Wednesday, May 15, 2013

Detailed BC Election 2013 Results - Yes, polls got it wrong, again.

In case you were asleep last night and just found out, the BC Liberals under Premier Christy Clark completely blew away the challenge from Adrian Dix's NDP. Not only did the Liberals defy most expectations and come to win, they actually managed to increase their seat count and take over about half a dozen previously safe NDP ridings. This was not a possibility foreseen by anyone, and if they tell you different they are lying.

There are going to be many questions asked by this election result, the most important of them to me being how we can trust pollsters to give us accurate numbers anymore. But more on that later.

The other question will simply be, well, how? How did Clark pull off this feat that, as Bryan Beguet mentioned to me last night, she only had a 14% chance of managing?

But first...

The Results

BC Liberal Party: 723,113 votes (-28.5K) - 44.4% (-1.4%) - 50 seats (+1 seat)
New Democratic: 643,069 votes (-48.5K) - 39.5% (-2.7%) - 33 seats (-2 seats)
Green Party: 130,245 votes (-4,145) - 8.0% (-0.2%) - 1 seat (+1 seat)
BC Conservative: 77,770 votes (+43.3K) - 4.8% (+2.7%)
Others: ~27,000 votes (-1,000) - 3.3% (+1.6%) - 1 seat (Huntington)
Total: ~1.60-million (-~400,000)

You can't get much clearer of a win than that. The Liberals decimated their opposition, losing only a faction of their vote compared to the NDP dropping almost 50,000 voters and the Greens about 4,000. The Conservatives, who jumped up 43,000 votes, clearly have a tremendous impact on the race, but it also means that, assuming most of those voters are former BC Liberals, that the NDP may have bled voters to the incumbents. Either that, or we assume their turnout dropped significantly while the Liberal's gained new voters. Those are your only two options, as the math cannot add up any other way.

Lets go quickly through the regions, and we'll see what happened. Remember however that these numbers are preliminary, and the percentages may change by very slight amounts in the future.

First, the North:

Northern BC - Election 2013
BC Liberal Party: 50.2% (+1.4%) - 5 seats (=)
New Democratic: 33.9% (-4.9%) - 3 seats (=)
BC Conservative: 8.1% (+6.2%)
Green Party: 2.7% (-4.0%)
Others: 5.1% (+1.2%)

The Liberals held on pretty strongly in the North, losing no incumbents but gaining no seats. This despite the Conservatives jumping up a good sum, though most of that is due to Kurt Peats' candidacy in Peace River South. The NDP lost voters, and nearly lost a relatively safe seat in Skeena.

My projection had shown the Liberals winning about 40% to the NDP's 43%, based on much of the regional polling done here. I had also expected the Independent candidate Arthur Hadland to have a much better showing in Peace River North than he did.

Interior BC - Election 2013
BC Liberal Party: 48.4% (+4.0%) - 13 seats (+2 seats)
New Democratic: 36.3% (-2.6%) - 3 seats (-2 seats)
BC Conservative: 6.9% (+0.1%)
Green Party: 4.3% (-4.4%)
Others: 4.2% (+3.0%)

The result in the Interior was a bit unsurprising, given that the Liberals had some strong momentum in the region. Yet, once again, this kind of blowout was unexpected. The Liberals gained roughly 4% in terms of support, as well as two seats - Fraser-Nicola and Cariboo North - with the NDP dropping from their 2009 numbers, as did the Greens who lost roughly 4%. The Conservatives stayed almost exactly stable, gaining the tiniest amount, even though this should have been their best region. And if they were stable, guess where the extra Liberal support came from...

My projection had given the Liberals roughly 43% to the NDP's 42%, with the Conservatives at 8%. Luckily, I had Cariboo North as a possible pick-up for the Liberals and Fraser-Nicola was closer than expected, so I didn't do too badly here with my seat calls.

The Lower Mainland, given the disparity of results within the region, requires I break it down, going from east to west.

