What Single Member District Plurality Voting Systems Do Wrong And How To Fix Those Problems

Single member, plurality voting systems, also called first past the post systems (FPTP for short), are used almost exclusively in the U.K. and Canada, and are by far the predominant system in the U.S.

In FPTP systems when voting for a multi-member legislative body, or for a single jurisdiction wide office (like the Presidency or a Governorship), there is one seat to be filled in each district in which candidates run. The candidate who gets the most votes in that district wins, whether or not that candidate gets more than 50% of the votes cast for that office.

One of the pros and cons of a FPTP system is that it encourages the overall political system to develop two dominant political parties.

There are pros and cons to a two party political system, as opposed to the multi-party political systems that proportionate representation systems, such as one where you vote for party only and that party gets a percentage of seats in a legislative body proportionate to the percentage of the votes cast for that party that a picked from a list of candidates for each party.

The practical differences between these a two party system and a multi-party system in terms of what policies result from it in the long run once people deliberately use political tactics designed to win in that system can be overstated.

In a two party system, people make coalitions to form a majority before the election. In a multi-party system (which proportional representation encourages), people make coalitions to form a majority after the election. But it is a bedrock of any representative democracy that decisions are made by legislative majorities whether those coalitions are made before or after the election.

There are still, however, important differences between how proportional representation systems work, and how single member district systems work.

The single member district plurality voting system in the U.S. naturally favors a two party system (unless a third-party is well focused geographically, rather than evenly spread out, which is why the U.K. and Canada which have the same system still have nationalist, local autonomy or independence oriented third-parties), because third-parties in this system naturally hurt the main party closest to them in ideology. The Green Party hurts the Democrats. The American Constitution Party hurts the Republicans.

In the U.S., the main solution to this problem has been to have political primaries in each party in advance of the general election (basically creating a two stage runoff election with some structure to it), and disadvantaging the ability of third-parties to gain ballot access. But because the U.S. still allows independents and third-parties to gain ballot access without overcoming too many hurdles, this is still an imperfect solution to the spoiler effects of a FPTP system. In Canada and the U.K., where parties internally decide who to run as candidates in each district without true political primaries and ballot access for third-party and independent candidates is easier to obtain, the spoiler effects are even more problematic.

But while in the long run, a single member district plurality voting system favors the continuation of a two party system, it does so by now and then producing elected officials who are contrary to the majority sentiment, because most voters vote non-strategically and some of them allow the best to become the enemy of the good and emotional have temper tantrums causing them to vote for a third-party because they convince themselves that the two major parties are the same even though they don’t really believe that.

You can avoid the spoiler effect of FPTP by having either a true runoff system, like Louisiana or Georgia or France, where it takes a majority to win a single member district and the top two candidates face off in a second round if spoilers have prevented the plurality winner from doing so. You can also compress that process in essentially the same way with Instant Runoff Voting which is used, for example, in Ireland and Maine.

Between the two, I prefer actual runoff voting to instant runoff voting. The average person (and hence the average voter) isn't good at making hypothetical decisions like who their second choice for a political office would be in a multi-candidate race. And, it takes less work for a voter to research only enough to determine which candidate is their first choice without continuing to parse which of the remaining candidates is their second choice. In an actual runoff, voters whose preferred candidates don't make it to the runoff can research who their first choice of the final two candidates is only if that decision really matters, so so they will make better second choice decisions than voters research a second choice that may not actually matter to them will. Reducing the burden of voting in the 90%-95% of elections that don't end up in runoffs is a valuable benefit that counterbalances the cost and delay of holding a runoff election 5%-10% of the time. And, I am convinced that the quality of the decisions made by voters for candidates that don't vote for one of the top two candidates in an actual runoff is better, on average, than the quality of a decision about their second or third choice made by voters in an instant runoff system. Still, either system profoundly reduces the spoiler effects of third-parties that plurality voting facilities and that is sometimes (as recently as the November 3, 2020 elections as documented in media reports) often used as an intentional dirty trick in close races in which one party deliberately but secretly funds a third-party candidate (maybe even one who isn't all that serious) similar to their opponent's views to win an election that their party's candidate would lose without a third-party candidate in the race.

Runoff and instant runoff systems do make it somewhat easier to run a third-party candidate since the third-party candidate isn’t toxic to his natural political allies. But, what runoff and instant runoff systems don’t do, which many other forms of proportional representation do permit, is allow someone who doesn’t have the backing of a majority of the public somewhere to hold office.

