Wednesday, December 30, 2009

Projection Trends, 2009

Seat projections are not useful only as snapshots of public opinion, but also as indicators of trends. This chart plots all the projections made since the creation of Canadian Election Watch last summer, according to the date of the projection.

During the second half of the summer, nothing much happened: the Tories clung to a narrow lead, and the projections were reminiscent of the 2006 election results. There was an NDP rise over the month of August, but those gains were lost at the beginning of September.

Then, during September and the first half of October, in the wake of Michael Ignatieff's decision to try to bring down the Harper government, the Conservatives climbed steeply, attaining a high of 153 projected seats on October 16, just shy of a majority. The Liberals fell even more rapidly, as the Bloc numbers rose slightly. The NDP was flat during this period.

The second half of October and first half of November were marked by a gentle retreat of the Conservative numbers, and a continued, though much slower, fall of Liberal fortunes. The Bloc continued making slight gains, and the NDP made steep ones, eventually attaining their high of 39 seats.

The second half of November and first half of December were flat for all parties, but a small Liberal rebound in voting intentions produced a modest gain in its seat projection at the end of the year.

Finally, a few thoughts about what these projection trends reveal about the methods that generated them:

Anyone making projections faces a tradeoff between stability and responsiveness. I am rather satisfied with the balance between the two here at Canadian Election Watch. The formula was responsive enough to reflect the changing political scene in September and early October (as opposed to heavily backward-looking formulae that might have taken until November to fully reflect the Tory rise). Yet, it did not produce wild fluctuations (a formula that projects past trends into the future would have given the Tories a comfortable majority in mid-October, only to get back to a minority just a couple of weeks later).

The alternative approaches outlined in parentheses above yield projections that are also predictions, assuming either that future behavior tends to revert to past behavior, or that momentum tends to carry on. Both of these approaches have their merits, but they often produce drastically different results, as was witnessed this fall. I've decided to remain agnostic, and thus Canadian Election Watch's projections are really just that: my best guess at what election results would have been during the latest polling period, with no attempt to account for likely future changes.

And with that, I wish you all a Happy 2010!

Maps! Last Projection of 2009
















I've been too lazy to map the projections that I've been making, but I thought that it'd be nice to visualize the state of things at the end of the year. So here it is: the second mapped projection of Canadian Election Watch!

Comparing the Latest Projection to the 2008 Results

So where are we 14.5 months after the last federal election? Lots of things have happened: a parliamentary crisis with threat of coalition and, ultimately, a controversial prorogation; a new Liberal leader, his honeymoon through the spring and great fall in, well, the fall; the biggest federal deficit in history (though this is misleading since it's only true not taking into account inflation and economic growth); the NDP deciding to temporarily support the government; the Afghan detainee scandal; and Canada's declining reputation internationally due to the climate change issue.

It seems, though, that in terms of voting intentions, these have pretty much all canceled out. Here are the seat changes from the 2008 election (counting André Arthur and Bill Casey as Conservatives) to the most recent projection:

CON -1
LIB -1
BQ 0
NDP +2

So essentially, we're back where we were on Election Night 2008. It doesn't get much more exciting when you break it down by region either:

British Columbia: This is the region, along with Atlantic Canada, where there has been the most movement, but that's not saying much. Conservatives are on -3, with Liberals gaining 1 and NDPers gaining 2. Stéphane Dion's departure (Dion being extremely unpopular in BC) and the HST probably explain these changes.

Prairies and North: Nothing. Kind of fitting for this region, no?

Ontario: NDP loses 1 seat, while Liberals gain 1. Out of 106 seats, that's essentially no change as well.

Québec: Tories lose 1 seat, while Grits gain 1. Again, out of 75 seats, that's also pretty much no change.

Atlantic Canada: Tories gain 3, NDP gains 1, while Liberals lose 4. These changes are significant, and the Tories now lead the seat count there (but just barely). One explanation is that Dion was not being horribly unpopular there. Also, Atlantic Canada seems to react a bit more slowly than the rest of Canada to political developments (again fitting: as they say in French, une heure plus tard dans les Maritimes), so one can speculate that while the Grits have marginally picked up in the rest of the country, they are still at their November lows in the Atlantic provinces. Remember that when Ontario swung to the Tories in September, Atlantic Canada held good for the Liberals, only to then swing in October.

What will happen in 2010? My humble guess is that the Liberals will recover somewhat from their lows this Fall. But I don't think that they will retake the lead from the Conservatives before an election campaign. Of course, once a campaign is underway - which may or may not happen in 2010 - all bets are off!

Nanos December Poll

Nanos released today its December voting intentions poll. This one was actually conducted around the same time as the EKOS and Angus polls reported in the last two posts, and the most recent day of polling included is still EKOS' December 15.

Nanos' poll has some interesting results: it shows the NDP first in Atlantic Canada (essentially tied with the Tories, and a few points ahead of the Grits, who should be worried there), the Liberals within 3 points of the Bloc in Québec and ahead of the Tories in Ontario, and the Conservatives with a very strong lead in BC (and of course in the Prairies too, though that's not "interesting").

This poll is a good one for the Liberals: while it has them still 9.3% behind the Tories, it's the first poll since the first week of September to show the Grits above 30% nationally (a few more polls in September, as well as Nanos' October, had them at 30%). More importantly, those gains have disproportionately (read: almost all) come from the electorally fertile Ontario and Québec. The Liberals therefore end the year on an upward trend in the aggregate projection, though still at a low level:

CON - 144
LIB - 76
BQ - 49
NDP - 39

Thursday, December 17, 2009

New EKOS, and Other Polls

Here is a new EKOS poll done over the past week. Not much movement was recorded, but including this poll did influence some marginal projections.

There was also a Manitoba poll done by Probe Research (props to ThreeHundredEight.com for finding it) that confirms the status quo there. Finally, some media have been reporting the national and Québec result of a Harris-Decima poll, but I will wait until more regionals are reported before including that poll in my projection.

So the new aggregate count is:

CON - 145
LIB - 72
BQ - 52
NDP - 39

Tuesday, December 15, 2009

New Angus Reid

Angus Reid provides us with our first non-EKOS national poll in 3 weeks. Compared with the last two EKOS polls:

- The Tories are doing better in BC, which usually happens when comparing Angus Reid polls to EKOS polls. Also, their Québec result is similar to the one in the latest EKOS poll, which is 5 points lower than in previous polling. Copenhagen?

- The Liberals are doing better in Atlantic Canada, Québec (picking up votes from the Tory drop?) and Alberta (useless).

- The Bloc is also up.

- No significant change for the NDP.

- The Greens are way down, but again that's probably due to the different polling methods of EKOS and Angus Reid. The only two pollsters that tend to produce reliable results for the Greens are Nanos and Angus, and both have them at just 6% nationally in their latest poll.

In terms of seats, the Liberals retake the seat they lost to the NDP in the last update, and the Conservative drop in Québec makes them lose two seats to the Bloc. Thus:

CON - 145
LIB - 72
BQ - 53
NDP - 38

Thursday, December 10, 2009

New EKOS

The first national poll in two weeks is out, and it is... completely underwhelming. Basically, nothing changed. Due to the scarcity of polls lately, this one was weighed almost 50% in my projection. But despite half the inputs being new, the only change was in Newfoundland and Labrador, where the Grits lose a seat to the dippers. So the new totals are:

CON - 147
LIB - 71
BQ - 51
NDP - 39

This is a new low for the Liberals, and a new high for the NDP. This trend is unlikely to persist, however: the Liberals simply don't have much room left to slide, and the NDP is really milking out every possible seat in this projection, given that their national numbers are actually still below where they were in the 2008 election.

