Why electoral systems matter: a thought experiment

What’s this? The same votes cast through different electoral systems can elect completely different representatives?

Imagine a mythical city in the heartland, which is seeking a better method to ensure that all of its residents in their “multi-everything” city feel like they have adequate political representation and a vested interest in participating in a healthy society. Let’s call this city:  America.

In this bright shining city of America, there are three politically polarized factions: the Jeffersons, the Hamiltons and the Douglasses. These three factions have sharply divergent political viewpoints, and neither faction ever votes for a member of the other faction. Moreover, the Jeffersons, the Hamiltons and the Douglasses are evenly distributed throughout the city, such that in every neighborhood 60% of the people are Jeffersons, 30% are Hamiltons and 10% are Douglasses.

The city of America is governed by a city council of 10 members. So the question arises, and becomes hotly debated: how should this “multi-everything” municipality elect its council to ensure that all three political factions are adequately represented, and that the city does not polarize amid ongoing bitterness and dysfunction?

The leading framers of this new vision for America realize they have a few choices, in terms of which electoral method to use.

One choice is to elect the city council “at-large” by “plurality” (that is, citywide elections with no division into districts, and the 10 highest vote-getters win). This is a seemingly simple method, in which each voter would have 10 votes and can use as many or as few votes as she or he wishes. The 10 most popular candidates would be elected, which seems to be not only simple but to comport with much-revered bedrock values of representative democracy.

The problem with this method, however, is that the Jeffersons are virtually certain to capture all 10 seats on the city council. The Jefferson majority will reliably vote for members of its own faction, and its voters’ 60% can easily outvote the 30% Hamiltons and 10% Douglasses every time, for every seat. Even if the Hamiltons and Douglasses joined forces, they would not have enough votes for either of them to win.

The only hope for a Hamilton or Douglass candidate to get elected would be if the Jefferson faction ran too many candidates and its voters split their votes on spoiler candidates. But the Hamilton and Douglass factions would have to be very disciplined and only run a single candidate, have all of their voters “bullet vote” for that candidate, and then pray for a lucky roll of the dice.

The “plurality at-large” voting method is the worst of all for ensuring broad representation. Of ensuring “no taxation without representation.” And yet it is the most widespread method in use today. Some say it is simple to use, but in reality it is primitive, and certainly inadequate as a foundation for representation of today’s “multi-everything” populations.

Representation by districts?

To prevent such a one-sided monopoly on representation, the city of America instead decides to adopt district elections – dividing the city up into ten geographically-delineated zones that are equal in population and will elect one council member each.

But in America, dividing the city into 10 districts made no difference in the composition of the city council. That’s because the Jeffersons, Hamiltons and Douglasses are distributed evenly throughout the city, so no matter how the districts are drawn, each district still has a large Jefferson majority, once again shutting out the Hamilton and Douglass minorities from winning representation.

Under any of the systems discussed so far, broadly called “winner-takes-all” voting, it is almost certain that the Jeffersons will in fact “win all,” and the Douglasses and Hamiltons will fail to capture any seats on the council, even though they each constitute a significant minority of the city’s population.

Now suppose that the 30% Hamilton and 10% Douglass minorities are not dispersed equally throughout the city, but each is substantially concentrated in one or two areas. In this case, if the district lines are drawn just the right way – called “gerrymandering” – so that the Hamilton and Douglass minorities become a majority in one or possibly two districts each, then this method could result in a city council consisting of seven or eight Jeffersons, one or possibly two Hamiltons, and maybe one Douglass.

But the minority voters would have to make sure they don’t split their vote among too many of their own candidates, otherwise those candidates would all spoil each other. Consequently, a political organization might have to play the role of a “boss machine” that decides which candidates get to run.

The districting plan’s specific lines become all important, and the line-drawing process would likely be controlled by the Jefferson majority. Using modern computers and voter databases, the biased line-drawers would be able to split either or both the Hamilton and Douglass territories into separate districts, so that they don’t have a majority in any one district. That would deny them fair representation.

Or, looking to blunt criticism of its political monopoly, the Jefferson majority might concede to draw the lines in a way that elects one Hamilton and one Douglass, giving them symbolic but ineffective representation.

However, even if the gerrymander is successful in creating a “majority-minority” district or two for each of the Hamilton and Douglass minorities, it’s only those Hamilton and Douglass voters living in the right district who will be able to vote for a winning candidate. All the other Hamilton and Douglass voters in the other eight Jefferson districts become what is known as “orphaned voters” – voters without an electoral home where they can elect a candidate that reflects their viewpoint.

For that matter, the Jefferson voters in the majority-minority Hamilton and Douglass districts also become orphaned voters. Over time, orphaned voters realize that their candidates can’t ever win, and so they quit participating. Voter turnout declines, apathy reigns.

