Oh, Split!

So, did the Greens really split the vote in last night’s BC election? That’s the claim being made by many today. There are a couple of ways to consider this question. The simplest is to examine the margin of victory in the election, and compare it to the Greens’ result. The GP captured just over 130,000 votes, or 8% of the total; in comparison, the margin of victory was about 80,000 votes, or 4.9%. If we simply take the GP total and add it to the NDP, the Dippers come out on top.

Click to make bigger.

We can get a little fancier and repeat the exercise on a district-by-district basis. There were 61 districts with Green candidates running, and in 23 of them, their vote share exceeded the margin of victory. 12 were won by Liberals, 10 by NDP, and 1 by the GP itself. (Congratulations Dr. Weaver!)  If we repeat the process from above and add the Green votes to the NDP totals, the revised result has the NDP gaining 13 seats, narrowly winning the election with a total of 46 seats to 38, with one independent. Again, vote splitting seems to be at work, right?

Not so fast.

Both the above approaches assume that, absent a Green Party, all Green voters would simply move en masse to vote for the NDP. This seems unlikely, however. Green parties at both provincial and federal levels regularly tout their ability to attract support from across the political spectrum, and to activate voters who have no strong ties to any of the other parties.[1] Accordingly, while some Green supporters probably would vote NDP, others would likely vote for the Liberals, and some might stay home.

Absent good polling data on BC Green voters’ second best options (and frankly, we don’t seem to have good polling data on anything right now), we can’t know for sure what percentage of Green supporters would do what in an alternate Green-less reality. What we can do, however, is calculate how many would have had to switch in order to make a difference.

For instance, let’s assume for a moment that the Green Party disappeared on the eve of yesterday’s election, and that all Green voters in fact either voted NDP, or else stayed home (or ate their ballots, or voted for the Work Less Party, or basically did anything but vote for the Liberals). This amounts to the easiest case for the vote splitting argument. In that case, what percentage of former Greens would have had to switch over to the NDP to flip the result? By my calculations, it’s right around 62%.[2][3] That is, three in every five Green Party partisans would have had to vote NDP, with none voting Liberal, to get the NDP to a 43 seat majority.

Conversely, we can assume that every single Green voter who doesn’t switch to NDP bolts instead to the Liberals. Under those conditions, obviously, a higher percentage of Green voters would have to vote NDP to make a difference. How high? My results indicate that more than 80% would have had to vote Orange to get the NDP to a majority.

The truth probably lies somewhere in between those two numbers, but at least we have a range in mind: somewhere between 60 and 80% of Green voters would have had to defect in order to tip the balance. While possible, it seems unlikely. Figure 1 illustrates.

As a final exercise, it’s worth thinking about the above in the context of vote splitting on the right. Had the Liberals and the NDP been the only two choices, with Conservatives siding with the former, and Greens voting with the latter. In that case, the Right would have won last night, 45 seats to 39.

So what’s the bottom line? While we can’t say anything definitive, we have clarified what we’re assuming when we talk about vote splitting. We are making assumptions—potentially strong ones, at that—about why voters vote, and how they view the options available to them. It may be the case that voting has tipped elections in the past (ahem), but it’s hard to argue that that’s what happened in British Columbia last night.


[1] Indeed, I’ve written elsewhere that this tendency is tantamount to a fatal flaw for Green Parties in a majoritarian system like ours, where voters depend upon a stable political “identity” in parties to guide their choices. To the extent that voters remain unclear what the Green Party’s governance “philosophy” is, they are unclear on what the party would do if actually elected to power. Would they raise taxes? Fund health care sufficiently? Devolve power to the provinces? Absent a reliable way to impute answers to such questions, most voters remain unlikely to trust Green Parties sufficiently to grant them even a share of power. That’s an argument for a different day, however.

[2] Basically, I’m assuming a constant rate of defection from Greens to NDP, with no counter-defection to the Liberals. This approach ignores regional variations, in that there are probably some areas of the province much more likely to break for NDP absent a Green option. In my defence, modeling that would be really hard, and this is just a blog post.

[3] Being satisfied with a minority would not have helped much at all. In fact, under these assumptions, there’s only one possible scenario in which the NDP could have achieved a minority: by winning exactly 42 seats. If they had done that, the Liberals would also have been reduced to 42. In that case, the tie-breaking Independent, Vicki Huntington, would have wielded the real power. As it happens, this would have been exceedingly unlikely outcome under this model, given that both Burnaby North and North Vancouver-Lonsdale flip between 61% and 62%.)

photo credit: pfos via photopin cc

Advertisements

2 thoughts on “Oh, Split!

  1. According to our Vote Compass data (take it for what it’s worth), 77% of people leaning Green had the NDP as a second choice. I’m goign to think about all this a bit more and come back…

  2. Interesting. I suppose we would need data on how strong that second preference was, or some other way to measure what proportion of Green leaners would follow through on that second choice.

    Out of curiosity, does Vote Compass publish data on if/how the sample is weighted to achieve representativeness?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s