Latest national poll median date: October 20
Projections reflect recent polling graciously made publicly available by pollsters and media organizations. I am not a pollster, and derive no income from this blog.

Sunday, October 30, 2011

Bill C-20: Representation

Well, moving back to Canada, settling down, and starting a new job have taken their toll, but given my past posts (see here and here) on this issue, I wanted to say a few words about Bill C-20.

My comments will refer to the following five redistribution schemes, all of which ensure that no province loses seats relative to the 1988/1993 elections:

1. Current law: divide 279 seats proportionally among the 10 provinces, and then make sure no province loses seats.

2. My proposal: make the size of the House such that Québec is proportionally represented with 75 seats. Divide the remaining seats proportionally among provinces other than Québec. Then make sure no province loses seats, and divide the remaining seats proportionally among the remaining provinces. Iterate until no province loses seats.

3. Second* Tory proposal: Québec gets 75 seats, and every province gets seats in proportion. Then make sure no province loses seats.

4. Third Tory proposal: each province's number of seats is its population divided by the electoral quotient, but no province loses seats. The electoral quotient is 108,000, and then increases with the 10-province population growth.

5. Current Tory proposal (C-20): each province's number of seats is its population divided by the electoral quotient. The electoral quotient is 111,166, and then increases with average of the population growth in each of the 10 provinces. Then make sure no province loses seats. Then add seats to Québec such that it is not underrepresented relative to its share of the 10-province population.

The first 4 schemes use raw census counts. C-20, however, uses population estimates based on census counts. The advantage is that this corrects for some people not counted by the census. The disadvantage is that figuring out how many people the census missed is somewhat arbitrary, so whoever comes up with those estimates (Statistics Canada bureaucrats, perhaps under pressure by the government) has the power to influence the number of seats.

*If you remember, the first Tory proposal sought to disadvantage Ontario with respect to Alberta and BC due to Ontario being a "large" province. Obviously, that went nowhere.


Short-run comparisons (what happens after the next redistribution)
- All of the following are estimates, based on current population estimates for July 1, 2011 (which haven't yet incorporated 2011 census data), and on the 2006 census' estimated undercounts.
- Under all proposals, provinces and territories other than BC, AB, ON and QC keep their current seat count, for a combined 63 seats.


BCABONQCTotal
Current law373010875313
My proposal383111175318
Second Tory423512275337
Third Tory423411975333
Bill C-20423412178338

- Under C-20, relative to current law, the House expands by 25 extra seats. Over half of that gain (13) goes to Ontario, with 5 accruing to BC, 4 to AB and 3 to QC.

- Under C-20, Alberta could very possibly end up with 35 seats, while BC and ON may well end up with just 41 and 120 seats respectively - this will depend on how the census plays out.

- Under C-20, relative to the previous Tory proposal, Québec obviously gains, but not all other provinces lose. Indeed, the change is roughly neutral for Ontario and Alberta, at least if census undercoverage remains similar as in 2006: it was around 4% in ON and AB, but just 3% in Canada as a whole. Thus, using population estimates rather than census counts gives Ontario 1-2 extra seats, and pushes Alberta closer to that 35th seat. This roughly compensates for the slightly larger size of the House.

- C-20 is a little worse than my proposal for Québec even though both seek to make its seat share equal to its population share. The reason is that census undercoverage tends to be lower in Québec - just 1-2% in 2006 - so using raw census data would give Québec a higher share than using population estimates. This is, of course, just a measurement issue, and in principle, C-20 treats Québec in the same way as my plan, so I view it as fair. Props to the government on that.

- Overall, I'm ambivalent towards the main difference between C-20 and my plan: the size of the House. On the one hand, C-20 makes it quite a bit bigger, with all the associated expenses. On the other hand, given that the 6 small provinces' seat numbers are fixed, the only way to reduce their overrepresentation is to have more MPs. I think I'd be willing to swallow the pill if future growth in the House size were controlled under C-20. Is it the case?


Long-Run Comparisons

All five plans have the following common-sense feature: the size of the House doesn't change if all provinces grow at the same rate. But this condition is far from holding in Canada! In each plan, the House grows if historical trends continue.

1. Current law: the rate of expansion of the House depends on the difference between BC, AB and ON's population growth rate and the 10-province population growth rate. 6-7 seats in the past two redistributions, and would probably have been 5-7 this time.

2. My proposal: the rate of expansion of the House depends on the difference between the national population growth rate and Québec's population growth rate. This difference has shrunk dramatically recently: just 0.1-0.2% per year in the past three years. That translates into a 3-to-7-seat increase in the size of the House per decade. In the late 90s, the gap was instead 0.5-0.6% per year, which would mean a 16-to-20-seat increase per decade.

3. Second Tory proposal: the rate of expansion of the House depends on the difference between BC, AB and ON's population growth rate and Québec's population growth rate. Since this difference would only affect BC, AB and ON's seats, the House would grow roughly as fast as under my proposal. (The fact that BC, AB and ON's seat shares are slightly below their population shares should be roughly canceled by the fact that the smaller provinces grow, on average, a bit slower than Québec.)

4. Third Tory proposal: essentially the same as current law.

5. Bill C-20: the rate of expansion of the House depends on the difference between BC, AB and ON's population growth rate and the average of the population growth rates of the 10 provinces. The latter is likely to be significantly lower than the 10-province population growth rate because smaller provinces send to grow slower. As a result, the House would grow more quickly under Bill C-20 than under current law or the previous Tory plan. This discrepancy was only about 0.1% in the past three years because of healthy growth rates in SK, MB and PEI, but it was 0.4% as recently as in 2007, and 0.5% from 1991 to 2001. Since BC, AB and ON would have about 60% of seats, Bill C-20 is likely to increase the House size 0.7-3% per decade more quickly than current law. This would mean a total increase of 7 to 18 seats per decade, about the same as under plans 2 and 3 above.

What we see is that Bill C-20 combines the largest immediate seat increase and the largest House growth rate of the 5 plans. It's an expensive way to correct representational inequities, necessitated by the six small provinces being unwilling to recognize the logical consequence of their reduced population weight...

Monday, August 22, 2011

R.I.P. Jack

My thoughts are with his family. Coping with the loss of any family member is difficult. But from recent experience, I can attest that when the decedent is a public figure, there are even more things to arrange and take care of.

A sidenote: the CPC is now the only party in the Commons with a permanent leader and more than one MP.

Monday, July 25, 2011

Layton Has Second Cancer, Takes Break

Layton suggests that Hull--Aylmer MP Nycole Turmel be interim leader of the NDP.

Good luck, Jack.

Wednesday, June 22, 2011

PQ In-Fighting Opens Door to Legault

A Québec provincial poll suggests that the centre-right party proposed by François Legault has 40% support against 26% for the Liberals, 17% for the PQ and 8% for the ADQ. Just like for the NDP, such a result would give Mr. Legault a commanding majority.

Of course, as Legault gets more scrutiny, his support may wane, though that didn't happen with the NDP. Still, these results imply that 74% of the Québec electorate is ready to vote for a right-of-centre party. This is consistent with the analysis that Québec's NDP vote on May 2 was for change rather than for a left-wing agenda.

Even without Legault's new party, the PQ would lose by 6% to the Liberals (29-35). It goes without saying that the recent crisis has greatly damaged Marois' party.

Note: Sorry for the lack of posting in the past month - three major life events that normally only occur every few years and lots of traveling have taken their toll. Posting will likely remain limited for the next three months, but hopefully there won't be any more whole-month gaps.

Monday, May 23, 2011

Election Results Are Official

The election results from every riding have been validated, and the four recounts have been completed. As previously noted, the validation process moved one Québec riding from the Tories to the NDP. None of the recounts produced a further change in the winner.

Wednesday, May 18, 2011

Economy, Economy, Economy

For any government, the state of the economy is of paramount importance. But it is even more so for the Harper government over the next four years. After all, it is its perceived competence in that matter that drove Torontonians to give it a majority.

The Conservative economic and fiscal platform was basically: we'll make sure the recovery continues; this will eliminate most of the deficit, and we'll make some not-too-painful cuts to eliminate the rest.

