Analysis – Ontario: 2013 by-election results versus 2011 results

The ballots have been counted in Ontario’s five provincial by-elections: London West, Etobicoke-Lakeshore, Scarborough-Guildwood, Windsor-Tecumseh, and Ottawa South. Pundits are telling the public what they think the 2013 by-election results – considered in isolation – mean, but I have a different analysis for you which, I think, reveals what is actually happening to the parties in Ontario. Below, I compare the results from the 2013 by-elections to the results from those ridings in 2011, each adjusted for the riding’s change in voter turnout.

The big conclusions – described in more detail after the data from the five ridings are broken down below – are these:

1. Support for the PCs is more or less stagnant, right across the province: the PCs were unable to improve their standings even in the face of a province-wide Liberal melt-down.

 

2. Freedom Party of Ontario made a big breakthrough in the riding of London West, with the candidacy of Al Gretzky, where PC candidate Ali Chahbar’s campaign flopped.

 

3. Polling proved to be misleading. In particular, in many cases, at times grossly exaggerating anticipated PC fortunes, or underestimating Liberal strength.

 

4. The PCs quite possibly have lost the machinery they need to turn ‘likes’ into votes: the NDP and Liberals appear still to have the polling-day workforce needed to get people out of their houses and into polling stations (probably due to volunteer support from public-sector unions).

 

And now, the analyses:

LONDON WEST

In London West, 35980 ballots were cast in the 2013 by-election. This compares to 49,525 votes cast in the 2011 general election. Therefore 72.7% of the number of ballots cast in 2011 were cast in 2013. That’s a 27.3% reduction in total ballots cast.

In 2011, the ballots cast were as follows:

NDP: 10,757
PC: 14,603
Liberal: 22,610
Green: 1,194
Freedom: 300

Therefore to do as well in 2013 as they did in 2011, as a proportion of the total ballots cast in 2011, the parties needed to bring in the following numbers of votes (i.e., 72.7% of the ballots each received in 2011):

NDP: 7,820
PC: 10,616
Liberal: 16,437
Green: 868
Freedom: 218

Therefore, as compared to the ballots each party needed to repeat their 2011 performance, the parties increased or decreased their performance as follows (2013 votes minus votes needed to do as well as they did in 2011):

NDP: 15,063 – 7,820 = +7,243 votes (+92.6%) better than breaking even
PC: 12,122 – 10,616 = +1,506 votes (+14.2%) better than breaking even
Liberal: 5,866 – 16,437 = -10,571 votes (-64.3%) worse than breaking even
Green: 1,575 – 868 = +707 votes (+81.5%) better than breaking even
Freedom: 1,838 – 218 = +1,620 votes (+743.1%) better than breaking even

And, in respect of Freedom Party, at 5.1% of the total ballots cast in London West, we had our best ever result in Freedom Party’s history.

Conclusions re: London West

1. In London West, with the exception of the Liberals, all parties did better in the 2013 by-election than they did in the 2011 general election.

 

2. As a percentage of the votes each party won in 2011, at 743.1%, the Freedom party’s support increased more than the increase enjoyed by the NDP (92.6%), Greens (81.5%) and PCs (14.2%).

 

3. In absolute terms, the number of votes gained by each party, as compared to 2011 (adjusted for lower 2013 voter turnout) was greatest for the NDP (7,243 votes), followed by the Freedom Party (1,620 votes), followed by the PCs (1,506 votes), followed by the Greens (707 votes).

 

4. As compared to the other parties that gained at Liberal expense, the PCs were relatively unable to capitalize on the implosion of the Liberals, which suggests that – even during a by-election, when each party’s workforce can be concentrated in just a number of Ontario’s ridings – the PCs are relatively unable to improve their standings. This suggests that the PCs have little potential for improving their support in London West. They appear to have maxed-out their support. In a general election, when all parties’ workforces are stretched thin, PC support may well turn out suffer a reduction.

 

5. Clearly, the unions either did not come out to help Coran, or were not able to convince voters to vote Liberal. And, clearly, unions were out in force in London West getting out the vote for Peggy Sattler (as I had told Andy Oudman and others at the beginning of the by-election).

 

ETOBICOKE-LAKESHORE

In Etobicoke-Lakeshore, 34584 ballots were cast in the 2013 by-election. This compares to 43,448 votes cast in the 2011 general election. Therefore 79.6% of the number of ballots cast in 2011 were cast in 2013. That’s a 20.4% reduction in total ballots cast.

