In the absence of hard data, today’s Parkdale-High Park by-election is really too close to call. To many, that statement alone is a victory for the NDP. For a seat that the Liberals won 57%-to-16% in 2003, this seat should not be as close as it is. Everyone knew that it would be much closer this time, since wildly popular MPP Gerard Kennedy was replaced with controversial city councilor Sylvia Watson. But for the NDP to overcome a 40-point gap would represent a colossal defeat for the governing Liberals (the NDP actually finished third in 2003, just behind the PCs).

So what should the results result look like? Let’s walk through it. Let’s start with the 2003 results: LIB 57%, PC 16%, NDP 16%, GRN 11%. We can estimate that the loss of Gerard Kennedy itself will likely cause a 10-point drop in the Liberal vote, virtually all of which would benefit the NDP given the riding’s history, bringing the totals to: LIB 47%, NDP 26%, PC 16%, GRN 11%. Now if Parkdale-High Park followed the shifts in support we have witnessed in the province since 2003, that would cause another 8-point drop for the Liberals, with the PCs picking up 1, the NDP 5 and the Greens 2 (note, however, this province-wide data is about 2 months old, but assuming things haven’t changed much over the summer, it’s probably close what it would be today). From past elections, however, we know that the Liberals in this riding tend to drop a little more (5%) than the province as a whole, while the NDP gains considerably more in this riding (20%) than the province as a whole. And we know from polling that where the NDP gains more, usually the Greens gain a bit less. So let’s assume the Liberals drop 8-points, the PCs gain 1, the NDP gain 6, and the Greens gain 1, bringing the totals to: LIB 39%, NDP 32%, PC 17%, GRN 12%. But the perception that the NDP could win the seat will probably shift 15-20% of soft Green supporters towards the NDP, bringing the totals to: LIB 39%, NDP 34%, PC 17%, GRN 10%.

So, even when accounting for the loss of Kennedy, the shifts in support provincially, and the stronger NDP presence locally, by all rights, the Liberals should edge out the NDP by about 5%. However, the big unknown is the impact of the Liberals negative campaign against NDP candidate and United Church Minister Cheri DiNovo (the Liberals called into question her values because of questionable comments her has made during her unorthodox sermons in the past). Negative campaigning usually works, and it is likely that some people will shy away from DiNovo because of her statements. So, let’s assume that 2% shift their support because of DiNovo’s statements. More than likely, however, the PCs will benefit from this shift as much as the Liberals. So let’s assume they split that shift, bringing the totals to: LIB 40%, NDP 32%, PC 18%, GRN 10%. On the other hand, there are many people who are likely to shift their support away from the Liberals because they perceive the negative campaigning as an unjustified character assassination (rightly or wrongly). In this case, I would estimate that more people will shift their vote away from the Liberals than move towards the Liberals due to the backlash against the negative campaign. Let’s assume that shift is 4% and the NDP gains most, but not all, of that shift. This would put the final tally at: LIB 36%, NDP 35%, PC 19%, GRN 10%.

As you can see, the NDP and Liberals are too close to call. A 1-2% shift either way for any of the above reasons and the outcome will be different. What we can say is this: the Liberals have the advantage, but if there is a larger backlash against the Liberals due to their negative campaigning, then the NDP could very well edge out the Liberals. More than anything, it is a referendum on whether the benefits of negative campaigning outweigh the backlash it generates. We shall see when the results come in…