Roy Minet

Roy Minet

Plurality voting is widely used in most elections. It’s simple: The candidate with the most votes wins.

Unfortunately, that’s too simple. Around the time of the American Revolution, two French scholars, Jean-Charles de Borda and Marquis de Condorcet, pointed out serious flaws and kicked off the quest for a better voting method.

Plurality’s first problem is that it doesn’t allow voters to provide enough information to reliably identify the correct winner in many elections. More than just each voter’s first choice of the candidates is required to consistently do a good job.

Plurality’s second big problem is that the first choice information it does collect is often bogus. Voters lie. All too often, there are two very polarizing “main” candidates. Many voters do not really like either, yet they feel strongly compelled to insincerely “vote for the lesser of the two evils,” even when they may actually prefer some other candidate. Worse, this effect is so powerful that most voters don’t even bother to learn about other candidates who may be running.

To see how a better voting method would work, let’s start with the simplest possible situation. We just wish to decide whether or not to change our neighborhood’s trash pickup day from Wednesday to Thursday. The obvious solution is to hold a referendum wherein each affected party can vote either “yea” or “nay.” If there are more “yeas” than “nays,” the change is approved. Each voter gets to weigh in either for or against. This is a very straightforward, simple and serviceable decision mechanism.

When one candidate is being chosen for an office from several candidates on the ballot, what works really well is to hold a separate referendum on each of the candidates. The candidate who wins his or her referendum by the largest majority is then the overall winner. Voters are able to weigh in both for and against the candidates, just as they did on the trash collection issue.

It has turned out that the magic ingredient that enables making better decisions in all kinds of elections is to allow voters to express both their approval and their disapproval — with approvals and disapprovals for each candidate offsetting each other — just as they do in any referendum.

Especially note that candidates with “high negatives” have a much harder time winning, thus strongly encouraging the nomination of candidates more broadly acceptable to most voters.

Each voter’s choice of the best candidate and the worst candidate are the most important pieces of information, enabling the correct winner to be identified. Thus, voters may vote “yea” in the referendum of the candidate they think is best and “nay” in the referendum of the candidate they think is worst. Allowing additional inputs opens the door to “strategic” or “garbage” data that can only degrade decision-making.

Plurality’s first big problem is now well solved, but one additional feature is necessary to prevent strategic voting for the lesser evil.

If voters fear that their best choice may not win and are tempted to instead vote for the lesser evil, they may designate an “alternate best choice.” An “alternate” does absolutely nothing unless and until the candidate marked “best” is eliminated. Candidates are eliminated one by one, always eliminating the weakest candidate, until only the strongest (the winner) remains. If a voter’s “best” candidate is eliminated, the “alternate” candidate (if any) is thereafter counted exactly as though it had originally been marked “best.” Votes are retotaled after each elimination.

The acronym BAWV (Best/Alternate/Worst Voting) is descriptive of this method. Some details were omitted because of space limitations, but they are available at royminet.org/voting-elections.

During the 250 years since Borda and Condorcet, a huge amount of effort has gone into finding the best voting method. Hundreds have been proposed. But the problem is a whole lot trickier than anyone expected it to be. What has recently resulted in rapid progress is the power of modern computing. By simulating millions of elections, various voting methods can be evaluated and compared in all possible types of elections.

It is crucially important that voters choose the best politicians to wield awesome government power over all our lives. It’s impossible for voters to reliably make good choices using plurality. Expert consensus is that plurality is a truly awful voting method. Most think it is a contributing factor to increasing polarization which is reaching uncomfortable, perhaps dangerous levels. It is quite important to switch to a greatly improved voting method as soon as possible.

It is not certain that BAWV is the best possible voting method. But it very well might be the best one that is practical — meaning acceptably simple and understandable. We do know that BAWV is very significantly better than all other methods evaluated. We also know that it is so good that, even if a better method is found, it couldn’t be a whole lot better.

The critical question now is how long will it take to get BAWV into service.

Roy Minet, a Mount Joy resident, was the Libertarian Party’s nominee for Pennsylvania auditor general in 2016.