In 2011, the ballots cast were as follows:

NDP: 6,713
PC: 12,705
Liberal: 22,169
Green: 1,164
Freedom: 174
Libertarian: 172

Therefore to do as well in 2013 as they did in 2011, as a proportion of the total ballots cast in 2011, the parties needed to bring in the following numbers of votes (i.e., 79.6% of the ballots each received in 2011):

NDP: 5,344
PC: 10,113
Liberal: 17,646
Green: 927
Freedom: 139
Libertarian: 137

Therefore, as compared to the ballots each party needed to repeat their 2011 performance, the parties increased or decreased their performance as follows (2013 votes minus votes needed to do as well as they did in 2011):

NDP: 2,705 – 5,344 = -2,639 votes (-49.4%) worse than breaking even
PC: 16,130 – 10,113 = +6,017 votes (+59.5%) better than breaking even
Liberal: 14,513 – 17,646 = -3,133 votes (-17.8%) worse than breaking even
Green: 780 – 927 = -127 votes (-13,7%) worse than breaking even
Freedom: 57 – 139 = -82 votes (-59.0%) worse than breaking even
Libertarian: 157 – 137 = +20 votes (14.6%) better than breaking even

Conclusions re: Etobicoke-Lakeshore

1. In Etobicoke-Lakeshore, with the exception of the PCs (and, to a marginal extent, the Libertarians), all parties did worse in the 2013 by-election than they did in the 2011 general election.

 

2. As a percentage of the votes each party won in 2011, at -59.0%, the Freedom party’s support technically decreased more than the decrease suffered by the NDP (49.4%), Liberals (17.8%) and Greens (13.7%).

 

3. However, the percentage of the vote cast in favour of Libertarians or Freedom was so small that it is pretty much impossible to read anything into the changes in those parties’ fortunes relative to 2011. At 100 seats between the two parties, that’s not a difference of any importance: its a bus accident’s worth of difference.

 

4. Clearly, the candidate for the PCs in Etobicoke-Lakeshore was what turned an almost certain Liberal win into a PC win.

 

SCARBOROUGH-GUILDWOOD

In Scarborough-Guildwood, 24,706 ballots were cast in the 2013 by-election. This compares to 31,894 votes cast in the 2011 general election. Therefore 77.5% of the number of ballots cast in 2011 were cast in 2013. That’s a 22.5% reduction in total ballots cast.

In 2011, the ballots cast were as follows:

NDP: 6,194
PC: 9,137
Liberal: 15,607
Green: 413
Freedom: 136
Libertarian: 407

Therefore to do as well in 2013 as they did in 2011, as a proportion of the total ballots cast in 2011, the parties needed to bring in the following numbers of votes (i.e., 77.5% of the ballots each received in 2011):

NDP: 4,800
PC: 7,081
Liberal: 12,095
Green: 320
Freedom: 105
Libertarian: 315

Therefore, as compared to the ballots each party needed to repeat their 2011 performance, the parties increased or decreased their performance as follows (2013 votes minus votes needed to do as well as they did in 2011):

NDP: 7,010 – 4,800 = +2,210 votes (+46.0%) better than breaking even
PC: 7,606 – 7,081 = +525 votes (+7.4%) better than breaking even
Liberal: 8,852 – 12,095 = -3,243 votes (-26.8%) worse than breaking even
Green: 532 – 320 = +212 votes (+66.3%) better than breaking even
Freedom: 80 – 105 = -25 votes (-23.8%) worse than breaking even
Libertarian: 118 – 315 = -197 votes (-62.5%) worse than breaking even

Conclusions re: Scarborough-Guildwood

1. In Scarborough-Guildwood, the Greens (66.3%), the NDP (46.0%), and the PCs (7.4%) improved their performance over 2011 (adjusted for 2013 voter turnout); the Libertarians (62.5% decrease), Liberals (26.8% decrease), and Freedom (23.8% decrease) suffered reduced returns. That said, the Green, Freedom, and Libertarian numbers arguably are too small for those percentage-change figures to mean much: consider that the one Independent who ran in this by-election got 195 votes, as compared to Freedom’s 80, and the Libertarians’ 118.

 

2. The NDP appears to have benefitted most from Liberal support loss, with the PC candidate picking up the minority of the loss. A lot of campaigning was done by the NDP in Scarborough, for “star” candidate Adam Giambrone, and that – combined with a probably-large contingent of unionists getting out the vote – accounts for the NDP’s increased support in the riding. The Liberals pulled off a win *despite* losing so much support, because they were so far ahead in the riding in 2011.

