Thursday, December 13, 2012

Your Vote Counts

Does it? Ever since we moved from an older Edmonton neighbourhood to much more conservative suburbia, from the garage in the back to the garage in the front, I haven't felt like my vote really mattered at all. I tend to vote on the left side, or what is now usually called "progressive". In the past few years I have voted for Liberal, NDP, Green and Alberta Party candidates, none of which had a snowball's chance of winning. It gets a little demoralizing lining up at the voting station knowing more or less who will win, especially if it's not who you are voting for. My voting strategy is usually some combination of supporting the candidate most likely to beat the incumbent (ie: polling second to the conservative) and supporting a party's overall vote count in order to provide them some kind of encouragement for the next election.

Canada's first-past-the-post system means that the more parties and candidates there are, the less percentage of the votes is likely to be required to win. Almost half of the last federal election seats were won with a less than half the votes, including a couple where the winner got less than one-third. With 40% of the vote Harper won 54% of the seats and a majority government; the Liberals won 11% of the house of commons with 19% of the total vote. The current system ensures the winners win disproportionately more seats than the popular vote indicates.

Sometimes a majority government, even one with a mandate from a minority of voters, is not such a bad thing. We look on in amusement at gridlocked and often ineffective systems in other countries resulting from multiple parties and some kind of coalition minority government. But the Canadian experience has been mostly positive: minority governments here are by and large effective and balanced, and tend to govern more reasonably under the constant threat of being toppled. A more proportional system should allow for the number of elected officials to resemble the popular vote, which would virtually guarantee minorities based on current polls. A succession of minority governments is not really desirable, as it can lead to voter fatigue if the co-operation isn't there and coalitions can't hold. Plus, I'm not sure how a proportional voting system (like Fair Vote) would even work - would you still vote for a local candidate, or just a national party?

But there seems to be something inherently unfair - even undemocratic - when a small minority can elect a majority. The recent by-elections in November inspired some discussion about how we vote, progressives vs. conservatives, and party mergers. The Calgary Centre race in particular was a great sounding board for all sorts of ideas. The Conservative candidate Joan Crockatt won with 37% of the vote, the first time in five elections where there was not a clear majority winner there. Andrew Coyne proposed a one-election non-competition pact for progressive candidates. Liberal MLA Kent Hehr wrote that the other 63% of people voting for the Liberals, NDP and Greens actually agree on 95% of policy issues, and wouldn't it make sense to join forces. It would, and Hehr deserves some credit for actually saying so in public. But in politics agreeing on 95% means arguing vehemently over the other 5%, which makes any merger option subject to petty bickering and thus a longshot. [Edit: that's hilarious - literally as I was writing this, I was sent a press release from the Liberalberta Party president saying they will not be merging with the NDP and throwing Hehr under the bus. I guess that proves the point. Let's add "infighting" right after petty bickering in the previous sentence.] Of course, the real problem with merging political parties is it reduces the diversity of voices and ideas available with a multitude of platforms; the U.S. two-party state is the logical end point of the merger discussion.

In the Calgary Centre by-election a group called 1 Calgary Centre appeared with what sounded like a realistic goal: unite the progressive vote. As the results show, if progressives had settled on a single candidate, he would have won. Unfortunately, this 1 Calgary Centre never really got much momentum (even though there was a lot of talk of strategic voting prior to the election), and worse, they never made a clear recommendation for voters; instead, by trying to be polite and politically correct, visitors to their website had no clue which of the three progressive candidates might be a consensus choice. Many people see strategic voting as a cop out, or that it is somehow distasteful since you are voting against a person or party instead of for someone. I disagree, but there never seems to be more than a small minority of voters willing to switch their vote to a different party just to defeat another party. And it's always nice to be able to vote your conscience.


Preferential voting could be the alternative. Also known as instant runoff voting, this is where instead of a single mark on the ballot, you indicate your first, second and third (or more) choices depending on the system and how many candidates are running. The Alberta PCs use a preferential ballot to choose their leader, and this system was the reason Ed Stelmach became premier - he was everybody's second choice. Let's walk through the mechanics of how a simple preferential ballot could work, using the Calgary Centre returns as an example:

Party Candidate Votes Percent
Conservative Joan Crockatt 10201 37.2%
Liberal Harvey Locke 9034 33.0%
Green Party Chris Turner 7090 25.9%
NDP Dan Meades 1063 3.9%
Total
27388 100.0%

Just to keep things simple I will ignore the other two candidates whose combined 262 votes were less than one percent of the total. In the current system, Joan Crockatt had more votes than any other candidate, so she wins - end of story. But let's say every voter was allowed to mark a first, second and third choice. A first choice would be required for a valid ballot, but a second and third would be optional. If there were only two candidates you would be willing to support, just mark a 1 and a 2. Or just choose one and only one candidate if that's how you feel.