Fraser Valley - Election 2013
BC Liberal Party: 50.1% (-3.4%) - 9 seats (+1 seat)
New Democratic: 30.9% (-4.2%) - 0 seats (-1 seat)
BC Conservative: 7.7% (+4.3%)
Green Party: 6.5% (-0.6%)
Others: 4.9% (+4.0%)

The Liberals swept all nine ridings in the Fraser Valley, taking away the only incumbent the NDP had from 2009 - Maple Ridge-Pitt Meadows - as well as Chilliwack-Hope, which they won in a by-election. The Liberals also beat out their former colleague, John Van Dongen, in Abbotsford South by a good sum, as well as relegating BC Conservative leader John Cummins in Langley to a terrible third-place finish.

I had expected the NDP to do much better than the 31% they managed, possibly getting to around 40% and making play of some of the more margin seats such as Langley. The Conservatives, I thought, would also manage to get around 12% - yeah, not so. The Liberals crushed everyone here.

Surrey - Election 2013
BC Liberal Party: 48.4% (+2.6%) - 5 seats (+1 seat)
New Democratic: 41.5% (-5.6%) - 3 seats (-1 seat)
BC Conservative: 5.6% (+5.0%)
Green Party: 3.2% (-3.1%)
Others: 1.3% (+0.9%)

Here is another example of the NDP likely losing voters to the Liberals, rather than simple turnout depression. The preliminary numbers I have show the Liberals with 75,000 voters, compared to 68,000 in 2009 - the NDP dropped almost 6,000 in the meantime. This despite an increase in the Conservative's vote, and a drop in the Green's. I also don't think anyone expected numbers like these. The one pick-up was BCL Peter Fassbender winning over incumbent NDPer Jagrup Brar in Surrey-Fleetwood.

My projection had given the NDP almost 53% support in Surrey, to the Liberal's 35%. This was a big miss on my and the pollster's part.

Richmond-Delta - Election 2013
BC Liberal Party: 47.5% (-4.7%) - 3 seats (+1 seat)
New Democratic: 27.5% (-2.4%) - 0 seats (-1 seat)
BC Conservative: 5.9% (+4.1%)
Green Party: 5.4% (=)
Others: 13.7% (+2.9%) - 1 seat (=)

Richmond-Delta featured only one close race, that being Delta North where the NDP lost a previously safe seat to the Liberals. Despite that, both major parties decreased while the Conservatives and "Others" (mostly Independent Vicki Huntington, who kept her seat) jumped up, and the Greens were stable. It just didn't translate into much support for any of the smaller parties.

This is one of the regions where I did relatively OK, as I expected the incumbent Liberals and Huntington to hold on to their seats with relative ease. However, I obviously missed Delta North - I think everyone did. The Liberals didn't run a star candidate in Scott Hamilton (and in fact had issues before with their previous star candidate, Sukh Dhaliwal), and this was by all accounts a safe NDP seat. What happened on the ground there?

Vancouver Suburbs - Election 2013
BC Liberal Party: 45.4% (-2.4%) - 8 seats (=)
New Democratic: 42.3% (+0.6%) - 6 seats (=)
Green Party: 7.3% (-0.6%)
BC Conservative: 2.5% (+1.3%)
Others: 2.6% (+1.2%)

The suburbs around Vancouver didn't see too much change, though both the Liberals and the NDP lost seats to each other. The Liberals lost Burnaby-Lougheed to the NDP's Jane Shin, while the Liberals picked up Coquitlam-Maillardville with Steve Kim. The Liberals also won Port Moody-Coquitlam back from the NDP, who won it in a by-election previously.

My projection had expected the Liberals to lose big in the eastern suburbs, losing all of their seats except maybe Coquitlam-Burke Mountain, which was a Too Close To Call race. Clearly that did not happen. The Liberals also held on to all of their northern suburb seats, which I had said would happen with the exception of North Vancouver-Lonsdale, which I said leaned NDP.

Vancouver - Election 2013
New Democratic: 45.9% (+0.8%) - 7 seats (+2 seats)
BC Liberal Party: 43.1% (-2.6%) - 4 seats (-2 seats)
Green Party: 8.7% (-0.4%)
BC Conservative: 1.0% (+1.0%)
Others: 1.4% (+0.4%)

The Liberals lost within the City of Vancouver itself, losing two incumbents - Margaret MacDiarmid in Vancouver-Fraserview, and Premier Christy Clark in Vancouver-Point Grey. George Heyman and David Eby won those races for the NDP respectively. However, the NDP failed to win other low-hanging fruit, such as Vancouver-Fraserview or Vancouver-Langara, to former of which I really expected to go NDP. Despite this, the numbers overall barely changed, and Conservatives made zero impact anywhere - they didn't even hold the difference in Clark's race, with their candidate only winning 350ish votes (there was an over 700-vote gap).

