Tuesday, May 6, 2014

2014 Ontario election projection - Week 1

Week 1 projection map
Last Friday, Ontario NDP leader Andrea Horwath announced that she would no longer be supporting the governing Ontario Liberal Party's budget. This triggered Premier Kathleen Wynne to call an election for the province which will be held on June 13. Officially, the writs will drop on Wednesday, and the campaign will officially begin.

The Liberals have governed the province since 2003, and have been led by Premier Wynne since February 2013 when provincial Liberals elected her leader, replacing the previous leader Premier Dalton McGuinty, who had resigned. Since the last election held in 2011, the Liberals have faced many scandals, such as the “gas plant scandal”. The scandals, as well as a number of corruption controversies threaten their 11-year reign. In the last election held, voters returned a Liberal minority government, with the Liberals winning just one seat shy of a majority. However, subsequent by-election losses have eaten into this minority. Since 2011, the Liberals have counted on the support of the left wing NDP to help keep the government going. On Friday though, the NDP had enough of the scandals and alleged corruption and withdrew their support. The opposition Tories, led by Tim Hudak, have not supported the Liberals since the election. In recent polls, the Progressive Conservatives have traded with the Liberals for the lead among Ontario voters. Up to about a year ago, they consistently led in the polls. Since the last election, the NDP have also polled quite well, compared to their historic third place showings. They even led in one poll conducted in January 2013. Just in the last two weeks however, the NDP have fallen back to their 2011 election numbers in the low 20s.

The beginning of a provincial election campaign means that once again I will attempt to do weekly projections of what the polls are showing, and then map the results to predict what might happen on election day. For my first projection, I will be using two recent polls conducted by EKOS and Forum Research. The EKOS poll released on May 2, showed the Liberals in the lead with 34.7%, the Tories in second at 31.6%, the NDP at 22.2% and 9.4% for the Greens. Forum Research released a poll the next day showing very different numbers. They have the Tories at 38%, the Liberals at 33%, the NDP at 22% and the Greens at 6%. What this means is that the Liberals and Tories are in a statistical tie at the moment, and the parties are close to where they were at in the last election.

Pumping the regional numbers from both polls into my new Ontario election model, I get a very narrow PC lead, with the Tories at 46 seats, the Liberals at 43 and the NDP at 18. This would result in a minority Progressive Conservative government in the 107 seat legislature.


The basis of my model extrapolates the results of the 2011 election using regional polling averages from recent polls. If a party is polling at double the level they won in 2011, then that party would see its support double in each riding in the region my model. I have also taken into account recent by-elections, but projecting the results of the by-elections backward to the last election based on poll numbers from around the time of each by-election. That is, if a party did 5 points better than polls indicated they would've won in a by-election, then I have made their 2011 result 5 points more than they actually received in 2011. (I've only done this for the NDP, the PCs and the Liberals, except for Thornhill where I only did it for the PCs and Liberals).

In the coming weeks, I will be doing further tweaks to my model based on a number of factors, such as candidate strengths. However, candidates are still being nominated, so they have been ignored for the time being. You can say that this model reflects polling data as if all candidates were the same as in 2011 (or recent by-elections).

Odd results

I should also mention some strange results in my model that I hope will be corrected before my final projection on election day. Ontario, being my hope province, is the province that I know the most about in terms of its political geography. I like to think that I have a good knowledge of how ridings will vote, even without a model helping me. In 2011, without the aid of a true mathematical model, I correctly predicted the winner of 92% of the ridings. (threehundredeight.com which did use a mathematical model only correctly projected 85% of the ridings). This knowledge will allow me to identify any flaws in my model so that I can come up with “tweaks” to correct them. In this first map, the three strangest results in my opinion are:

* Barrie. My model has the Liberals ahead here thanks to the EKOS poll showing the Liberals ahead in the 705 area code. However, I think the Tories will easily win the seat. The seat is currently held by the Tories, and in the 2011 federal election, the Liberals won just 16% of the vote. In the last election, the race was close, so any strength the Liberals have in the 705 is going to give Barrie to the Liberals in my model. The Liberals haven't nominated a candidate yet, so they don't seem to be prepared to win this one.

* Scarborough—Guildwood. My model has the Tories ahead, thanks to a close by-election result from last August. However, I believe the Liberals will win it. The candidates for both the PCs and the Liberals will be the same as in the by-election, but the wild card is the NDP. In the by-election, the NDP ran city councillor Adam Giambrone who was popular enough to eat into Liberal support in the riding. The NDP have yet to nominate a candidate, and are unlikely to nominate anyone of Giambrone's strength, so Liberal support is likely to go back up to normal levels. This riding has a Liberal history, and is one of the few ridings the federal Liberals were able to win in the 2011 election. I don't see the Tories winning any seats that their federal counterparts didn't win in the last federal election.

* Kenora-Rainy River. Small sample sizes in the 807 area code from the EKOS poll, and strong Tory strength in Forum Research's “Northern Ontario” region (which likely takes in all of the 705 area code which includes large swaths of Tory friendly Central Ontario) are to blame here. Currently, my model shows the PCs ahead in this riding, but its history clearly indicates the riding will likely be won by the NDP. When the NDP won the riding in 2011, it came as somewhat of a surprise to me, as I predicted the Tories would win. After all, the riding overlaps with the federal riding of Kenora which the NDP have failed to take back from the Conservatives in recent elections. However, the NDP ran a strong candidate in Sarah Campbell, who had worked for the riding's previous MPP, former Ontario NDP leader Howard Hampton. Hampton had represented the area since 1987. I suppose the Tories could win it, but it will be a bit of a long shot.

Gains and losses

Compared to the 2011 election, my model shows the Tories picking up 11 seats: Etobicoke-Lakeshore, Scarborough-Guildwood, York Centre, Brant, Kitchener Centre, Glengarry-Prescott-Russell, Ottawa-Orleans, Ottawa South, Ottawa West-Nepean and Niagara Falls (from the Liberals) and Kenora-Rainy River (from the NDP). It has the Liberals picking up two seats: Trinity-Spadina (from the NDP) and Barrie (from the Tories), and it has the NDP picking up three seats: Kitchener-Waterloo (from the Tories) and London West and Windsor-Tecumseh (from the Liberals). The Tories would have a net gain of nine seats, the NDP would have a net gain of one seat while the Liberals would have a net loss of 10 seats. Some of these changes aren't really gains or losses due to recent by-elections (like in Etobicoke-Lakeshore, Kitchener-Waterloo, Windsor-Tecumseh and London West, while in Niagara Falls, the Tory gain is actually from the NDP who won the riding in a recent by-election)

As always, I will be taking in suggestions from readers about ways I can improve my model. So please share with me your opinions. I will also be monitoring electionprediction.org to see what submitters to that site think about how each riding will go.

Here are my projected numbers for each seat. Ridings are coloured by how they voted in 2011.


  1. "This would result in a minority Progressive Conservative government in the 107 seat legislature"?? Come now, you know better than that.

  2. You're right, I do know better than that. Perhaps I'll be more careful with my wording next time.