You're already using the best voting system

Oct 6th, 2024
by makeworld

Here in North America we vote for our politicians using plurality, also called First Past The Post (FPTP). It’s a bad system, mainly because of vote splitting. It’s been obvious for a long time that change is needed, but what voting system should we use instead?

IRV

Let’s get this out of the way first: instant-runoff voting (IRV, RCV) is not the solution. There are many issues, but I’ll boil it down to two: complexity, and the center squeeze.

Complexity

IRV is significantly more complex than FPTP, and this has several negative effects.

• Voters find it harder to vote, discouraging participation
• Election results/processes are harder to understand, increasing distrust & animosity
  • Multiple rounds increase confusion and delay
• Voting machines must be replaced or fully reprogrammed

If you’re a social choice theorist or other form of nerd, you don’t care about any of this, which is why groups such as Debian vote using possibly the most accurate and complex voting system, the Schulze method.

But here in the real world these concerns matter a lot. Please try explaining how IRV works to your oldest relative. Seriously, I tried and it didn’t go over great. Or imagine explaining it to your least favourite local politician. Being able to understand the voting system of your country really matters!

Changing voting machines feels like a small thing but it’s probably not. At least it’s a cost that would be worth eliminating if possible.

The center squeeze

Worse than being complex, IRV has a glaring technical issue.

The center squeeze is a specific kind of vote splitting. From Wikipedia:

In a center squeeze, a majority-preferred and socially-optimal candidate is eliminated in favor of a more extreme alternative. Extreme or polarizing candidates who focus on appealing to a small base of core supporters can thus “squeeze” broadly-popular candidates trapped between them

Understanding that the usage of “extreme” doesn’t mean extremist and “center” doesn’t mean centrist, it’s easy to recall examples of this. Here in Canada there is often vote splitting between the Liberal Party (center-left) and the NDP (left) that results in the right-wing Conservative Party winning seats. Even if the majority of voters would prefer the Liberals win over the Conservatives, they may lose due to votes lost to the NDP. This discourages voters from voting for the NDP in the first place (strategic voting), resulting in two-party domination.

You are familiar with this for FPTP, but it unfortunately occurs with IRV also. Imagine some results like this:

It is hopefully clear that “Middling” should win, as it is the party that appeals to the most people. But with IRV, Middling is eliminated immediately for having the fewest first-round votes. In the end “Worst” wins. This once agains encourages strategic voting where Best fans put Middling first to prevent Worst from winning, see here for details.

Empirically, this has happened before, for example in Alaska. There are few countries that use IRV nationally: Australia is one and they end up having a two party system. Ireland’s president is elected using IRV and this is pretty much a two party deal as well.

IRV discourages candidates with broad appeal, and tends towards two-party states.

How you vote already

Before talking about the fix for this, let’s look at a type of voting I do much more often than election voting.

Probably you’ve done the same! It’s just the easiest way to pick event dates, the movie to watch, where to go for dinner, etc. You can select everything you want, not just your favourite, and the results are easy to determine.

This method of voting happens to be superior to IRV, and you already know how to use it. It’s called approval voting.

Approval Voting

In approval voting, you select all the options you approve of, instead of choosing one as with plurality. The option/candidate with the most votes wins. It’s that simple. This is better than FPTP, avoids the issues with IRV, and doesn’t introduce novel issues.

Let’s revisit our criticisms of IRV from before and see if they apply:

• Voters find it harder to vote: not really, and voting for only one candidate like before is still allowed
• Election results are harder to understand: nope, votes are just tallied
  • There is only one round
• Voting machines must be replaced/reprogrammed: nope, only minor code differences should be necessary
• Center squeeze: doesn’t happen with approval voting in practice! Votes are not split among candidates since voters can vote for more than one candidate.

From a more scientific perspective, approval voting has been tested in voting simulations and it does quite well (click for sources):

For that last graph, the author Jameson Quinn wrote:

[Approval voting’s] VSE [voter satisfaction efficiency] is around 89-95% for most levels of voter strategy. That’s not the best of the methods I tested, but it certainly is the best “bang for the buck”; a simple reform, with basically no downsides, which improves outcomes hugely.

That just about sums up how I feel about it. There is a massive electoral improvement just sitting around, waiting to be adopted. And once you consider the complexity that theoretically better methods entail, approval voting stands out as a clear winner to me.

Usage

You can take a look at Wikipedia for details, but approval voting in practice is mainly limited to two cities, both in the USA: St. Louis and Fargo. If you’re in the US and want to improve this, the Center for Election Science is a good place to start.

Unfortunately I’m not aware of any Canadian-specific resources. The best I can find is that Ontario MPP Ted Hsu wrote about it one time, and there was a passionate submission to Trudeau’s 2016 Committee on Electoral Reform by a computer science student. Clearly there is work to be done.

If you’d like to create approval voting polls online, one site I’ve found is strawpoll.vote, which supports approval voting under the name “multiple choice”.

You can find more resources regarding voting systems in my digital garden.

Conclusion

IRV is better than FPTP, but why stop there? Approval voting is simply better than both. I will not be endorsing or promoting IRV in the future, and will instead try to raise awareness of approval voting.