Tuesday, December 21, 2010

November Best and Worst Case Scenarios

I'll try to be a bit more timely with these monthly averages and best/worst case scenarios in the future. But for now, it's time to look at the best and worst case scenarios from the month of November.

These best and worst case scenarios calculate each party's best and worst projection results in each region.

For example, if the Conservatives had their best result in the western provinces in an Angus-Reid poll, their best result in Ontario in a Nanos poll, their best result in Quebec in a Léger poll, and their best result in Atlantic Canada in an EKOS poll, I would take each of these bests and combine them.

In other words, these projections are the best and worst possible results each party could have gotten had an election taken place last month, based on the available polling data.

These best and worst case scenarios are in terms of total seats only, and not necessarily about how a party would fit in with the others in Parliament.

Let's start with the New Democrats. They had a decent month, and improved their range over October's best and worst case scenarios. They are still, however, stuck in a fourth-place spot in the House of Commons.

With 24.1% of the vote, the best case scenario for the NDP is a total of 50 seats. The Conservatives would still end up with the greatest number of seats, with 122. The Liberals would win 85 and the Bloc 51.

The NDP would win 16 seats in British Columbia, one in Alberta, five in the Prairies, 19 in Ontario, two in Quebec, and seven in Atlantic Canada. It would put them in a good position to form a coalition with the Liberals, but together they would still be short of a majority.The worst case scenario for the NDP would result in the party earning only 13.9% of the vote and 22 seats, while the Conservatives would form a minority government with 139. The Liberals would win 96 and the Bloc 51 seats.

The NDP would win six seats in British Columbia, none in Alberta, two in the Prairies, 11 in Ontario, one in Quebec, and two in Atlantic Canada.

This puts the New Democratic range at between 13.9% and 24.1% of the vote and between 22 and 50 seats. That's an improvement over October's range of 13.6% to 20.9% of the vote and 21 to 43 seats.

The Liberal best case scenario is just short of a plurality: 110. The Conservatives, meanwhile, would still win 111 seats. But such a close result would give each party an equal chance at forming government. The NDP would win 36 seats and the Bloc would win 51.

With 31.3% of the vote, the Liberals would win 12 seats in British Columbia, two in Alberta, six in the Prairies, 49 in Ontario, 16 in Quebec, and 23 in Atlantic Canada.The worst case scenario would be slightly worse than the party's result in 2008. With only 23.2% of the vote, the Liberals would win 73 seats. The Conservatives would still be short of a majority with 140 seats, while the NDP would win 41 and the Bloc 54.

The Liberals would win four seats in British Columbia, none in Alberta, one in the Prairies, 37 in Ontario, 13 in Quebec, and 16 in Atlantic Canada. It would be a bad result for the party, as even the tiniest step backwards from 2008 would be a complete failure.

The Liberal range for November would have been 23.2% to 31.3% of the vote, and 73 to 110 seats. The range has gotten larger compared to October, when the Liberals had a range of 24.5% to 31.1% of the vote and 75 to 108 seats.

Now, the Conservatives. Still no majority for them in the cards, as their November best case scenario would only give the party 147 seats. The Liberals would win 85 seats and the NDP 27, while the Bloc would win 49 seats.

With 38.4% of the vote, the Conservatives would win 20 seats in British Columbia, 28 in Alberta, 23 in the Prairies, 54 in Ontario, nine in Quebec, and 12 in Atlantic Canada. It would put the Tories very close to a majority, and so they would likely be able to hang on to power against a weakened NDP and an only marginally larger Liberal caucus, but it would be the same-old situation.The Conservatives' worst case scenario would result in only 107 seats, but it would still give them a six-seat margin over the Liberals. The Bloc would win 53 seats and the NDP 47. It would be very difficult for the Tories to remain in power with this result.

With 29.2% of the vote, the Conservatives would win 13 seats in British Columbia, 26 in Alberta, 19 in the Prairies, 38 in Ontario, five in Quebec, and five in Atlantic Canada.