Vancouver Island - Election 2013
New Democratic: 43.9% (-6.1%) - 11 seats (+1 seat)
BC Liberal Party: 34.2% (-4.4%) - 2 seats (-2 seats)
Green Party: 17.2% (+6.8%) - 1 seat (+1 seat)
BC Conservative: 4.0% (+3.7%)
Others: 0.7% (-0.1%)

Obviously the big change on Vancouver Island was the increase of the Greens, who came out to win Oak Bay-Gordon Head with 40% of the vote. In fact, the Greens won almost 27% in Greater Victoria itself, just behind the Liberals, who also lost Saanich North & the Islands to the NDP in a tight three-way race with the Greens. Outside of Victoria, however, the Liberals held on to their incumbent seats in Comox Valley and Parksville-Qualicum with little difficulty, while almost taking Cowichan Valley from the NDP. The Conservatives also managed a somewhat impressive number outside of Victoria, about 7%, mostly thanks to the popular local candidate in Parksville-Qualicum.

Amazingly, except for Oak Bay-Gordon Head (an incorrect call I expected), my projections did pretty well in Vancouver Island, calling every riding correctly with the ranges included.

So, with the results out of the way, what is there to say?

The polls got it wrong. Why they did so will be something I need to toss over in my head for a few more days. Was it a shy-Liberal effect? Did expectations of an overwhelming NDP win lower their turnout? Did we get undecideds completely wrong? All questions to be answered in time.

What I can speculate on is why Clark won, and that is, essentially, the fact that her opponents were second-rate at best. For all of Clark's shortcomings, she was a known quantity that people began to appreciate when thrown into direct comparison with Adrian Dix, the mousey policy wonk who proposed tax increases and could apparently kick a dog and still win the election. Clark was clearly the more impressive and experienced leader, a fact really shown by the April 29 debate, while her opponents were fumbling amateurs (Dix and Cummins) or just irrelevant (Sterk). The authors of Clark's victory was as much them as it was herself and her charisma.

There is also much to say about negative campaigning, and the fact that you need to respond to it forcefully. Dix pulled defeat from the jaws of victory because he allowed Clark and the BC Liberals to define him as a flip-flopping, tax-increasing, backroom boy behind the NDP's previous fiasco governments in the 1990's. Dix, admirably trying to take the high road, doomed himself by not responding forcefully to such ads. He pulled a Dion and Ignatieff, except on a much worse scale given that neither of them were expected to win government.

In the end, I don't think this is confirmation of the BC Liberal's governing agenda, as much as it was a rejection of Adrian Dix and John Cummins. Voters, both the right-leaning voters and the centrists that moved from the NDP to the Liberals (they clearly exist), were convinced that they were worse than the Devil they knew.
That is saying a lot, considering how much they dislike Clark.

The future of my projections, frankly, are in doubt. Not because I got it wrong. In fact, had I known the previous numbers I gave above, my model would've come out with 51 Liberals to 33 NDP and an Independent. That's pretty damn accurate. The pollsters, and the data they gave us, were clearly incorrect. Sadly those are the only numbers I have to go on, so where do I really go from there? Until my confidence in their numbers is restored, how can I put out any projections and say this is (relatively) accurate? How can anyone?

As I said before, questions to be answered in time.


  1. I think there's a demographic factor that played into the Liberals' hands. There was a 48% turnout. Older voters with a stronger sense of civic duty probably have more influence in low-turnout elections. They also seem to be the most susceptible to fear-mongering and attack ads.

    1. True, but pretty clearly in many ridings the Liberals gained new voters compared to 2009. Where did these new voters come from?

    2. I think a lot of these new voters for the Liberals came from centrist voters who voted NDP last time because Gordon Campbell was to right wing for them but felt able to support the more centrist Christy Clark.

    3. Hm, possible. But then, how did she lose her riding? That is a big federal Liberal area.