Thank you for inviting me to appear before this committee.
The way we translate citizen preferences into votes, and then those votes into seats, is fundamental to how democracy works in Canada. A serious discussion about the way we have been and could be doing this is long overdue. I can't think of a better way to spend an evening in August than to sit and talk about electoral systems.
As the chair pointed out, my research lies in two areas that might be of interest to the committee. First, in what seems like the increasingly distant past, I wrote my doctoral thesis on the use of the alternative vote and single transferable vote systems in Alberta and Manitoba between 1920 and 1955. I've read some of the testimony you've heard, so you've heard this already. These provinces used STV in the big cities and the alternative vote in the rural areas. These cases, along with British Columbia in 1952 and 1953, which used AV, represent the only uses of electoral systems other than the first past the post electoral system in provincial and federal elections. The use of STV, I think, is particularly interesting because it represents the only use of proportional electoral system in a context with political parties in Canada. I think we can, should, and do learn a lot from things around the world, but these examples provide some important domestic cases to help us understand how these systems might work.
I'm sure you're all going to be running out to read my doctoral thesis when this is done, but to save you the joys of reading 300 pages of turgid graduate student prose, I'll give you my major conclusions.
The second area where I do research that might be of interest to the committee is around Canadians' online political activity. My colleagues at four other Canadian universities and I have been surveying Canadians for the last two and a half years about their online political activities. In our surveys, we've asked a few questions about Canadians' attitudes toward online voting. This research was funded by the Social Sciences and Humanities Research Council of Canada, and part of what they asked us to do was to try to come up with knowledge that would be useful and inform public debates. We've only begun to start analyzing some of that data, so I'm going to give you some preliminary findings quite briefly at the end.
My research into the uses of preferential balloting in Canadian provincial elections has led me to conclude that the alternative vote is probably not the best option for Canada. The historical experience with AV suggests that it results in election outcomes that differ little from those we would encounter under the first past the post system, and it wouldn't do anything to address the most serious shortcoming of first past the post: the failure to produce a legislature that accurately reflects the preferences of Canadians. In Alberta and Manitoba, the system had no impact on proportionality, which is how political scientists measure the correspondence between seats and votes. It had no impact whatsoever.
Even if we look at the district level, the alternative vote produces results that differ little from first past the post. In the entire experience of Alberta, with over 30 years of using the alternative vote, fewer than 3% of all the seats contested would have turned out differently under AV than first past the post. In Manitoba, the number is less than 2%. In other words, in 97% to 98% of the cases, the person ahead on the first count ends up winning, and that person would have won under the first past the post system.
A big reason for that is that over half of Manitoba's voters, and nearly half of Alberta's voters, in the AV districts only bothered to indicate a first preference. Even though they can rank all the candidates, what we find is that many voters—and in Manitoba's case most voters—only indicate a first preference. They don't even bother with a second preference, let alone a third or subsequent preferences.
When we imagine how the alternative vote might work in Canada, we often cast our eyes toward Australia, but one of the things we often overlook in the Australian case, especially nationally, is that in Australia voters are legally compelled to rank each and every candidate. That's a big difference. In Canada, we didn't do that. There are, I would argue, some moral issues around that. Would we want to compel people to perhaps contribute to the election of a candidate or party that they fundamentally oppose? It might be their third choice out of four candidates. In Canada, we didn't do that, and when we leave voters to their own devices, many voters don't seem to want to indicate more than one preference. We tend to imagine a greater use of preferences than might be the case.
Another finding was that the alternative vote resulted in a significant increase in rejected ballots in all three provinces where it was used. In some of the cases, this was the result of various stringent ballot marking rules. In the case of Manitoba, they didn't have those stringent ballot marking rules, and we saw a tripling in the rate of rejected ballots. We might think it doesn't ask a lot of voters to write one, two, and three rather than an X, but there is some circumstantial evidence—and we don't have access to the ballots to know exactly what the problems were—that there were problems for voters, at least at that time.
By contrast, I'd argue the STV system used in Edmonton, Calgary, and Winnipeg performed much better. It much more accurately reflected the wishes of voters in legislative representation.
We tend to focus on the fact that a move to a proportional electoral system would likely mean the end of majority governments, and I know this has been pointed out to the committee several times by now. But one of the things we often overlook is that such a system can ensure that we end up with an opposition that's sizable enough to do the job. I come from Alberta where provincially that's been a problem for much of its history. We don't tend to have large oppositions. The use of STV in Edmonton and Calgary during the Social Credit years actually was instrumental in ensuring that there was some opposition in the legislature. While Social Credit swept nearly all the rural seats, after they got rid of STV and AV after the 1955 election, Social Credit swept nearly all the seats in Edmonton and Calgary.
One of the purported benefits of STV is that voters can jump between candidates of different parties. But to what extent did they actually do this? Theoretically, you're right: they could pick, and I'll use the current federal context. They really like the Liberal candidate, then a Conservative, then a New Democrat, then a Green, then a Liberal, then a Conservative. They could go in whatever order they want to. But what I found, when we looked at how voters actually used it in Alberta and Manitoba, is that they tended to vote along party lines. If there was another candidate from that party available, 60% to 90% of transferred ballots stayed within the party. Once the last candidate from a party was eliminated, 35% to 40% of the votes were just by people who wouldn't indicate any more preferences after that. This suggests to me that pure voters take advantage of that freedom to jump between parties more than we might anticipate. This is one of the arguments made to support STV over other PR alternatives.
There are some costs with STV, which is that the ballots tend to be longer and more complex. We may not actually see as many voters taking advantage of the benefits as we might expect.
I'll just quickly say there was no impact on voter turnout in either system. Often this is cast as a potential benefit. I could find no evidence that it made any difference. There was also concern expressed that proportional representation could lead to a proliferation of parties, and I also found no evidence of that.
To me, the big take-away, looking at these cases, is to echo something that André Blais told you last month, that claims about other impacts of electoral systems on things like party formation and voter turnout should be taken with a bit of caution, as those kinds of things tend to have multiple causes. The electoral system can play a role, but it's not the only role.
The one thing that the electoral system does is alter the math of the translation of votes into seats. PR systems like STV or MMP or list PR do that much more accurately than first past the post, or the alternative vote. I'd argue that this is really the fundamental basis on which a decision about electoral reform should be made.
Quickly, when it comes to online voting, we find some interesting things. Our survey of the research shows that a significant proportion of Canadians have concerns about its safety, but they still are very interested in doing it. In our survey, 36% of Canadians thought that Internet voting was risky; 42% thought it was safe; and the rest weren't sure one way or the other. What I found very interesting, though, is that a third of the people who said it was risky still said they were very likely to vote online if it were available to them. I think it's because the risk doesn't accrue to them personally; it's a risk to the system, not to them personally.
That said, those who think it's safe are considerably more likely to vote online if they can. I think that if there were to be any move towards that, there would need to be some work done to reassure voters of the safety and security of it.
If it were an option, 55% of our respondents said they were very likely to vote online; 22% said they were somewhat likely. That's pretty staggering.
One caution: our survey was done through an online panel.