Election predictions, 2009
Posted May 10, 2009 by Sacha Peter - Link
Category: Predictions
Comments (35)
Here is a list of various people with their election predictions (note this will be a “live” entry, which will change as predictions come in up to May 11). Note the focus is on seats and not popular vote:
Real people with real names
Jordan Bateman – BCL 49, NDP 36
Nick Boragina – BCL 52, NDP 33 (Link to seats predictions: 1, 2, 3)
Milton Chan – BCL 46, NDP 39
Keith Freeman – BCL 42, NDP 43 (Link to seats predictions: 1, 2)
Freddy Hutter – BCL 54, NDP 31
Eric Lanoix – BCL 57, NDP 28 (Link to seats predictions: 1)
Trevor Loke – BCL 49, NDP 35, Ind 1
Vaughn Palmer – BCL 47, NDP 37, Ind 1
Bernard von Schulmann – BCL 55, NDP 30
Andy Shadrack – BCL 46, NDP 39, Ind 1 (note: 86 seats is a bold prediction!)
Kennedy Stewart – BCL 52, NDP 33 (Link to seats predictions: 1)
If you know of any other people posting an election prediction, feel free to add them to the comments, with a link to their website. I will only link to people with real names. My prediction will come out on May 11th. (and no, my prediction won’t be influenced by what other people are predicting!)
The only other prediction mechanism that I would have loved to have seen was the ESM, but alas, they were not open this time around.
Milton Chan’s electionprediction has 7 TCTC, not 8 (as there are only 85 seats ;-) It appears that he has simply not called the close seats is the Vancouver suburbs/Fraser Valley yet. Hey has called all the Interior, Island, and Vancouver Ridings.
I think he might be the closest one if he assumes that all the close ones in the Vancouver suburbs stay Liberal. Kennedy Stewart’s numbers are I think a best case scenario for the BC Liberals. The others seem too lopsided for a race that, like 2005, is probably closer than most polls suggest. I hope I’m wrong.
“Eric Lanoix – BCL 57, NDP 28
Bernard von Schulmann – BCL 57, NDP 28″
Even Freddy Hutter makes this statement: “based on our long term momentum overlay, the Liberal Party is poised for a 57-28 Majority victory over the NDP on May 12th.”
So three prognosticators make the exact same prediction.
I’m of the view that the Libs will win between 55 – 60 seats myself. (derived from opinion polls, momentum, and other anecdotal evidence)
I’m also of the view that the popular vote spread will be in favour of the Libs by a margin of 10%.
The question in the Ipsos poll, regarding 66% who confirmed that they are “absolutely certain to vote” resulting in a 11% margin in favour of the Libs (an additional 3%), sealed the deal in my mind, notwithstanding the 9%, 8%, and 9% spread in favour of the Libs by Mustel, Ipsos, and IRG respectively.
ARS is simply the outlier in this modelling.
I think its a lot closer than it appears, and my prediction is for a 4% spread with the Liberals squeezing by 45 to 40, with a few recounts. It can be ugly.
On my blog (link in the website section) I am predicting
BCL – 52
NDP – 33
riding by riding out very shortly.
Riding by Riding is out.
Note that we use a mathematical arithmetic uniform ratio method. That means that, in short, if a party took 40% in the last election, and we think they will take 44%, a 10% increase, we increase their assumed votes in each riding by 10% (and then re-balance each riding to total 100%)
I think that 2005 is much closer to 1983 and that this election could end up looking like either 1979 or 1986 in vote spread between the Liberals and NDP. I am not in a position to predict because I do not have an intuitive feel for what is happening on the ground.
My gut tells me that the voters are not ready to return the NDP to power from the penalty box, but that neither is Campbell very popular, especially after that condescending remark to Carole in the tv debate.
That said the wild card for me is that I know there are some very tight individual races around the province like Bill Bennett in Kootenay-East and I know that Michelle Mungal is not Corky Evans in Nelson-Creston. In 1986 Corky was seen as too radical and lost to Socred Howard Dirks by a mere 27 votes. This time Liberal Josh Smienk has been a regional district Director since he was 25 and he helped found and headed up the Columbia Basin Trust.
In Powell River-Sunshine Coast one has to ask where will all the votes Adriane Carr recieved in 2005 go in 2009. So this election might boil down to a number of individual constituency races and I am awaiting the final Advance numbers to see how many constituencies are between 12% to 15% of the registered vote.
