With the general election less than two weeks away, and with the Liberals and Conservatives still tied in the polls, turnout could decide the result. Here’s why:
In 2011, when Stephen Harper’s Conservatives won a majority government, 61 per cent of eligible voters cast a ballot. But in 2015, which delivered a Liberal majority government for Justin Trudeau, turnout shot up to 68 per cent, a level not seen in more than two decades. The reason, according Elections Canada, was increased participation by younger voters.
To study this phenomenon, I looked at seven ridings, chosen more or less at random, within the 905, the band of suburban ridings surrounding Toronto. All seven went Conservative in 2011, then switched to the Liberals in 2015. (Thanks to Alice Funke of Pundits Guide for providing 2011 results redistributed to 2015 boundaries.)
In every riding, the number of people who voted Conservative in 2015 was about the same as in 2011. In a few cases it even increased slightly. Yet still the Conservatives lost. Why?
There were two reasons. First, the NDP vote in every riding declined between 2011 and 2015, sometimes by as much as half. It’s reasonable to assume that NDP deserters migrated to the Liberals.
More important, the total number of votes cast in each riding in 2015 increased, and these new voters were Liberals. In Burlington, the Liberal vote more than doubled. In Whitby it quadrupled.
Bottom line: In 2015, the Conservative vote held, but was overwhelmed by new voters and NDP defectors voting Liberal.
What does that mean for those ridings in 2019? Since the Conservative vote has been stable for two elections, and since there has been no Tory surge in the polls, we can reasonably predict that the Conservatives will get the same level of support this time out.
Whether those ridings turn Conservative or remain Liberal will depend on whether the voters who came out to vote for Mr. Trudeau in 2015 come out again.
The good news for the Liberals, says Richard Johnston, a political scientist at the University of British Columbia who is one of Canada’s leading authorities on voting behaviour, is that the most reliable indicator of whether someone voted in any given election is whether they voted in the previous one. "This means that people who came out in 2015 are likely to come out again this time,” he believes – though he adds this doesn’t mean they’ll vote Liberal.
For pollsters, accurately assessing who will turn out to vote is crucial to predicting election outcomes. The approach used by Darrell Bricker at Ipsos Public Affairs is to create a portrait of the civic literacy of those being questioned. (Mr. Bricker and I have written two books together.) Respondents are asked whether they know the date of the general election, whether they know the location of their polling station, and how closely they are following the campaign.
They are also asked behavioural questions, such as: How do you feel about elections in which you don’t vote? How would not voting this time make you feel?
Based on each respondent’s knowledge and attitudes, Ipsos assigns a score, and then runs turnout models and seat projections based on such-and-such a percentage of voters turning out, which means the intentions of those with a certain score or higher are counted.
But what will that percentage be? “It’s the hardest thing for us to work out,” Mr. Bricker says. He believes that soft Liberal support could suppress turnout below 2015 levels.
Peter Loewen, a professor of political science at the University of Toronto, believes that the turnout in 2011 was a low-water mark, with unpopular party leaders and a disaffected public, while 2015 was a high-water mark, especially among younger voters. He suspects turnout this year will be somewhere in between.
The bad news for the Conservatives, he says, is that they been unable to grow their vote under Conservative Leader Andrew Scheer. The bad news for the Liberals is that young people who voted for them in 2015 might stay home or that they might show up and vote NDP. “For the Liberals, it’s a double-edged sword,” he said.
But when it comes to predicting turnout, he adds, best guesses remain just guesses. “It’s a mug’s game.”