Projection: McGuinty to win third term in Ontario

Two months ago, I was looking at opinion polls thinking that this election was Tim Hudak's to lose.  Now, three days away from election day, it looks like he has done just that.


Applying a regional swing model on a weighted average of the various opinion polls that came out over the past few days, I came up with the following seat projection (change from 2007 result in brackets):

LIB: 55 (-16)
PC: 33 (+7)
NDP: 19 (+9)
Required for majority: 54

Taking the margin of error into account and assuming that the polls perform better than they did in the federal election earlier this year, the Liberals will certainly be the largest party in the legislature, and the odds of them being able to form a majority government are about 50:50.  This is a stark contrast from two months ago, where the Progressive Conservatives were sitting in the 65-seat range.

With the usual caveat that this is no better than a guide, you can see seat-by-seat projections here.

But the polls have them virtually tied!

Those following opinion polls may be wondering how the Liberals could win a majority government while being statistically tied in the popular vote with the PCs (indeed, the central forecast has the Liberals ahead by only 0.5%).  Before we do that, we need to understand the concept of "swing", which is the degree by which voters change their votes from one party to another.  For example, 5% of the population voted Liberal last time and plan to vote PC this time, we call that a 5% swing from LIB to PC.  If we then assume that the New Democrats remained unchanged, we'd also call that a 2.5% swing from LIB to NDP, because that's the effect that such a swing would have on LIB-NDP marginals.

Now back to how the Liberals could win a majority.  In the 2007 election, the Liberals finished about 10 percentage points ahead of the Conservatives in the popular vote, which means a 5% swing would be required from LIB to PC in order to tie.  However, as you see on the below "Swingometer" (a visualization popularized by the BBC), the number of seats the Liberals won with a 10%-or-less margin (and hence would be won on a 5% swing) works out to the magnificent number of 6.  However, just beyond this point, there's a large number of seats that the PCs would be able to pick up, the majority of which are in the Greater Toronto Area.

In most regions of the province, the swing is between 5 and 6 percent.  The two notable exceptions are in the City of Toronto, where the swing is only around 3%, and in northern Ontario, where it's above 10%.  Northern Ontario is virtually useless for the PCs in terms of seat potential; besides Nipissing, the closest Liberal-PC marginal there was won by nearly a 30% margin.  If we were to even the swing between northern Ontario and the GTA, the PCs could pick up another six seats or so.  This may not have been enough to put them ahead of the Liberals, but it would have allowed us to rule out a Liberal majority, which would be Tim Hudak's worst nightmare considering where he stood two months ago.

So what happened?

Unfortunately, having not been in the province over the course of the campaign, it's hard for me to say for sure.  The downhill slide for the Conservatives started around the time when Hudak accused McGuinty of offering tax credits for hiring foreign workers (the tax credit is actually for hiring new citizens).  I also understand that various interest groups have been airing anti-PC ads en masse across the province, as it is not regulated at the provincial level as much as at the federal level, which along with an otherwise-eventless campaign can only do damage.

Were you in Ontario throughout the campaign?  Did you notice a major blunder that could have cost the Progressive Conservatives a majority government?  Comment below!