(Version française disponible ici)
The April federal election outcome was dramatically different than what many had been predicting – in fact, assuming – only months earlier. What was expected to be a coronation for Conservative Leader Pierre Poilievre instead turned into a hotly contested, close race that led to a new minority Liberal government.
There was one key difference between the parties. The Liberals had a new leader, Mark Carney, who won a sweeping victory in a party leadership race after his deeply unpopular predecessor Justin Trudeau announced in January he was stepping down.
Carney seemed like a game-changer. Coupled with a rapidly shifting international environment caused by U.S. President Donald Trump, his leadership appealed to enough Canadians to earn the Liberals even more seats in the April election than they held prior to the dissolution of Parliament.
That leaders can play a role in their party’s success or failure is not a new idea. From classic to more contemporary models of voting behaviour, leaders, along with party identification and issue preferences, are always listed among the most important factors in predicting voting.
Significant research has focused on developing a better understanding of what aspect of leadership affects voters’ preferences the most – competence or personalities. Other work has considered how the influence of leaders may be growing over time – the presidentialization thesis.
Yet while leaders have an influence and indeed are often pivotal to vote choice, their actual substantive impact on an election outcome can be minimal. Sometimes, as with Carney, changing leaders really helps a party. Sometimes, as with John Turner following Pierre Trudeau in 1984 and Kim Campbell following Brian Mulroney in 1993, it does not.
This leads to a question: Can leaders actually make as much of a difference on voting as it seems to a casual observer? Or is there a disconnect between perception and reality?
We think the impact of leaders on party fortunes (their vote share) really comes down to the way they influence the party brand. Drawing on the idea that party brands matter for attracting and retaining supporters, we argue that leaders shape how voters view the parties themselves.
Returning to the case of Carney, the platform put forward by the Liberals in 2025 was starkly different than the one Justin Trudeau introduced when he was first elected in 2015. “Sunny ways” gave way to the perception of Carney as a prudent manager.
We hypothesize it is at least partially through this impact on their party’s brand that leaders influence vote choices. To test this, we turn to data gathered as part of the Democracy Checkup run by the Consortium on Electoral Democracy from May 21 to June 12, 2024.
In particular, we posed hypothetical questions to more than 1,000 respondents: how would they rate the political parties (Liberal, Conservative, NDP) on a scale of 0 to 100 if they had different leaders?
For the Conservative Party, we asked about Leslyn Lewis and Jean Charest, who had both run for the leadership against Poilievre in 2022. For the NDP, we asked about Charlie Angus and Nikki Ashton, two of the party’s more prominent MPs. For the Liberals, we asked about Carney (in a nod to rumours circulating at the time of our survey in 2024 that he might run for the party) and Chrystia Freeland, deputy prime minister from 2019-24.
We focus here on voters outside Quebec because the strength and regional nature of the Bloc Québécois (along with the Quebec-focused issues it tends to champion) complicate comparisons among choices between national parties and candidates across provinces.
We weight the results by age, gender and province. We can compare these results to ratings of each party with its actual leader at the time to see how much of a difference a hypothetical leader change – even without a corresponding change in party platforms – might make.
Figure 1 shows the effect of changing leaders would be substantial, in many cases leading to a significant change in the overall rating of a party. But, the effects vary in direction.
The Conservative and NDP parties were rated the highest under their then-leaders. The Liberal Party, on the other hand, fared its worst under Trudeau and the party would have been rated significantly higher – by 4.5 points – if Carney had been the leader at the time. (When the data were gathered, Carney was not yet a serious contender for the leadership).
These results are based upon respondents of all parties and include non-partisans – those who indicated they did not identify with any party.
However, for electoral outcomes to change as dramatically as they did in 2025, parties must earn (or lose) support from outside their traditional base.
To see if there is evidence of this with a hypothetical leader change, we break the results down by the partisanship of respondents. That is, we consider the effect of a leader swap on a party’s supporters, but also on the supporters of other parties and non-partisans.
Figure 2 shows the results for the Liberal Party. Its supporters preferred Trudeau over either Carney or Freeland, but that wasn’t the case for Conservatives or non-partisans, who had a strong preference for the party under Carney, or NDPers who preferred Freeland (but not significantly more).
It is notable that Liberal ratings almost doubled among Conservative supporters in this hypothetical 2024 world with Carney as leader. These data suggest that the switch from Trudeau to Carney could have led to lower ratings by Liberal partisans and perhaps a loss of support from the party’s base, while at the same time increasing its appeal to other voters.
Figure 3 shows the Conservative Party would be rated much higher by Liberal and NDP supporters if Charest were leader, but much lower by Conservative partisans. There was no significant effect for non-partisans, unlike the Liberal results in Figure 2. Conservative partisans preferred Poilievre.
Figure 4 shows NDP supporters rated their party the highest when led by Singh. Liberals and non-partisans also preferred Singh while Tory supporters preferred Angus, although the magnitude of the differences was relatively muted among the non-NDPers.
These results confirm our contention that leaders can greatly affect how a party is rated by its own supporters and those who prefer other parties, too.
This is particularly important. In a close election such as 2025, a party that can appeal to other parties’ supporters has a huge advantage, despite a potential significant decline in support from its own base.
Our results suggest that even a year before the 2025 election, there was already evidence the Liberals might perform much better among some opposition supporters and non-partisans under Carney specifically. Although such a change was associated with potential losses among Liberal supporters, it was offset by a significant gain in support from others.
On the other hand, a change in the Conservative leader would have had the opposite effect – moderate gains among opposition supporters and significant losses amongst Conservatives. For the NDP, a change would have only harmed the party among its followers without attracting new voters.
While a tiger might not be able to change its stripes, it seems that a political party can get more flexibility if it gets a new leader, at least in some cases.
