Setting the record straight on preferential voting – again

Evan Solomon’s recent article in Macleans looking at the electoral reform promised by Canada federal Liberal Party contains a number of rather bombastic statements that demonstrates, yet again, the general misunderstanding surrounding preferential voting. For example, Solomon asserts that:

in this system, Liberals could solidify power and still fulfill their democratic reform promise. For a party with no natural allies, like the Conservatives, it could be a fatal blow.

Regular readers must be tiring of my repeated attempts to clarify how preferential voting works in the real world, but please bear with me as we go through the facts one more time (and I doubt it will be the last time).

Let’s start at the very beginning, where Solomon writes:

The current winner-take-all system has few defenders. After all, the Liberals won a little more than 39 per cent of the popular vote, but took 184 seats, or almost 55 per cent of the House of Commons. It was no better under Stephen Harper’s majority.

I get it that a lot of people don’t like our current voting system. However, if you are going to write about it, at least explain it properly. The Liberal Party was not elected on 39% of the popular vote, rather, 184 individual Liberal Party candidates were elected in 184 different constituencies, thus giving the Party a majority of the seats in the House of Commons. That is how our system works. We elect individual MPs, not governments. That is why the national percentage of the popular vote is irrelevant. It is not one, single, national election to elect a government, but rather 338 individual constituency elections to elect 338 individual members to sit in the House of Commons. This is very important distinction. The national percentage of the popular vote doesn’t matter. We don’t elect governments; there is no single box on every ballot nationwide for individuals to tick to indicate their preferred choice for government. There are only individuals, usually representing a party but sometimes not, contesting for a single seat in the House of Commons, and this happens in each and every one of our 338 constituencies.

That is why our voting system is actually called Single Member Plurality (SMP), because that is what we are doing. We are electing a single Member, and that individual needs only a plurality of the vote to be elected. First-Past-the-Post is a misnomer; there is no “post” to be passed. There is no set number of votes or percentage of the vote that a candidate has to achieve — they need only win one more vote than the candidate in second place.

In the UK General Election held in May this year, similar outrage arose over the results of some parties. The UK Independence Party (UKIP), for example, received over 3.8 million votes nationwide, yet only one UKIP candidate was elected. This led to many headlines and graphics claiming that it took 4 million votes to elect one MP, which is patently false. It took 19,642 votes to elect the one UK MP. Actually, since the second-place candidate in the Clacton parliamentary constituency received 16,205 votes, it took only 16,206 votes to elect the UKIP candidate — the rest were excess votes. The fact that almost 4 million other individuals across the country voted for UKIP doesn’t matter — what matters is that in only one constituency did enough voters vote UKIP to secure a victory for that party’s candidate. In every other constituency, more voters preferred different parties.

Now that we’re clear on what our voting system is called (SMP), and how our system works — we are voting to elect one individual, not a government, let’s move on to some of the claims Solomon makes about preferential voting:

In the basic ranked ballot system—also called “alternative vote” or “preferential ballots” (there are multiple variations)—you mark down your first, second and third choice of candidate. If a candidate gets 50 per cent of the vote, they win outright, but if no candidate gets 50 per cent, then the candidate with the lowest vote total is knocked off the list and their vote is split to the others according to each voter’s preference. That happens until one candidate reaches the 50 per cent threshold.

This passage is not entirely wrong, but not very well stated. There isn’t a “basic” preferential voting system. There are essentially two approaches, full preferential and optional preferential. Under full preferential, which is used in Australia federally to elect its House of Representatives and also in some Australian states to elect their lower chambers, the voter must rank every single candidate on the ballot paper for the ballot to count. If there are 10 candidates on your ballot, you have to rank them all in order, from 1 to 10. Under optional preferential, voters can choose to rank as few or as many candidates as they wish. This means that they are also free to use the ballot as an SMP ballot and rank only one candidate. I have never heard or read of anyone proposing full preferential for Canada, and when, in the past, preferential voting was used in BC, Alberta and Manitoba, they all used optional preferential. Consequently, for the rest of this discussion, we will work on the assumption that if preferential voting is adopted in Canada, it would be optional preferential.

