In 2015 and before the election in 2016, hardly anyone believed Donald Trump would become president, not even Donald Trump. But he did and a wide variety of explanations have been offered for this turn of events. The media was made this happen by chasing ratings; they repeated endlessly everything Trump said or did, making Trump ever more famous. But no, the reason was really the Constitution which gave outsized political power to rural states. Or no, the problem was the lazy or ignorant voters. But no, it was the Green Party and Jill Stein. Oh wait, it was surely because of Putin.
There is merit to these rationales, but there is something more fundamental at play. Even though the polls showed Trump (and Clinton as well) had more opposition than support these became the nominees of the two major parties, giving both a weak opponent. How can candidates win an election with more opposition than support? A hint is in another question. Why don’t we even ask voters which candidates they oppose? This is a fundamental problem with how we vote. And how we vote is surely fundamental to our democracy.
In the 2015 Republican primaries, there were generally more than just two candidates. This is important because it is so well understood that when there are three or more candidates, vote-splitting might easily corrupt our election outcomes. This flaw in the primary elections was key to how Trump could win the Republican party nomination. Sadly, we may now experience such a failure yet again.
The plurality voting system that generally we use in the United States (and even in most of the world) only allows a vote to specify one single candidate to support. That is sufficient so long as there are just two candidates; that is what gives us the rational for clinging to our two-party system. That is why we make it so difficult for third parties to gain ballot access. Allowing for only two points of view may help to make plurality voting more tolerable but it is no way to let a democracy flourish.
But when, as occasionally even happens even in our general elections, there are more than just two candidates, plurality ballots cannot possibly paint a clear picture of voter sentiment. Even if all voters tried to be honest in identifying their favorite candidate, the election could not take account for any degree of willingness to compromise among several top candidates. And other voters may be primarily worried about the possibility of electing a candidate who seems to be a terrible choice. These voters may choose to lie for strategic reasons, perhaps choosing randomly to support one of several other candidates. It becomes impossible to determine from a ballot what the voter was trying to express. Perhaps it genuinely expresses support for a candidate, but maybe it was meant as a stand in opposition to some other, unidentified candidate. Can we really claim that the winner is the true choice of the electorate? This is a process of course, but this is a pitiful way to choose someone for a powerful position; surely, a better way must be possible.
Nonetheless, media pundits and successful politicians tend to talk as if elections were flawless. We hear phrases like, the voters have spoken, suggesting that voters are even capable of speaking clearly. Unfortunately, with plurality voting they do not enjoy that luxury unless elections are limited to only two candidates. But what kind of democracy could tolerate a voting system that only allows for two candidates?
As mentioned before, with plurality voting, a voter chooses just one candidate to vote for. There are two fundamental problems hidden within this very short description. One is that with plurality voting, voters must choose just a single candidate and express an opinion about only that one candidate. The second problem is that the only opinion the candidate can express is of support for that candidate. With several candidates running for a single office, there may well be some candidates that a voter feels ambivalent about but there will likely be some that the voter finds to be unacceptable. And as well as one or more candidates that the voter supports. Is there some reason that opposition is not just as important as support? Is it not possible for a voter to fervently want to oppose one or more of them but to support none of them? Is it reasonable to force such a voter to feign support for some candidate that in fact the voter does not truly support? Would it not be better to allow the voter to explicitly say so when they oppose a candidate?
Even though plurality voting fails to allow a vote of opposition, it is not uncommon to hear a voter say they plan to vote against a candidate. They no doubt would like to vote against that candidate but the closest they can come to this is to choose some other candidate to support with their vote. That action amounts to a feeble attempt at an opposition vote, but their words reveal what they prefer to do.
A sensible voting system should address both of these fundamental problems and there are several voting systems that do that. Simplest among them is a system that has been adopted in Latvia. Some call it balanced approval voting, but the system also goes under the name yes/no/abstain voting or sometimes as combined approval voting . In this voting system, a voter is provided with the list of candidates and asked to indicate, for each candidate, either support or opposition; quite necessarily, a voter is allowed to skip over some candidates who they neither support nor oppose. The net-vote for a candidate (which could be negative) is the number of support votes for that candidate minus the number of opposition votes and the winner is then identified as the candidate with the largest net-vote. Notice how unlikely it would be for a candidate with a zero or negative net-vote to win election. Though the zero or negative possibilities make clear information that is helpful, some people dislike the possibility of such outcomes. The implementation of this system in Latvia avoids the use of negative numbers by adding the total count of voters to the net-vote for their vote tally.
It is illuminating to consider why a voter might abstain. It would likely happen when the voter really does not know enough about the candidate to have strong feelings one way or the other. For example, a voter might feel unable to either support or oppose a candidate if the voter does not even recognize the name. But conversely, votes of support or of opposition for a specific candidate will be drawn from the pool of voters who are reasonably informed about the candidate. This will be most voters when the candidate is famous.
A famous candidate will generally have more votes of support but also more votes of opposition than will another candidate who is relatively unknown simply because the famous candidate will have fewer abstentions. Any voting system which tallies only votes of support while failing to compensate for votes of opposition as well, will bias its elections unfairly in favor of a very famous candidate. This is a failure of plurality voting, but also of most voting systems that are widely discussed. This may also be what accounts for Trump’s nomination in 2016.
Another virtue of balanced approval voting (BAV) is that it mitigates against polarization. An easy way to see this is to imagine a political system in which there are only two parties of roughly equal size; let’s call those parties R and D. And imagine that the R’s consistently support the R candidates but consistently oppose the D candidates. And further imagine that the reverse is true, that the D’s support the D candidates but oppose the R candidates. With BAV the net-vote for both the R and D candidates will be at or near zero. This allows a third-party candidate or an independent candidate who is widely supported by those relatively few voters who are informed about the candidate, a realistic possibility for winning election.
Would that be such a bad outcome for democracy? In subsequent BAV elections, voters would no longer dismiss third parties so easily and neither could the news media. A multi-party system with ideas from many points of view being considered and discussed could finally emerge. Less famous candidates would gain more access to media and so earn fewer abstentions.
I should now add that there are some difficulties with presidential elections, not only for introducing BAV but for adopting any new voting system. The problem lies in the Electoral College which presumably would continue to use plurality voting. But that should not prevent us from improving the great many other elections, importantly even including even the primary elections for presidential nominations.
