an election with more than three to six seats and dozens of candidates, most voters are unable to sensibly rank beyond their two or three top choices (or even top three jams at the market- Schwartz). Additionally, the election’s outcome is a calculation, not a simple totaling of votes. Despite the fact that only high school math is required to determine the winner in a ranked election, for many, this will be viewed as a less-than-transparent result.3Cumulative voting does not violate the “one man one vote” precept, as every voter has the identical ability to influence the outcome of the election. Indeed, since voting participation rates often vary by a factor of two between districts, such that one voter has double the impact in a low vs. high rate district, the at-large nature of cumulative voting erases this disparity between districts. It also erases any issues with population imbalances between districts as well as mitigating the “efficiency gap” of McGhee and Stephanopoulos.
A reasonable alternative is Cumulative Voting with a Flexible Slate of Candidates. Here, candidates run “at-large”, state-wide. In a Cumulative schema, each registered voter is entitled to the same number of votes as there are open seats. They may split their vote among all candidates, or focus multiple votes on a single candidate3, such that all votes are eventually cast4. Focused voting assures a minority viewpoint, embodied by a single candidate, could receive enough votes to win a seat.
For example, if the majority party, holding 51% of the votes backs a slate of ten candidates for ten seats, and all majority party members apply their votes evenly to the slate, each candidate receives 5% of the total electoral votes. In contrast, a minority community of interest could pool all their votes towards one candidate, and if their community encompasses 6% of the electorate, their candidate would receive 6% of the total votes and therefore capture one seat from the majority. A seat they are unlikely to ever hold in ten separate district-based elections.
The advantages of Cumulative Voting (fairvote.org) are well established, and has a long history of producing unbiased outcomes. Occasionally, it is wielded as a remedy for corrupt or discriminatory elections. Unlike rank-ordering schemes, a simple vote tally determines the winners.
4A well designed computerized voting machine can simplify this process, reducing errors and minimizing disqualified ballots. Stranded votes are avoided by indicating how many votes remain to be cast, and offering to apply the remaining votes proportionately to those already cast (subject to voter over-ride, of course). An appendix will illustrate possible voting screens, both computerized and paper-based. This proposed system augments Cumulative voting with “Slates” – a bloc of candidates choosing to appear together on the ballot. Why slates? In a low-population state with only 2-3 open seats and perhaps a dozen candidates, people have the time and mental acuity to differentiate between the leading contenders. The candidates have the financial resources and opportunity to campaign effectively as individuals. On the other hand, in a large state with 30 open seats and hundreds of candidates, informed voting is problematic. While it is not the role of government to ensure voters make informed decisions, any fair voting system must not impede traditional decision methods, or create new barriers. Historically, party affiliation was relied upon by many voters as a surrogate for detailed analysis, and many voters trust slates to provide this indirect guidance-- guidance they are free to ignore in the privacy of the voting booth5.
Slates offer many other advantages. While geographic districts only capture a narrow (and often artificial) community of interest, practical benefits accrue when individual representatives campaign locally and assume specific geographic responsibilities after the election. Although representatives owe allegiance to all voters in their state, a Slate would be free in the campaign to indicate that specific candidates would have additional responsibilities as the “point person” for certain localities. Which is no different than a congressman sitting on the Intelligence committee having a different set of responsibilities than a member of the Science Committee.
5In reality, people split votes unevenly, vote preferentially for the most popular or well-advertised candidate (Reeves), and so on. A party may sponsor a sham candidate with a similar name to their antagonists, to try and split the vote. But these tricks already bedevil district voting, and will be easier to mitigate at the state-level. We expect slates will be diverse and fluid. Political parties may propose a slate for all the open seats, expecting loyal voters to close ranks. Or, slates for a “green party” may run just one or two candidates, hoping to attract enough votes for a single win. Slates may cross party lines- advocating “pro-life” or “isolationism” or “fair taxes”. They may differ from year to year, in response to the issues of the day. They draw strength from flexibility and adaptability.
Switching from electoral districts to at-large cumulative voting will be disruptive and initially perplexing. But just because something is hard, doesn’t mean it shouldn’t be done. Voting technology will adjust, and the candidates will learn new ways to get their message out. Political parties will lose some power, as slates (or individuals) with less loyalty to the “machine”, mange to peel off an occasional win. We believe this diversity is healthy and elections will be harder to manipulate than in the current system.
In summary, voting by geographic districts denies citizens the rights of free assembly and equal protection. Gerrymandering is a partial, but highly flawed and easily manipulated, remedy. A superior remedy is Cumulative Voting with a Flexible Slates of Candidates.