The Conservative range for November was between 29.2% and 38.4% of the vote and between 107 and 147 seats. It isn't a huge amount of change from October's range of 29.6% and 37.6% of the vote and between 113 to 147 seats.

The NDP and Liberals have a lot to gain with these numbers - the NDP would win a huge number of seats and have a lot of influence in Parliament, while the Liberals would have a good chance at forming a (weak) minority government. But Michael Ignatieff could also preside over the party's worst-ever performance, while Jack Layton could bring his party back to pre-2004 levels.

The Tories didn't have much to gain in November. A handful of extra seats and a handful of extra votes, but no majority. They could also risk losing their government to the Liberals or a combination of the Liberals and New Democrats.

I've already crunched the numbers for December, but in case a poll is released between Christmas and the New Year (though that is extremely unlikely), I'll hold off until early January. But from what I've seen, December has been a very bad month for the New Democrats and a very good month for the Conservatives.


  1. A different way of looking at best and worst case based on the 2008 election results:

    There were 111 ridings in the 2008 election which were won by less than 14.9%.... This is the amount that The Liberals won Vaughan by.

    The following is the count of close ridings won by party. These would be the seats a party might lose with a relatively small change in national vote intention; the second number is the safe seats by party (ie the worst case)

    Cons – 39 - 104
    Liberal – 38 - 39
    NDP – 21 – 16
    Bloc – 13 – 36

    The second number should be considered the worst case scenario. It would take a major event for the parties to lose seats that they won by more than the Liberals won Vaughan last election. It will happen but there will be extraordinary events there, No incumbent, real scandal etc.

    The count of strong seconds by party follows. The second number was the elected plus the strong seconds (the absolute best case)

    Cons – 46 – 189
    Liberal 44 – 121
    NDP 14 – 51
    Bloc 6 – 55
    Green – 1 – 1

  2. Hi, I found this blog somewhere. Forget where. But its awesome to see Nate Silvers work emulated in other countries. Count me in as a growing follower. I cant wait to learn more about Canadian politics, which seems quite twisted up at the moment. (For that matter, the whole Anglosphere seems that way.)

  3. BC Voice of Reason,

    Why do you always mention Vaughan but not Winnipeg North?

  4. BCVOR:

    I'd be very interested in some of the reasoning you use when you reinterpret polling results.

    For example, selecting and mentioning Vaughn as your benchmark for close ridings seems to make no sense. Fantino is a star candidate so it could logically be argued that in ridings where the Con-Lib race was within 15 points and the Con are running a star candidate that it could be considered a close race.

    Vaughn also had other irregularities that won't apply to a general election.

    For example, it was a byelection with a star candidate who was regularly getting national headlines. In a general election few local candidates will get any national coverage let alone regular coverage. What effect does that coverage have? I don't know that is another discussion but my instinct tells me it has some effect.

    Also the race was Con-Lib the whole way with few votes being cast for other parties. Some ridings the winning candidate won't crack 40% let alone two parties finishing above that level. You would think a stronger ndp vote would have benefitted the Conservatives as some of the left splits so if there were any ndp vote in Vaughn it would have been a huge conservative win?

    Also just because the Conservatives picked up Vaughn with a 15 point make up I don't think you can extrapolate that for other parties or other races. What I mean here is that seats where it is Conservatives v BQ or NDP couldn't be considered "close" if they were within 15 points. Even if Vaughn creates a new electoral paradigm where a 15 point race is somehow considered "close" it would only apply to Con-Lib races.

    You pose some thought provoking scenarios - I liked your post a while back talking about a leadership index and adjusting polls as one example. I thought the numbers you proposed were worth a debate perhaps but I thought it was a very interesting was to think about polling. I get the feeling from you recent "15 points is a close race" posts that you are really just trying to irritate Liberals rather than pose serious ideas.

  5. There must be significantly better ways to do this tahn finding one particular poll and running with it. Again, how about Monte Carlo simulations? I know it requires building a model rather than just scanning a list of polls, but it would be so much more useful!


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.

Note: Only a member of this blog may post a comment.