 

3. Footnote: this is former Libertarian leader Sam Apelbaum’s riding. It’s not clear why he did not run as the candidate (Heath Thomas did), but it is worth noting that he was not re-elected party leader in November 2012.

 

4. As in London West, the PCs appear to have their support largely capped. In London West, they grew only 14.2% and in Scarborough only 7.4%, in the face of huge Liberal losses in each riding. As in London West, this suggests that the PCs have little potential for improving their support in Scarborough-Guildwood They appear to have maxed-out their support. In a general election, when all parties’ workforces are stretched thin, PC support may well turn out suffer a reduction.

 

WINDSOR-TECUMSEH

In Windsor-Tecumseh, 25,596 ballots were cast in the 2013 by-election. This compares to 37,231 votes cast in the 2011 general election. Therefore 68.7% of the number of ballots cast in 2011 were cast in 2013. That’s a 31.3% reduction in total ballots cast.

In 2011, the ballots cast were as follows:

NDP: 12,238
PC: 7,751
Liberal: 15,946
Green: 830
Libertarian: 476

Therefore to do as well in 2013 as they did in 2011, as a proportion of the total ballots cast in 2011, the parties needed to bring in the following numbers of votes (i.e., 68.7% of the ballots each received in 2011):

NDP: 8,408
PC: 5,325
Liberal: 10,955
Green: 570
Libertarian: 327

Therefore, as compared to the ballots each party needed to repeat their 2011 performance, the parties increased or decreased their performance as follows (2013 votes minus votes needed to do as well as they did in 2011):

NDP: 15,693 – 8,408 = +7,285 votes (+86.6%) better than breaking even
PC: 5,149 – 5,325 = -176 votes (-3.3%) worse than breaking even
Liberal: 3,057 – 10,955 = -7,890 votes (-72.0%) worse than breaking even
Green: 934 – 570 = +364 votes (+63.9%) better than breaking even
Libertarian: 398 – 327 = +71 votes (+21.7%) better than breaking even

Conclusions re: Windsor Tecumseh

1. In Windsor-Tecumseh, the NDP (86.6%), the Greens (63.9%), and the Libertarians (21.7%) improved their performance over 2011 (adjusted for 2013 voter turnout). The Liberals (72.0% decrease), and PCs (3.3% decrease) suffered reduced returns. That said, the Green and Libertarian numbers arguably are too small for those percentage-change figures to mean much.

 

2. This was Freedom Party’s first ever appearance on the ballot for that riding, and the Family Coalition Party did not run a candidate in Windsor-Tecumseh, so there can be no change figures for those two parties. Every vote those to parties received in Windsor-Tecumseh was an increase in their respective provincial reach and support.

 

3. Excluding relatively tiny gains by the Greens (364 votes) and Libertarians (71 votes), the NDP was the sole beneficiary of the Liberal collapse in Windsor-Tecumseh. Liberal voters swung over to the NDP.

 

4. As in London West and Scarborough-Guildwood, the PCs appear to have their support largely capped. In London West, they grew only 14.2% and in Scarborough only 7.4%, in the face of huge Liberal losses in each riding. As in those two ridings, this suggests that the PCs have little potential for improving their support in Windsor-Tecumseh. They appear to have maxed-out their support. In a general election, when all parties’ workforces are stretched thin, PC support may well turn out suffer an even great reduction than the 3.3% reduction that they suffered in this by-election.

 

OTTAWA SOUTH

In Ottawa South, 35,247 ballots were cast in the 2013 by-election. This compares to 44,707 votes cast in the 2011 general election. Therefore 78.8% of the number of ballots cast in 2011 were cast in 2013. That’s a 21.2% reduction in total ballots cast.

In 2011, the ballots cast were as follows:

NDP: 5,988
PC: 14,945
Liberal: 21,842
Green: 1,442
Libertarian: 252
Party for People with Special Needs: 103

Therefore to do as well in 2013 as they did in 2011, as a proportion of the total ballots cast in 2011, the parties needed to bring in the following numbers of votes (i.e., 78.8% of the ballots each received in 2011):

NDP: 4,719
PC: 11,777
Liberal: 17,211
Green: 1,136
Libertarian: 199
Party for People with Special Needs: 81

Therefore, as compared to the ballots each party needed to repeat their 2011 performance, the parties increased or decreased their performance as follows (2013 votes minus votes needed to do as well as they did in 2011):