Whether the recovery continues or not is, alas, largely outside the incoming government's control. Obviously, Canada has little impact on international factors such as the U.S. recovery. But even on the domestic front, most of the story has already been determined, and we're just waiting to see how things play out.

Indeed, on macroeconomic issues, a government's performance often depends more on the previous government's actions than on its own. For example, Mulroney's poor showing had much to do with Trudeau's economic mismanagement, while foundations of the Liberal success in the 90s were laid by the Tories with the GST and FTA. Similarly, the continued success of the Canadian economy under Harper was mostly a product of Chrétien and Martin's sound policies and Canadians' sacrifices in the 90s.

In many ways, over the next four years, we will find out how good the Tories' management of the economy has been over the past five years. We will have a better sense of whether:
- the government's stimulus plan succeeded in producing a sustainable recovery;
- the Tories spent outside the country's means prior to the recession;
- there is a housing bubble which the Tories should have tried to prevent.

How these issues play out will have a direct impact on Canadians' well-being and on the government's agenda. If there is no housing bubble, the recovery is sustained, and past spending increases were reasonable, then the Conservatives can deliver on their plan to balance the budget with moderate cuts that most Canadians won't notice. However, if we find out that there is a housing bubble (through it bursting), if the recovery runs out of steam, or if the spending growth of the good times was excessive, then the government will have to introduce unpopular measures, show up to the 2015 election with a deficit, or both.

Barring major events in Québec or a major scandal, the Conservatives' fate in the 2015 election may have already been largely determined.

The same cannot be said of the Liberals and the NDP. Obviously, if the economy does well, they have little chance of winning power in 2015. But the NDP can durably squeeze out the Grits if they manage to establish economic credibility over the next four years. Conversely, if the NDP fails to do so, then the Grits could regain the position of government-in-waiting if they get their act together. Of course, since the election, both parties' performances have ranged from disappointing to laughable...

Sunday, May 15, 2011

Districts of Different Sizes: Who Benefits?

It is well-known that the size of electoral districts (whether measured by population, electorate, or votes) varies wildly, not only across provinces, but also within provinces. When we put everything together, which party benefited from these discrepancies in 2011?

One measure of this would be to compare a party's actual voting share to what I'll call its effective voting share, defined as the average of that party's share of votes in each riding.

For example, suppose that a country has two electoral districts, with 40 votes in district A, and 60 votes in the district B. A party gets 50% of the vote in A, and 30% in B. Clearly, that party is favoured because it's doing well in a less populated district. Its actual national vote share is 38% (38 votes out of 100), but because the districts have equal weight, the party's effective voting share is 40% (average of 50% and 30%).

For each of the major parties, below is its effective voting share, with its actual voting share in parentheses:

CON: 39.7% (39.6%)
LIB: 19.3% (18.9%)
NDP: 30.8% (30.6%)
BQ: 5.6% (6.0%)
GRN: 3.7% (3.9%)

CON: 39.9% (39.8%) in 307 ridings with candidates
BQ: 23.2% (23.4%) in Québec
GRN: 3.8% (4.0%) in 304 ridings with candidates

We immediately notice that all three main parties modestly benefit from some votes being more important than others. This may be slightly surprising in the case of the Tories, whose support is concentrated in underrepresented BC, Ontario and Alberta. However, this is more than countered by Conservative support being concentrated in rural ridings, which tend to be less populated, and by Albertans voting less than other Canadians.

The Bloc and the Greens, on the other hand, are hurt. The Bloc is disadvantaged mainly because Québec votes are worth less than the average vote elsewhere due to a larger number of electors per district, while the Greens' support is concentrated in underrepresented BC.

The lesson I take from this exercise is that this distortion is rather minimal: no party's effective voice is changed by more than 0.4% (1 seat if we had a proportional system). So while the variation of district sizes may be unfair to certain constituencies, it does not significantly affect party representation in the Commons.

Saturday, May 14, 2011

Safe Seats

Safest Seat in Each Province

NL - St. John's East (NDP by 50.3%)
PE - Egmont (CON by 23.3%)
NS - Central Nova (CON by 32%)
NB - Acadie--Bathurst (NDP by 53.5%)
QC - Gatineau (NDP by 46.7%)
ON - Wellington--Halton Hills (CON by 47.3%)
MB - Portage--Lisgar (CON by 66.2%)
SK - Souris--Moose Mountain (CON by 55.3%)
AB - Crowfoot (CON by 74.8%)
BC - Abbotsford (CON by 44.8%)
North - Nunavut (CON by 21.2%)

It's quite striking that the safest seat in Québec is a turnover. Also, the safest seat in Nova Scotia was projected incorrectly (though I'm in good company): MacKay was far from safe in 2006, and of course it was hard to say what to make of 2008 with May as a candidate.

Seats Won by over 50%

Outside Alberta
66.2% - Portage--Lisgar, MB (CON)
55.3% - Souris--Moose Mountain, SK (CON)
53.5% - Acadie--Bathurst, NB (NDP)
52.7% - Provencher, MB (CON)
50.3% - St. John's East, NL (NDP)

In Alberta (obviously all Conservative)
74.8% - Crowfoot
70.0% - Wetaskiwin
68.5% - Vegreville--Wainwright
67.1% - Macleod
66.0% - Calgary Southeast
65.7% - Westlock--St. Paul
64.0% - Yellowhead
63.5% - Wild Rose
63.2% - Calgary Southwest
60.8% - Red Deer
59.6% - Peace River
58.6% - Fort McMurray--Athabasca
58.5% - Medicine Hat
57.7% - Calgary--Nose Hill
55.3% - Edmonton--Spruce Grove
53.3% - Calgary East

The only riding outside of Edmonton or Calgary that was decided by less than 58.4% was Lethbridge, where Conservative Jim Hillyer won by 29.3% (he still carried 56.5% of the vote). This is the mystery candidate known for going to the bathroom when cornered by a reporter.

The median margin of victory in Alberta was 56.5%...

Addendum: The largest margin of victory for a Liberal was 31.9% in Humber--St. Barbe--Baie Verte; outside of Newfoundland, it was 19.1% in York West. The largest Bloc margin was 10.5% in Haute-Gaspésie--La Mitis--Matane--Matapédia (new MP), less than Elizabeth May's 10.7% cushion in Saanich--Gulf Islands.

Just under half the ridings went by over 20%: 106 to Tories, 41 to Dippers and 2 to Liberals.

Ontario and Toronto

Despite a larger swing away from the Liberals in the Toronto area, the region remains much friendlier to the Grits than the rest of Ontario, which is a complete Liberal wasteland. Below are some telling stats comparing the 39 ridings entirely within the Toronto CMA to the 67 other Ontario districts.

Seats won
Toronto CMA: 24 C, 8 N, 7 L
Elsewhere: 49 C, 14 N, 4 L

Came in first of second
Toronto CMA: 36 L, 31 C, 11 N
Elsewhere: 66 C, 45 N, 21 L, 1 G, 1 I

[Aside: Without looking up election results, can you guess the only Ontarian riding outside Toronto CMA where the Conservative candidate failed to make the top two?]

Won, or within 15% of winner
Toronto CMA: 31 C, 29 L, 13 N
Elsewhere: 56 C, 19 N, 10 L

Clearly, even though the Liberals won fewer seats than the New Democrats in the Toronto area, they were still the Tories' main opponents. In fact, in every single one of those 39 ridings, a Liberal won, was second, or was within 15% of the winner. (The comparable figures were 31 for the Conservatives and 14 for the NDP.)

Indeed, within the Toronto area, the Liberals caught lots of bad breaks. Looking at the number of wins and the number of losses by less than 10%:
CON: 16-4
NDP: 2-3
LIB: 3-18

[Incidentally, yes, 21 of 39 races within the Toronto CMA were decided by less than 10%. The same statistic was 12/32 for the Atlantic, 21/75 for Québec, 9/67 elsewhere in Ontario, and 18/95 in the West and North. Thus, despite having less than 13% of the ridings, Toronto accounted for almost 26% of the close races nationwide.]

Elsewhere in Ontario, however, the Grits were almost completely out of it. The only ten ridings where a Liberal won or came within 15% were:

- Guelph
- Kingston and the Islands
- Kitchener Centre
- Kitchener--Waterloo
- London North Centre
- Nipissing--Timiskaming
- Ottawa--Orléans
- Ottawa South
- Ottawa--Vanier
- Ottawa West--Nepean

All of these are mainly urban, except for Nipissing-Timiskaming.

Friday, May 13, 2011

NDP Win Confirmed in Montmagny--L'Islet--Kamouraska--Rivière-du-Loup

One judicial recount is complete. In Montmagny--L'Islet--Kamouraska--Rivière-du-Loup, the NDP won by 9 votes, four more than the initially validated results. The preliminary results had the Conservative candidate as the winner by 110 votes on election night, but it was subsequently discovered that a ballot box's NDP votes were mistakenly given to the Greens.

This confirms that the NDP won 59 seats in Québec, and the Tories, 5. Despite coming fourth in the popular vote, the Grits have the second-most seats (7), and of course the Bloc only got 4 seats.

Tuesday, May 10, 2011

Democratic Reform

My guess is that the most significant democratic reform we will see under this Conservative majority is a change in the seat allocation formula for the House. Perhaps the government will prove me wrong by moving forward on Senate reform or other issues, but I doubt it. Beyond the fact that it is now in the Conservatives' interest to keep the Senate as is and to stick with first-past-the-post, there's also the fact that reformers don't agree on what the final product should look like. As often in these situations, the status quo prevails.

Take Senate reform, for example. You'd have to decide on at least the following issues:

1. The form of the Senate:
a) Appointed
b) Elected
c) Abolished

2. If you answered a or b to Q1, the seat allocation principle:
a) None (PM appoints from anywhere, or pure nationwide proportional voting)
b) By province
c) By region, as is
d) By region, some other way

3. If you answered b to Q1, the voting method:
a) Nationwide proportional
b) Proportional by province/region/large constituency
c) First past the post
d) Something else

And of course, your answer on these might depend on what we do with the voting system for the House. I'm pretty sure that no single combination of answers to just these three questions would be picked by a majority of Canadians, and my guess is that none would come even close.

My question to you: if you could change the way in which we choose are MPs and Senators to your liking, what would you do? (My preferred solution involves abolishing the Senate and moving to a sort of MMP system, modified to make majorities easier to achieve.)

Monday, May 9, 2011

One Person = One Vote: What Is a Person?

Before reading the rest of this post, ask yourself: when we say 'One person = One vote', what should we mean by person? Do we mean:
a) resident of a constituency;
b) eligible voter; or
c) actual voter?

In the debate about seat distribution, it has been assumed that the first definition is the one to use. It is indeed the current basis for seat allocation at the federal level. But is it fair that a citizen's vote counts more just because he lives next to more people that can't vote (e.g. immigrants, children)?

I suspect that at least a few of you believe definitions b) and c) make more sense. Even if you don't, I hope that you agree that they are not crazy.

So let's take a look at the Conservative redistribution bill (C-12), which will likely be reintroduced and passed in the new House, under the lens of definitions b) and c).

We know that under Bill C-12, Ontario will likely have at least 119 seats, Alberta 34, and BC 41. (These numbers could be as high as 122, 35 and 42, but let's just take the low numbers, since they work against the argument I'm making here.) Here is the number of electors per riding:

QC: 81,479
BC: 74,739
ON: 74,729
AB: 72,865
National: 72,204

And here's the number of valid 2011 votes per riding:

QC: 50,646
ON: 46,481
BC: 45,650
National: 44,339
AB: 41,098

As you can see, Bill C-12 is extremely unfair to Québec voters, singling them out for unfavourable treatment. Québec voters would be worth 9% less than voters in any other province.

The reason for this inequity is simple: Bill C-12 reduces Québec's weight in the House, but Québec votes and voters are already undervalued! Indeed, in the 2011 election, Quebecers accounted for 25.5% of the country's eligible voters, and 25.8% of actual valid votes. Yet, Québec's weight in the House is only 24.4%, and would be reduced to 22.3-22.6%. That's below even its population share of 23.2%.

Now, look at our current system. Here's the number of valid votes per riding in the 2011 election:

ON: 52,182
BC: 51,991
QC: 50,646
AB: 49,905
National: 47,794

Compare these to the numbers above. Which are more equitable?

The Albertan self-righteousness in this debate would put Duceppe to shame.

Sunday, May 8, 2011

It's Not a Good Sign...

... when the featured article of Québec's main federalist newspaper's website is titled, "Newspapers in English Canada Demand the Marginalization of Québec".

Columnists like Lorne Gunter might consider that, in fact, that a majority of Canadians outside Québec voted for what he calls a "hysterical approach to the environment." Indeed, if he truly wants various groups' influence on policy to be proportionate to population, well, 60% of Canadians voted for left-of-centre parties.

I've blogged about this before (yes, part IV of that post is now obsolete), but at this point it's relevant to remember that Bill C-12 would make Québec's seat share in the House of Commons disproportionately low. It'll be interesting to see if the NDP puts up a fight, now that it is the main custodian of Québec's interests.

(Here are the facts: taking into account the most recent population estimates and census undercounting, Bill C-12 would likely give Ontario 13-16 extra seats, Alberta 6-7, and BC 5-6. This means a House size of 332-337, implying a Québec weight of 22.3-22.6%. Under current law, Québec's weight in the House would decrease to around 23.9%, while its population share is 23.2%. Therefore, under Bill C-12, Québec's weight in the Commons will be as far from its population share as it would be under current law - just in the opposite direction.)

Saturday, May 7, 2011

Liberal Party: Uncertain Future

After the 1953 election, when facing a united Right, the Liberals only won more non-Québec seats than the Conservatives twice: 1968 (Trudeaumania) and 2004 (Right still gluing pieces together).

If the Liberals want to become a party of government again, it will have to regain Québec (unless the Right self-destructs again...). That's not too hard to imagine: if the NDP's performance as Official Opposition is unsatisfactory - this first week has pretty much been a disaster - and if the separatists decide to disband the Bloc and focus on the PQ, then the Liberals have a chance of getting Québec back in 2015.

Still, it's equally easy to imagine a scenario where the NDP becomes Québec's default party for several elections. If the Bloc manages to find a bit of money, it could also win Quebecers back next time. Moreover, if Stephen Harper steps down in, say, 2014, and the Tories find a more moderate leader, they'd have a shot at Québec as well.

At this point, all of the following seem plausible to me:

- Québec stays with the NDP for several elections, and the Liberals become a permanent third party, sometimes wielding the balance of power when the Left and Right are equally matched;

- Québec stays with the NDP for several elections, and the Liberal Party disintegrates, leaving Canada with a two-party system;

- Québec returns to the Liberals, and we are back to the situation of the 60s and 70s, but with the Tories stronger due to the increased weight of the West;

- Québec picks the Tories or Bloc in 2015, and the Liberals and NDP, two severely weakened parties, have no choice but to merge;

- Québec sticks with the NDP, but the Liberals regain the GTA in 2015, and a merger/coalition is considered.

All of the above assume that Québec does not hold a separation referendum where Yes support exceeds 50%. However, it is a real possibility that the PQ wins the next provincial election during a period of widespread discontent with an ideologically incompatible federal government. If the PQ can stoke the sentiment that Québec is being oppressed by a majority imposed by Ontario and the West, it may yet get a majority of the votes in a referendum. If that happens:

- in case of a "clear" majority, game over, Québec secedes - this is very unlikely;

- otherwise, all hell breaks loose: separatists will want to hold yet another referendum, or worse, unilaterally declare independence; Tories will be seriously questioned about potentially "losing the country"; the NDP will face large internal dissent; the Bloc might be back stronger than ever, totally withdraw from Ottawa, or both (win 50+ seats in 2015 and have its MPs not show up). For the Liberals, the implication would be very murky: it might be the party Canada turns to in order to sort out this mess, or might be pushed into complete oblivion as all of the attention is focused on the others.

All this to say that at this point, virtually anything could happen to the Liberal Party. Its best chance of averting oblivion is Québec, but whether it'll even get a chance depends more on how the other parties - NDP, Tories, PQ - perform than on what it does. For this reason, what appears to be the Liberal strategy of lying low for now seems like a good one.

Thursday, May 5, 2011

The Number of this Election: Four

- Ignoring intervening by-elections, the Tories only gained a net four seats outside the Greater Toronto Area. Inside the GTA, they gained 19 and now hold a majority.

- This is just the fourth time in Canadian history that a party wins a majority with such a small vote share. The others? 1867, 1874 and 1997.

- West of Guelph, the Liberals have only four seats, and in each case, the margin of victory was below 5%. In fact, outside of Newfoundland, no Liberal candidate reached 50%.

- The Liberals also only have four seats outside the Atlantic and metropolitan areas with over 1 million people: Kingston and the Islands, Guelph, Winnipeg North and Wascana. The former two are university towns, while the latter two had very personally popular Liberal candidates.

- Of the four ridings won by the Bloc, three were due to the presence of a strong (> 25%) non-NDP federalist candidate that divided the vote. Only in Bas-Richelieu--Nicolet--Bécancour did the Bloc win a two-way race.

- Outside Québec and the City of Toronto, the NDP won just four seats carried by the Liberals in 2008: St. John's South--Mount Pearl, Dartmouth--Cole Harbour, Newton--North Delta and Esquimalt--Juan de Fuca. In fact, relative to 2008, the NDP gained no net seats outside Québec, the City of Toronto and Greater Vancouver.

- With just under four per cent of the vote, the Greens not only carried a smaller share of the popular vote than in 2008, but they are also down compared to 2006 (Update: and 2004, a reader points out).

- This was Stephen Harper and Jack Layton's fourth election as party leaders.

Tuesday, May 3, 2011

Model Performance: Comparison with Other Projections

Updated to reflect change of lead in Montmagny--L'Islet--Kamouraska--Rivière-du-Loup

This post will compare the accuracy of my final projection with that of others that were freely available during the campaign. For a region-by-region analysis of how my projection fared, click here. Once again, the websites mentioned here are linked through the left bar.

Here are the eight final projections that were based on an average of polls (C-N-L-B-G-I):
148-100-44-14-1-1 (The Mace)
152- 94-46-15-0-1 (Canadian Election Watch)
142-114-39-12-0-1 (Riding by Riding)
151- 91-47-18-0-1 (Calgary Grit)
144- 98-51-15-0-0 (LISPOP)
155- 86-47-20-0-0 (democraticSPACE)
143- 93-58-14-0-0 (Too Close to Call)
143- 78-60-27-0-0 (ThreeHundredEight.com)

First, I would like to congratulate democraticSPACE for being the only model projecting a Conservative majority. On that front, I fared honourably by having the second highest Conservative seat count. Four websites estimated the probability of a Conservative majority. Obviously, democraticSPACE had the highest, right around 50%. Canadian Election Watch said it would occur with a 45% chance, while Calgary Grit had 28%, and The Mace, 13%.

Although the Conservative seat count was the most politically relevant figure to project, to get a sense of the overall quality of a projection, one needs to look at the other numbers as well. Below is half of the total deviation of each of the above eight projections. (This is the sum of the absolute value of the difference between the projected and the actual result for each of the parties and independents. It is divided by 2 because, by definition, it is always even.)

21 (The Mace)
24 (Canadian Election Watch)
25 (Riding by Riding)
28 (Calgary Grit)
28 (LISPOP)
29 (democraticSPACE)
34 (Too Close to Call)
49 (ThreeHundredEight.com)

Once again, Canadian Election Watch comes in a strong second, this time behind The Mace. democraticSPACE actually did not do too well, placing sixth. Interestingly, the two prognosticators with newspaper columns fared worst... (Of course, Éric and Bryan still have great websites with interesting content.)

If you average the above two measures of success (Conservative seat count and half total deviation), Canadian Election Watch comes in first! I'm not going to use this to declare victory, but I think there's a strong case for saying that I gave at least as accurate a portrayal of the overall situation as any other projection.

I also note that EKOS, which projected 138-113-41-15-1, has a half total deviation of 28, in the same ballpark as the projections above based on multiple polls. Update: Steven Britton also projected on the EKOS poll alone, and got 150-116-37-5, which has a half total deviation of 17, better than the above projections. However, his official call, below, was less accurate.

Now, it is true that seat projections based on polls did not do too well due to poll inaccuracy. Would relying on intuition and other information have been better? To look at that, I compiled the following relatively better publicized projections that were obtained via other methods:

151-86-45-24-1-1 (Andrew Coyne)
156-76-46-30-0-0 (Steven Britton)
146-83-55-22-1-1 (Dan Arnold, aka Calgary Grit)
152-71-52-31-0-2 (Bernard von Schulmann, aka BC Iconoclast, April 25)
146-65-63-33-0-1 (Election Prediction Project)
156-46-60-46-0-1 (Glen McGregor, April 29)

Two of these called for a bare Conservative majority, but in both cases, the NDP was very low, and the Bloc was absurdly high. The half deviations were as follows:

32 (Andrew Coyne)
38 (Steven Britton)
40 (Dan Arnold, aka Calgary Grit)
47 (Bernard von Schulmann, aka BC Iconoclast, April 25)
59 (Election Prediction Project)
69 (Glen McGregor)

The striking thing here is that everyone did worse than most projections based on polling averages. Thus, while relying on polls is far from perfect, it still gives us a better idea of what's going on than letting "gut" and "instinct" cloud one's judgment. Obviously, some prognosticators probably did predict 166 Tory seats (one of the commenters here was close, with 170). However, the above numbers suggest that in most cases, taking poll numbers seriously, even when they're significantly off, is still helpful.

What about riding-by-riding predictions? As far as I know, 7 of the above projectors bothered making a call for each of the 308 races (democraticSPACE also did so for most of them, but did not call some close races). Here is the number correct for each:

259 (Canadian Election Watch)
255 (Riding by Riding)
250 (-1?) (Steven Britton)
242 (Too Close to Call)
234 (ThreeHundredEight.com)
234 (Election Prediction Project)
217 (Glen McGregor)

Once again, Canadian Election Watch performs strongly, this time coming out on top. I'm particularly proud of my projections for BC, where I registered 34/36. The two predictions that relied the least on polls were the worst. In fact, the four most widely known ones came out at the bottom. So while everyone, myself included, did pretty poorly, we still did better than the media will give us credit for.

So where does this leave us? As I've emphasized above, while these results are far from satisfying for election projectors, they still point to the value of relying on the hard data provided by imperfect polls. None of the 7 "soft" (i.e. not based on a numerical model) projections had the Bloc below 22 seats, while 5 of the 8 projections using models had it at 15 or fewer.

As for myself, I'm obviously not thrilled about the absolute result, but very satisfied about how I fared compared to others. Moreover, I'm glad that some of the major issues emphasized on this blog and little discussed elsewhere, such as massive Conservative gains the GTA area, the efficiency of the NDP surge in Québec and the ballot box penalty for the Bloc, all came to pass.

It was my goal to provide you with the most accurate and up-to-date coverage of voter intentions and seat implications for the 41st General Election, and I believe that in many ways, I have succeeded. I hope that you agree, and that you will check back sporadically over the next few years for more political coverage and insights. In fact, please check back over the next few days: I will be analyzing the possible implications of these results as I digest them.

2011 Result Maps

Updated to reflect change of lead in Montmagny--L'Islet--Kamouraska--Rivière-du-Loup

Here are the maps of the results of the 2011 General Election. Notice how little red is left in the Greater Toronto Area: the Liberals went from 32 to 7 seats in the area, a far steeper drop than the final projection's call for 18 GTA Liberal seats.

I remember saying to a commenter that complained about my low Ontario projection for the Liberals that many people would be surprised at how many seats the Grits lose around Toronto. I had no idea I would be this right...


















Model Performance: Region by Region

Updated to reflect change of lead in Montmagny--L'Islet--Kamouraska--Rivière-du-Loup

The election is over, and we have a Conservative majority. In this post, I will discuss where the projection went wrong, and why. A major reason for most projectors' failure to foresee a Conservative majority is that outside Québec, the polls underestimated Tory support by at least 5% in every single region. This performance is frankly embarrassing. My vote share adjustment allowed me to be closer to the mark by bumping the Tory vote estimate by 1-2%, but that wasn't nearly enough.

In a subsequent post, I will compare my final projection to those of other models using poll averages, as well as to predictions made using other methods. It turns out that although projections based on polling averages didn't fare well, they did better than those based on other methodologies. Moreover, among projections based on polling averages, Canadian Election Watch performed very well.

Atlantic Canada (Projected|Actual)
CON - 12|14 ; 32%|38%
LIB - 12|12 ; 28%|29%
NDP - 8|6 ; 34%|30%

Here, we can see that the polls overestimated the NDP vote share, and underestimated the Conservative one. My vote share adjustment from the NDP to the Tories was not enough to compensate for the bias. Had the actual vote shares been known, the model would have been exact. (There would still have been ridings called incorrectly, but overall numbers would have been correct.) As I had predicted, a 14+ seat performance in Atlantic Canada would point to a Conservative majority.

Québec
NDP - 44|59 ; 40%|43%
LIB - 7|7 ; 15%|14%
CON - 8|5 ; 17%|17%
BQ - 15|4 ; 25%|23%
IND - 1|0

The polls in Québec were mostly on the mark. They were a little low for the NDP and the Conservatives, and a bit high for the Bloc. My vote share adjustment appropriately compensated for the bias against the Tories, reduced the Bloc problem, but exacerbated the NDP inaccuracy. Unlike in Atlantic Canada, the projected result would have differed from the actual one even if the actual vote shares had been known: the Bloc would still have been given around 10 seats.

Would simply applying a uniform swing, without taking riding polls and potential swing variability into account, performed better? Well, yes and no: as I said previously, uniform swing on polling numbers would have given the Bloc just 5 seats, which would have been very prescient. However, uniform swing on the actual results would have left the Bloc with just one seat, a significant underestimate. Thus, taking other factors into account and increasing the Bloc's predicted count was correct - however, it should not have been done to the extent that everyone, including myself, did.

If you've been reading this blog, you know that I've been fighting against two serious misconceptions of the Bloc vote, namely that it turns out well and that it is efficiently distributed. In fact, Bloc voters, as usual, turned out less than polls suggested, and the Bloc vote was shockingly inefficiently distributed - even much more than I had thought. Indeed, most of the NDP victories were not tight as the model suggested: the Bloc didn't even come close except in a handful of cases. Many political commentators outside Québec either don't know basic electoral facts about Québec (sovereigntists almost always underperform polls) and/or can't process numbers correctly (how can the NDP win only 5-7 seats with 41% in the polls?).

Ontario
CON - 62|73; 40%|44%
NDP - 22|22; 27%|26%
LIB - 22|11; 26%|25%

Like in Atlantic Canada, the polls failed in Ontario. While the former can be attributed to small sample size, the latter is somewhat embarrassing for the polling industry. The actual result was well outside the margin of error of the last polls from Forum, EKOS and Léger, and it was right on the upper edge for Angus Reid. Other pollsters were probably saved by their small sample size, which implied wider confidence intervals: no pollster came within 3% of the Conservative tally in Ontario. My vote share adjustment only mitigated the problem by a tiny bit.

On actual results, the model with the GTA adjustment would have done quite well: 72-20-14. It would have been farther off without the GTA adjustment, which on actual results, shifted 3 net seats from the Liberals to the Conservatives. This adjustment has been heavily emphasized on this blog, while it was barely mentioned elsewhere.

Manitoba/Saskatchewan
CON - 21|24 - 50%|55%
NDP - 5|2 - 29%|29%
LIB - 2|2 - 15%|13%

Once again, the pollsters severely underestimated the Tory vote, and my pro-Conservative vote share adjustment was not nearly big enough. On actual popular vote, the projected count would have been closer, at 23-4-1.

Alberta
CON - 27|27 - 63%|67%
NDP - 1|1 - 18%|17%

The projection was, unsurprisingly, correct in Alberta, though the pollsters underestimated the Conservative vote here as well.

British Columbia
CON - 21|21 - 41%|46%
NDP - 13|12 - 33%|33%
LIB - 2|2 - 16%|13%
GRN - 0|1 - 8%|8%

Once again, the Tories did much better than predicted. Here, however, the polling inaccuracy did not significantly impact my projection. On the actual provincial split, the model would have given 22-12-2, though I might have made it 23-11-2 via a risk adjustment.

Overall
CON - 152|166 - 37.3%|39.6%
NDP - 94|103 - 30.5%|30.6%
LIB - 46|34 - 20.0%|18.9%
BQ - 15|4 - 6.3%|6.0%
GRN - 0|1 - 4.9%|3.9%
IND - 1|0

Overall, the polls underestimated the Conservative vote by 3.7% - and, I repeat, that was over 5% everywhere outside Québec. They would have been even worse had they accounted for the fact that turnout is significantly lower in Alberta than elsewhere. My vote share adjustment shrank the gap to 2.3%.

On actual vote splits, I would have projected 166-167 Conservative seats - bang on! The Liberal count would have been 36, which is very close as well. The Bloc would still have been overestimated (though by less), and the NDP would still have been underestimated because although I kept repeating that the NDP vote is efficient in Québec, even I did not grasp the full extent of it. Outside Québec, the projected count would have come within 3 seats of the actual count for all parties. Based on these observations, I believe that the model was, in fact, very solid.

As for the riding-by-riding calls, I was correct in 259 cases out of 308 = 84.1%. (Etobicoke Centre and Westmount--Ville-Marie flipped late yesterday, both away from the projected winner. I am, however, very pleased at the change in my childhood riding. Update: Montmagny--L'Islet--Kamouraska--Rivière-du-Loup also flipped away from the projection.) This is worse than how such a model usually performs - you would have expected about 280 correct calls - but obviously that was caused by the inaccurate polls and the great uncertainty in Québec.

Monday, May 2, 2011

Thank You!

Wow, over 22,500 pageviews on Election Day, and consistently over 10,000 per day during the last week of the campaign! Thank you so much for reading. Come back tomorrow for some analysis, and after that, please do check back once in a while even though there likely won't be another federal election until 2015.

2015 Fall General Election Calendar

October 1: Ontario
October 5: NWT
October 5: PEI
October 6: Manitoba
October 13: NL
October 19: Federal
November 2: Saskatchewan

Election Night 2011: Comments Thread

10:00p: Conservative majority projected.

10:08p:
NL: 1-2-4 as projected, but Tories take Labrador, not Avalon.
NS: 4-3-4, NDP doesn't take Central Nova or South Shore--St. Margaret's.
PE: 1-0-3, Tories don't take Malpeque.
NB: 8-1-1, Tories also take Madawaska--Restigouche.
Atlantic: 14-6-12, Tories 2 above projection, NDP 2 short.

10:17p: Bloc leads in just 4 out of 72 Quebec ridings.

10:19p: Iggy trailing.

10:26p: So to all those that said I had the Tories too high and the Bloc too low. Well, it's the opposite.

10:31p: Tories about 5% higher than polls virtually everywhere. Pollsters appear to have failed miserably.

10:40p: Have we just witnessed the death of the Liberal Party of Canada?

10:43p: With the polls this wrong, obviously none of the projectors did well. In fact, if someone came very close, that means their model converting votes into seats has a problem! But it looks like I'm among the least bad...

10:48p: Conservative majority, Bloc loses official party status. Only suspense left: will the Greens elect MPs? Nothing yet from Saanich--Gulf Islands or Vancouver Centre.

10:53p: 3 of 5 leaders ahead in their ridings. Iggy not among them...

11:02p: It is conceivable that the Liberals could be down to one seat west of the GTA (Ralph Goodale in Wascana), and the Bloc could be down to one seat overall!

11:06p: It's the opposite of 2004, when many Americans considered moving to Canada. Several of my friends are already considering moving South...

11:23p: If the Conservatives end up with 41-42% of the vote, that would be within the margin of error of the last poll of none of the 11 pollsters that have released a national survey during the campaign. Everyone was too low, except for COMPAS, who would be too high. We'll see if that's where things end up... One also wonders if this whole campaign was fake, and the Conservatives were in fact always comfortably in majority territory and toying with the media.

11:37p: All 28 Alberta races have been called, and went as projected.

11:39p: Interesting tidbit: Tories win popular vote in PEI, but just 1 of 4 seats.

11:51p: Will having your supporters intimidate the media and just taking 5 questions a day that you don't answer anyway be the standard way to campaign in Canada from now on? Will asking for emails of the government's critics and removing social scientists' tools to evaluate policy be the standard way to govern Canada from now on? Clearly, Canadians don't give a damn.

11:57p: The only Bloc candidate declared elected up to now is Jean-François Fortin, in the open seat of Haute-Gaspésie--La Mitis--Matane--Matapédia. The reason why he won is that the Liberals presented a very popular candidate, who split the vote with the NDP.

12:04a: Elizabeth May declared elected in Saanich--Gulf Islands!

12:14a: Michael Ignatieff lost in Etobicoke--Lakeshore!

12:37a: All 14 Saskatchewan ridings have been called, and the outcome is the same as in 2008, so the projected NDP gain did not materialize. In the 6 provinces where all races have been called, I was correct in 67/74. This score will certain decrease when the 4 other provinces, Ontario, Québec, BC and Manitoba, are included.

1:02a: All 14 Manitoba ridings have been called, and the Tories make a net gain of two on the NDP. They take Elmwood--Transcona from the Dippers and Elmwood--Transcona from the Liberals, who avoid being shut out by holding onto Winnipeg North, which they won in a by-election.

1:09a: Well, it's time to call it a night! For now, if one assumes candidates leading at this point will win their ridings:
QC: 6-59-6-4, 58/75 correct
ON: 72-22-12, 90/106 correct
BC: 21-12-2-1, 34/36 correct
North: 2-1, 2/3 correct
Canada: 166-103-34-4-1, 262/308 = 85%

Tomorrow, I will post maps of the results and compare my projections to those of the 7 other websites using poll averages.

Things to Watch for Tonight, Part 2

This is continued from a previous post.

3. How blue does the GTA become?
This is the main piece of the puzzle for figuring out whether the Tories win a majority. My projection calls for 11 GTA seats that were Liberal in 2008 to turn Conservative this time (one of these is Vaughan, which is already Tory from a by-election). A strong Liberal performance in the GTA could limit losses to 3 or 4, while a weak one might lead them to a catastrophic 18-20-seat loss. There is a lot of variability here, so this will be an area to watch all night.

4. How many Liberal seats west of Greater Toronto?
The projection calls for the Liberals to carry only six of these, as they are projected to lose Esquimalt--Juan de Fuca (sure thing), Newton--North Delta (likely) and Vancouver South (very tight). Of the remaining six, Wascana and Yukon should be fairly safe, but Vancouver Centre (NDP and Green), London North Centre (Conservative) and Winnipeg South Centre (Conservative), while not tossups, could slip away. I think Vancouver Quadra is somewhere between the two groups in terms of precarity.

5. How docile are the Prairies?
The Prairies are the Conservatives' fortress, and not much is expected to change. The NDP's surge is expected to finally give them a seat in Saskatchewan (they got 25.5% of the vote, but no seats last time), and Winnipeg North will likely revert to the Dippers.

A few more holes, however, could appear in the Tory armour. The NDP has shots at gains in Palliser and Edmonton East, and while I have no clue what's going on in Edmonton--Sherwood Park, conservative Independent James Ford could unseat Tory incumbent Tim Uppal. If losses in Saskatchewan and Alberta are what prevent the Conservatives from reaching majority territory, Harper might feel betrayed.

6. How large is the NDP surge in BC?
We know that the NDP is up in BC. But while some polls suggest a modest gain that might only yield an extra seat or two, others have them nearing 40%, which could lead to a gain of eight seats (and a near, or possibly even complete Liberal wipeout). If the East plays out nearly as predicted, this is the question that will keep us up for a long time.

7. Can the Greens finally win a seat?
The Greens are mounting strong challenges in Saanich--Gulf Islands and Vancouver Centre. In the former, leader Elizabeth May is up 7% according to an internal poll released to the media. Can May really suck up virtually all of the Liberal and NDP votes? This one should be close. Vancouver Centre could be a three-way race between the Grits, Dippers and Greens. (The model says Liberals and NDP, but there's buzz that the Greens have a shot.) Heck, if they split the vote evenly, even the Tories might come close. What should be a safe Liberal riding could host the most exciting race tonight!

And there you have it, seven things to look out for. At 10pm, I will start a thread where you can post comments and where I will share my thoughts as results roll in. Happy election watching!

Things to Watch for Tonight

The polls are already closed in Newfoundland, and will close shortly in the Maritimes. Here is an incomplete list of interesting things to look out for tonight:

1. How many Tory seats in Atlantic Canada?
By the time the polls close in most of the country, we will either know, or have a very good idea, of the seat distribution in the Atlantic. The Conservatives are projected to win 12 seats there. If they get 14 or more, a majority becomes more likely than not. But if they win 10 or fewer, Harper may have to start thinking about whether he wants to compromise, and about his political future in general.

Of the 8 seats changing hands that I predicted (remember, these are relative to 2008, so Cumberland--Colchester--Musquodoboit Valley is classified as a change, even though the change really occurred during the by-election), I am of course least confident about Central Nova. Frankly, I was tempted to override this projection by the model, but I left it there because even with that seat, the NDP wins fewer seats than both other parties while leading the popular vote.

I also had problems with the two Halifax area Liberal seats: the Atlantic numbers suggest the NDP could take both, but a Halifax area poll suggested that those are safe. I split the difference, giving Dartmouth--Cole Harbour to the NDP and calling for a hold in Halifax West, but I have little confidence in these projections.

In PEI, while by 2008 results, Malpeque is overwhelmingly the most vulnerable Liberal seat in the province, the buzz is that the two other Liberal ridings are likely Tory pickups. Because I want my projection to be data driven, I didn't take the hearsay into account, but don't be surprised if the Grits drop Cardigan and/or Charlottetown while keeping Malpeque.

Update: I forgot to mention that the Tories could also take Random--Burin--St. George's. I guess Atlantic Canadians are finding this out right now, but most of us will be in the dark for a little longer.

2. Can the Bloc hang on to official party status in Québec?
Many of you thought I'm crazy for having the Bloc down to 15. But uniform swing without balancing risk or taking ridings polls into account actually gets them all the way down to 5. It is therefore a real possibility that the Bloc could be almost wiped off the map.

I can see the Bloc get anywhere from 4 to 26 seats. There are likely to be dozens of tight Bloc-NDP races in Québec, right across the province. You have two parties, both of whose support are distributed quite evenly. Thus, for the Bloc, just a few points can make the difference between keeping half of its seats or near annihilation. Furthermore, while I think 15 is a very reasonable estimate, I have no idea where those 15 seats are going to materialize. I'm guessing they won't be in Montréal, where the NDP's left-of-centre platform is most likely to appeal, but I could be wrong.

The list of ridings that I'm uncertain about is too long to list - virtually every race in Québec but Outremont is unpredictable, which is pretty much the opposite of what one would have said at the outside of the campaign. (OK, Beauce and Roberval--Lac-St-Jean are also safe, but that's really about it!)

More coming later tonight.

Missed Poll: Forum: Tories by 3

I missed the last Forum Research poll, conducted yesterday, as it was not reported by the Hill Times. It has the Tories jumping by 4 points in Ontario, but losing the lead to the NDP by 2 points in BC.

The final projection might have been different for Etobicoke Centre (that's the seat that I moved from Tories to Grits to balance risk in Ontario), Kamloops--Thompson--Cariboo (very narrowly projected Conservative over NDP), and perhaps Vancouver South (ditto over LPC).

Final Projection: Conservative Minority

Canadian Election Watch's final projection was posted on the left sidebar around 6:45am EDT, and it calls for a strong Conservative minority government, with the NDP as the Official Opposition. I encourage you to visualize the projection with these maps (Update: now posted). The expected national seat count and vote share for each main party are as follows:

152 seats, 37.3% for the Conservative Party of Canada
94 seats, 30.5% for the
New Democratic Party of Canada
46 seats, 20.0% for the Liberal Party of Canada
15 seats, 6.3% for the Bloc Québécois
0 seats, 4.9% for the Green Party of Canada
1 seat held by an Independent

Because the Independent, André Arthur, normally supports the Conservatives, this projection implies that the Tories are just 2 seats short of a working majority. Here are some unscientific estimated probabilities of various events based on the record of similar models in past elections (not that the projection model is totally rigorous, but the probabilities are even more ad hoc):

Conservative working majority: 45%
New Democratic win (most seats): 2%
Liberal Official Opposition: 1%

Of course, if the Tories actually finish only two seats short, they could still form a majority if they manage to convince MPs from the Liberal party to cross the floor.

As a reminder, the projection model is a uniform swing model (dividing Canada into six regions) adjusted for riding polls and by-elections (such as Elizabeth May running in Central Nova last time). Additionally, the following tweaks are made:

- Different swings are applied to the Toronto Census Metropolitan Area and other areas in Ontario, based on EKOS data.
- The Conservative polling averages are slightly increased, and the Bloc's, Greens' and NDP's slightly decreased, as is consistent with past history and commitment measures. (The magnitude of these adjustments, around 0.5-1.5% in each case, is admittedly ad hoc.)

Both of these changes favour the Conservatives by a few seats, and account for most of the difference in the Conservative seat count between this projection and that of some other websites using a similar methodology.

At this point, it is important to remember that the idea behind a uniform swing model is that mistakes on one side and the other tend to cancel out. Thus, this model is not designed to correctly predict the winner is every riding. Rather, it is supposed to give a good idea of the expected total count for each party. As a result, the riding-specific projections should be considered unreliable (Atlantic, Québec, Ontario, West). But, hey, they're fun, and Lawrence Cannon, Peter MacKay and Gilles Duceppe are all projected to be swept away by the orange wave (though all these races are very tight, so none of the three may come to pass).

This final projection is a little different from the ones made throughout the campaign because it has an extra ingredient. Rather than just projecting as the winner whoever the model points to and adding up the ridings, I also took a look at how many close races each party is projected to win or lose. I then manually adjusted the outcome in certain ridings in order to come up with an aggregate projection with roughly equal upside and downside risks.

For example, the Bloc Québécois would have won only 5 seats on a uniform swing model. Because my model also includes riding polls and the regional effects they suggest, it had the Bloc winning 13 seats. However, the Bloc was still losing more close races than it was winning, so I shifted two additional seats to the Bloc from the NDP. The other beneficiaries of this adjustment are the Liberals, who get one seat from the NDP in Québec and one seat from the Conservatives in Ontario.

Furthermore, I took another look at Portneuf--Jacques-Cartier, which is handled outside the model due to the presence of an Independent incumbent. I decided to move that seat back into André Arthur's fold, though of course it is a tossup.

Had the above changes not been made, the final projection would have been 153 C, 98 N, 44 L, 13 B. This is what appears on the campaign trends graph (Update: now posted) in order to maintain consistency.

Later today, I will post a more detailed analysis of the projection, which could also serve as a quick (and incomplete) guide to election night.

Final Projection: Maps

Here are the maps for the projection. The one for Montréal is the most striking.








Final Projection: Campaign Trends

Here is the complete trend graph for this campaign. As previously explained, the discontinuity is due to a change in methodology, so the continuous parts of the lines reflect actual variations in the parties' fortunes. Also, as indicated in the main post about the final projection, the numbers plotted as the final point of each line differ from the final projection, which adds a step in the data crunching.

We can see that the Conservatives started the campaign barely in majority territory, but fell out of it due to a strong start by Ignatieff. (Yes, the Grits were actually doing well in the first week. Weird, huh?) For most of the campaign thereafter, they were teetering between a minority and a majority (if you mentally shift the Tory line before the discontinuity up). Throughout the NDP rise, the Conservatives were quite stable, losing just a handful of seats, as I had predicted. This lands them just outside majority territory in the final projection.

The NDP was in the low 30s for the first 60% of the campaign. It was a few points higher in Québec than in 2008, but that wasn't enough for any seat gains. In the rest of Canada, the NDP was actually lower than its 2008 results through the first half of the campaign! As the NDP wave hit Québec, its projected seat count started increasing at a moderate pace until it got to around 50, just 6 days ago. Then it hit the zone where a big chunk of Québec seats swing almost in unison, and the NDP seat count skyrocketed. The rest is history.

It's striking how the Liberal and the Bloc lines are almost parallel, though in the first half of the campaign, the Grits were either on the rise or flat, while the Bloc was already trending downward. While Gilles Duceppe's drop can be at least partly attributed to an uninspired and ill-conceived campaign, Ignatieff didn't do too bad of a job. Unfortunately for him, Jack Layton was much better, and enthusiasm for the NDP crowded out the Liberals.

Final Projection: Seats Changing Hands, West

These are seats changing hands relative to the 2008 General Election, ignoring any intervening by-elections.

Manitoba/Saskatchewan
CON 21 (-1), 50%
NDP 5 (+1), 29%
LIB 2 (=), 15%
Conservative to NDP: Saskatoon--Rosetown--Biggar

Alberta
CON 27 (=), 63%
NDP 1 (=), 18%

British Columbia
CON 21 (-1), 41%
NDP 13 (+4), 33%
LIB 2 (-3), 16%
Liberal to Conservative: Vancouver South
Liberal to NDP: Esquimalt--Juan de Fuca, Newton--North Delta
Conservative to NDP: Surrey North, Vancouver Island North

Final Projection: Seats Changing Hands, Ontario

These are seats changing hands relative to the 2008 General Election, ignoring any intervening by-elections.

CON 62 (+11), 40%
NDP 22 (+5), 27%
LIB 22 (-16), 26%

Liberal to Conservative (12)
Ajax--Pickering
Bramalea--Gore--Malton
Brampton--Springdale
Brampton West
Don Valley West
Eglinton--Lawrence
Kingston and the Islands
Mississauga South
Mississauga--Streetsville
Richmond Hill
Vaughan
York Centre

Liberal to NDP (4)
Beaches--East York
Davenport
Guelph
Parkdale--High Park

Conservative to NDP (1)
Oshawa

Final Projection: Seats Changing Hands, Québec

These are seats changing hands relative to the 2008 General Election, ignoring any intervening by-elections.

NDP 44 (+43), 40%
BQ 15 (-34), 25%
CON 8 (-2), 17%
LIB 7 (-7), 15%
IND 1 (=)

Liberal to NDP (7)
Brossard--La Prairie
Honoré-Mercier
Hull--Aylmer
LaSalle--Émard
Laval--Les Îles
Notre-Dame-de-Grâce--Lachine
Westmount--Ville-Marie

Conservative to NDP (3)
Beauport--Limoilou
Charlesbourg--Haute-Saint-Charles
Pontiac

Bloc to Conservative (1)
Montmagny--L'Islet--Kamouraska--Rivière-du-Loup

Bloc to NDP (33)
All remaining ridings, EXCEPT:
Bas-Richelieu--Nicolet--Bécancour
Berthier--Maskinongé
Chicoutimi--Le Fjord
Gaspésie--Îles-de-la-Madeleine
Haute-Gaspésie--La Mitis--Matane--Matapédia
Joliette
Laurentides--Labelle
Manicouagan
Montcalm
Montmorency--Charlevoix--Haute-Côte-Nord
Québec
Repentigny
Richmond--Arthabaska
Rivière-du-Nord
Sherbrooke

Final Projection: Seats Changing Hands, Atlantic Canada

These are seats changing hands relative to the 2008 General Election, ignoring any intervening by-elections.

