If you live in the US, Canada, or the UK, you’re electing your representatives wrong. The voting method you use — called “choose-one plurality with single-member districts” or, more succinctly, “first past the post” (FPTP)—could almost have been designed to waste votes and reduce your meaningful choices.
Is your preferred major party not the biggest in your area? So sorry, you will have a “representative” who’s basically your sworn enemy. Would you actually prefer a “third” party, one that’s not top-two in your area? If you actually vote for that favorite, instead of for your preferred frontrunner, you’re essentially helping your sworn enemy win. Or maybe you live in a district where almost everyone prefers the same party as you? Your vote is wasted on somediv who would win easily without you; you’d have more voting power if you moved to an area where the balance is closer to even.
Maybe you’re getting angry that I’m even using the phrase “third party”. Some people hate those words because they’re all just parties and privileging two of them is unfair. Others hate those words because giving any attention at all to anyone but the “correct” major party only helps the greater evil win. Either way, if you’re angry, you should be; but the problem is not me, but the voting method.
The upshot is that FPTP routinely wastes half or more of all votes. “Taxation without representation”? That’s how it works for most English voters.
And that’s just the voters! In addition, there are all the people who don’t even bother to vote, because choosing one of 2 impoverished options and then having a better-than-even chance their voice will be ignored anyway isn’t important enough to them.
What would a solution look like?
A perfect solution would:
- Keep wasted votes to a minimum — at most a tiny fraction of ballots should be unrepresented.
- Give voters meaningful choices; not just two options or a few parties.
- Keep voting easy.
- Keep accountability clear. Voters should be able to point to their representatives, and representatives should know who are their constituents. Parties should generally reflect the will of their voters.
- Be politically feasible.
- Favor representatives who can deal reasonably with each other on a range of issues, not those exclusively focused on all-or-nothing views on single issues.
Points 1, 2, and 3 would combine to boost turnout; points 4 and 6 would keep politics healthier; and point 5 would help us get there.
Does such a perfect solution exist? Yes! But before I describe it, I want to explain a few others that are good-but-not-perfect. In particular, I want to talk about proportional representation (PR).
Problems with past proposals
PR means voting methods that ensure that X% of the voters can elect about X% of the representatives. That means that by their nature, they are paying attention to all the votes; few wasted votes, so they meet point 1 above.
There’s over a hundred years of theory around PR methods, with dozens of methods and variants that have been proposed and tried over the years. The most popular PR methods today are list methods (which come in two flavors, open or closed); mixed member proportional (MMP); and single transferable vote (STV).
As I said, all of these methods would meet point 1 above, nearly eliminating wasted votes. How do they do on the other points?
No better than decent.
- All the widely-used methods are stuck in a tradeoff; either, like STV, they have meaningful choices for voters (point 2 above), or, like MMP and list methods, they keep voting simple (point 3 above), but not both. STV asks voters to rank all the candidates—often over a dozen serious contenders and dozens of also-rans—in order of preference; hardly simple. In MMP and list methods, once you’ve chosen a party, you have very little meaningful choice about which factions of that party you want to help or hurt.
- Clear accountability (point 4 above)? All of the existing PR methods give this up to some degree.
- Feasibility (point 5 above)? Well, on the bright side, New Zealand managed to reform their electoral system to use a PR method — MMP—in a process that began in 1992 and culminated in 1996. But imitating that success in the US, Canada, or the UK seems quite elusive. In New Zealand, it took support from both the governing party and the voters in a referendum to pass. But though there is clearly a major party in each of US/CA/UK which would benefit from a transition to PR, any of the PR methods I’ve listed would still be a serious threat to the actual incumbent politicians from that party; when it comes down to a conflict with self-interest on one side and the interests of both party and country on the other, too many politicians will choose self-interest.
- Governability (point 6 above)? MMP and list methods tend to push representatives to be relatively-interchangeable servants of their party; while STV tends to push them to splinter off into single-issue parties. None of these methods has the perfect balance of incentives so that parties remain reasonably large, with comprehensive platforms and appeal to various groups, but representatives still exercise independent judgment, without blindly following the party line.
GOLD: a better way
GOLD voting (I’ll explain the acronym later) is a PR method that was designed to be “perfect” in all the ways listed above. Essentially, it’s a kind of hybrid of STV and FPTP, with the meaningful choices of the former and the ballot simplicity of the latter.
Voting is mostly like FPTP: the ballot lists the candidates in your district/riding/constituency, and you can choose one. Many voters will simply do exactly that. But there are a couple of extra possibilities. First, instead of just choosing a local candidate, you may vote for any one candidate running in any constituency by entering their code and party; there’s a list of codes and parties for all candidates in each voting booth.
And second, you may choose one of two “transfer methods”, which determines how your vote is transferred if your chosen candidate does not win or if they get more than the minimum votes needed to win. Essentially, the transfer option turns your simple FPTP ballot into a full STV ballot, while still keeping the voting and the tallying simple. The two options are partisan transfers, in which your vote is split among same-party candidates; and delegated transfers, in which public, predeclared ratings provided by your chosen candidate are used to determine how your vote is transferred. If you don’t choose a transfer method, a reasonable default is used (partisan transfers if you chose a local candidate, and delegated if you chose a nonlocal one).