Monday, November 30, 2009

Léger Québec Poll

Link
No big news (on the federal scene, at least) in terms of top-line numbers, but the Tories seem to have retaken the lead in Quebec City, where they were a few points behind the Bloc for several polls in a row. (And that's a long time, since Québec pollsters, the only ones that provide a breakdown for Quebec City, typically poll at most once a month.) Of course, that subsample is very small, but the 12-point lead suggests that the Conservatives now have a good chance of keeping all their Quebec City seats. This is not surprising, considering that they've been back up near their 2008 election levels for a while.

As a result, the Tories take a seat from the Bloc in the aggregate projection:

CON - 147
LIB - 72
BQ - 51
NDP - 38

Thursday, November 26, 2009

New EKOS Poll: Flatline

Here is the now bi-weekly EKOS poll. The news? That there is no discernible movement. So little movement that despite this poll counting for over 35% in my projection (due to its huge sample size and the small number of other recent polls), it generated no change. Not even regional changes that canceled out! Nada. So for the third time in a row, the projection is:

CON - 146
LIB - 72
BQ - 52
NDP - 38

Saturday, November 21, 2009

New Ipsos

On the heels of a bad Angus poll for the Grits, here's a bad Ipsos poll for them. Just like the Angus poll, this one shows the Tories receding from their October highs, so the main beneficiary of the Liberal slump is the NDP. In fact, this Ipsos poll has the Dippers at a high 19% nationally (just 5 behind the Grits), an excellent 21% in Ontario, and a whopping 34% in BC.

In the aggregate predictions, these results made the NDP take back a seat in Ontario from the Tories, but they lost the Saskatchewan seat that they had just gained yesterday. Thus, no change in the national numbers:

CON - 146
LIB - 72
BQ - 52
NDP - 38

Friday, November 20, 2009

New Angus Reid Poll

Now that EKOS is bi-weekly, updates will come even less often... Here's the latest Angus, which shows the Grits at a disastrous 23% - down below Dion election levels, to Dion coalition levels...

Accordingly, in the aggregate projection, the Liberals lose 3 seats. Those losses come mainly in Québec, due to the Bloc's strong result in this poll. I'm guessing, though, that this Bloc resurgence is just statistical noise, but we'll have to wait and see.

CON - 146
LIB - 72
BQ - 52
NDP - 38

Bad news for the Liberals, who are down to their lowest projection (tied to October 29), and who are the only party projected to lose seats. Fortunately for them, all of the other parties' projected gains are tiny, and with the Holidays and the Olympics coming, the earliest an election campaign could get underway is probably in March 2010.

Saturday, November 14, 2009

November Nanos

The November Nanos is out, and finally gives us some confirmation of what EKOS has been observing for the past few weeks: the Tories are now back to where they were in 2008. This comes after a few weeks of being slightly above that level, and flirting with (though not quite getting to, at least according to my projections) a majority.

As usual, relative to EKOS, Nanos' figures scale up the support of all parties at the expense of the Greens, with an extra little bump for the Liberals. It's interesting to note that while this week's EKOS and Nanos both show the Bloc at 35.6%, EKOS has it at 8.8% nationally, while Nanos shows 9.3%. The latter probably takes into account that Québec turnout is typically slightly higher than national turnout at general elections.

Very little change in the aggregate projection, though we now get nice round numbers for the top three parties:

CON - 145
LIB - 75
BQ - 50
NDP - 38

The CPC's rise in Atlantic Canada lagged its rise in Ontario by a few weeks. Now that the Tory wave seems to be subsiding in Ontario (I believe they got as high as 58 seats at one point), one can ask whether the same will occur in Atlantic Canada in the coming weeks.

Wednesday, November 11, 2009

Weekly EKOS

In the end, 3/4 ridings were "holds" on Monday's by-elections, if you count Cumberland... as a hold. I'll go out on a limb and say that Montmagny... would also have been a hold given general election level turnout. Indeed, the Tories carried Roberval... by 27% in the 2007 by-election, only to win it by just 4% in the 2008 general election, while in the case of St-Hyacinthe..., they went from losing by 5% to 26%. While there was a large swing against the Tories in Québec between September 2007 and October 2008, it wasn't nearly as large as 21-23%. My guess is that relative to the Bloc, the Tories get a 5-10 point bump in rural Québec by-elections.

It's been two very quiet weeks on the polling front: only EKOS has put out new polls during this time! Here is the most recent one. The new aggregate projection is:

CON - 145
LIB - 75
BQ - 51
NDP - 37

This is the lowest that the Tories have been in the past 5 weeks, though the main beneficiary of this decline has been the NDP, whose popularity increased in BC and the Prairies, and not the Liberals, who are still mired at Dion levels. Counting both independents elected in 2008 as Conservatives (since they align themselves that way in the House of Commons), here are the changes since 2008 in terms of voting intentions and seats:

CON: -0.1%, 0 seat
LIB: +0.3%, -2 seats
BQ: -1.4% in Québec, +2 seats
NDP: -1.8%, 0 seat

The storyline: very, very little change. The Liberals are losing seats due to weakness in Atlantic Canada, while the Bloc is gaining seats due to the Tory vote becoming less efficient in Québec. Indeed, polls with a regional breakdown all suggest that Conservative support is less concentrated around Quebec City than in 2008, which helps the Bloc. The NDP and Conservatives trade seats in Ontario (Dippers down) and BC (Tories down).

Sunday, November 8, 2009

By-Elections Tomorrow

The projection model on this website is not designed to project by-elections, whose results are rather unpredictable due to low turnout and a paucity of polls just before election day, among other things. That said, the model is currently showing all 4 ridings as "holds" (counting the Conservatives winning Cumberland--Colchester--Musquodoboit Valley as a hold), though New Westminster--Coquitlam is close, so the Tories have a chance to take a seat from the NDP. Also, the departure of popular Bloc incumbent Paul Crête could allow the Tories to take Montmagny--L'Islet--Kamouraska--Rivière-du-Loup.

Given their current numbers, the Liberals are probably happy that they do not have any chance of winning any of the by-elections, so they don't have a way to underperform.

Thursday, November 5, 2009

Weekly EKOS: Conservative Retreat?

The Tories have lost some ground in both Ontario and Québec over the past week according to this week's EKOS poll. Is this just a statistical blip, or has Ignatieff finally turned the corner? Maybe we will find out next week...

For now, not much change in the aggregate projection:

CON - 148
LIB - 74
BQ - 52
NDP - 34

The projection has now been quite stable for four weeks, with the Conservatives at the door of a majority (they never actually got there though, topping out at 153), and the Liberals squarely in Dion territory.

Thursday, October 29, 2009

New Angus Reid Poll

Here's the most recent Angus Reid poll. It's mostly in line with recent polls, except in MB/SK where the Conservatives got a huge number, and take the last Liberal seat in Manitoba in the aggregate projection. Not a good poll for the Liberals in Atlantic Canada and Québec, and they lose a seat in each of these regions. The Grits are no longer projected to make any gains in Québec.

The aggregate projection is now:

CON - 149
LIB - 72
BQ - 52
NDP - 35

The Bloc is in a sweet spot where the Tories and Liberals almost evenly split the federalist vote. Québec poll averages are almost identical to the 2008 results, but I still have the Tories losing a couple of Quebec City seats because their numbers there, in polls with regional breakdowns, are consistently worse than what they got in 2008.

There has been a bit of a slingshot effect during this Liberal slide (and it is a Liberal slide rather than a Conservative uptick, since the BQ and NDP have gone up as well): first, Ontario moved while the rest of the country stayed put. Now, Ontario has been very stable for a while, and the rest of Canada caught up.

Wednesday, October 28, 2009

Two Good Polls for the NDP

First, the weekly EKOS concurs with Environics in giving about 29% to the NDP in BC. Otherwise, there is little change from last week.