With optional preferential, you as a voter, would not have to “mark down your first, second and third choice of candidate” as Mr. Solomon writes; you could mark down only your first choice, as we currently do using SMP, or you could rank two candidates, your 1st and 2nd choices, or you could rank every candidate on the ballot in order of preference, or any variation in between. This is important to remember because the number of people who opt to not use any preferences affects how the system works.

When ballots are counted, only the first preferences on every ballot are counted. If a candidate secures 50% of the valid ballots cast, he or she is declared the winner. This is where preferential voting differs from SMP — with preferential voting, there is an actual “post” that the candidate must pass — and that is the 50% mark. To win, a candidate requires not a simple plurality of the vote, but a majority. If no candidate hits the 50% mark on the first count, Mr. Solomon correctly explains that the candidate with the lowest vote total is dropped, and his or her votes are redistributed among the remaining candidates based on the indicated 2nd choice preferences. This process continues until a candidate secures 50% of the votes still in play. This number will most likely be lower – even significantly lower – than the total number of votes cast because many voters will not indicate any preferences beyond their first, some will indicate perhaps only two choices, etc., therefore the pool of votes still in play will usually be smaller with every single round of counting that is required.

Solomon’s assumption is that such a system would hurt the Conservative Party; in fact, he writes, as I cited above, that it would be a “fatal blow” to the party. He is not the first to put forward this argument. A couple of years ago, for example, Eric Grenier of 308.com published an article in the Globe and Mail that argued much the same thing, that using a preferential ballot “would limit the ability of the Conservatives to win elections”. The underlining assumption many make is that all Liberal, NDP and Green party supporters would necessarily converge their preferences behind those parties, leaving the Conservatives out in the cold.

What both Mr. Solomon and Mr. Grenier (and many, many others) fail to understand is that the preferential ballot, in particular optional preferential, always favours the party which receives the most first preference votes – at least going by Australia’s long history, at both the state and federal level, with this form of voting. Yes, it is true, based on the few such polling sample available, that the Conservatives are the least favourite second choice among Liberal, NDP and Green party supporters, but that doesn’t mean they don’t get any second choices. All of these polling samples do show there are Liberal, NDP and Green supporters who do rank the Conservative Party as their second choice ahead of any other non-Conservative Party. And if a Conservative candidate is ahead on the first count, some second preference support is all that is needed.

Going by Australia’s experience with preferential voting, what matters is how close a candidate is to the 50% after the first count. The closer they are to that mark, the fewer preference transfers they require. Consequently, a Conservative candidate with 45% of the vote on the first ballot count would most likely still win the seat because they need far fewer preferences to boost them over the 50% mark, while a second-place NDP or Liberal candidate who has 40% of the vote would require a far greater number of preference transfers to boost them over the 50% mark. Even if they were further from the 50% target, say at 42%, but the 2nd place candidate was well behind, say at 30%, the Conservative would still most likely win. Only in instances where two candidates were quite literally neck and neck on the first ballot, or in ridings with neck-and-neck three-way races, would the outcome be up in the air and open up the possibility of a candidate other than the one one ahead on first preferences winning.

Solomon writes:

It is a hard system to defend. As Kelly Carmichael of Fair Vote Canada—an advocacy group for proportional representation—told me, more than nine million Canadians in this past election did not get to elect any representative.

“We have to get rid of these false majorities,” Carmichael said, referring to the fact that neither  Harper nor Trudeau got close to a majority of Canadians voting for them. Should a government that gets less than 40 per cent of the vote get to lead as a majority? Shouldn’t the seat totals equate with the popular vote total?

Carmichael believes the ranked ballot is just the same first past-the-post cake, but with more icing sugar, a system that still favours one party over another at the expense of voter choice. “It really doesn’t solve any of the fundamental problems, and likely helps centrist parties such as the Liberals,” she said.

