On The Cusp: State of the Canadian Race 29/09

This is a look at the state of the race in Canada's General Election judged by the polls as we enter the week of the leaders' debates, with particular attention to the prospects of a majority government.

The first hypothesis is that in using the polls to estimate the Conservatives' chances of a majority government, the Green vote can be ignored entirely. It can be subtracted from the total and the results reallocated among the 3 major parties. The Greens will win at most a handful of seats, their strength is a nuisance to the Liberals and the NDP, and the nature of their demographic appeal can affect Liberal/NDP races. But the number of seats the Conservatives will win in a given region is best predicted by comparing their poll standing with only the other parties likely to win seats. That's why people who talk about a supposed "40% threshold" for a majority government are wrong. As support for minor parties who are not going to win seats goes higher, the threshold percentage of the total popular vote needed to win a majority becomes lower.

We'll look first at most up to date national poll numbers -- all with the Green vote (and those of other 6th parties) being removed. In surveys which leave undecideds out of the party totals we will assume that they will break in the same proportions as decided voters. The dates are the dates that the polls were taken, where that can be determined. We will call these numbers the Toral-adjusted totals.

Nanos Research C 40 L 28 NDP 22 BQ 11 (Sept. 26-28)
Harris-Decima C 41 L 29 NDP 21 BQ 9 (Sept. 25-28)
Ekos Research C 38 L 29 NDP 22 BQ 11 (Sept. 25-28)
Angus Reid/Toronto Star C 43 L 23 NDP 23 BQ 11 (Sept. 25)

The next complication is that the Bloc will be highly effective in turning its votes into seats, because of the concentration of its vote. Let's give the Bloc 40 seats for the purpose of this exercise. Are these majority government numbers for the Tories? The Tory numbers are more impressive than they look because of the size of the differential between the Conservatives and their closest challengers. All the numbers except those of Ekos Research would suggest a Tory majority government; Ekos would have them right on the cusp. Perhaps this may explain some of Prime Minister Harper's recent willingness to come out and ask for a majority government, where he had previously been trying to dispel fear that one might happen. They're assured enough of at least a minority, and just uncertain enough about the possibility of a majority, that the time may have seemed right to stop pretending that they didn't want anything for Christmas and come out and ask for what they want.