Second, a CROP Québec poll has the NDP at 16% in La Belle Province, while the Liberals are now 14 points behind the Bloc. The latter is consistent with the latest results from Rest-of-Canada pollsters, and is very bad news for the Liberals: last month, CROP and Léger had the Grits trail the Bloc by only 7 and 3 points, respectively. So the recent poor Québec results for the Liberals seem not to be due to Rest-of-Canada pollsters undercounting the Liberal vote in Québec, as they are believed to sometimes do.

The aggregate projection shows an uptick for the NDP. While the Conservatives have dropped, they remain close to a majority:

CON - 147
LIB - 75
BQ - 51
NDP - 35

This seat count has every party within 5 seats of their 2008 result.

Monday, October 26, 2009

Environics Poll

ThreeHundredEight.com has a preview of an Environics poll. The strong NDP result in BC bumps its number up, and the aggregate projection is now:

CON - 149
LIB - 75
BQ - 51
NDP - 33

New Ipsos

ThreeHundredEight.com has the details of the new Ipsos poll. The new aggregate projection is:

CON - 151
LIB - 75
BQ - 51
NDP - 31

Thursday, October 22, 2009

New EKOS and Nanos

EKOS and Nanos

Pressed for time today: no analysis, but to point out that this is the first time the Conservative seat projection has decreased in about 7 weeks.

CON - 148
LIB - 78
BQ - 50
NDP - 32

No update until Monday.

Monday, October 19, 2009

Election Day

Today is the day on which, by law, we should have had the 40th Canadian general election had the first Harper minority not been defeated. Of course, we all remember that it wasn't defeated - the Prime Minister simply decided to break his own law and call an election last year.

Interestingly, current standings in polls are quite similar to the results from 2008. But of course, had an election not occurred last year, who knows what the polls would show now. A Harper majority? Maybe - we're close to that anyway. A Dion minority? Not totally impossible, but, um, no.

Friday, October 16, 2009

New Angus Reid and Incomplete Harris-Decima

After Strategic Counsel and EKOS, Angus Reid now also has the Tories above 40%, in majority territory. However, Harris-Decima produced a poll (only partial details available through CP) with the Conservatives still far shy of a majority.

All the polls released in the past two weeks agree on one thing: the Liberals are below 30%.

The main disagreement in the polls is about Ontario. Angus Reid, Strategic Counsel and EKOS have the Tories up 16, 16 and 13.1 respectively in their latest polls. However, Ipsos Reid and Harris-Decima both peg the Conservative lead at just 4% there. That's the difference between a majority and a minority.

Averaging up everything, I get the following projection:

CON - 153
LIB - 74
BQ - 50
NDP - 31

The Conservatives are now projected to increase or maintain their seat count in every region, except in Québec, where they're still expected to lose two marginal ridings. Also, for the first time since this website was launched, the Conservatives are projected to win more than twice as many seats as the Liberals.

The Grits may be doing a bit better than my projection suggests in Québec: we haven't had numbers from CROP or Léger in the past couple of weeks, and those generally favour the Liberals, counterbalancing the fact that many national polling firms tend to undercount them in Québec. Still, that's no more than a 5-seat difference, which would still leave the national Liberal seat count firmly in Dion territory. And seat switches between the Liberals and the Bloc obviously have no bearing on the Tory seat count, which is the relevant number nowadays.

Wednesday, October 14, 2009

EKOS: Tories above 40%

More good news for the Tories here. I believe that this is the first poll since Dion's coalition-talk fiasco where the Tories are ahead of the Grits and Dippers combined. This has never happened at an actual election during my lifetime, and not so long ago it was the Liberals getting more than the PC and Reform/Alliance combined. Is this just a blip, or a sign of a durable shift in Canadian politics?

Compared to last week's EKOS, this poll shows the Conservatives increasing their lead everywhere except in Québec and BC. Based on this poll alone, the Tories would win a majority with a half-dozen-seat cushion, while the Liberals would win about 66 seats - Stockwell Day's performance in 2000 (and that was out of 301).

The aggregate projection does not quite put Ignatieff at Day levels yet - he remains on par with Stéphane Dion (what's with the initials S.D. anyway?):

CON - 150
LIB - 76
BQ - 49
NDP - 33

The NDP's national poll average may be below 15%, but its seat count remains respectable due to its number holding up in BC and to Liberal weakness in Ontario.

Will the Tories get the last 5 seats that separate it from a projected majority? Stay tuned!

Tuesday, October 13, 2009

Ipsos: Conservative Lead in Ontario down to 4

Here's the writeup by ThreeHundredEight.com. This poll is a bit odd, giving the Tories excellent results in B.C. and the Prairies, while Ontario is now again a statistical tie. This weirdness introduced some regional changes in the aggregate projection, but they mostly cancel each other out, yielding:

CON - 149
LIB - 76
BQ - 50
NDP - 33

The Liberals are now projected to do just worse than under Dion. Their expected gains in Québec have almost entirely evaporated: the poll average I'm using puts them 12% behind the Bloc, barely less than the 14% registered at the 2008 election. However, some of this may be a temporary blip due to the recent events surrounding Denis Coderre.

Putting Québec and Ontario together, the Grits are now the third party in Central Canada, well behind the Tories, but also behind the Bloc. Atlantic Canada is now playing two important roles: (a) preventing the Tories from forming a majority; (b) preventing the Bloc from approaching Official Opposition territory.

Thursday, October 8, 2009

EKOS: CPC by 14

The weekly EKOS is out, and is disastrous for the Grits: they have lower support nationally than in the 2008 election. In this poll, the Conservatives take the lead in Atlantic Canada, pass the Liberals in Québec, more than double their lead in Ontario from last week's EKOS, while the NDP passes the Liberals in BC.

This polls also asks about party leaders' handling of their jobs. The following net approval ratings speak for themselves: Harper -3.6%, Ignatieff -31.6%, Layton +2.6%.

My first reaction when I saw this poll was that these numbers, if borne out, would result in a solid Conservative majority. Not so after crunching the numbers though! Even such a meltdown for Ignatieff would only put Harper on the cusp of what he's desired for so long: 154-155 seats.

Nevertheless, EKOS' large sample size, combined with the relative lack of recent polls (Harris-Decima and Nanos have been AWOL and receive no weight for now), means that this poll substantially moved the aggregate projection:

CON - 149
LIB - 77
BQ - 48
NDP - 34

For the first time, the Conservatives are projected to gain seats if an election took place in the past week. They are now very close to a majority. The Grits have retreated to their 2008 levels, while the Bloc and NDP post very modest losses.

The problem for Harper: much of his party's recent swell in the polls was probably a result of Ignatieff's (feigned) eagerness for an election. So if the Conservatives engineer their own defeat, a good chunk of this goodwill may instantly evaporate. Besides, many believe that Harper will be under pressure to leave if he delivers yet another minority - however strong it may be. So perhaps he's not exactly burning to roll the dice quite yet, since a majority is of course no sure thing.

The Liberals though, look less and less like a government-in-waiting by the day. Musing about gimmicks like high-speed rail and defending Suaad Hagi Mohamud are not valid substitutes for a platform. Actually, don't even mention platform - we now have absolutely no clue what the Grits stand for anymore, beyond the fact that they're not Harper and not socialists. Or are they? After all, Ignatieff seems to like the oil industry in Alberta, and the 360-hour EI threshold is so far to the left that the NDP didn't even bother coming up with a more lefty proposal.

Ironic that one of Canada's greatest intellectuals is now in danger of becoming a failed politician due to a lack of credible ideas. The silver-lining: these things can turn around very quickly, so Ignatieff still has time, especially since Québec is nowhere close to tipping to the Tories. If that happens though, the following storyline would write itself: "Canada has opted for a Conservative majority, and is just waiting to go to the polls." So Mr. Ignatieff - stop tempting fate, OK?