“Broad representation” methods can make a difference

The new framers of America realize they have a problem. The two methods most widely used – “plurality wins all” at-large and winner-take-all districts – really don’t provide adequate representation for a population as diverse and polarized as their city. Using such methods will only further the bitter polarization, underrepresent vast swaths of voters, and increase apathy and decrease voter participation. Their local democracy will wither.

So they decide to think “outside the box.” America is a place that likes to think of itself as innovative, and also likes to be self-sufficient and doesn’t usually look outside of itself for better methods. But the new framers have heard a rumor – that most of the world’s established democracies do not use these winner-take-all election methods. Instead they use different voting methods under the category of what is called “proportional representation” or “full representation.”

Under these electoral methods, the city of America would elect its 10 city councilors from multi-seat (rather than single-seat) districts, either city-wide or a smaller subset (such as districts with anywhere from 3 to 5 seats). Different methods can be used to ensure that the Jeffersons, Hamiltons and Douglasses all win their fair share of representation on the city council. Here is a description of three of these methods.

Limited voting. Under a limited voting system electing all 10 city council seats citywide, each Jefferson, Douglass and Hamilton voter would be permitted to vote for only five candidates. Unless the Jefferson majority can limit its number of candidates and skillfully distribute its votes among those candidates, the Hamilton or Douglass minority will have a good chance of capturing at least one or two seats, because the Jefferson majority won’t have enough aggregate votes to win all 10 seats. This method is “semi-proportional” because it would provide new opportunity for the Hamilton and Douglass minorities, but they must be disciplined and strategic, concentrating their votes around a small number of their own candidates, i.e. bullet voting. This method has been used to settle voting rights lawsuits.

Cumulative voting. In this system, each voter is allowed 10 votes for the candidates running citywide for the 10 city council seats. The voter is allowed to distribute their votes in any combination she or he wishes, i.e. all 10 votes can be given to a single candidate, one vote to each of 10 candidates, or any combination in between. Voters are able to strategically use their votes to express a strong preference for one or more candidates. By cumulating their votes for their own candidates, the 30% Hamilton minority would have a good chance to elect two or three council members and the 10% Douglasses to elect one council member. However, this method also requires their minority faction to be strategic and disciplined in how many candidates they run, and how they instruct their voters, i.e. bullet vote, so that they don’t split their vote among too many of their own candidates. This method also is “semi-proportional,” and has been used to settle many voting rights lawsuits.

Proportional-ranked choice voting (P-RCV). With this ranked choice voting form of full proportional representation, if the Hamiltons win 30% of the vote they are guaranteed to elect 30% -- three -- of the 10 seats. If the Douglasses win 10% of the vote, they are guaranteed to elect one of the 10 seats. P-RCV is a nonpartisan form of proportional representation (though it can be used in partisan elections too) based upon the concept of “transferable ballots.” Voters pick their favorite candidates, ranking their top choice first, next choice second, next choice third, and so on. If their top choice doesn’t have enough support to win, their vote transfers to their second choice, and keeps transferring until it can help elect one of their choices.

Also, if their top choice has already been elected, their surplus vote transfers to their next choice (though at a reduced value), and so they may see two or more of their favorites win. Strategic, i.e. “bullet voting,” like the kind used in plurality at-large, limited or cumulative voting, or political machines deciding which candidates can run, like in district elections, is not necessary and does not give an advantage to candidates or voters.

All votes are transferred until all city council seats have been filled. The Jefferson, Hamilton and Douglass voters will all win representation in proportion to their voting strength at the polls. P-RCV is also known by political scientists as “single transferable vote” or “preferential voting.” It has been used for decades in many countries (such as Ireland and Australia), and it provides the fairest and most flexible method for ensuring full and broad representation, reducing polarization and not wasting any voter’s vote on spoilers.

The new framers of this mythical city called America finally realize that proportional ranked choice voting is the best method for their multi-everything city. They realize that a “plurality-wins-all” at-large method or winner-take-all districts might seem simple at first, but in reality there are significant negative consequences to using these primitive methods, especially considering the deep divisions and diverse demographics of their city.

Proportional ranked choice voting may seem more complicated at first, but the rules of professional baseball or football are far more complicated than the rules of P-RCV, yet millions of people master those. In reality, it’s just unfamiliar and so seems complicated. But I like to think of it as being more sophisticated in the benefits it provides, and in the efficiency with which it does not waste votes and provides broad representation. It is truly “state-of-the-art” democracy.

For America to reconnect with that shining city on the hill, for the land of “We the People” to reengage with those self-evident truths that “all are created equal,” we must re-embody the spirit of democratic innovation that inspired and guided the hand of those late 18th-century framers and founders. We know a lot more today about which voting methods provide the best representation. Indeed, proportional voting methods had not yet been invented in the late 18th-century.

The way to keep faith with the brilliance of the founders is not to worship them or what they created but to imitate their genius of reinvention, meeting the democratic challenges of our times, just like they met the challenges of their times.

Steven Hill, @StevenHill1776