Seat Projections:
Liberal: 41
NDP: 41
Green: 2
Independent: 1
So what happens in a tie, besides a minority government, that we need and deserve to keep Campbell in check. Obviously the incumbent party governs until it is defeated on non-confidence and then there will be a James government.
Sacha – how reliable is the ESM? Markets also have a dynamic of their own and they are subject to network and bubble forces.
Has there been a study comparing ESMs to Polling lets say?
I also miss the UBC ESM. I think it was taken off due to political reasons. After all, UBC is one giant state institution.
This predictor is weird. It says UBC ESM on May 2 is predicting 66 Lib 19 NDP!
http://www.nodice.ca/elections/britishcolumbia/projections.php
How is it that in this election there is no talk about an MLA’s voting record?
In the UK elections, this is a central issue, and each and every vote cast by each and every representative is closely followed and statistics and simulations are garnered.
I don’t recall any talk of this in this election.
Hello everyone,
BJ predicts a spread of 10% between the LIBs and the NDP. That could be true, as long as Angus Reid has been consistently really wrong over their last 3 polls (yeah, their methodology IS different from other firms). The jury is out until Tuesday night on that one… That’s why I predict a spread way above Angus Reid and just below the other polls that came out lately (7.3%). With that, I get 57LIB/28NDP. With a 10% spread, the most likely outcome from my model (HiPDaTA) would be 64LIB/21NDP…
Thanks a lot to Nick Boragina for sharing his riding-by-riding predictions. Mine are also available on the website. I look forward to looking at your predictions in more detail tomorrow.
Also, I don’t think the “paper-napkin” model will perform so well this time because the electoral map and the party dynamics are very different from 4 elections ago. For example, a repeat of the ‘96 results where the NDP got a legislative majority, but lost the popular vote by 3% is just not in the cards this time, given the new electoral map. In fact a 3% spread in favour of the LIBs gives a most likely outcome of 47LIB/38NDP.
If I did not screw up anywhere, the only way the NDP could win Tuesday is if Angus Reid was actually over-estimating the LIB vote (and all others were way off!?!). That is a very very long shot… It’s the first time I use this model, so for me it is a learning experience…
Glass asks if there has been a study of the reliability of the UBC ESM in predicting the election outcome. I did that analysis a few months ago. The results are available at http://www.reallifemath.ca/Site/Blog/Entries/2009/2/16_Error_for_Several_Election_Forecasting_Methods.html The ESM ranked 1st out of 3 (that I know of) for the ‘05 BC election and 2nd out of 6 (that I know of) for the ‘08 Canada election in the category of “Overall Error”. Not bad!
I apologize for the long posting, I just enjoy talking about this kind of stuff. A great week to everyone and good luck Tuesday night :-)
EL
Glass – the ESM is fairly good because there are market participants that know how to price political contracts properly, for economic benefit. The reason why I was looking forward to the UBC ESM was because it has been a consistent net generator of cash for me whenever election time comes.
The Nodice 66/19 ESM projection is clearly incorrect.
Finally, voting records are nearly irrelevant in BC due to the majority government – they’re better for attendance records than anything else. The only vote which stood for anything was Lorne Mayencourt’s lone vote against repealing the initial MLA pay hike.
During the campaign, the BC Liberal line of “(insert policy plank), and the NDP voted against it” is exceedingly tiring to listen to.
Another constitutional crisis anyone? ;)
In my opinion, the only way to predict an outcome is on a riding by riding basis; with regional polls being some help. This was truly the lesson of ‘96 if you will. I don’t put much stock in province-wide polls or paper napkins…
Also, there will be surprises on election night, the question is– where?
Not based on any other prediction, but aided by the updated ridings Sacha has profiled here, I get a final count of NDP: 43, BC Liberal: 42.
Honestly I expect the BC Liberals to win this election, so the prediction was surprise to even myself until I counted the ridings up at the end.
If I were to just pick numbers without going riding by riding, it would probably be something like 46-39 BC Lib… (but I’m not hedging, I’ll stick to the riding by riding prediction!)
Most criticisms of the predictions will surely be directed towards the Burnaby ridings (NDP 3-1), but most of my picks I think are fairly reasonable to anyone who isn’t a “hack”, particularly.
Well, at least one person is going out on a limb here, even if he didn’t expect to be. If just for laughs, here they are— :~..
http://keefer.typepad.com/predictions1.jpg
http://keefer.typepad.com/predictions2.jpg
(note: I left out Maple Ridge-PM, which I’ll put under NDP by 3 per cent)
By the way, I appreciate the amount of work that has gone into this site. It’s by far the best online resource ever created by an individual in advance of a provincial election.