Tuesday, October 6, 2009

Strategic Counsel: CPC by 13, 16 in Ontario

Numbers here. Aggregate prediction is unchanged, though there was seat trading in Ontario and Québec:

CON - 142
LIB - 84
BQ - 48
NDP - 34

Of course, taking this poll's numbers at face value would probably yield a Conservative majority. However, the Strategic Counsel doesn't provide numbers for Atlantic Canada, and only gives combined figures for the West, so all I can do is speculate...

Friday, October 2, 2009

Angus Reid: CPC Leads by 10

Numbers here. New aggregate projection:

CON - 142
LIB - 84
BQ - 48
NDP - 34

Posts over the next 10 days are going to be short and sweet like this one...

Thursday, October 1, 2009

Three New Polls Included

Not much time today, but I've incorporated to my aggregate projection the following 3 polls published over the past few days: Ipsos, CROP and EKOS. The Québec projection finally moved, in favour of the Conservatives.

CON - 142
LIB - 85
BQ - 47
NDP - 34

This looks quite a bit like the current parliament, despite the fact that the Tory lead is now almost 5% smaller than in the 2008 election. The main reason is Ontario, where the Conservative has actually increased. Liberal gains elsewhere are for now strikingly inefficient.

Monday, September 28, 2009

Léger Marketing National Poll

Here is the survey report. This is a large poll, and mainly confirms the trends in Ontario, and west of it. However, contrary to most polls by non-Québec firms, this Léger poll still has the Liberals within 3% of the Bloc. Also, since mid-September, the Liberals have been doing very well in Atlantic Canada, and the Tories poorly - maybe one region of the country did appreciate Ignatieff's stand after all.

The aggregate projection shows a further strengthening of the Tories, but I believe that things will now stabilize:

CON - 140
LIB - 87
BQ - 48
NDP - 33

Given Layton's plans to support the government on the Liberal non-confidence motion, Ignatieff will likely have a few months to get the Liberal number to a more respectable level.

Sunday, September 27, 2009

Angus Reid: Tories Lead by 14 in Ontario

This Angus Reid poll has the Conservatives leading 44-30 in Ontario. But the Liberals lead 57-22 in Atlantic Canada! The latter is obviously a sampling quirk, but the former reinforces the previous Angus Reid poll that showed the Tories up by 12 in Ontario. Also compared to the previous Angus Reid poll, the Tories jumped from 13% to 21% in Québec, while the NDP jumped from 6% to 17% in Alberta. The latest EKOS had also picked up a Tory uptick (albeit much smaller) in Québec, but actually had the NDP dropping in Alberta...

This poll alone puts Harper a half-dozen seats away from a majority, while the new aggregate projection shows:

CON - 138
LIB - 88
BQ - 48
NDP - 34

The Liberals are still the only party that would gain seats, but now only a handful of them.

Friday, September 25, 2009

Cauchon in Outremont after all

If Ignatieff had made this decision at the very beginning, things would obviously be better for the Liberals. But as they say, better late than never!

As a result of this, I'll revert to projecting Outremont normally, which for now means that the NDP would lose it. Thus we now have:

CON - 136
LIB - 92
BQ - 48
NDP - 32

Thursday, September 24, 2009

Ignatieff Offers Jeanne--Le Ber to Cauchon

Earlier this week, I had actually thought of this as a possibility, but didn't post about it since I thought that if Ignatieff wants both Cauchon and Le Prohon to run, having Cauchon in Outremont and Le Prohon in Jeanne--Le Ber is much more electorally efficient. Indeed, it currently appears that Outremont will be a tougher battle than Jeanne--Le Ber. Since Cauchon is a former Outremont MP, and is in any case much better known than Le Prohon, the logical choice would have been to nominate Cauchon in Outremont.

It looks like the reason why we now have it the other way around is to avoid embarrassing Coderre too much. Ignatieff is attempting to cut the pie in half by saying, "Denis, you can put your preferred candidate in Outremont, but you cannot prevent Martin from running as a Liberal." Unfortunately, this may achieve the "worst" of both worlds: Coderre is still disavowed since Cauchon is allowed back, and the Liberals may have lost Outremont in the process. (I put 'worst' in quotes since that's from Ignatieff's perspective - I personally have no problem with embarrassing Coderre to punish him for overreaching.)

EKOS Confirms Conservative Strength

After a whole week without polls, reliable EKOS comes through, and shows a 7.1% Tory lead. However, the Liberals kept their support steady from last week. Rather, it is the NDP (surprise, surprise) that has bled support, reaching 13.8%, their lowest level in a non-Ipsos poll in at least 6 months (according to ThreeHundredEight.com's list of polls).

Ontario is the big news of the poll: results in other parts of the country are actually par for the course. In Ontario, the Conservatives lead the Liberals by 6.7% according to EKOS, more than the 5.4% margin of the 2008 election. Normally, this would bring the Tories very close to a majority, but they do not have commensurate strength in Atlantic Canada or BC (not to mention QC, where the Conservatives are still weak even though this poll shows an uptick for them). As a result, even based on this poll alone, the Conservatives are no closer to a majority than they currently are.

The updated aggregate prediction shows the Liberals losing two seats to the Conservatives:

CON - 136
LIB - 91
BQ - 48
NDP - 33

I gave Outremont back to the NDP due to Denis Coderre's shenanigans in that riding, and Ignatieff's lack of leadership on this issue. While the 33-seat figure doesn't look bad at all for the NDP, many of those seats are extremely vulnerable. The province to watch for them is BC: while the NDP has been losing strength everywhere else in the country, BC has held up for them. If that changes, and Ontario keeps giving poor numbers, the much feared losing of a third of the current NDP caucus could well become reality.

Wednesday, September 16, 2009

Two Bad Polls for the Liberals

Angus Reid and EKOS have the Conservatives up by 7 and 5.2 points respectively over the Liberals. This confirms the anti-Liberal tendency observed since Ignatieff declared that he no longer supports the Harper government. This also makes the most recent Ipsos poll less of an outlier, and suggests that something odd, beyond methodological differences, was going on in the mid-August Ipsos poll (Really Conservative data points? Rogue interviewer?)...

Although the Angus Reid and EKOS polls give the Tories similar national leads, the Angus Reid poll is by far more devastating to the Liberals: it has the Conservatives 12% ahead in Ontario, compared to just 4.6% according to EKOS. For this reason, the Angus Reid poll - just like the Ipsos - would actually imply that the Tories are very close to a majority (despite having only 36% nationally), while the EKOS poll - just like the Harris-Decima - suggests that they are nowhere near.

All this makes the Liberals lose 5 seats in my projection, mostly going to the Tories:

CON - 134
LIB - 93
BQ - 48
NDP - 33

It will be interesting to see how the NDP fares in the next round of polls now that it has all but declared that it would support the government long enough to avoid an election in 2009. Will left-wing Canadians reward Layton for "making Parliament work," or will they punish him for allying with the Devil? This looks like a make-or-break moment for the NDP. Fortunately for them, it's in Harper's interest not to damage the NDP too much, in order for the Liberals to remain weak. So perhaps the Prime Minister will try to keep the number of confidence votes to a minimum, in order to avoid embarrassing the NDP each time. Heck, he may even decide to move to the center just to make Ignatieff look like the bad guy! But if Canadians punish the Liberals too much in the polls for election mongering, Harper may decide to introduce a poison pill that would force the NDP to defeat him.

I feel that although the NDP may not feel too much public opinion pain in the short run due to public relief at avoiding an election, this is a tremendously dangerous strategy for Layton: if he indeed carries through supporting the government for months, he risks being Dionized, and losing a good number of left-wing voters come next election. These voters will eventually get over their relief of avoiding an election - probably, and inconveniently for Layton, just as the next campaign begins.

The numbers look bad for the Liberals for now. But if the next few months indeed feature a Conservative government propped up by the NDP, both will lose their main attack line against the Liberals. Who's forming an alliance with the socialists now? And who's the chicken? Ignatieff is getting a pretty good deal out of this, provided that Canadians do not hold it against him to have almost provoked an election this fall. Historically, voters don't care about such things once a campaign is underway. Will this time be different?