It’s actually a very easy system to defend because it lends itself perfectly to our system of government. It’s only a “false majority” if you misrepresent how SMP and our system of government work. It’s a false majority if you look at the percentage of the national vote each party gets, but that is not how our system works. It’s not “a government” getting 40% of the vote — we aren’t electing a government, we are electing 338 individuals in 338 constituencies. It’s not the national percentage of the popular vote that matters, but the results in each and every individual constituency. One-hundred-and-eighty-four Liberal candidates were elected in their individual constituencies, and under our system of government, which requires that a government command the confidence of the House, that result ensures that a Liberal government could be formed that would command that confidence.

I disagree with Kelly Carmichael’s assessment that preferential voting as “doesn’t solve any of the fundamental problems, and likely helps centrist parties such as the Liberals.” These are two separate issues, and we’ll start with the first, that preferential voting doesn’t solve any fundamental problems.

It is true that preferential voting will not address any of the “problems” identified by those who want some form of proportional representation. However, preferential voting does have the potential to at least minimize one “problem”, and that is the issue of a candidate winning a seat with less than majority support. Because Canada has three competitive major parties, four if you add in the Greens, though they lag significantly behind the other three, a majority of elected candidates win their seats without majority support. In theory, at least, adopting preferential voting would minimize that since it requires that a candidate obtain 50% of the vote, and achieves this through vote preference transfers if they fail to achieve it on the first count.

This ties into what I believe is the point of the second part of Ms. Carmichael’s comment, that preferential voting “helps centrist parties such as the Liberals”. This is a frequent argument made against preferential voting, that in order to attract preferences from other parties’ supporters, parties will either recruit only bland, moderate candidates (grey candidates), or that parties will further move their platforms to the mushy middle. From what I have seen of Australian politics, this simply isn’t the case. Another example to counter the grey candidate myth: London Mayor Boris Johnson. Preferential voting is used to elect the London mayor — no one has ever described Boris Johnson as bland.

Preferential voting would also allow voters to vote for the party they really want to vote for, rather than focus as much on so-called strategic voting. Under SMP, many supporters of smaller parties are reluctant to actually vote for their party of choice, opting instead to support a different party in an attempt to prevent a third party from perhaps winning that seat. With preferences, this sort of strategic voting wouldn’t happen any more — or at least, would happen differently.

In the end, however, the take-away lesson here is that preferential voting doesn’t produce results dramatically different from SMP. It favours the candidate in first place after the first count, and it most certainly would not deal a death blow to the Conservative Party. It does have its own special problems, however, particularly optional preferential, which is most likely what would be adopted in Canada. That is the issue of plumping — using a preferential ballot like an SMP ballot and voting for only one candidate. If a few people do this, it’s not that big a deal. If a lot of people refuse to indicate preferences, you’re left with a de facto SMP system. The Australian states which use optional preferential have seen a huge increase in the percentage of voters who now mark only one preference — almost two-thirds of voters now mark only one candidate. There is no easy answer to this, short of adopting full preferential, but it too can be quite problematic. By forcing people to indicate false preferences, you end up with a much higher rate of informal ballots (ballots which won’t be counted because they were filled in incorrectly — not all candidates were ranked).

Solomon makes a most disingenuous comment when he writes: “Most people never get their vote counted.” That is an absolutely atrocious thing to say. Every single vote is counted. And every single vote counts. It may not produce the result that the person who cast it wanted, but it still counted. In “safe” seats, where one party is dominant election after election, it is true that some might feel casting a vote for any other party to be a rather futile act. However, other constituencies are very competitive, the vote almost equally split between two or three candidates. I know that what is meant by such facile statements is that a significant number of people end up with a representative they did not personally vote for, but guess what? That it is always going to be true for some people regardless of what electoral system is adopted, even proportional systems. Most proportional systems, for example, require that a party attain a certain vote threshold (e.g. 5% of the popular vote) before they can qualify for top-up seats. In the 2015 and 2011 federal election, that would have excluded the Green Party from getting any top-up seats as the party garnered only 3.45% and 3.9% of the vote in those elections. Even smaller parties, and independents, would still be left out. Lower vote thresholds pave the way for truly fringe parties to exert disproportionate influence over minority governments or in coalition governments — it that really more democratic? There isn’t a voting system out there that will make everyone 100% happy and ensure 100% that their vote “counted”.