The Antweiler UBC ESM placed 8th of 16 projections in the recent October Federal Election: http://trendlines.ca/electcanada.htm#2008 and 2nd of six in 2006.
We use the Antweiler Voter Matrix in our long term tracking both Federally & Provincially when there are no models contributing results. It was the best model during the recent Federal election albeit its results were not publicly available. EKOS had the best “available” forecast.
Our running average chart attempts to include all models that are making periodic web-available projections. Dave MacDonald’s nodice efforts seem to reflect only single poll conversions, whereas others mostly use a weighting or avg of several. Because it doesn’t reflect the sentiment change over recent days, we chose to omit it from the final AVG.
I disagree with K4 on Bill Bennett. I think Troy Sebastian will take it with a 5% margin of victory.
I also think Nelson-Creston will be 46%-44%, but I do not know who will be on top Mungall or Smienk.
I will not be surprised if Mckee beats Norm McDonald in Columbia River- Revelstoke.
I think Penticton will be tighter for the Liberals.
One of the surprises of the night could be Vernon-Monashee, where vote splitting could actually see an upset for the NDP.
Finally Peace River-North, do not assume the NDP will come second there.
I believe Nodice is using the UBC ESM forecasting tool and what the outcome would look like if one uses a current set of polling numbers.
As to the voting record, the BC Liberals are the first party in generations in BC not to demand that all MLAs vote the party line all the time. Though very few of them have taken advantage of this over the last eight years. Blair Lekstrom and Tom Christensen are two examples of people who voted against the government at times when they were not in cabinet.
http://www.langleypolitics.com/2009/05/bc-prediction-thread.html
Take it the bank, Sacha:
Seats – BC Liberals 49, NDP 36
FPTP 51%, BC-STV 49% (STV needs 60% approval to go ahead)
Bernard, the problem with MacDonald’s reliance solely on the Environics poll is that altho its L-47% mirrors later polls, their 36% for NDP sentiment was the lowest of all the May polls. It is certainly not a rogue poll, but use of it alone explains his 66-MPs award to the Liberals … much higher than consensus. BTW he uses the UBC-Antweiler VMM … not ESM.
By not averaging the other polls, inclusion of Nodice forecast would skew the Avg for all practitioners, an incorrect methodology. As we saw subsequent to the Environics release, the other modelers revised downward their projections as they considered, weighted and converted the five later polls.
My own application of VMM indicates 57 Liberals. As mentioned VMM was the best forecast model for 2008 Federal Election. OTOH, the TrendLines multi-model AVG (53) has the best forecast record over the past 4 Elections (Federal & Ontario). I have low confidence in our long term momentum indicator (57) method due to its shaky record.
All things considered, a result between 53 & 57 seems assured.
Liberals: 50
NDP: 35
Popular vote: 46% – 40% The Greens get 10%, the BC Conservatives get 2%, and independents and others get 2%. James and Campbell are fine. Sterk barely comes 3rd, Huntington comes 2nd, and Hanni comes 3rd and costs Bennett his seat. Referendum gets 50% support. (My riding of Saanich South goes to Robin Adair of the BC Liberals.)
Rationale: Within the margin of error of all recent polls, and fairly similar to last time, including the understated NDP vote. This is because NDP will again be the best at getting the vote out, the BC Liberals not quite as good, and the Greens will have the least reliable voters. Green supporters will be less inclined to vote for the NDP than 2005. Less of a backlash effect than in 2005 i.e. people saw that the BC Liberals were likely going to win and wanted to ensure a strong opposition. Again, the NDP vote will be more efficiently distributed, so they won’t “lose” seats, but their proportion of seats will drop slightly. BC Liberals pick up in the Fraser Valley, GVRD and my riding but lose a couple rural ones.
Referendum loses support. Last election, people were still mad that the NDP only had 3 seats, but this time the legislature is more balanced. Plus many people are more inclined to vote against something that was already defeated, even if it originally had popular support. It’s beating a dead horse. Much like the Charlottetown Accord referendum after Meech Lake failed.
My riding-by-riding predictions here: http://scaryfundamentalist.blogspot.com/2009/05/bc-election-predictions.html
Seats: Liberals 45 NDP 40
Predictions! Harumph. No 4th day Advance poll results posted yet and it is nearly 3.30 PM. Well here goes.