The last place finisher, Dan Meades, drops off the list and all votes for the NDP are redistributed. This is my guess of how NDP voters might have selected a second choice, after voting first and foremost for their own candidate:

Second Choice
VotesPercent
Conservative 1st choice NDP 53 5%
Liberal 1st choice NDP 213 20%
Green Party 1st choice NDP 425 40%
No Second Choice 1st choice NDP 372 35%


1063 100%

We take Mr. Meades' 1063 votes and add 691 to the respective second choice parties. The other 372 didn't mark any other preference, so these votes are discarded. After adding the second choice to the primary numbers, we get this:

Party Candidate Votes Percent
Conservative Joan Crockatt 10254 38.0%
Liberal Harvey Locke 9247 34.2%
Green Party Chris Turner 7515 27.8%
Subtotal
27016 100.0%
Uncounted
372
Total
27388

As there is still no candidate with over 50% of the votes, we repeat the process of eliminating last place; this time Chris Turner drops off. Again, my guess as to how Green voters might feel about a second choice, although in reality Green Party supporters are fairly unpredictable. Also, 425 of these are originally from the NDP, so those ballots would now look at their third choice:

Second Choice
Votes Percent
Conservative 1st choice Green 1127 15%
Liberal 1st choice Green 3006 40%
NDP 1st choice Green 2631 35%
No Second Choice 1st choice Green 751 10%


7515 100%

The Conservative and Liberal votes are fine, but since there is no longer an NDP candidate on the ballot at this stage, those 2631 ballots need to go to the third choice:

Third Choice
Votes Percent
Conservative 1 Green, 2 NDP 657 25%
Liberal 1 Green, 2 NDP 1579 60%
No Third Choice 1 Green, 2 NDP 395 15%


2631 100%

So the PCs pick up 1784 votes from Green voters' second and third choices, while the Liberalbertans add another 4585. 1146 votes are thrown out here. The final tally:

Party Candidate Votes Percent
Conservative Joan Crockatt 12038 46.5%
Liberal Harvey Locke 13832 53.5%
Total Counted
25870 100.0%
Uncounted
1518
All Votes
27388

We have a winner. A preferential system here would have changed the outcome, given these 2nd and 3rd choice assumptions. Note that the overall progressive vote count was 17187 or 62.8% of the total, but a full runoff with reasonable assumptions leaves the final progressive candidate with 53.5% of counted votes. This preferential result is probably more accurate of the overall voter intentions of Calgary Centre than the first-past-the-post result. What I like is that most of the votes for the third and fourth place parties do not have to be wasted. If you are a card-carrying party member, you might only be willing to vote for your candidate and never pick a second or third choice. But most of us would have little problem prioritizing two or three candidates.

Doing a quick spreadsheet calculation based on the last federal election: if you could combine all the Liberal, NDP and Green votes together, this merged fictional entity would have 54% of the overall vote, and won 53 more ridings than the individual parties did. But doing the same thing for the 2012 Alberta election would have barely moved the needle, with only 4 seats switching from conservative to progressive. Without crunching the numbers, a preferential ballot would likely have not allowed Chretien three successive majorities as a result of the Conservative-Reform split. However, voters can vote differently depending on how their vote gets counted. A real preferential system has a lot of nuance, and would be a good challenge for the Nate Silvers of the world.

I believe a preferential model would be more democratic, not because it may or may not change the outcome of any one race, but because it allows but does not require a greater level of engagement and participation. Only slightly more complicated than marking a single X on a piece of paper, it is simple and straightforward enough to work. It's also not too difficult to tabulate the results. I wonder if it would even increase turnout, especially in traditional strongholds which likely are home to greater numbers of discouraged voters who have simply stopped bothering. Calgary Centre was accurately forecast to be a close election, and still over 70% of eligible voters stayed home.

In Calgary Centre it's fair to say that most people who did come out to vote were not conservatives. The question is do people see one right-of-centre party and three somewhat similar left parties, or are there in fact four very different choices with little overlap? We probably won't know for sure until we try it, and electoral reform is hard because the status quo system is how the current government won in the first place. But things can and do change occasionally, and this is one issue that deserves some serious consideration. It's not about changing outcomes necessarily, but rather making every vote count.