Strategic Voting

By Jean-François Daoust (Université de Montréal) and Damien Bol (King’s College London)

What Is The Story?

The concept of strategic voting is widely used by political parties and the media. It is assumed to be a widespread behaviour because Canada has a “winner-takes-all” electoral system. But what is the actual proportion of strategic votes?
To tackle this question, we need to define rigorously what a strategic vote is. There are two very simple conditions for a vote to be qualified as strategic: first, a voter must not vote for her preferred party, and second, she must do so because her preferred party has little or no chance of her constituency.


Data from the Making Electoral Democracy Work (MEDW) surveys are well suited to analyse the extent to which people vote strategically as they provide indicators of individual preferences, expectations about the various parties’ chances of winning in the respondent’s constituency, and vote choice.
Preferences were tapped by asking people how much they like each party on a scale from zero to ten (from ‘really dislike’ to ‘really like’). For expectations, respondents were asked to evaluate the chances of the various parties winning in their constituency, also on a scale from zero to ten (from ‘no chance at all’ to ‘certain to win’). Finally, vote choice is measured by a self-reported answer in the post-electoral wave.

We use data from two provincial elections and the last federal election.

The Method

The method has two steps. First, we identify respondents who were potential strategic voters, i.e. those whose preferred party was not one of the top two contenders (the two parties perceived to have the highest chances of winning) in their constituency. Second, we identify within this subgroup of potential strategic voters those who voted for their most preferred party among the top two contenders. These voters casted, following our definition, a strategic vote.


Table 1 displays the extent to which Canadians voted strategically.



Note: Entries are non-weighted proportions. Weighted proportions using actual results for each political party are into parentheses.

Firstly, less than one fifth of the electorate had incentives to cast a strategic vote (i.e. were potential strategic voters). Among those who did have incentives to do so, around 40% desert strategically. At the end of the day, around 8% of Canadians voted strategically in the three largest provinces in the last federal election. We see a relatively substantial increase in Ontario (compared with the provincial election). There appears to be a slight decrease in Quebec.


Measuring strategic voting requires indicators of preferences, expectations and vote choice. In order to ascertain how many citizens act strategically in an election, we must focus on the subgroup of voters with incentives to vote strategically and then analyse if they desert their preferred option in favour of the most preferred party between the top two contenders in their constituency. When we do so, we find that about 20% of voters face a situation where they have to decide whether to vote strategically or not and that about 20% of them (and 8% of all the voters) cast a strategic vote.

This entry was posted in Blog. Bookmark the permalink. Both comments and trackbacks are currently closed.