I understand that SMP is not the most popular electoral system out there, but is it too much to ask that people talk about it correctly? And if you’re going to talk about preferential voting, at least study how the system works in Australia before jumping to all sorts of conclusions.

Related Posts:

Radical Centrist

  • Mark Henschel

    Hi

    Good post.

    Some observations:

    I absolutely agree that the way the language of electoral reform is used and abused it’s a wonder we can have any meaningful discussion. Proportional representation may be the central premise of advocacy but it’s poorly understood and “PR” is… anything they want it to be… mostly some episode or other of MMP. And the terminology for a ballot that can be ranked has pretty much all been co-opted to mean IRV that there is no way to concisely say what you mean, mean what you say or be understood by anybody. We all suffer. What we have in our reform advocacy is failure to communicate.

    Legitimacy is the one important issue that Fair Voters don’t seem to think matters (amongst others)…. whereas it is actually fundamental. During this election cycle — and indeed since Michael Chong started championing MP empowerment with his “Reform Act” — there’s been an growing emphasis on re-balancing power as it extends from citizen/voter through MP to party and government… a power shift. Both the Liberal and Green platforms on democratic reform and the principles of democracy strongly support a reorientation of representational focus: we elect them they represent us… to their parties and to parliament. My response to Chong was that the Reform Act necessitated a “second act” — electoral reform that ensured that MPs who were being freed up to speak and vote on their own recognizances had the sufficient electoral mandate from the voters to do so.

    The thing is, AV is so weak in addressing this problem. 50% is a bare minimum figure and where the point of an election should not be who gets represented and who doesn’t, but rather picking the team who will represent us all… where most if not all of us get a rep, that’s too many voters left without a voice. This wouldn’t matter if, after the election, there was nothing left to say, no issues left to debate, nothing unforeseen on the horizon. But that’s simply not the case.

    Paris terrorism days after the election.

    And it’s called “parl”iament for a reason. Our MPs are there to talk first then make decisions. The thing is, at the talking stage all the stakeholders deserve a voice and a chance to speak. In a democracy the stakeholders are… the voters. We all should have our commensurate seat at the table. It’s the only way to redress inequality now and sustain it in the long run… to make it hard for legislators to make decisions that benefit a tiny elite at the expense of everybody else.

    But the preferential ballot idea is not a bad one. It can be leveraged to perform better… with multi-member districts. STV’s Droop Quota defines increasingly higher percent mandates with every additional MP elected in a riding. And more voters get more and better representation in the House and (incidentally) the parties are represented more fairly too.

    PR is like a canary in a coal mine or the walk sign at an intersection. Whether the walk light is on or the canary is still breathing is only an indication: what’s important is that a pedestrian doesn’t choose to walk in front of a truck that blows through a red light and the miners aren’t asphyxiating.

    Our electoral dysfunction isn’t a matter of party politics so much as a matter of data collecting and communication. What we have here is failure to communicate. Our ballots need to be much better generators of useful data — Justin should follow the re-institution of the long-form census with the implementation of a “long-form ballot” — and this data needs to be faithfully abided in the translation and transmission into representation in the House.

    Oops. Out of time. More later perhaps.

  • JD Mussel

    Well put, although I don’t understand your argument against Grenier’s projection, which is based on polls on people’s second and third preferences. I think those projections clearly show the Conservatives, while remaining the official opposition, would indeed be dealt a blow, and would seriously need to adapt (probably by moving closer to the centre) to have any chances of regaining government in the coming decades, if AV were adopted.