Nationalist third-parties, that can evolve in FPTP systems, like parties seeking Scottish and Quebec and Northern Ireland independence in Canada and the U.K., are almost invariably fairly centrist in terms of policy, because they need to win majorities in the places that are seeking autonomy.

In contrast, a party that has widely dispersed 10%-20% support, but majority support nowhere, gets lots of voice and the ability to be a kingmaker in multi-party parliamentary systems with proportional representation, but is shut out completely in FPTP, runoff and instant runoff systems, all of which are systemically biased against extremists. A party like that can be much more extremist than one that needs to win a majority in some district.

Germany’s proportional representation system artificially addresses this problem by having a 5% threshold to get seats in parliament. 

This can also be achieved in a less obviously arbitrary way by breaking up the total legislative body's representatives into multiple, multi-member districts, with a de facto threshold of the percentage of the vote needed to get a seat in a multi-member district. For example, if each multi-member district has ten seats and representatives are chosen in a pure proportional representation system, a party needs about 10% of the vote, in some multi-member district, to get a seat in the legislature, which excludes extremists to a similar degree that the arbitrary 5% threshold in Germany does.

Election systems with no threshold and proportional representation, like Israel, Italy and Weimar Germany, gets lots of true extremists in parliament with a prominent voice and legitimacy, and with kingmaking potential. In Italy, which had a similar system, the five big non-extremist parties constantly entered unstable coalitions amongst themselves to keep the far right (basically neo-Nazis) and far left (basically Soviet style Communists) out of power. Both Israel and Italy are plagued with political instability as a result on their need for either extremist support in a ruling coalition, or supermajority support among non-extremist parties, which forces ruling coalitions to have a larger tent with more internal dissent. Weimar Germany, of course, presents the worst case scenario of what can happen when an extremist party gets a voice and king making potential in the legislature (i.e. it allowed Hitler and the Nazi party to rise to power).

Single member district plurality or majority systems don't just prevent extremists from getting legislative majorities even as kingmakers in a particular election. These systems also, over time, delegitimatizes extremist political voices by keeping them entirely out of elective office until the extreme reaches a point where it can capture a majority of one of the two major parties (which natural rebalance themselves to approach 50-50 in the overall outcome over time by adjusting coalitions and positions, so basically about 25% overall support), in a place where the extremism seems more desirable to the party it has gained a majority of than the other major party. Of course, by the time a faction is approaching 25% overall support, almost by definition, it isn’t so extreme any more.

If true extremists never gain political power, then the political culture learns instinctively, that extremism is not a winning strategy. So, in the long run, this takes the steam out the sails of anyone trying to start a new extremist political movement and builds a society better suited to democratic governance.

From a real politick perspective, a political force, even a substantial 10%-20% political force, that is a minority everywhere, isn’t much of a threat to the legitimacy and power of the parties that do win everywhere by majorities (collectively). They are always outnumbered until they can chance a lot of hearts and minds.

Thus, the benefits of single member district first past the post systems aren’t really seriously impaired by runoff or instant runoff systems, which reduce spoiler effects that can elect people only favored by a minority. And since political systems naturally gravitate towards a 50-50 balance in the system as a whole, a few spoiler driven upsets can dramatically shift the nation’s entire policy set towards minority views and that’s pretty much definitionally bad.

A runoff or instant runoff voting system isn’t a perfect solution, however. Any single member district system is inherently prone to gerrymandering which is a powerful distortion that isn’t solved with runoff or instant runoff voting systems. Proportional representation can fix that.

One good way to split the difference is with mixed member proportional representation. First, have a single member district election (plurality, runoff, or instant runoff). Then, count the voters that were cast for each party in those elections, discarding votes cast for members of parties that received no seats (and thus preventing extremism). Then award bonus seats to parties that receive fewer seats in single member districts than they received in percentages of votes cast for parties that got at least one person elected to make the percentage of legislative seats match the percentage of the votes cast for members of that party (this also creates a turnout incentive for supporters of secondary parties in safe single member districts) to counteract the effects of gerrymandering.

If you a worried that at least one district could still elect someone who is too extremist, you can limit the amount of benefit that can be received from bonus members, for example, to not more than the number of single district members elected on that basis, which is sufficient to overcome even moderately severe gerrymandering effects, but not so strong as to permit a party with a widely dispersed minority support to gain its true proportionate share of representation.

If you also have some meaningful standards in place to limit the extent to which gerrymandering is possible, this kind of limitation is good enough to solve the only moderately permitted gerrymandering in the system, without providing a voice or a vote to extremist political factions that are a majority almost nowhere, even if they are a majority in a few isolated places. And, good progress has been made on ways to limit extreme, but not moderate, gerrymandering effects.