Friday, October 3, 2008

Election preditions on October

Based on more polling data and information on individual ridings, here is where I see things at the moment.

Party

BC

AB

SK

MN

ON

QC

NB

NS

PE

NF

Nt


CPC

25

28

13

10

57

14

6

5

2

2

2

164

LPC

2



2

31

6

3

3

2

4

1

54

NDP

11


1

2

18

2

1

2


1


38

Bloc






52






52

Ind






1


1




2

GP















One thing that happens in every election is that almost everyone over estimates the power of incumbency - it is a factor but only a measurable difference in your first election as an incumbent for the party.

Another things that happens is that most people assume one election will look like the next. That someone that polled at 10% in a riding can not win, that someone that won by more than 10% over the number two party can not lose. These are constantly shown to be false in Canadian elections. These two assumptions lead to people making significant errors in how many seats parties will win in an election.

How do I think the election will end? I still believe we will see a large Conservative majority. Harper will be able to build a larger majority than Chretien did in 1997 with the same popular vote because the main opposition party in this election is their support much more dispersed. Preston Manning of the Reform party had his vote concentrated in one region, much as the Bloc has their vote only in region. The Liberal party is all over the country. This means the Liberals are on course to lose a lot more seats than anyone expects in this election.

Another way to measure the relative support level of the parties is to look at the election as if the Green Party vote simply did not exist. If we do this, the Conservative party has about 41% of the vote in the election and the Liberals about 27.5% and the NDP 21%. Looking at the election as a three party race to win seats outside of Quebec means the Conservatives are comfortably into majority government territory.

Another factor to consider in the election is the impact of voter turn out. The supporters of the Conservatives are the most likely to come out and vote, next highest are the core New Democrats and the core Bloc voters. The Liberals have a much lower level of core do or die supporters. The Green vote on election day is very, very hard to estimate because their supporters are young and not nearly as attached to the party or voting as the others.

I see the impact of turnout being a negative for the Liberals and Greens and a benefit for the other parties. Based on that, I am estimating the final election day support for the parties to be:

  • Conservatives - 39.0%
  • Liberals - 23.1%
  • NDP - 20.7%
  • Bloc - 9.8%
  • Greens - 7.4%

Taking these numbers across the nation we get the following result:

Party

BC

AB

SK

MN

ON

QC

NB

NS

PE

NF

Nt


CPC

26

27

12

10

66

12

6

6

2

2

2

171

LPC




1

20

6

3

3

2

4

1

40

NDP

12

1

2

3

20

3

1

2


1


45

Bloc






53






53

Ind






1






1

GP














The biggest change from my first prediction for October 14th has been the revival of the Bloc. I now believe they are the party most likely to for the official opposition in Ottawa.

I am now comfortable in predicting a shut out for the Liberals in BC and the higher number of Conservative seats in the Maritimes.

I keep putting my numbers for Ontario through various algorithms and keep coming up with large Conservative numbers - I am at odds with most other people on this and I am not sure what to do about this. I put it down to basis in favour of the previous election results. The Liberals are down a long way in Ontario, the math just keeps pointing towards a lot of loses for them.

No comments: