Canada: When to Leave

Beleaguered Alberta premier Ed Stelmach is stepping down.

It's interesting news: we've had three premiers retire or resign in the past few months. Danny Williams of Newfoundland retired with levels of support the likes of which we rarely see in Canada . The question there is how much the PC's sky-high levels of support depend on the personage of Williams himself - in other words, how much of those astronomic poll numbers are now available for the Liberals and (snort) the NDP to cherry-pick.

Gordon Campbell and Ed Stelmach wish that was their parties' main concern at the moment. You get the sense that Campbell understood that the BC Liberals were in a real free-fall, and that they had yet to bottom out. You sense that Campbell 'took one for the team', stepping down for the good of the party in order to take the blame personally for BC's current problems and let the party carry on without him. Amazingly, due also to similar stories in the NDP opposition, it may actually have worked. Time will tell, but on the day of Campbell's resignation he was looing at polls so terrible that an NDP victory at the next election seems like a done deal.

Alberta's somewhere in the middle, and this is where it intrigues so much. Its history is one of single-party dominance, where one party dominates politics for a generation and then in a single election is crushed into insignificance by another party, who holds the reins for a generation. Since 1935, these have been parties on the right: Social Credit until 1967, when they were challenged by the reemergent Progressive Conservatives, who have held the reins ever since.

By March of 2010, Stelmach's PCs were in a freefall, and the Wildrose Alliance had overtaken them in the polls. Projections based on the polls at the time showed that the WRA would have decimated the PCs, overtaking them as the party of government and as the main party of the right. And if we can understand anything from Alberta's history, that change might have been permanent.

Were Stelmach Cambell, that is the point at which he would have resigned. But he's not, and interestingly in the past 10 months or so, polls suggest the PCs have come around: surpassing the WRA or at least drawing even with them. "The inevitable" suddenly seemed a good deal less so.

I'm genuinely curious and unable to guess what this means for the 'battle for the right' in Alberta. The captain abandoning a sinking ship might be the single event that pushes Danielle Smith's Wildrose Alliance to electoral success (for thirty-some years, perhaps). Or does Stelmach see in Campbell's departure an example to imitate? Will Albertains forgive the party now that Stelmach is gone? One thing I can say for sure is that it's a huge gamble, one that I never saw coming.

Another thing I can say for sure is that while Dalton McGuinty is probably nonchalantly pretending not to care, Jean Charest is almost certainly in the curious position of wonder if there is anything - anything at all - that Québec can learn from Alberta.

Canada and the UK: Lessons From Britain

What we know, heading into 2011, regarding the Canadian political scene is this: there might be an election, but there might not. And if there is an election, it might change a lot, but it might not. Not much to go on, really.

The interesting thing, though, is that those among us who sense that the stasis that has set in regarding our political landscape might be shaken up (might need to be shaken up) have by-and-large ignored the Westminster System just across the Atlantic that has a half-year leg up on the changes we might well be seeing in Canada in 2011.

If we really do enter an era of coalitions, we shouldn't be looking at countries like Germany or Italy that know them well: we should be looking at the UK as it stumbles through its first proper coalition. The other way the UK presents a model for us is in the existence of Caroline Lucas, the UK's first-ever elected Green MP. This matters because (for now, anyway) the UK has the same problematic FPTP system that we do. So if Nick Clegg's reform referendum fails, sticking the UK with its current dysfunctional system, Caroline Lucas's successes and failures will tell us a lot about the feasability of having a Green Party in a Westminster FPTP system.

The comparison is not entirely valid. For instance, the three-party system that the UK has right now might seem to match ours quite closely (with the same colours, even), but since the mid-nineties, the non-right of the political spectrum in the UK has been really confused. The Liberal Democrats, whose participation in the current coalition has been disastrous for their popular support, appeared to be an option more progressive than the Labour party - an appearance not borne out of the histories of the two parties but on recent trends in parliament, current platforms, and to a large degree on the personages of Nick Clegg and Gordon Brown. Nick Clegg was the story of the 2010 election, and his party seemed like a vital alternative to the two stagnant main parties.

Things are different here. Setting aside the Greens and the BQ (who, of course, correspond to the SNP and the PC), we're quite clear on who our left, our centre and our right are - to the extent that the actual beliefs and policies of the NDP, the Liberals and the Conservatives respectively would do little to change that popular perception. A Conservative/NDP coalition seems even less plausible than a Conservative/LibDem one, which means that Nick Clegg's noble aspiration, as the third party, to negotiate with the first party before talking ot the second party might be dead on arrival domestically. It's a pity, because it's a mature and intelligent working system in a 'coalition era' of politics. But the concept of coalition has coalesced so firmly around the Liberals and the NDP that nothing else is even considered. Realistic in the short-term, I realise, but dangerous in the long.

Dangerous, perhaps, for the NDP. Whatever the rank-and-file might think, I think the NDP as a party has grown comfortable with its role as kingmaker: in Ontario it has been kingmaker a few times before and in 2011 there is a decent chance it will be again. Federally it was kingmaker in the 1970s, and was more comfortable in the role under Martin than it has been under Harper. If the scales tipped just enough this year to give the Liberals more seats than the Conservatives, I think the NDP would be happy to 'prop up' a minority - if they could introduce some key legislation during that time (say, regarding Afghanistan), then it might even help them. If the Conservatives still lead in total seat count (which seems likely), then the NDP are stuck where they were before - either discussing the so-called 'coalition of losers' that so inflamed popular opinion in 2008 or carrying on as they have been for five years now.

A genuine Liberal-NDP coalition presents an interesting challenge for the NDP: as we saw in the 1970s, failures would taint the NDP brand while successes would not necessarily vindicate them. It's tough for a junior partner to take much credit in a coalition, and it's here also that we need to look at the LibDems.

The numbers in the UK are horrifying: scant months ago at the polls, the LibDems took 23.0% of the vote to Labour's 29.0 and the Conservatives' 36.1. This happened after a heady post-televised-debate bubble that had all three parties polling equal numbers; there were even four polls that put the LibDems out front - on 24 April, YouGov posted the LibDems at an unbelievable 34%.

It was unbelievable. Unrealistic and unsustainable. YouGov today posts the following numbers: Labour 43%, Conservatives 37% and the LibDems 9%. The cost of coalition has been astonishing for the LibDems, though you could argue (quite rightly) that the LibDems' numbers were artificially inflated by wavering Labour supporters who have since returned home. Why they have returned home is important, and I don't think it's merely a distaste for the government side of Commons (though there's that). Yet think this way: of any five people who voted LibDem just a few months ago, only two remain faithful. More than half have wandered away. Obviously the tuition fiasco is of great importance. But it reflects the strange situation the LibDems, historically a centrist party, find themselves in: are they responsible to their party loyalists or to their most recent electors? If their intentions have been misunderstood by a large number of the people who cast votes for them (though the tuition fiasco is a bit different - it does seem to be a genuine turnaround), is that really their fault? And how grateful should they be for the votes they managed to secure last time around, if many of those votes were cast against the other two parties as opposed to genuinely for them?

For the moment at least, the LibDems and Labour are tied to each other: one party's success is the other party's failure. One wonders how well a Labour/LibDem government would have performed - and one wonders how well the LibDems would be polling were Nick Clegg Ed Miliband's junior partner. One also wonders where the UK Greens - a more soundly left-wing party than the Green Party of Canada - fit into all of this (quite well in the short-term, at least - were an election to be held tomorrow, Caroline Lucas's seat would seemingly be secure and perhaps joined by a few others). All this matters over here because whatever fluidity can be squeezed from the stone of Canadian politics falls clearly on the left side of the Conservatives-vs.-everyone-else axis. Coalition, merger, collapse, whatever: it's only a shakeup in the relative positions of the four non-Tory parties that can really change the composition of parliament. The signals from the UK are mixed at best. But we need to be reading them much more carefully than we currently are.

2008 Election: Provincial Microcosms #1: Québec

So I'm still looking at the 2008 election. I got to thinking primarily about Québec, and how the 'provincial averages' tell so little of the story, as one riding could have numbers remarkably different from another. I got to wondering where in Québec the 'average riding' might be - where 'average' can be defined as having vote percentages closest to the provincial averages. So if province-wide 38.1% of people voted BQ and 23.7% voted Liberal, in this riding as close as possible to 38.1% of voters voted BQ (sending a Bloc MP to Ottawa) and as close as possible to 23.7% of voters voted Liberal.

So for each riding, I calculated the per-party 'deviation' from the provincial average and added them together. So, in this example, if a riding saw a 36.4% BQ vote and, say, a 29.0% Liberal vote, that would be a 1.7 point deviation for the BQ and a 5.3 deviation for the Liberals, totalling 7.0 - before calculating other parties. Well, before calculating the Conservatives, the NDP and the Greens. I didn't bother with minor parties and I didn't bother with independents - which obviously screws up Portneuf-Jacques-Cartier. Actually my numbers for that riding would be doubly screwed up, since the lack of a CPC candidate doesn't mean an additional 21.7 deviation (as it probably should) - it merely means that there are no CPC numbers for the calculation at all. In any case, the extent to which votes for André Arthur can be seen as distinct from CPC votes in debatable. There are a small handful of Québec ridings that didn't have a Green candidate either.

So anyway, here are the results:

• The winner for "least deviant Québec riding" - or "best Québec microcosm"... is a tie. Both Shefford and Compton-Stanstead deviate from the Québec norm by a total of only 9.6. Those are pretty impressive numbers, really. Compare:
• BQ province-wide: 38.1%. In Shefford: 42.8%. In Compton-Stanstead: 41.9%.
• Liberals province-wide: 23.7%. In Shefford: 21.4%. In Compton-Stanstead: 22.5%.
• CPC province-wide: 21.7%. In Shefford: 19.6%. In Compton-Stanstead: 19.4%.
• NDP province-wide: 12.2%. In Shefford: 12.5%. In Compton-Stanstead: 11.3%.
• Greens province-wide: 3.5%. In Shefford: 3.7%. In Compton-Stanstead: 4.9
• Both of these ridings are in the Eastern townships, an area whose linguistic demographics are rather typical of the province as a whole (despite a minimal allophone population). Both have gone BQ in the previous two elections, and before that both supported a PC candidate in 1997 who crossed the floor to the Liberals for the 2000 election (both Jean Charest and Paul Martin are native sons, incidentally). Shefford has existed as a riding since confederation, so it might be interesting to look at older elections to see how much of a microcosm it really is. Another day, though.
  

Rounding out the top five for 'least deviant Québécois ridings' are the following:

• Vaudreuil-Soulanges at 10.7. This Montérégie riding has traditionally been a pretty safe seat for the Liberals, so it's impressive that star candidate Marc Garneau managed in 2006 to lose in this riding to the BQ candidate.
• Gaspésie-Îles-de-la-Madeleine at 12.0. This is the easternmost riding south of the St. Lawrence, and it's a bit tough to trace its history, since the Gaspé has been carved up several times recently. But it seems to be a pretty competitve area with no real historical loyalty to anyone, having elected the Liberals, the (Progressive) Conservatives and the Bloc Québécois in recent history.
• Alfred-Pellan at 12.3. This riding in the city of Laval has tended to go BQ since the creation of that party, though like Laval as a whole has been a two-way BQ-Liberal race.

Though the two least deviant ridings are both in the Eastern Townships, it's interesting otherwise how wide-ranging the top five are, geographically. They are by necessity ridings that have to be competitive in more than merely two-horse races, since Québec's numbers as a whole are quite disparate - which hides a very regionalised electoral map where you can find two-way BQ-CPC races (Roberval-Lac-Saint-Jean), two-way BQ-Liberal races (Papineau) and even two-way Liberal-CPC races (Pierrefonds-Dollard).

By contrast, here are the five most deviant ridings. It doesn't mean much, really: it either shows loyalty to one star candidate or another or otherwise shows location in Montréal. Still, they're five interesting ridings, wildly atypical at present:

• Beauce, at 83.2. Beauce's deviation is due to its remarkably strong support (62.4%) for local Conservative hero Maxime Bernier (which leads, of course, to relatively dimmer support for his competitors). Beauce is also an Eastern Townships riding, though it's closer to Québec City than the others we've looked at. It has also existed since Confederation and has a long and proud history of lending strong support to newsmaking MPs, including Bernier's father Gilles, elected twice as a PC and once as an independent, Fabien Roy, the last major leader of the Social Credit Party, and three-time independent MP Raoul Poulin. It has never returned a BQ MP.
• Jonquière-Alma at 79.9. This Côte-Nord riding, like Beauce, has a strong Conservative MP in Jean-Pierre Blackburn, who got 52.6% of the vote. This riding is noteworthy for having almost no Liberal presence in recent years: in 2008, they got 5.16% of the vote, and in 2006 before that, they got less than 3%, the lowest Liberal performance in the whole country.
• Mount Royal at 76.6. Réal Caouette's famous comment that a mailbox could win an election in Mount Royal (merely because it's red in colour) holds true. Mount Royal is the safest Liberal seat in the country, having returned Liberal MPs in every election since 1940. It was Pierre Trudeau's own riding. An amazingly diverse riding where Francophones are a smaller population than either Anglophones or Allophones and the largest religion is Judaism, it has never seen a BQ candidate get more than 7% of the vote (in 2008, the BQ candidate finished fifth). Current MP Irwin Cotler was first elected in a 1999 by-election with 92% of the vote.
• Outremont at 74.3. Outremont is Mount Royal's neighbour on the island of Montréal, and like it tends to return Liberal MPs, though not with such resoundng majorities, as the Bloc tend to do well too in this riding that straddles East and West Montréal. A 2007 by-election saw both parties lose support to the NDP's Thomas Mulcair, who in 2008 became the first NDP candidate ever to win a Québec riding in a general election. Obviously it is his high numbers that skew this riding.
• Westmount-Ville-Marie at 73.7. Neighbour to both Mount Royal and Outremont, this riding occupies a very 'deviant' area of Montréal. Again it's a Liberal stronghold, with high Liberal numbers by Québec standards, but it also brought in the province's second-best NDP numbers in 2008. The two parties combined for almost 70% of the vote. At 7.3%, the Bloc are no threat in this riding., where Anglos are a larger number than the statistically-tied Francos and Allos. This is where Marc Garneau in 2008 finally managed to get elected.

I've also done the very same thing for Ontario and BC. We'll see those numbers soon.

2008 Alternate Scenarios #1: D'Hondt, Part One

Image via WikipediaSo I've decided to look back at the 2008 election and crunch some numbers to consider how the votes cast in 2008 might have brought a different parliament had Canada had a different voting system than it currently does. I'm starting with the D'Hondt Method, a proportional representation system that delivers MPs in a ratio approximating the ratio of votes they recieved. It's a 'party list' system in principle, meaning you vote for a party and not an individual, and the MPs are chosen from lists. It might not necessarily have to be that way, but that's how it's applied generally.

Now, 2008 gave is the following result:

CPC 143, Lib 77, BQ 49, NDP 37, Ind 2.

A simple nation-wide proportional representation calculation using the D'Hondt method would have returned this:

CPC 118, Lib 82, NDP 56, BQ 31, Green 21.

Which is a huge decrease for the Conservatives and the BQ but a huge increase for the NDP and Greens. I'm surprised to find the Liberals actually gaining seats in a prop-rep simulation.

I can't really see a system like this flying with the Canadian people, mind you. It's too anonymous, and it sees Canada as one huge mass as opposed to a federation of provinces. It might go over better to apply the D'Hondt method to the results of each province (retaining the current number of seats per province). If we do that, we get the following:

Newfoundland and Labrador: Lib 4, NDP 2, CPC 1.
Prince Edward Island: Lib 2, CPC 2.
Nova Scotia: Lib 4, NDP 3, CPC 3, Green 1.
New Brunswick: CPC 4, Lib 4, NDP 2.
Québec: BQ 29, Lib 18, CPC 17, NDP 9, Green 2.
Ontario: CPC 42, Lib 37, NDP 19, Green 8.
Manitoba: CPC 7, NDP 3, Lib 3, Green 1.
Saskatchewan: CPC 8, NDP 4, Lib 2.
Alberta: CPC 20, NDP 3, Lib 3, Green 2.
British Columbia: CPC 17, NDP 9, Lib 7, Green 3.
Yukon: Lib 1.
Northwest Territories: NDP 1.
Nunavut: CPC 1.

Which gives us a grand total of:

CPC 122, Lib 85, NDP 55, BQ 29, Green 17.

Not radically different from the nation-wide D'Hondt: a few extra seats for the Conservatives and Liberals, a few fewer for the Greens - closer to reality, maybe, except for a further BQ drop. But you can see how having this additional distribution of numbers might make for a more convincing system: Prince Edward islanders, for example, would be able to identify the two Conservative MPs and the two Liberal MPs who were 'theirs'. The North, incidentally, remains exactly the same, since a prop-rep region of one seat is effectively a FPTP region.

Ontario and Québec in particular, however, remain pretty impersonal. How would someone from, say, Trinity-Spadina or someone from, say, Nickel Belt know which MP was 'theirs'? Which one to write a letter to for whatever reason it is that normal citizens write MPs? Sure, with 105 riding and 105 MPs, we could arbitrarily allocate one MP to each riding, but that would be a sheerly bureaucratic decision. Which eight ridings would get Green MPs?

The smaller a D'Hondt region gets, the less useful the D'Hondt system is. But if you broke up larger provinces into smaller regions, you might have results that gave an increased sense among the public that the MPs actually represented a constituency and not merely represented their parties. That's what I'd like to do, but it'll take a bit of time, and number-crunching...

Canada: King Jack

Image by mattjiggins via Flickr
Of the several ways that Jack Layton is very truly Ed Broadbent's heir, one of the most obvious is that he is personally much more popular than his party. One wonders if this is not a prerequisite for leading the NDP (or, thinking of Carole James, Darrell Dexter and Bob Rae, among others I should say a prerequisite for leading the federal NDP). One almost wonders if people don't compensate for not supporting the NDP by tossing their leader a few bones.

The extent, though, is almost comical. Angus Reid does this kind of stuff best. In addition to simple 'approval/disapproval' questions, they also ask what characteristics best describe the four main pary leaders (Elizabeth May is not included, for one reason or another). On this particular front, Layton has always performed well, but Angus Reid has new numbers and they're even better than usual. I thought it would be interesting to quantify just how well.

A random example of these numbers might be 'open' (as opposed to 'secretive'). Harper, Duceppe and Ignatieff are seen as 'open' by 8%, 9% and 10% respectively of the public polled by Angus Reid. Layton, on the other hand, is seen as 'open' by fully 29% of the public. 29% is not a stellar number, but it's more than the other three combined. You might not believe that that was a random selection, but let me put it this way: of the nine positive traits on Angus Reid's poll (down-to-earth, open, efficient, compassionate, honest, strong, exciting, in-touch, and intelligent), Layton scores better than the other three on every one except 'efficient' and 'strong' (in both cases, he comes second behind Harper). Of the nine negatives, primarily opposites to the positives (arrogant, secretive, inefficient, uncaring, dishonest, weak, boring, out-of-touch, foolish), Layton gets the best scores (obviously, the lowest percentages in these cases) on every one except 'weak', where he is beaten by both Harper and Duceppe.

Putting these numbers into action, I decided to add together the 'negative trait' values of each of the leaders and then the 'positive trait' totals. The numbers are interesting.

Stephen Harper's total negative-characteristic score is 258. That means that the total of the percentages of people who attributed each of those nine negative values listed above to Harper is 258 - his number is highest, which doesn't mean he's the least-liked leader, merely the one the most people have an opinion on. Ignatieff comes next with 245, quite near to Harper. Duceppe's negative values total 178, though I should mention that Angus Reid only asks about Duceppe in Québec. Nationwide this number would presumably be higher. Jack Layton leads the pack with 121, which is less than half of Harper and Ignatieff's numbers.

Looking at those nine positive traits, Ignatieff can muster only a meagre 82 points across them. Considering that that's the total of nine different percentages, that's a lousy number. 'Intelligent' is the only thing that a decent number of Canadians have positive to say about Ignatieff, and even at that Harper and Layton beat him (Ignatieff's numbers are bottom of the pack in every other positive trait: only one in fifty Canadians, for example, describes him as 'exciting'). Duceppe gets an also-weak total of 105. Harper may have the largest negative total, but his positive total is high too - 146 points, reflecting again the fact that more people have opinions about him (one way or another) than the other three. Still though, Layton pulls far ahead with 215 points. The average Canadian can find almost three good things to say about Layton for every one good thing they can find to say about Ignatieff.

If you subtract the negative totals from the positive totals, the results are striking: Ignatieff fails miserably, with a total of -163. Harper does only a little better, with -112. Duceppe remains in the negative, with -73, but Layton not only has a net positive rating, but quite a decent one: +94. That's an amazing difference, really. By this particular measure, Layton is head-and-shoulders above the competition.

It's arguable how much that actually means: Ipsos Reid also has numbers out that show Harper leading the competition on the question of 'who would make the best prime minister': 43% say Harper, 33% say Layton and 24% say Ignatieff. It's tough to be sure how to analyse that, except to say that Canadians like Layton more than they actually want him to be prime minister. That's an awkward position for Layton to be in, but all in all these polls clearly look good for him. Especially if he ignores polls that discuss voting intent.

Related articles

• Contrasting New Democrats (theglobeandmail.com)
• Lessons from successful politicians: How Canada's junior coalition partner went on to win more votes and seats (solutionfocusedpolitics.wordpress.com)
• Harper best choice for prime minister: poll (nationalpost.com)

Ontario and Other Kingmakers

So there's a belief that you come across every now and then that Ontario determined federal elections; that what Ontario wants, Ontario gets. Historically, this claim has most often been made in connection to so-called 'Western Alienation', the feeling that the Western provinces are left outside the halls of power. In fact, as often as not that claim is made not merely of Ontario but of Ontario and Québec in combination as 'Central Canada' or the whole of the country east of Manitoba as 'Eastern Canada'. As those geographic blocs, however, form a majority of seats in parliament, I'm not sure how meaningful it is to talk about them dominating politics. To take my point to an extreme, if someone from the North complained that 'the provinces' held too much power, they would by any definition be factually right, but it would be tough to argue it as a failing of democracy.

Breaking the country into four blocs - the West, Ontario, Québec, and the Atlantic - is practical as the four contituent parts each hold a large amount of seats (at present 92, 106, 75 and 32 respectively), and as they're each terms with historical merit ('The West' is the most nebulous term, really, consisting of four provinces each with very distinct voting trends, but when totalled, trends do emerge).

In this exercise, I've gone back to Diefenbaker's 1958 landslide, and in each case I've looked at the total number of seats per party returned in each of the four regions, and I've also considered what Parliament would look like if it were composed of the whole country minus the region in particular. In other words, a "minus-Ontario" result would be the combined seat totals for the West, for Québec and for the Atlantic (the North as well - in each case I include the North's three seats in relevant national totals, but otherwise ignore them); as if Ontario had seceded from the union prior to the election.

The point of this math is to consider scenarios where the actual government formed by the election in question (which party, and whether it was a majority or a minority) differed from the "-Ontario" result; if removing Ontario from the equation would produce a different government than actually occurred.

The overview results are interesting themselves: In 17 elections, the nation as a whole has returned 6 Liberal majorities and 4 Liberal minorities, 3 PC majorities and 2 PC minorities, and two Conservative majorities. So Liberal domination, but a relatively healthy bipartisan mix. Looking at the regions, though, shows quite a difference: the Atlantic considered alone returned 9 PC majorities, 8 of them straight, between 1958 and 1984; after that, they switched teams decisively, electing 7 liberal majorities, interrupted by their only plurality: the surprising result in 1997 when the Atlantic returned a plurality of PC MPs.

Québec looks similar. Though they went PC for Diefenbaker and Mulroney (being largely responsible for Canada's only two near-consensuses in recent history), otherwise it was Liberal pre-Mulroney (7 straight majorities, one plurality) and BQ post-Mulroney (6 straight majorities). The West, despite an amazing volatility in BC and intermittently strong NDP results in Saskatchewan and Manitoba, have been remarkably consistent in their affections. Trudeaumania took over BC in 1968 enough to return a Liberal plurality, but otherwise it's been non-stop PC majorities from 1958 to 1988, non-stop Reform/CA/Conservative majorities since.

That leaves Ontario, which perhaps is the Bellwether, based on how consistent the other three regions have been. Ontario has returned 11 Liberal majorities, 2 PC majorities, 2 PC pluralities and 1 Conservative plurality.

Overall, the West has swayed the winning party on three occasions (1962, 1979 and 2006), while Ontario has done so on three occasions (1979, 1997 and 2004), and Québec on four occasions (1963, 1965, 1972 and 1980 - remarkably a Québec-less Canada would have returned the PCs every year from 1958 to 1988 except in 1968). The Atlantic since 1958 have not once swayed an election.

Looking down the years, then:

• In 1958, Diefenbaker won a decisive majority. Each region gave a majority of seats to the PCs, so obviously each 'minus parliament' also did so. It's interesting that while the Liberals tend to be 'Canada's natural governing party', the only true landslide sea-to-sea consensuses we've seen have both been PC.
• 1962 brought a PC minority, the first of three minorities. The coasts stayed majority-PC, while Ontario returned a Liberal majority and Québec a Liberal plurality. 1962 definitely shows that the 'Central Canada' thesis has holes in it: Ontario and the Atlantic had no affect, while Québec held the PCs to a minority (a minus-Québec parliament would have returned a PC majority) and the West actually chose the PM (a minus-West parliament would have been Liberal minority).
• 1963, though, shifts the numbers just enough to prove the West's point: a Liberal minority brought about by the coasts returning majority PC MPs and the 'centre' returning Liberal majorities. But the minus-numbers tell an interesting story. Again, Ontario and the Atlantic had no effect, while the West this time kept the Liberals to a minority and Québec turned the tides, from a PC minority to a Liberal one.
• 1965 was another Liberal minority, and everything I've said for 1963 applies here too, except that a minus-Atlantic parliament would have been an even split: 116 Liberals to 116 'others'. I think technically that's a Liberal minority, but I'm not entirely sure how to classify it.

• 1968 is Trudeau's big year, and a Liberal majority. It's interesting to note, looking back, that Trudeau's party-poopers were not the West so much as the Atlantic. The West brought a threadbare Liberal minority, while the Atlantic decisively stuck with favourite son Robert Stanfield, giving his PCs 25 in 32 seats. 'Central Canada', of course, went deep red. This was a Liberal majority, and no single region could have changed it from a Liberal majority.
• Trudeau won 1972 too, but by two seats. The Atlantic kept its PC majority, the West returned to a PC majority (never again to waver), and Ontario went to a PC plurality. Thus, it seems like 1972 was Québec's year, with their decisive Liberal majority keeping Trudeau at 24 Sussex. And indeed, Ontario and the Atlantic had no effect, while the West held Trudeau to a majority, and a minus-Québec parliament would have been a PC majority.
• In 1974, Trudeau got his majority back, and the numbers are similar, just shifted in the Liberals' favour. PC majorities in the West and Atlantic, Liberal majorities in Québec and Ontario, and the notion that 'Central Canada' runs the country quite firmly established by now. This is certainly the case in 1974, with neither the West nor the Atlantic having any sway, with Ontario enough to shift it from a Liberal minority to a Liberal majority, and with Québec's vote enough to move from a PC minority to a Liberal majority.
• 1979's PC minority was interesting. This time out Ontario joined the West and the Atlantic in returning PC majorities, while Québec returned another Liberal majority. Only the Atlantic had no sway this time: were it not for the West, the PC minority would have been a Liberal majority. Were it not for Ontario, the PC minority would have been a Liberal minority. And were it not for Québec, the PC minority would have been a decisive majority. No wonder the government was doomed to failure.
• 1980 brought Trudeau decisively back, with another Liberal majority. With the West returning only two Liberals and Québec only one PC (PC majority and Liberal majority respectively), it increasingly looks like Ontario really is the kingmaker, its fickle preferences dictating results when the West and Québec are so decisive in their support for their 'home teams'. In fact, both it and the Atlantic 9for the first time) went Liberal-majority. Only Québec can lay claim to holding sway, though, as minus-parliaments for the other three regions still return Liberal majorities, while a minus-Québec again brings a PC minority.

• 1984 is our second landslide, with Mulroney's PCs gaining a majority in each region (in each province and territory, in fact) and reducing the Liberals to 40 seats. There's little to say here - electing Mulroney was a nation-wide effort, and no individual region could have stemmed that tide (or wanted to, as it transpired).
• 1988 was another PC majority, but less decisively. Ontario gave the PCs only a plurality, while the Atlantic returned a Liberal majority. With the West handing the NDP their best-ever result (while still going PC-majority, though), it seems like Mulroney had Québec to thank, proving that after years of kingmaking for the Liberals, La Belle Province could play kingmaker for the PCs too. Interesting, then, that after this election Québec will decisively turn away from affecting federal outcomes, remaining by and large on opposition benches henceforth. As it turns out, though, no single region kept Mulroney in power. A minus-Québec would have still been PC, but minority, while the other three regions had no effect.

• 1993, of course, is where everything changed. I imagine most Canadians old enough to remember could tell you that Ontario and the Atlantic sent decisive Liberal majorities, while the West sent a Reform majority and Québec a BQ majority. What all this means, though, is that ultimately not one region could be said to have turned this election, however it appears. A minus-West election and a minus-Québec election would have unsurprisingly returned Liberal majorities. So would have a minus-Atlantic parliament. A minus-Ontario parliament would have returned a minority, but the Liberals would still have won it.
• 1997 was perhaps even more interesting. The only regional overall change from 1993 was that the Atlantic went plurality-PC (thus each of the four regions preferred a different party). Like last time, neither the West nor Québec nor the Atlantic could have prevented a Liberal majority, but this time the majority was all Ontario's - two-thirds of Chrétien's MPs were from Ontario. A minus-Ontario parliament would have returned, of all things, a Reform minority. But it would have been almost completely dysfunctional, with 60 Reform MPs, 54 Liberals, 44 Bloquistes, 21 New Democrats and 19 PCs. God know what kind of bills, if any, that parliament could pass.
• 2000 was 1993 again: Liberal majorities in Ontario and the Atlantic, a barebones BQ majority in Québec, and another majority for the Party-Formerly-Known-as-Reform in the West. A Liberal majority, that without Ontario would have been a minority, but otherwise no region could have changed.
• The Liberals won in 2004 too, if a minority mandate can truly be called 'winning' and 'merely not losing'. The regions once again still look the same, though that western party is now called the Conservatives. Neither Québec nor the Atlantic could have swayed a Liberal minority, but but for the West it would have been a Liberal majority, and but for Ontario it would have been a Conservative minority. As indeed it was about to be.
• 2006 brought Harper's Conservatives a minority, elected through regional majorities: the BQ in Québec, the Liberals in Ontario and the Atlantic, and the CPC in the West. So while all four regions have returned the exact same results in each of the previous three elections, the result has shifted from Liberal majority to Liberal minority to Conservative minority. And quite clearly the West were kingmakers this time, as no other region could have prevented a Conservative minority, but a minus-West parliament would have been another Liberal minority.
• By 2008, the Conservatives had been able, with a pickaxe, to slowly pick away at Liberal support in Ontario to the point that this time out they'd tipped the balance - a CPC plurality in Ontario. The other three regions have stuck with their favourites for the sixth time in a row (excepting the Atlantic in 1997) though, suggesting again that Ontario is where elections are made and broken. In fact, though, we have Québec to thank for the current government makeup. None of the West, Ontario or the Atlantic could have changed a CPC minority, but were it not for Québec, Harper would have had his long-dreamed-of majority.

