Wednesday, April 23, 2014

Conservatives ahead in new Angus Reid poll

For only the third time in the last year and out of the last 49 polls, the Conservatives have placed first in a national survey. Seasoned poll-watchers know that when something so out of the ordinary has happened, it is usually an anomaly that is quickly corrected. But until a new poll emerges to confirm or refute the findings of Angus Reid Global, we can only speculate as to whether this is a blip or a sign of things to come.

Angus Reid was last in the field in early March. Since then, the Conservatives picked up four points to lead with 32% support, putting them two points up on the Liberals, who fell three points to 30%. The New Democrats were down one point to 26%, while the Greens and Bloc Québécois were unchanged at 6% and 5%, respectively.

As I wrote in my piece for The Globe and Mail yesterday, these are unusual numbers. They are not completely implausible, though, as a swing of three or four points between the Liberals and Conservatives is not enormous. Nevertheless, these sort of numbers have been extremely rare over the last 12 months, as the chart below shows.

Polls since April 2013, with Angus Reid highlighted
Each dot represents a poll that has been released since Justin Trudeau became Liberal leader. The poll highlighted in green is this Angus Reid survey. As you can see, the Conservative result is higher than any recent survey. The Liberal result is lower than any recent survey. Is it a wobble or have the Liberals really taken a hit to the advantage of the Tories? We will have to wait and see.

Angus Reid also reported support among likely voters, which they determine based on a respondent's historical likelihood of voting and past voting behaviour. This is a good model for estimating likely turnout if turnout does not change to a great degree. If the assumptions the model is based on end up being off (if, say, young people turn out in larger numbers in 2015) the estimate will be wrong. This happened in the last US presidential election for a few pollsters. They assumed that the voting patterns of 2008 were anomalous and that turnout would revert to pre-2008 patterns. That did not happen, and so some pollsters, for instance, told Mitt Romney that he was going to win.

Nevertheless, according to Angus Reid's estimation of likely voters, the Conservatives have the support of 34%, a gain of three points since March. The Liberals dropped three points to 29%, while the New Democrats were up one point to 27%. At the regional level, the major changes were that the Tories placed first in the Prairies (42% to 34% for the Liberals) and extended their lead to 17 points in Ontario (43% to 26% for both the Liberals and NDP). The NDP's advantage in Quebec also increased (36% to 30% over the Liberals).

That the Conservative lead grew among likely voters is in line with what Angus Reid has found to be the case in the past. In February, a five-point national lead for the Liberals among all voters turned into a one-point deficit among likely voters. In March, that lead of five points decreased to just one.

Among eligible voters in this new poll, the Conservatives led in Alberta and Ontario. They had 57% support in Alberta, against 20% for the NDP and 17% for the Liberals. The Conservatives led with 41% support in Ontario (a gain of nine points) while the Liberals dropped eight points to 28%. The NDP was at 23%.

The surge in Conservative support in Ontario explains most of the gain the party made nationwide. It is what put them in front. But here again the numbers are unusual, as the Conservatives haven't been pegged at over 40% or with a lead of 13 points or more in the province in over a year.

The Liberals led in the Prairies and Atlantic Canada, with 36% support in Saskatchewan and Manitoba against 33% for the Conservatives and 18% for the NDP. In Atlantic Canada, the Liberals had 57% support to 21% for the Tories and 17% for the NDP.

The New Democrats led in British Columbia with 34%, a gain of seven points, followed by the Liberals at 31% and the Conservatives at 26%. In Quebec, the NDP was at 33% to 29% for the Liberals, 19% for the Bloc, and 14% for the Tories.

With these regional levels of support among all eligible voters, the Conservatives would likely win 139 seats, with 100 going to the Liberals, 96 to the New Democrats, two to the Bloc, and one to the Greens.

The Conservatives are far from majority territory due to their poor showings in British Columbia, the Prairies, and Atlantic Canada. The Liberals are hamstrung due to their distribution of the vote in Quebec, while 71% of the NDP caucus comes from either B.C. or Quebec.

Among likely voters, the Conservatives get much closer to a majority with 157 seats, with 98 being won by the NDP, 82 by the Liberals, and one by the Greens.

Angus Reid included some leadership questions in its poll. Stephen Harper topped the list for best Prime Minister with 27%, up four points from March, while Trudeau fell four points to 20%. Thomas Mulcair was at 16%, Elizabeth May at 4%, and André Bellavance at 1%.

Mulcair had the best approval rating of the leaders, at 46% against 35% disapproval. Trudeau had 45% approval to 44% disapproval, while Harper split 38% to 55%.

Comparing these numbers to Angus Reid's March poll is problematic. That poll showed increases of about 10 to 30 points in the number of 'unsures' compared to the previous poll in February, while this current poll shows a reset, with the number of 'unsures' dropping by 16 or 17 points for Mulcair and Trudeau and by 21 to 26 points for May and Bellavance. I'm not sure why that happened, but it makes it wiser to compare the approval rating results in this April poll to the last set of numbers in February.

When we do that, we certainly see that Trudeau has taken a step backwards. His approval rating fell by six points, while his disapproval rating increased by six. Harper and Mulcair experienced only insignificant shifts. When asked if their opinion of the leaders had changed in the last three months, 30% said that their opinion of Trudeau had worsened, against 21% who said it improved (those numbers were 41% to 8% for Harper and 11% to 17% for Mulcair, respectively). So, the drop in support for the Liberals nationally is corroborated by the drop in support for Trudeau - though, of course, if the sample was unusually anti-Liberal the results would be repeated throughout.

One interesting aspect was how support for who would make the best PM did not always follow voting intentions. In British Columbia, for example, the NDP outpaced the Liberals by 34% to 31%. But Trudeau outpaced Mulcair by 24% to 20% on this question. On the other hand, in Quebec the margin between the NDP and Liberals was relatively narrow (33% to 29%), but Mulcair was the choice of 30% of respondents as best PM, compared to just 16% for Trudeau. Support for a leader and support for a party, then, is not always contiguous.

But this is one poll and an unusual one at that. We have to wait and see what other surveys show before we take these results at face value. Has the political landscape in Canada really flipped, or was this a blip? The death of Jim Flaherty, and the resulting outpouring of sympathy and analyses of his legacy, could not have had no influence on the results of this poll (the biggest Conservative jump was in Ontario, after all). We should have a better idea of the significance of this poll in a few weeks' time.