The Green Party has not dropped below it’s federal results since before 1991, but this might be the exceptional election year. The Green Party vote will be between 9% and 11%. It depends on whether Green supporters believe the NDP have a chance of forming government and how many traditional NDP voters are mad at Carole over single gender nomination races and her position on the carbon tax.
The Liberals and NDP are a replay of Socreds versus NDP. For the NDP to beat the Liberals in this election they would have to collect all their vote from the federal election plus all the federal Liberal votes. The Conservatives under Harper took 44.5% of the vote and one has to ask how many of those voters will support Carole James and the NDP, but more importantly one has to ask how many of those votes will be siphoned off to the 24 BC Conservative candidates. 2% to 2.5% tops, maybe lower now they realize they could help elect an NDP government.
As for “Other” that has dropped from 2.2% in 2004 federally to .7% in October 2008, but the wild card in all this is how many votes will Vicki Huntington and any other well known Independent get. Another wild card is how much influence former Socred Cabinet Minister, Rafe Mair, is going to have over his quest to stop IPPs. I am actually hearing that some federal Conservative and traditional provincial Liberal voters will vote NDP over this issue.
I think there is a possibility of an NDP sweep on Vancouver Island, but no change of status quo in the Lower Mainland, with the possible exception of Delta-South. This pushes the NDP up to 38, with the possibility of one “Other”. While Vancouver may not be hurting economically because of the Olympics the rest of BC has been hurting with the forest industry being in recession since at least late 2007 and now we have mining and oil and gas down too. Thus my prediction for the Liberals starts at 46 and goes down from there depending on what happens in the North and the Interior.
To form a majority government one party has to get 44 seats, less the Speaker equals 43. At a lower number of 43 seats, still a statistical majority, with one independent elected, things would start to get very interesting after May 12, as the split between the NDP and Liberals would be 41 to 42. Which way would a Vicki Huntington go? Would she go with the Party with the most seats or the largest popular vote? And finally if we end up with one party getting more seats than the other, and the popular vote is reversed as it was in 1996…BC-STV might fail in the referendum but do not bet that electoral reform will be off the agenda…especially if Vicki Huntington holds the balance of power as she is one of the endorsers of BC-STV on their website.
I altered my prediction slightly to Liberals 55 and NDP 30. Since this is such a status quo election, I boosted incumbency values and this pushed the NDP up a bit and shifted two seats
Using a formula based on a regional (e.g Okanagan, Fraser Valley, etc.) analysis of polls (which are averaged with a weighting formula based on sample size, recency, methodology, and accuracy in the last election), my predictions for this election are:
Liberal 50
NDP 35
There is a greater probability that the Liberals will win less seats than more, mainly due to the possibility of the NDP sneaking up the middle in a couple more ridings than expected. I have already accounted for that effect to some degree, otherwise, based on the polling average, the prediction would show a much more lopsided Liberal victory.
Looking at the numbers it may appear that the Liberals are walking away with this one, but it won’t take much for the NDP to upset. It may sound a bit cliche, and a lot of people have said it before, but it’s “closer than you think.”
New updated prediction:
NDP – 44
Liberals – 37
Greens – 3
Independent – 1
NDP: keep Fairview, take Burnaby Deer-Lake, Burnaby Lougheed, North Vancouver-Lonsdale, Delta North, Surrey-Fleetwood, Maple Ridge-Mission and Maple Ridge-Pitt Meadows, Boundary-Similk, Kamloops-NT, Stikine
Greens: take Vancouver-False Creek, West Vancouver-Capilano and Vernon-Monashee
Independent: Huntington slays Oppal
What I note from the Advanced poll numbers is that about half the seats on Vancouver Island have the potential to become battelground seats, the same for the North, the Kootenays and 3 of 5 in the Cariboo-Chilcotin.
In contrast only 5 of 47 seats in the Lower Mainland/Fraser Valley had turnouts above 12%. Does anyone know what is going on in Surrey-White Rock with an 85.9% increase in Advance turnout?
My seat prediction:
46 Liberals
39 NDP
1 Independent
Seats to watch the Green Party in:
Esquimalt-Royal Roads
Peace River-North
Penticton
Vernon-Monashee
At least one of these four candidates will place second and the average vote for the four will exceed 20% and may go as high as 30% in one.
I also predict that it will be a long night and we may not know who is in government before Wednesday morning…even with a vote spread of 3% to 4% overall.
