Be Careful Strategic Voting: A Look at Where Strategic Voting Matter and Where It Does Not

As the 2015 election enters its last couple of weeks, the tight race between the Liberals and Conservatives has many progressive voters considering voting strategically in order to keep the Conservatives out of power. In my last post I wrote about how strategic voting might have long term costs for strong supporters of the Liberals or New Democrats, in this post I will look at how first past the post systems can complicate strategic voting. Because parties have different levels of strength in different ridings, strategic voting for progressives is not simply a matter of selecting the party on the left most likely to win the most votes nationally. The Liberals may be well ahead of the NDP in national polls but there are still several ridings where the NDP have a much better change of beating the Conservatives than the Liberals do. There are also a large number of ridings that are safe and where strategic voting will make little difference to the election outcome, and a handful of ridings where the Conservatives have little chance of winning but where the Liberals and the NDP (or NDP and the Bloc Quebecois) are in a close race. Strategic voting requires careful consideration and research into the kind of riding that one is voting in. In this post I will look at several different types of ridings and the strategic voting incentives in each of them.

In this post I try to give a rough account of the number of ridings that fall into each category. I have estimated this by using Three Hundred Eight blogger Éric Grenier’s riding projections. I use projections that were made with data from October 9th because they were the most recent posted at the time of writing. As the polls change the riding projections and the number of ridings that fall into each category will change. The estimates for the number of safe and competitive seats in this post are rough estimates and one should expect some fluctuations in these numbers over the next week depending on the extent to which polls change. I consider a riding safe if the party leading has at least a 75% chance of winning. Ridings are considered multi-party races if the high-end of the predicted range for the third place party is very close or within the low-end of the second place party’s predicted range.

Projections for individual ridings should be taken with a grain of salt. Because the projections that Grenier does are based on national and regional level polling they have a limited ability to take into account riding specific factors (such as the presence of a star candidate). Three Hundred Eight is a valuable resource for determining the range of seats that a party can win and in most ridings is the best reference point available for determining the competitiveness of a particular riding, but for the ridings in which riding-level polling has been done, that data is likely better for determining which parties are competitive.

Safe Seats

The majority of seats in this election are safe seats. There are about 217 seats in Grenier’s projections in which one party is 75% likely to win. In these seats strategic voting is of little value. Because the outcome in these seats is not in question, there is little value to changing one’s vote in order to either take the riding from the Conservatives or to keep the Conservatives from winning power. There may be value, however, in indicating the strength of each of the parties in the riding. In these situations voting one’s conscience is likely the best thing one can do. A surprisingly strong showing by a minor party in a riding can lead voters to see that party as competitive in future elections and might lead the party to benefit from strategic voting in the future. Such a vote may not impact the outcome of this election, but it could matter to future elections. A vote for a non-competitive party can also improve its support nationally. Strong showings nationally can indicate to all party’s the importance of adjusting their policies in order to adjust their policies to try to win the support of that parties’ voters. For example, if the Green Party wins a large number of votes other parties may adopt stronger environmentally policies in order to take votes from it. Because strategic voting is unlikely to change the party that wins a safe seat, voters in these ridings are better off using their vote to send signals to parties about their policy preferences than to trying ensure the non-Conservative candidate wins.

Competitive Races

A significant number of seats in Canada are also projected to be races between non-Conservative parties. There are about 43 ridings where the Conservatives are projected to be well behind the progressive parties, but in which some combination of the Bloc Quebecois, Liberals, and NDP are in a close race to win the seat. Because there is little chance of the Conservatives benefiting from split voting on the left in these ridings, strategic voting provides little benefit to progressives trying to keep the Conservatives out of power. The races in these ridings are far more about determining which of the Liberals or NDP emerges as the strongest competitor the Conservatives than they are about keeping the Conservatives out of power.

There may also be value to environmentalist voters supporting the Green party in these ridings. A Liberal loss to the NDP in a riding with strong Green support may force the Liberals to rethink their environmental policy. The same may be true for the NDP if the Greens are strong in a substantial number of ridings that they lose to the Liberals. Voting Green as a way to push parties to be stronger on the environment comes at less of a cost in these ridings than it ones in which the Conservatives are competitive, because doing so carries little risk of contributing to a Conservative victory.

Multi-Party Races Where the Conservatives are Competitive

The number of seats projected multi-party races in which the Conservatives are competitive is small. There are approximately 11 seats projected to fall into these categories. In principle strategic voting in these ridings can make sense as a way to keep the Conservatives out of power, but the problem is that the progressive alternative most likely to beat the Conservatives is unclear. When the Liberals and NDP (or NDP and Bloc Quebecois) are as strong as each other it can be very difficult to strategic vote effectively. In these races vote splitting could very well hand the Conservatives a victory, but coordination on the left to prevent this from happening is difficult to impossible. One may be best off voting for one’s preferred party and accepting that there is a danger that the Conservatives could benefit from split voting and that there is not a lot that progressives can do about that.

Clear Two-Party Races Involving the Conservatives

This leaves around 67 ridings that are projected to fall into this category. 43 of these are projected to be Conservative-Liberal races, 23 are projected to Conservative-NDP races, and in Fort Saskatchewan-Sherwood Park an independent (ideologically conservative candidate) as an outside chance of defeating the Conservative one. In these ridings strategic voting by progressives can be an effective way of keeping Conservatives out of power. Voters have to be careful to pay close attention to which of the Bloc Quebecois, Liberals, or NDP is best placed to defeat the Conservatives. The Liberals might be leading the NDP in national polls, but that does not necessarily mean they are best placed to defeat the Conservatives in any given riding. “Strategically voting” for the progressive party leading the national polls in a riding where the nationally third place party is running second could actually contribute to a Conservative victory in that riding. Strategic voters in one of the close two-party races involving the Conservatives should pay careful attention to Three Hundred and Eight’s projections as well as organizations such as Vote Together that have been doing riding level polling in some of the ridings in which the Conservatives are vulnerable.


Progressives’ concern over the potential for a Conservative government has led to a great deal of discussion of strategic voting. In some ridings strategic voting can play a role in keeping the Conservatives out of government. In a large number of ridings, however, strategic voting has a limited ability to influence results. In only around 67 (or 66 if Fort Saskatchewan-Sherwood Park is not counted) of 338 ridings is there a clear progressive candidate whom strategic voters can flock to in order to defeat the Conservatives. That works out to about 20% of Canada’s ridings. That number will fluctuate over the last week of the election, and it is not insignificant. At the same time, most Canadians will find themselves in ridings where strategic voting has limited value. These ridings are either safe ridings, races between the progressive parties, or races in which it is not clear which of the progressive parties has the best chance of beating the Conservatives. Voters should be very careful about the way that they approach strategic voting in this election.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s