What to do with all this information? Well, what it shows me is that the West is more of a kingmaker than Westerners might think but that ultimately Québec wields the most power - especially before the birth of the BQ. It's also interesting to see just how little power the Atlantic actually have. Unsurprising, perhaps, in terms of overall numbers, but seeing how independent-minded they can be, still unexpected.

Related articles, as Zemanta would like you to know.

• ‘Cluster’ strategy puts Tories on track for majority, poll suggests (theglobeandmail.com)
• Ten years later, Trudeau divides us still (fullcomment.nationalpost.com)
• The Liberals shall rise again (fullcomment.nationalpost.com)
• Hebert: 10 years on, who will save Trudeau's legacy? (thestar.com)

The Progressives: A Further Fiction

Just playing around with the Uniform Swin model Excel file that you can download at the Paulitics blog.

I decided to try using the current projection at ThreeHundredEight, which is to say the current regional voting estimates. The Paulitics model crunches the numbers differently than ThreeHundredEight does, though, since the same numbers give pretty different results. The Paulitics model gives us:

• Conservatives: 129
• Liberals: 88
• NDP: 36
• Bloc Québécois: 52
• Green: 0
• Others: 3

It's a bit odd to see three 'others' seats, and I guess it's a problem with the model that it lumps all 'others' data together. Still, interesting numbers, that produce the same kind of result as any other poll or poll ananlysis out there: Conservative minority.

But that's just the starting point. I then decided to look at the numbers for my 'progressives' fiction, wherein the NDP, the Greens and the Bloc united, without shedding any support. In other words, I combined the current vote for the NDP and Green Parties, putting the combined total into the NDP box, for all of Canada except Québec, where I combined the total of the three and put that total in the BQ box. What did that give us?

• Conservatives: 113
• 'Progressives': 110
• Liberals: 82
• Others: 3

Interesting numbers. And interesting in that 'Progressive' gains would come pretty much completely at the expense of the Conservatives - the Liberals would stay more or less constant. The regional seat-by-seat breakdown would be as follows:

• BC: Prog 15, CPC 14, Libs 7
• Alberta: CPC 25, Prog 2, Other 1
• Prairies: CPC 20, Prog 5, Libs 3
• Ontario: CPC: 44, Libs 41, Prog 21
• Québec: Prog 61, Libs 12, CPC 1, Other 1
• Atlantic: Libs 18, CPC 9, Prog 4, Other 1
• North: Prog 2, Libs 1

I don't know how reliable these numbers are, but they're intriguing. BC, where the Greens are not an insignificant party, gets its numbers rearranged dramatically. Much of the rest of the country doesn't, however. I'm surprised to see the 'Progressives' so low in Atlantic Canada, for example.

This particular parliament would be pretty darn unworkable, wouldn't it? Or rather it would put a lot of power into the hands of the Liberals. It actually resembles the UK political dynamic.

But let's take it further: let's imagie that a united Progressive party would be enough to win over one in ten current CPC supporters and one in five current Liberal supporters. What then?

Well, we still wouldn't be in majority territory. There might be something off about my calculations, but now I get:

• BC: Prog 22, CPC 10, Libs 4
• Alberta: CPC 25, Prog 2, Other 1
• Prairies: CPC 17, Prog 9, Libs 2
• Ontario: CPC 46, Libs 30, Prog 30
• Québec: Prog 65, Libs 8, CPC 1, Other 1
• Atlantic: Libs 13, CPC 10, Prog 8, Other 1
• North: Prog 2, Libs 1

For a grand total of:

• Progressives: 138
• Conservatives: 109
• Liberals: 58
• Other: 3

A minority, but a significantly stronger minority than the one we currently have. Interesting that my 'step one' ate away at the Conservatives more than the Liberals but that my 'step two' has decimated the Liberals. Highly interesting (though not implausible) that my 'step two' actually increases the Conservatives' seats in Ontario. And interesting that despite all this number crunching (combining three parties and taking a significant chunk from the remaining two), the Progressives not only don't get a majority nationwide but they also lag far behind everywhere except BC and Québec.

Strange numbers. Expect me to use the Paulitics model for other stuff too.