I want BC-STV to exceed 50%, but I am not sure it will.
My prediction
49 Liberal (45.31%)
35 NDP (39.72%)
0 Green (10.58%)
1 Independent (Vicki Huntington)
Green ridings with over 20%:
Vancouver-False Creek
Nanaimo-North Cowichan
Esquimalt-Royal Roads
West Vancouver-Sea to Sky
Peace River North
Penticton
Vernon-Monashee
Andrew – your total adds up to 86 – there are only 85 seats up for grabs
Cheers
I think the disatisfaction with Campbell is underestimated. The NDP have some considerable momentum coming out of the debate (even the TV polls suggest so much).
The US election is also on some people’s minds. Many folks who’ve lost big on the stock markets are likely to blowback on the deregulation, privatization meme of the Liberals. The senior crowd is also very irate.
Only one prediction here seems even close to me, and that’s the one that puts the NDP up a seat.
Poverty and children’s issues is likely to put a lot of mom’s and dad’s in the lineups at the polls.
The NDP have also done a very good job on the internet, not only raising money, but framing debate. The young will be most affected by this. If a lot of young folk show up, look for the NDP to do very well.
Large turnouts in advance polling, from what I’ve seen over the years, usually indicate a change in government. If the voting day numbers are also up by 10 per cent, that change is even more likely.
I predict a narrow NDP win, with an independent or two getting in due to voter disatisfaction.
I, unlike some folks here, am not allied with any party, and am basing my opinion on conversations I’ve had around the province over the past few months.
I’m afraid the STV will get lots of support but not quite enough. Its been set up to fail, but won’t fail as bad as hoped by the parties.
My quick estimate after looking at 05 results and current popularity polls and doing a few calculations:
Liberals: 47
NDP: 38
or if Vicki Huntington gets in, which may be possible:
Liberals: 46
NDP: 38
Ind: 1
I believe it is important to take into consideration the green and conservative vote is various key ridings.
Tim Ell, a political science student at UVic, has:
Liberals 46
NDP 38
Independent 1 (Vicki Huntington in South Delta)
STV: 48% (and 36/85, or 42% of ridings, approve)
FPTP: 52% (and 49/85, or 58% of ridings, reject)
His spreadsheet is actually very good and I’d say highly accurate – he has been updating consistently throughout the campaign as well.
You can find his most recent spreadsheet at http://web.uvic.ca/~aallen/BCElection2009_Final.pdf . Someone is hosting it for him there.
Just for the fun of it:
44 Liberal
41 NDP
This is based on nothing more than gut instincts, so I can’t give anything in terms of popular vote, riding shifts, or anything technical like that.
Nelson-Creston will be a horse race, and other than local boy Andy Shadrack, nobody else has been paying attention to this one.
Nelson-Creston is NOT solid NDP territory; it was solid Corky Evans territory, and that’s very different. Corky was a cult of personality, who drew votes from across the spectrum.
Michelle Mungall is quite divisive, and represents some of the old-school, step-on-your-neck style politics that I don’t think resonates much anymore.
Josh Smienk has huge experience, way better name recognition, and did much better in all-candidate forums, interviews, etc.
That said, the way Michelle won her seat on City Council a few years ago was by working hard, and knocking on every door in town. She worked hard this election too, but my reports from friends and family in Nelson is that she didn’t do a great job of ingratiating herself.
I betting on Josh, which will be one of the stories of the night tonight. It’s a riding that the NDP think they have locked up, and they don’t.
When the Premier stumped there mid-campaign, it was one of the best, most energetic early campaign stops. About 150 supporters, lots of energy, no protests. Mark my words, it’s not the same old Nelson Creston.
OK, I will throw in my 2 cents as well.
I predict we will end up with:
BCL 52 seats
NDP 33 seats
Huntington will come close in Delta South – within 5%, but Wally will hang on.
STV will fail with 51%.
Popular vote predictions:
BCL 47%
NDP 38%
Green 12%
Other 3%
My guess is also that overall voter turnout will be lower than expected.
I’m going to say 58%
I spoke with a candidate this morning (not from either the BCL or NDP) and he predicted as many as 55 BCL seats.
do i get a Cuban cigar if i get it right ?could be completely off here but my reasoning is that there is more resentment in the interior and the Island than has been baked into the opinion polls.therefore watch Kamloops and the Penticton ,williams lake ridings.also Burnaby in the lower mainland.maybe Delta?also Comox. my prediction is Liberals 44 NDP41.