NDP: 5,030 – 4,719 = +311 votes (+6.6%) better than breaking even
PC: 13,631 – 11,777 = +1,854 votes (+15.7%) better than breaking even
Liberal: 14,925 – 17,211 = -2,286 votes (-13.3%) worse than breaking even
Green: 1,105 – 1,136 = -31 votes (-2.7%) worse than breaking even
Libertarian: 213 – 199 = +14 votes (+7.0%) better than breaking even
Party for People with Special Needs: 103 – 81 = +22 votes (+27.2%) better than breaking even

Conclusions re: Ottawa South

1. In Ottawa South, approximately 6/7ths of the Liberals’ lost votes went to the PCs, the remaining 1/7th having gone to the NDP. However, there was only a 13.3% loss of votes in the Liberal count, as compared to 2011, which is relatively minor compared to the bloodletting in ridings like Windsor-Tecumseh and London West. The Liberals suffered their smallest loss in Ottawa South, in terms of percentage of the 2011 vote (adjusted for voter turn-out).

 

2. The gains by the Libertarians and the Party for People with Special Needs arguably say little, again because the total ballots cast for those parties in Ottawa West was relatively small.

 

3. This was Freedom Party’s first ever appearance on the ballot for that riding, so there can be no change figures for Freedom Party. Every vote it received in Ottawa South was an increase in its provincial reach and support.

 

4. As in London West, Scarborough-Guildwood, and Windsor-Tecumseh, the PCs appear to have their support largely capped in Ottawa South. The PCs appear to have little potential for improving their support in the riding. In a general election, when all parties’ workforces are stretched thin, the modest increase in PC support may well be lost.

 

5. Note that the number of votes cast for the Party for People with Special needs was 103 in 2011, and 103 in 2013. This odd result suggests the possibility that an already-existing organization has simply registered itself as a political party, and has made sure that all of its members vote for the party (perhaps by way of Special/mail-in Ballot). Contributions made to registered political parties in Ontario are tax creditable.

 

CONCLUSIONS re: ALL FIVE BY-ELECTIONS

1. The biggest gainer: Freedom Party, which not only broke 5% and thereby exceeded any credible poll’s margin of error, but which also gained more votes (adjusted for voter turn-out) than any other party except the NDP in London West.

 

2. The province-wide big picture: The PCs are in trouble, as is its leader, Tim Hudak. The party appears unable to grow its support anywhere in Ontario. The exception to the rule in this by-election – Etobicoke-Lakeshore – is unquestionably due not to anything that the PCs or Tim Hudak brought to the table, but to the fact that a popular sitting member of the Toronto city council put his weight into the riding, along with Mayor Ford and his brother Doug. Mr. Hudak, clearly, was just along for the ride. The knives will be out, and he will be encouraged to step down prior to the next election, probably in May. That said, he has the protection of his party’s constitution: he cannot be forced to step down.

 

3. The biggest misleading factor in elections: robopolls. Robopolls, like the ones carried out via automated push-button telephone polling by Forum and by Campaign Research, have at best a mischievous effect upon the election. They are not predicting people how voters will vote. In some cases, they are, instead, providing a false impression of how people will vote and, I expect, those false impressions influenced how people actually voted, to some degree. For example, going into these elections, Forum and Campaign Research robopolls had the PCs winning London West, and winning Ottawa South by a landslide. The PCs lost all by-elections save Etobicoke-Lakeshore (the polls correctly suggested it would be a close two-horse race there between the Liberals and the PCs). In fact, in London West, it wasn’t even close: the PCs were gutted. This suggests sampling error, or a poorly constructed set of questions (for example, the Forum polls excluded any mention of Al Gretzky or Freedom Party in London West, yet Gretzky ended up with 5.1% of the vote, beating the Greens, who were listed as a possible response in the polls carried out there.

 

4. The biggest factor affecting election outcome: workforce size. The poor predictive validity of the robopolls also suggests a large disparity between voter intentions, and voter practice, attributable to the varying abilities of each party physically to get out there on polling day, to encourage people to get out of their houses and vote. If that disparity in workforce size does exist (I have no first-hand knowledge yet), this would suggest that the PCs now have a relatively small or weak team for getting out the vote, as compared to the NDP/Liberals (both of whom arguably are able to marshall volunteers from the unions that support those parties). Where the PCs appear particularly popular in a poll, that popularity would be based not upon an intention to go out and vote, but upon residual goodwill that some people retain from prior elections (i.e., people are telling pollsters who they like most or dislike least, not that they will actually go out and vote for who they like/dislike least). Each electoral loss for the PCs will erode that goodwill considerably.