Curt Eaton is an Emeritus Professor of Economics at the University Calgary, and a Fellow of the Royal Society of Canada. He resides in Kelowna.

Regrettably, in the coming federal election we will still be lumbered with first-past-the-post. That system works well when there are just two candidates. Voters have no trouble figuring out what to do — vote for their preferred candidate. And collectively, their behavior produces the democratic result — the candidate preferred by the majority wins.

But in other circumstances, the democratic outcome is not guaranteed. The naming of Thunder Bay illustrates how things can go wrong. In 1970, the city was created by the merger of Fort William, Port Arthur and three regional municipalities. The name of the new city was chosen in a first-past-the-post election with three names on the ballot – Thunder Bay, Lakehead, and The Lakehead. Thunder Bay got 40% of the vote, Lakehead 39%, and The Lakehead 21%. So, the new city became Thunder Bay.

Clearly, if voters had been given just two choices —Thunder Bay and either Lakehead or The Lakehead — Thunder Bay would have lost by a big margin. In this sense, it was undemocratic to name the city Thunder Bay.

Now let’s look for parallels in the last federal election in British Columbia. Is it possible that some of British Columbia’s MPs would have lost if they had run against the second-place candidate in a head-to-head contest? Or, to put it another way, is it possible that the outcome was undemocratic in the same way that the choice of Thunder Bay over some version of Lakehead was undemocratic?

Here are the facts. Forty-two MPs were elected in British Columbia in 2015: Liberal 17, NDP 14, Conservative 10, and Green 1. Eight were elected with more than 50% of the vote, and the legitimacy of these eight elections can’t be disputed.

But 34 MPs were elected with less than a majority of votes cast—just as Thunder Bay was the name chosen by just 40% of voters in the election to name that city. If these 34 MPs had instead run against the second-place finisher in a head-to-head contest, would they still have been elected?

We can’t say for certain, but we can get a sense of the possibilities by looking at how many votes went to the “also-rans,” the candidates who failed to come in first or second. In 20 of these 34 ridings, the also-rans got more than a quarter of the votes, and in 8 they got more than a third. What might have happened if the voters for the also-rans had instead been asked to choose between the first- and second-place finishers? In 12 ridings the second-place finisher would have won with 60% of the also-ran votes, and in five the second-place finisher would have won with 55% of those votes. If these voters had voted instead for one or the other of the first two finishers, their votes might very well have changed the outcome of the election. Their votes, collectively, were effectively throw-away—they just didn’t make any difference in the outcome.

There is a real sense in which the voters for also-rans were effectively disenfranchised.

In the 20 ridings where these disenfranchised voters accounted for more than a quarter of the votes, can the MPs legitimately claim to represent their constituents? We don’t think so. That’s the trouble with first-past-the-post—it calls into question the very legitimacy of our elections.

Unfortunately, because the politicians we elected could not agree on electoral reform, we are stuck with first-past-the-post for the upcoming election. So, voters need to take matters into their own hands. How can they solve the problem of effective disenfranchisement? The answer is by voting strategically.

What does that look like? First, voters must realize that they could be disenfranchised if they simply vote for their favourite candidate. Then they need to try to figure out which two candidates are likely to be the frontrunners in their riding. Finally, if neither of the two frontrunners is their favourite candidate, they need to hold their nose and vote for the frontrunner who looks best to them. This is the only reasonable course of action if the voter only cares about who wins the election.

Of course, voters may very well care about more than just who wins. For example, they may want to “send a message” that the positions taken by their favourite candidate deserve attention. But even in this case, it’s wise to weigh the value of their message against the probability that by sending it, they help to elect the candidate they like the least.

That, in a nutshell, is strategic voting. It solves, imperfectly it must be admitted, the problem of effective disenfranchisement, and it reduces the likelihood of an undemocratic outcome, one in which the winner could not defeat the second place candidate in a one-on-one election.