Then, an STV-like process is used to find a proportional set of winners, with the constraint that there must be exactly one winner per district. The weakest candidates are eliminated, and their votes transferred to stronger candidates, until there is the right number of strong candidates with 1 “quota” of votes each, where a “quota” is defined so that there is less than 1 “quota” of leftover, wasted votes.
(The acronym “GOLD” stands for “Geographic Open-List/Delegated”. “Geographic” because, like FPTP, it gives one winner per district; and “open list” and “delegated” for the two transfer options. Partisan transfers are called “open list” in the name because if every voter chose a local candidate and partisan transfers, GOLD would give the same results as the open list method.)
GOLD Example (skip to next section if you want to)
Say that voters in the great state of Triangle are electing 3 representatives to the national legislature. There are three parties: Urban and Rural (the major parties), and Belowground (the miner party). There is one candidate from each party in each district, so for instance candidate R2 would be the rural candidate in district 2. For simplicity, the same number of voters vote in each district, and all the voters for the same candidate from the same district use the same transfer method, indicated in parens: (p) for partisan or (d) for delegated. Here’s how the votes break down:
District 1: 45% U1 (d), 10% U2 (d), 25% R1 (p), 20% B1 (p)
D2: 40% U2 (p), 30% R2 (p), 20% R1 (d), 10% B3 (d)
D3: 30% U3 (p), 20% R3 (p), 50% B3 (d)
In other words, D1 is mostly Urban voters, but a notable fraction of them prefer the Urban candidate from D2; D2 is mostly Rural voters, but likewise some of them prefer the Rural candidate from D1; and D3 is mostly Belowground voters.
Since there are 3 seats being elected, a quota is anything over 25% of the total (that‘s 1/(S+1), where S is seats); or, in other words, 75% of a single district. Initially, candidates all but the top two in each district are eliminated. That leaves:
District 1: U1 45, R1 25+(20 from D2)+(10 from D3)=55
D2: U2 40+(10 from D1)=50, R2 30+(10 from D3)=40
D3: U3 30, B3 50+(20 from D1)+(10 from D2)=80
Now, B3 has more than a full quota of 75, so they win a seat. They have 5 extra votes, so 5/80=1/16 of each of the votes given to them are transferred. But the votes they got from other districts cannot be transferred, because they chose partisan transfers and there are no more Belowground candidates remaining. So, just over 3 transfer to the non-Belowground candidate they chose ahead of time as a favorite: U2.
When B3 wins, D3 already has a winner, so U3 is eliminated, giving 15 votes each to U1 and U2. Now we have:
District 1: U1 45+16=60, R1 55
D2: U2 50+15+3=68, R2 40
Now R2 is the weakest candidate. They are eliminated and their votes go to R1, making a total of 95. That’s plenty to win D1, so U1 is eliminated, passing votes to U2, and electing U2. So the final winners are R1, U2, B3.
Once winners are chosen, there’s an extra step to give better accountability: each party assigns each district where they didn’t win to be “extra territory” of one of their winning candidates. That way, even if your favorite party did not win locally, you are in the extra territory of some representative from that party. In this case, there’s just one winner from each party, so each has the other two districts as “extra territory”.
Note that even though U1 and R2 got more votes locally than the ultimate winners, R1 and U2 had more cross-district popularity, so they ultimately won. Just running in a “safe district” for your party doesn’t mean you can get away with being a below-average candidate for that party.
How does GOLD do on the desirable points I listed above?
- Proportionality: like any PR system, GOLD guarantees that only a small fraction of votes are wasted.
- Affinity: Since you can vote for any candidate from any district, you have the widest possible number of options. Even if there is no local candidate you are excited about, there’s probably some candidate somewhere else you like because of who they are or what they stand for. And even if your choice doesn’t win, as long as you chose “delegated” transfers, they get to exercise real political power; other candidates will have to speak to the issues they care about to court their transferred votes. So no matter where you live, you have a good reason to get out and vote.
- Simplicity: Yup, voting is simple; you can vote as simply as FPTP if you want to, and still be confident your vote won’t just be wasted.
- Accountability: As you saw in the example, there are no “safe seats”; candidates are always accountable to voters. And because of the “extra territory” mechanism, just about every voter can point to a representative, from a party they sympathize with, who will listen to them.
- Political feasibility: True, GOLD is a new proposal, with less of a track record than MMP or STV. But on the plus side, its similarities to FPTP make it less of a threat to incumbent politicians, helping its chances to pass.
- Governability: Because no candidate can win without making at least a reasonably-strong second-place showing in their local area, the impetus is to form reasonably big-tent parties with comprehensive platforms, not to splinter into tiny extremist parties that only win seats by getting all the single-issue voters spread across the state.
So GOLD may not be exactly perfect, but it’s close enough.