Tuesday, September 15, 2009

New Harris-Decima Poll

Nothing much to report on this poll: the results are almost identical to the Harris-Decima from last week. The very strong MB/SK result for the NDP, and a more decent BC number for them mean that the Dippers gain 2 seats at the expense of the Tories in the aggregate projection:

CON - 130
LIB - 98
BQ - 48
NDP - 32

The four most recent polls as measured by their midpoint date (Ipsos, Harris-Decima, EKOS and Strategic Counsel) all imply a Liberal seat total under 100 in my model, as opposed to the vast majority of polls this summer. Now that the Bloc has announced that it will support the government on Friday's confidence vote, the Liberals might be secretly hoping that either the Bloc or the NDP will vote against their non-confidence motion in 2-3 weeks. Of course, the Bloc has little to lose from an election: it can make up for most of its expected losses to the Liberals by gaining seats from the Conservatives. So all eyes will likely turn to Jack Layton as the Opposition Day approaches.

Ipsos Reid Poll, and NDP Woes

Unsurprisingly, the new Ipsos Reid poll (more details via ThreeHundredEight.com) has the Conservatives far ahead, 39% to the Liberals' 30%. The NDP, however, has dropped to 12% from 14% in the previous Ipsos poll. In my projection model, that 2% and the way it's distributed across the country really hurt: the NDP drops from 30 to 21 seats in one-poll projections. Also, based on this poll alone, I have the Tories winning exactly half the seats in the House - literally on the cusp of a majority.

Incorporating this poll into the aggregate projection gives CON 131, LIB 99, BQ 48 and NDP 30. However, I've also decided to change the way in which I project Nunavut, in light of the great attention that it has received from the Harper government, and that Leona Aglukkaq will enjoy the first-time incumbent advantage. So the new projection is now:

CON - 132
LIB - 98
BQ - 48
NDP - 30

In one week, the NDP has lost 7 seats, and the Conservatives have gained 10. The NDP may feel that things are bad now, but it's not hard to see how things could get even worse if they prop up Harper enough times to avoid a fall election (i.e. both on Friday's ways and means motion, and on the Liberal non-confidence motion in 2 weeks). After all, I'd guess that a good chunk of 2008 NDP voters were actually Liberals voting for an MP that would not roll over on confidence votes; if the NDP starts supporting the government, those voters are gone. In addition, many Dippers could become demoralized and stay home if they see Layton vote confidence enough times.

Without fatigue at a dozen years of Liberal government and the sponsorship scandal like in 2006, and without a disorganized Liberal party with a weak leader like in 2008, the NDP is likely to return to 2004 levels of support. Their current saving grace is that they are the party that opposes the Conservatives, but that credential could evaporate over the next month.

In fact, according to polls, the NDP's popular support is already back at 2004 levels. I still have them at 30 seats (vs. 19 in 2004) because their vote has become much more efficient (e.g. they got 7 seats in Ontario with 18.1% in 2004, and 17 seats with 18.2% in 2008). But with Jack Layton not as new and exciting as back then, a toothless NDP could dip further, thus returning their seat total to 2004 levels.

If the NDP supports the government on Friday, they will have until the Liberals' motion to choose: do they accept to go into an election where they will likely sustain moderate losses? Or do they postpone the election in the hopes that things get better, but also risk losing all of the gains they made over the past 5 years?

If the NDP does get Harper through the fall, and if its support indeed collapses as a result, we might get a semi-stable Parliament, with the NDP wanting to avoid an election at all cost. Of course, then, each vote for the government risks pushing Layton's band further into irrelevance.

Sunday, September 13, 2009

Commenting Format Changed

Comments are invited as always, but I had to change the format from an embedded form to the somewhat less inviting pop-up window. The reason is that upon trying to comment, I realized that the embedded form doesn't work in my Firefox browser... Does anyone else using Firefox have this problem with Blogger?

Friday, September 11, 2009

Sooner than we thought?

If this happens, Election Day could be as early as October 26. That's two weeks earlier than the originally contemplated scenario - the government falling on the Opposition Day following the EI panel's report.

For now, I have to admit that I've somewhat tuned out due to the lack of policy ideas lately. But with the Conservative EI proposal coming up, not to mention all the parties' platforms in an eventual campaign, there is sure to be much to talk about later on.

Meanwhile, I ask: what is each leader's minimum performance to hold on to his/her position? I would suggest the following:

- Harper: form government. How tall of an order this is depends on the NDP and the Bloc's willingness to support the Liberals without a formal coalition, which Ignatieff is unlikely to agree to.

- Ignatieff: prevent a Conservative majority. This should be pretty easy, and the bar is pretty low since it's his first crack at it, and he's not Stéphane Dion.

- Layton: 25-30 seats. Above 30, and he's still above his 2006 result. But below 25, and he'll have lost a third of his party's seats, after which no leader is safe.

- Duceppe: 40-45 seats. Above 45, and it's pretty much par for the course. But below 40 would give something close to the Bloc's worst ever result of 38 seats from 2000.

The current projected outcome would bring little change unless Ignatieff seizes power with the support of the NDP and Bloc. But much can change during a campaign, and if the Liberals pick up just another 3-5% nationally, it's not hard to see a scenario where Harper and Layton would both be toast. (On the other hand, the Tories would still be short of a majority with an improvement of 3-5% over their current poll position.)

And Harris-Decima seems to be patting itself on the back for its new website design instead of, you know, actually updating the content...

Thursday, September 10, 2009

Three Polls

Now that Labour Day has passed, new polls are coming in at increased frequency. Since the last update, three were published: Nanos, Harris-Decima and EKOS. All three have the Tories leading the Liberals by between 3 and 4.1%, but the regional numbers differ significantly: Harris-Decima has a favourable breakdown for the Grits (5% lead in Ontario, and down only 5 to the Bloc in Québec), while EKOS has a bad one (down 1.9 in Ontario, and down 12 to the Bloc in Québec). The common denominator in these polls is a low NDP result: between 14.8% and 15%.

These results mean that my projection went up significantly for the Conservatives, down a lot for the NDP, and down a little for the Liberals. Note that the Prairies and Atlantic breakdowns are not yet available from Harris-Decima, so these numbers may be updated later today, but I don't expect major changes:

CON - 127
LIB - 101
BQ - 48
NDP - 32

This is the best projection for the Tories since this site was launched on July 21, and in fact we're back to almost exactly the same numbers as back then. Basically, the Conservatives' slow slide over the second half of the summer was reversed by election talk over the past 10 days.

Tuesday, September 8, 2009

Strategic Counsel: Tories Lead in Ontario, Bloc Strong

Here is the poll.

This poll is bad news for the NDP: terrible Québec (6%) and Ontario (11%) numbers, and not so great out West either (22% vs. 24% for the Liberals). The Grits can't be happy with being at 23% in Québec, and behind the Tories in Ontario (39% vs. 41%) either.

For the Conservatives, this poll is a mixed bag: they're still weak in Québec, but have their first lead in Ontario since early July in a non-Ipsos poll. However, 43% out West is actually very low for them.

The Strategic Counsel consistently overpolls the Bloc, and I'd be very surprised if their support were anywhere near 49%.

The updated projection is:

CON - 121
LIB - 103
BQ - 47
NDP - 37

Friday, September 4, 2009

New Angus Reid and Léger Polls

Here are the Angus poll and the Léger writeup. Almost everything in these polls are in line with recent data; the only new thing is that in the Angus Reid poll, the NDP is virtually tied with the Conservatives in BC. Accordingly, my projection just adds one NDP seat and subtracts one BC seat there:

CON - 120
LIB - 104
BQ - 46
NDP - 38

The Liberals got good numbers this week in Alberta, in both the Angus and EKOS polls. This doesn't net them any seats in my projections (and almost certainly wouldn't give them more than a couple in reality). This makes their vote less efficient, which cancels the beneficial effect of the growth of the Ontario gap (now back to about 5%). Thus, they remain 16 seats behind the Tories, even though they're only about 1% behind in the national poll average.

Wednesday, September 2, 2009

EKOS: Exact Tie!

32.6%-32.6%

Good news for the Liberals: for the first time in 5 weeks of EKOS polling, they have polled even with the Tories. Furthermore, they hold a 7.3% Ontario lead, and are only 1.4% behind the Bloc in Québec. However, this is a disappointing poll for them in MB, SK and BC, and their high number in AB probably doesn't help much seatwise.

Conversely, this is the lowest that the Tories have been in an EKOS in Ontario since late June, but the first time in 4 weeks that they're above 35% in BC.

Also, the Bloc's 32.3% in Québec is their worst non-CROP poll of the summer - and if ThreeHundredEight.com's list of polls is complete, their worst non-CROP since January!

Nothing surprising on the NDP or the Greens' side.

Given these regional figures, this poll implies something close to a tie, and pulls my aggregate projection even tighter:

CON - 121
LIB - 104
BQ - 46
NDP - 37

It'll be interesting to see next week's EKOS, now that the Liberals have announced that they will start voting against the Conservatives on confidence issues.

Tuesday, September 1, 2009

Pulling the Plug

After giving contradictory signals concerning their intention to defeat the government, the Liberals have apparently decided to indeed pull the plug. Given the strong and specific language used this time, it is hard to see how the Liberals can afford backing off once more - unless the polls take a dramatic turn for worse.

The official campaign would take place in October and early November, with November 9 being the most probable poll day. But no one will be surprised if September already starts to feel like a campaign. Stay tuned!

Friday, August 28, 2009

Angus Reid Poll

Angus Reid has put out a new poll giving the Conservatives a 34-30 lead over the Liberals. There are some interesting numbers from almost every region: the Tories are only at 20% in Atlantic Canada; the Grits and the Bloc are tied in Québec; the Greens are ahead of the NDP in Ontario (Cons and Libs par for the course there); the NDP is strong while the Liberals are weak in MB/SK.

But by far, the most striking numbers are in BC, where this poll has Conservatives 43%, NDP 28%, Liberals 19%, Greens 10%. The Tory result here is even better than the one in the Ipsos outlier (38%), while the Grits are even lower than the 20% they got in Harris-Decima's weird poll (that gave Greens 24% in BC). Incidentally, this is a low number for the Greens in BC, which is bad news for them since Angus and Nanos were the two firms that did not over-poll them in the last election (Nanos had them at a poor 11.1% in BC in its latest poll).

The BC numbers caused most of the action in my projection, which now gives:

CON - 122
LIB - 102
BQ - 47
NDP - 37

I have the NDP at the same number of seats as in 2008 (which is 8 more than in 2006) even though their national numbers are about 2% lower than in 2008 (and about 1% lower than in 2006). Essentially, the two provinces that matter for the NDP are BC and ON. In BC, compared to 2008, polls show the Tories losing ground and the Grits gaining support, while the NDP is roughly flat. Since most NDP contests in BC are with the Conservatives, this makes them gain seats. In ON, NDP losses are limited by the fact that many of their seats were won by a pretty comfortable margin in 2008. Furthermore, the fact that I use an arithmetic swing rather than a geometric swing helps the NDP retain most of their seats in my Ontario projection; I view this as reasonable since the NDP may benefit from being newly incumbent in most of their tight Ontario races.

Thursday, August 27, 2009

No November Election After All?

So it looks like, once again, the Liberal party is divided, though a previous report of Liberals wanting to pull the trigger was based on anonymous sources. Still, I do think that many Grits are very wary of Ignatieff being "Dionized" if he keeps bending, and tell themselves that while neither the numbers nor the state of the party's organization are great, they aren't disastrous either.

Keep an eye on that Ontario number in the polls. The Liberals will likely be happy to pull the trigger if they manage a consistent 5% lead there.

Wednesday, August 26, 2009

New EKOS and CROP

First, the weekly EKOS has come out, and doesn't show much movement relatively to last week. The two changes are that the Tories are back ahead in BC (by 3.3%, which is still worryingly low for them), while the Grits opened up a big lead (17%) in Atlantic Canada, but both these are likely due to sampling noise. Nationally, the Conservatives lead by 1.7%, while there's an exact tie in Ontario.

Second, CROP has published a Québec provincial poll, putting the Bloc and Liberals tied at 30%, followed by the NDP at a surprisingly high 18%, and with the Tories 4th with 17%. Unfortunately for the NDP, their 18% seems hopelessly uniformly distributed, which means that even if this result is borne out on election night, they probably would still get very few seats in La Belle Province. More importantly, according to this poll, although the Tories have lost less ground than the Bloc provincewide, they are now significantly behind in the Québec City area (Bloc 33, Lib 24, Con 23), which they carried last time. A CROP poll in June also had the Conservatives trailing (33-22) in that region. This means that apart from the seats in Chaudière-Appalaches and Louis-Saint-Laurent (Josée Verner), every Tory seat in Québec is in jeopardy.

Including these two polls and further discounting the Ipsos poll (now that it's even more likely to be an outlier) yields this new projection:

CON - 121
LIB - 104
BQ - 47
NDP - 36

This projection includes (at a discount) Angus, Nanos and Strategic Counsel polls that are several weeks old. But just including the last two EKOS and the last Harris-Decima would yield very similar estimates, so I'm comfortable with these numbers.

Monday, August 24, 2009

Harris-Decima Poll Details

The regional breakdown, except for MB/SK/AB, can be found in this writeup. Update: The full results are now available. It turns out that Harris-Decima continued weekly polls all summer, but played coy with the results, perhaps because the national numbers were mostly flat.

The most eye-catching number here is the 24% for Greens in BC. However, even with 24% in a four-way race, the basic arithmetic swing model gives them no seats. Of course, if the Greens truly have that much support, Elizabeth May would probably win her riding; furthermore, the support is likely to be unequally distributed, which would net the Greens some seats in the Lower Mainland and on Vancouver Island. Now, they probably aren't anywhere near 24% in reality: their best poll all year in BC had them at 16.3%.

Of greater consequence seatwise, the Grits are 6 ahead of the Tories in Ontario, which must be reassuring after Ipsos had them 12 behind. Bad news also for the Conservatives in Atlantic Canada (23%), Québec (12%) and BC (28%), as well as for the Liberals in BC (20%, probably due to the Green surge). The NDP gets solid numbers in Atlantic Canada (32%) and BC (26%), but does poorly in Ontario (13%).

Based on this poll alone, I get a virtual seat tie between the Libs and Tories, with the Bloc and NDP close to their 2008 results. My new aggregate projection is:

CON - 125
LIB - 100
BQ - 47
NDP - 36

My inkling is that the Liberals are actually a bit higher, and the Tories a bit lower. But I'll wait for more confirmation that the Ipsos poll is indeed a fluke before more heavily discounting it.

Half of Grits and Dippers Want May Elected

Here's an encouraging Harris-Decima poll for Greens: 40% of BCians, 49% of Libs and 53% of NDPers want to see May in the next Parliament. What does this mean?

For one thing, since this question isn't asked regularly, we don't know if this represents an increase or a decrease. I would interpret the result as follows: at most half of Liberal and NDP voters would switch their vote to the Green Party if May ran in their riding.

Greg Morrow at democraticSPACE has an excellent post about projecting the outcome in Saanich--Gulf Islands. Since that post, the polling in BC has worsened a bit for the Tories, and improved for the NDP. Also, I'd be a bit kinder to the Liberals at the expense of the NDP in that riding: some NDPers that switched over to the Liberal Party last time might decide to stick around. So if Elizabeth May weren't running in the riding, I'd say that things there now stand at about CPC 34, LPC 26, NDP 22, GPC 17.

If May indeed manages to draw half of LPC and NDP voters, she'd win. But of course, that's a bit unrealistic. My own very unscientific guesswork for Central Nova suggests that May drew 40% of NDP voters there in the last election. If she manages to draw 40% of Dippers and 35% of Grits this time around, she'd end up basically tied with Gary Lunn.

However, Greg Morrow, who has been involved with the Green Party and studied these issues more carefully, seems to think that May will only draw 25% of NDPers and 10% of Grits. The prognosis for May in that case is rather poor. This estimate is partly based on the fact that Ed Broadbent got an 8% bump in Ottawa Centre. I would venture that as a current party leader, May has more pull than Broadbent.

In summary, if May can "max out" and pull 50% of Liberals and Dippers, she has a good chance of winning. If she can only get 20% of them, it'll be a hard slug. For my projections, until a riding-specific poll comes out, I will use a proportion of 30% - somewhere between Greg Morrow's model and my back-of-the-envelope calculations. Currently, that's not enough for May to win, but she's close!

Also, Harris-Decima just put out a national poll (8/13-23, 2000+ respondents) today, but full details aren't available yet. From what I was able to glean from news websites, the headline numbers are: LIB 32, CON 31, NDP 16, GRN 11, BQ 9. In Québec, the breakdown is: BQ 37, LIB 28, CON 12, NDP 11, GRN 10. Unfortunately, I haven't been able to find the all-important Ontario breakdown... Hopefully, Harris-Decima will soon post the full results (or maybe the Canadian Press will decide to write a more detailed article on the poll), so that the projection can be updated. But at first glance, this poll might be the first indication that the recent Ipsos is indeed an outlier - we will have a better idea when the regional breakdowns and this week's EKOS become available.

A Weird Ipsos Reid Poll

Ipsos Reid came out with a poll that places the Conservatives (39%) far ahead of the Liberals (28%). Yup, an 11% gap! And the results suggest that the Tories have a 12% lead in Ontario (43-31), a clear break with the past 11 polls that have the Grits ahead there...

Projecting based on this poll alone gives a slim Conservative majority, with the three other parties losing about 5 seats each. Normally, the Tories need a bit more for a majority, but the unusual Ontario result changes that.

Incorporating this data into the projection obviously nudges it towards the Conservatives and away from the Liberals:

CON - 127
LIB - 99
BQ - 47
NDP - 35

However, if more recent polls from other firms confirm that this Ipsos poll is a strong outlier, I will depreciate it more quickly than I usually do.

Wednesday, August 19, 2009

EKOS: Small NDP Bounce

The weekly EKOS is out, and shows roughly the same results as last week, with the NDP up 0.8%, the Liberals down 0.8% and the Tories up 0.1%. The Conservatives lead overall by 2.6%, and the NDP polled a strong 17.3%.

The regional breakdown, however, is quite worrisome for the Grits. After several polls showing them ahead by 3-4% in Ontario, this one suggests that their lead is down to 0.9%. Now, given the smaller regional sample size, this could just be a statistical fluke, but given how consistent the Ontario numbers were since the start of August across all pollsters, it's a noticeable bump. If Ontario ends up anywhere close to being tied on election night, the Liberals can probably kiss their chance of forming government good-bye!

Now, the Tories didn't actually increase their support in Ontario: the Grits instead bled support to the NDP and the Greens. Moreover, the NDP has gone up for the 3rd week in a row in Québec, and is now back where they were last year - thus increasing Mulcair's chance of retaining his seat.

The only good news in this poll for the Liberals is in BC, where they lead the two other parties by about 5%. But at this point, I'd say that this is almost certainly statistical noise, though the Tory lead there does seem more and more fragile. And in any case, there are a dozen Alberta-like seats in BC that the Tories can't lose unless something totally freakish happens. That, combined with a few NDP strongholds, limits the Grits' potential for growth: even if they truly lead BC by 5%, they would probably still only get 10-12 seats. So Liberals shouldn't get too excited about this glimmer of light out West.

All this leaves the Grits down, and the Dippers up in my seat projection:

CON - 124
LIB - 103
BQ - 46
NDP - 35

If the top two numbers look familiar to you, it's because they're identical to the ones from the 2006 election. Looking at my projections over the past few weeks, the Tories are stuck around 125, the Libs bounce around 105, while the Bloc is flat at 46 and the NDP inched up from the low to the mid-30s. Will the fall bring us some significant movement?

Tuesday, August 18, 2009

Election Time

According to several Liberal strategists interviewed by La Presse, a fall election is on. That is, if the Bloc and NDP go along, and it's pretty hard to see how that wouldn't happen. This is of course unsurprising, given that for the Grits, the downside of keeping the government alive now seems to clearly outweigh the upside, and the polls indicate that they are likely to make major gains, and have a decent chance of winning.

A fall election would most likely take place on November 9.

Wednesday, August 12, 2009

New EKOS, Old Strategic Counsel

Two new polls are out: a Strategic Counsel poll from last week, and the EKOS poll for this week. They show an overall Conservative lead of 2% and 1.7% respectively, which fits into the recent trend. Sadly, Strategic Counsel polls are not very useful for projection purposes: they do not provide Atlantic Canada numbers, their Québec numbers tend to be out-of-whack, and they lump all Western provinces together. In Ontario, this poll shows a 3% Liberal lead, which is almost spot on the recent poll average.

This week's EKOS poll is good news for the NDP, who are having their strongest showing in a while in both Atlantic Canada and BC. It is also favorable to the Bloc, marking their third straight increase in EKOS polling (32.7, 34.2, 36.1, 38.5). But because Ontario, which fluctuated wildly in late Spring, is now stubbornly stable, my projection remains almost flat:

CON - 123
LIB - 106
BQ - 46
NDP - 33

The EKOS poll also asked if Canadians would prefer a two-party system. They were quite split on the question, with a slight majority (56%) answering 'no'. Variations across political parties weren't huge: unsurprisingly, the Tories were more open to a two-party system, while the NDP and Greens were less so. Liberal voters were also less prone to support a two-party system, perhaps because their base consists disproportionately of university degree holders, who strongly oppose such a system (66%). The surprise was that Bloc voters were actually the most receptive to a two-party system (51%)! Perhaps in their mind, a two-party system means BQ vs. Lib in Québec. Or maybe a lot of them would actually rather vote Liberal or Conservative, but are voting Bloc strategically.

The poll also gave the following approval ratings for the leaders: Stephen Harper is at -11% (36-47), Michael Ignatieff is at -9% (29-38), while Jack Layton barely eked out a positive number at +1% (34-33). Not surprisingly, approval figures for Harper and Ignatieff mirror party support across the country. Layton is at around 30% everywhere, except in Québec where he's the most popular federalist leader at 40%, even though the NDP will be hard-pressed to win more than 1 out of 75 seats there. Own party approval ratings are not great for any leader, and especially low for Ignatieff (Harper 79%, Layton 68%, Iggy 59%). Iggymania is definitely over, even among Grits!

By contrast, President Obama's net approval rating among Canadians is a whopping +62% (73-11). Liberals especially like him (86-7), while Conservatives are less supportive (still 66-17). Finally, 87% of Canadians believe that Canada has a better health care system than the U.S., while only 7% believe the contrary. That 87% is a combination of 80% in Québec and around 90% in all other regions (even Alberta). Like last week's EKOS poll, this one shows us that Quebecers are not as left-wing as their voting patterns or provincial politics suggest.

The problem for the Tories is that Québec conservatism is entwined with Québec nationalism, which necessitates government interventions to support the arts and preserve heritage; the Conservative base is probably the most hostile of any party's to both Québec nationalism and government subsidies. Meanwhile, I think the ADQ wouldn't have fizzled out if it had a more credible team - but that's of course a chicken-and-egg problem.

Monday, August 10, 2009

Nanos "Third Quarter" Poll

The first non-EKOS-or-Angus national vote intention poll in over a month just came out, from Nanos. The results (accessible here) show a 2.5% lead for the Liberals. Also, the NDP has a strong showing: 18.7%. Both these findings contrast with the EKOS poll's from last week, even though the two polls were done during roughly the same period (7/29-8/4 for EKOS, 7/30-8/2 for Nanos).

Regionally, as is customary with Nanos polls, the Grits are stronger than usual out West, especially in BC where they come out ahead of the Tories. However, this strength in Western Canada is mostly wasted on the Prairies and in the BC interior, so my model actually still spits out a small Tory minority based on this poll alone. Indeed, the 3.4% lead for the Liberals in Ontario is insufficient to put them ahead of the Tories, especially when the NDP is strong, as is the case in this poll.

My overall projection changed only slightly, with the Tories losing 2 seats out West, to the profit of the Grits and the NDP:

CON - 123
LIB - 107
BQ - 46
NDP - 32

Wednesday, August 5, 2009

EKOS: Tories Ahead by 3%

It's time for the weekly EKOS update! Just as last week's EKOS was good for the Liberals, this week's is favorable to the Tories, who are up 34.9%-31.9%. This is the largest lead in the EKOS poll for either party in the last 8 weeks. My projection, however, hasn't changed much as a result:

CON - 125
LIB - 106
BQ - 46
NDP - 31

This small movement is due first and foremost to the fact that I'm using the last 3 EKOS polls (though discounting the oldest), since only two firms are polling regularly over the summer. Thus the latest poll replaces not data from last week, but from mid-July, which wasn't that great for the Grits either. Also, although the Tories went from -1.6 to +3 in the national numbers from last week to this week, they only improved from -4 to -2.7 in all-important Ontario.

Other interesting regional snippets include the 19% result for the Tories in Québec (their best in a long time), and the 29% result for the Grits in Manitoba/Saskatchewan (confirming their strong showing in both of last week's polls).

The day-by-day numbers actually show the Liberals leading last Wednesday and Thursday, but falling behind on Friday, and trailing badly this Tuesday. I guess we'll have a better idea next week if this is a fluke or a trend, but can you think of anything drastic happening toward the end of last week or over the weekend?

The poll also asked Canadians to rate the economy. 74% say that it's still in a recession, and the rest are split almost evenly between depression (12%) and growth (14%). Quebecers are the most pessimistic (even though the downturn is likely quite a bit milder there than in Ontario), while Mabitobans/Saskatchewanians are the most optimistic. Bloc and NDP supporters are very downbeat, while Tories seem happier about the economy (still, only 19% of them see growth). My guess is that left-wingers tend to be more risk-averse, and thus more sensitive to small changes in income than right-wingers. This would explain both these sets of results.

Also, despite study after study from Statistics Canada, Canadians are almost twice as likely to believe that crime rate in Canada has increased rather than decreased over the past 10 years (48%-26%). Quebecers were most likely to be right, while residents of the Prairies and Atlantic Canada were most likely to be wrong. However, in all regions, more people believe that crime has increased than decreased. Unsurprisingly, more educated people were right more often, but even people with a bachelor's degree were slightly more likely to be wrong than right (37%-35%). The undecided voters were most paranoid, followed by (roughly tied) the Tories and the Greens (!), while Bloc voters had the most accurate perceptions, followed by (roughly tied again) the Grits and the Dippers.

When asked what to do about crime, Canadians favoured the Stick (more police, longer jail terms) over the Carrot (crime prevention, rehabilitation) by 53%-42%, though people with a university degree thought the opposite by 56%-40%. Regionally, BC, MB/SK (!) and ON voters were somewhat softer (roughly 50-45) than AB, QC (!) and Atlantic voters (about 57-38). In terms of vote intention, from toughest to softest, we get Conservative, undecided, Bloc, Liberal and NDP (basically tied), Green.

So I guess we learn from this that there's a silent hawkish majority on crime in Québec, especially within the Bloc. The latter might reflect the awkward fact that the Québec independence movement has been led by-and-large by intellectuals, even though it relies heavily on blue collar workers for support - it is not a grassroots/populist movement like the Conservative one, for better or for worse. (And, as a socially liberal Quebecer, I'd definitely say 'for better.')

Tuesday, August 4, 2009

NDP Name Change?


The NDP may soon decide to drop 'New' from their name. Attempt go get closer to the Obama halo? And would the new logo be an orange donkey?

Of course, if the name change goes through, and there's a coalition, it would likely be called the LibDem coalition. Now, I'm usually all for free trade, but importing names for political parties from foreign countries seems a bit much. It's also a bit ironic that Canada's most protectionist major party is the one that might make those imports happen...

(Update: True, "Conservative" and "Liberal" aren't homegrown names either, but those "imports" happened pretty much at Confederation, and the names are actually descriptive. Is it too much to ask that a new non-descriptive party name in a mature state be different from that of a major party in a closely allied country?)

Thursday, July 30, 2009

Super Thursday!

Angus-Reid also posted their most recent poll today, with results very similar to EKOS'. This poll is good news for the Liberals in Ontario, but not so good in BC. The national numbers (34-33 for Grits) are basically the same as in the EKOS poll, and my projection based on this poll alone also has the Tories winning by a very small number of seats. We are really in a dead heat!

Overall, it looks like if an election had occurred this past week, we would have basically had a rerun of the 2006 election outside Québec. In Québec though, the Liberals are a lot stronger, and the Conservatives are weaker, as is the Bloc.

Both polls from today suggest that the dominant opposition party in Manitoba/Saskatchewan has become the Liberals, by over 10%. The NDP beat the Grits by 7.6% in this region in the last election. This doesn't have big seat implications since the Tories are far ahead there, but it suggests that the Liberals are a true national party again.

My aggregate projection still shows the Conservatives ahead by 15 seats:
CON - 123
LIB - 108
BQ - 46
NDP - 31
That's because there are relatively few polls this summer (basically only EKOS and Angus), so I'm temporarily using more than one poll from each company. But if more data confirm today's polls, the numbers will keep moving in the Liberals' favor.

Weekly EKOS

Latest projection:
CON - 124
LIB - 106
BQ - 46
NDP - 32

Not much change for any of the parties. In this new poll, the Grits have taken a 34.1-32.5 (statistically insignificant) lead in the national popular vote. However, even looking at this poll alone, my seat projection gives the Tories a marginal lead. The Liberals have suspiciously good numbers on the Prairies in this poll, but these don't really help seat-wise. The Grit Ontario lead stays at 4%, which is very low considering their national lead.

This is the 7th EKOS poll in a row where the difference between the top two parties is statistically insignificant (keeping in mind that the uncertainty is higher when evaluating the difference between two parties than evaluating a single party's support).

Friday, July 24, 2009

First Mapped Projection











This is not a new projection; I'm just mapping this projection from July 23, based on polling through July 21. Mostly, relative to 2008, we have the Liberals gaining in close suburbs, but still not quite there in the outer ring of most cities. I will probably map another projection when there's significant movement, or when an election gets called. Stay tuned!

2008 Result Maps!











I have decided to spice things up on this blog by adding maps! In this post, you can view the results of the 2008 election, and if there's enough demand, I'll be happy to post maps of the 2004 and 2006 results as well (yes, including the different boundary between Acadie--Bathurst and Miramichi in 2004).

In the future, I will occasionally be mapping my projections. I will not always do so because:

1. While aggregate seat projections are a hazardous exercise, specific projections are even more prone to error: the idea behind the aggregate projections is that if some ridings swing more, and others less, things will roughly balance out in the total.

2. It takes time!

I will, however, certainly map the last projection before the election, and possibly other "significant" ones (e.g. at the start of the campaign) as well.

Enjoy!