CON 12 (+2), 32%
LIB 12 (-5), 28%
NDP 8 (+4), 34%

Newfoundland and Labrador
Liberal to Conservative: Avalon
Liberal to NDP: St. John's South--Mount Pearl

Prince Edward Island
Liberal to Conservative: Malpeque

Nova Scotia
Liberal to NDP: Dartmouth--Cole Harbour
Conservative to NDP: Central Nova, South Shore--St. Margaret's
Independent to Conservative: Cumberland--Colchester--Musquodoboit Valley

New Brunswick
Liberal to Conservative: Moncton--Riverview--Dieppe

Sunday, May 1, 2011

EKOS: Tories by 2.7

The final EKOS poll is now here. The Conservatives lead the Liberals by 11.4% in Ontario, which would imply roughly 60 Conservative seats. However, they trail the NDP by 2.1% in BC, which would seriously dent their chances of a majority. The Tories are also third in the Atlantic, and just 9.2% ahead of the NDP in MB/SK, both of which would imply seat losses.

I will post my final projection overnight, before the polls open in Newfoundland at 7 a.m. EDT. They will not change much from the current projection as, taken together, Nanos and EKOS suggest very modest shifts.

The projection will be accompanied by maps, a completed trend graph, a list of ridings changing hands, and some analysis. So check back tomorrow morning, or later tonight!

Nanos: Tories by 8.2 on Sunday, 5.5 over Weekend

The last Nanos poll is here, along with a daily breakdown of the national results here. Interesting, just like last year, the Tories' support increases on the last day before the vote, and last year, it's Nanos' last-day numbers that came closest to the mark. Will the Tories win by 8%, more than what every pollster (except COMPAS) has been showing in the past week?

The two notable movements in this poll are both good news for the Conservatives: their lead over the Liberals in Ontario more than doubled, from 5.1% to 11.3%, while the NDP lost 7% in BC and is now 13% behind.

I will wait for the last EKOS poll before making a new projection.

Small Mistake

Because of the Elizabeth May's presence in Central Nova in 2008, I made an adjustment for that riding, namely that it is projected using 2004 and 2006 election results. It turns out there was a sign mistake in the formula, and that riding should have been projected NDP instead of Conservative since today's first update. Those posts will be updated with the correct numbers shortly.

Harris-Decima: Tories by 6

Harris-Decima has released its final poll, which is favourable to the Conservatives. The main highlight of this survey is the 15% Conservative lead over the NDP in Ontario (17% over the Liberals), a big change from a 1% Liberal lead in last week's numbers. Also, Harris-Decima has the Tories with a 10% lead in Atlantic Canada, though this diverges from all other polls and is probably due to small sample.

There are no changes in the projection: The Grits lose one seat to the Tories in BC:

CON - 153 154 153
NDP - 97 98
LIB - 45 44
BQ - 13

The average Conservative national lead is 7.1%.

EKOS: Tories by 3.2

Another poll, this time from EKOS, suggests that while things have been stable for a few days, they are now moving again for the NDP, this time outside Québec.

This survey has the NDP surging to a large lead in Atlantic Canada, which some other polls have shown as well. In Québec, the Bloc has fallen to 22.8%, giving the NDP a 17.1% lead. In Ontario, the Tories have increased their lead over the Grits to 13.1%, and are 3% ahead in the GTA, which is also where my model has them.

While all previous polls have shown the Tories leading by at least 5% in BC, this one is the third one today to essentially show a tie in the province. In the end, this shift may be what costs Harper his majority.

In the projection, the NDP picks up two seats in BC, one from the Grits and one from the Tories:

CON - 153 152
NDP - 97 98
LIB - 45
BQ - 13

The Tory average national lead decreases slightly, to 7.0%.

EKOS will have another update around 10pm tonight with data from today, so we still have at least three polls to come.

One More Crazy Scenario that People Haven't Mentioned Much

What if the Tories get to 153 or 154 (after any possible Liberal defections), and the Greens hold the balance of power? Obviously, it's very unlikely that things work out exactly this way, but that's now within the realm of possibility.

I'm not currently projecting any Green seats and won't unless a non-partisan riding poll comes out, but you just don't know what's going on in Saanich--Gulf Islands and Vancouver Centre (which the Liberals are worried about) based on provincial numbers.

Forum: Tories by 2

Forum Research's last poll of the campaign is the tightest yet: 35-33. The NDP is up to 31% in Ontario, just 5% behind the Tories, and 43% in Québec (the writeup says 33, but there are clearly 10 points missing in Québec, and the NDP number would not reach 33% nationally if the Québec number were that "low"). In BC, the Tories are just 2% ahead of the NDP, so maybe Abacus was onto something. This poll has a large sample of 3,789.

It is becoming clearer and clearer that Ontario is turning its back to the Grits. This is great news for Stephen Harper's majority prospects (but see update 2 below). The Tories pick up two more GTA seats from the Liberals in the projection. If the Forum poll is right that the Conservatives lead the Liberals by 10% in the GTA, the Tories could win over 25 seats there, up from their current 8.

[Update: A word of caution: this poll says that the Tories lead the Liberals by 10% in the GTA, a huge change from 2008 when the Grits won the area by 8%. However, the poll defines the GTA as "the two-dozen ridings in the City of Toronto itself, and more than 30 ridings in the surrounding metropolitan area." That's at least 53 ridings, while usually the GTA only includes 42 ridings. Adding those 11 ridings means that, in fact, the GTA did not swing by 18% as might appear at first glance. Fortunately, I did not use Forum's GTA numbers for the projection.]

[Update 2: The GTA/non-GTA divide is very important. The Tories' main competitors in the GTA are still the Liberals, but elsewhere in Ontario, it's now the NDP. So if the Liberals are bleeding in the GTA, it's good news for Harper. But if they're bleeding elsewhere, that could hurt the Conservatives.]

The other changes in the projection are an NDP gain from the Liberals in Toronto and another NDP gain from the Bloc.

CON - 154 153
NDP - 95 96
LIB - 46
BQ - 13

The Tory national lead shrinks to 7.1%.

Abacus: Tories by 5

Abacus has released its final poll of the campaign, and it largely agrees with recent polling averages. Still, there are some differences:

- The Liberal collapse is especially apparent across the West: they score less than 10% west of Ontario, and no higher than the Greens in Alberta and BC.

- The Tories have a strong 16% lead in Ontario, which would provide them with a majority, except...

- The NDP catches up with the Conservatives in BC, where it's 40-40.

No changes in the aggregate projection result from this poll; things appear to have stabilized:

CON - 152 151
NDP - 93 94
LIB - 49
BQ - 14

The average national Conservative lead is also stable, at 7.6%. Still at least two three national polls (Harris-Decima is reportedly going to have one, and of course EKOS and Nanos are coming), and possibly several more to go today.

Nanos: Tories by 6.4

The latest Nanos poll shows a slightly smaller Tory-NDP national gap, but there is no big variation in terms of the regional splits. Interestingly, the NDP lead over the Bloc is down from 16.8% to 13.5%. While this is not a statistically significant change, if such a shift were real, it could cost the NDP several seats. The other important number to observe is the Tory lead over the Liberals in Ontario. Nanos has it at 5.1%, up slightly from 3.6%, but still nowhere near what Harper needs for a majority.

Adding this projection to the poll mix shifts one Tory seat to the NDP in Ontario due to the depreciation of older surveys. Update: It also shifts a Nova Scotia Tory seat to the NDP:

CON - 152 151
NDP - 93 94
LIB - 49
BQ - 14

The average Conservative national lead is 7.6%.

If I interpret this tweet correctly, Nanos suggests that the Saturday numbers were tight, and will release another poll at 9pm tonight containing data from today. This means that we can still expect at least three more polls today, EKOS, Abacus and Nanos, with more likely.

Saturday, April 30, 2011

COMPAS: Tories by 20

COMPAS has the Tories at 46%, the Dippers at 26%, and the Grits at 17%. Needless to say, this poll is grossly out of line with everything we've seen. Because the previous COMPAS poll was also pretty crazy, I will simply assume that COMPAS' methodology is flawed and ignore this survey.

Note that this decision will counteract a significant fraction of the Tory ballot box bounce included in the model. Of course, that bounce is relative to "normal" polls, which this one is not.