Canada: consider score or approval voting

By Ted Hsu & Warren D. Smith. This was published by on 8 January 2016, but unfortunately the editor, without consulting us, changed our title to "Yes, we can have a referendum on electoral reform – but not just any referendum."

As Prime Minister Justin Trudeau's government pursues reform of Canada’s first-past-the-post system for electing federal governments, many observers have been insisting that any changes be approved by a general referendum before going forward.

But any such referendum on electoral reform should, at the very least, be free of the sorts of problems and flaws that are driving the desire for electoral reform in the first place.

How should such a referendum be set up? We’ll get to that in a minute. First, let’s ask a few basic questions about the need for electoral reform that avoid the trap of subjective language, such as what is “fair”.

What reasons do people give for wanting a new voting system? Many Canadians find themselves in a familiar bind in a general election: They want to support Party A, but feel that they have to vote for Party B instead because they’re convinced A can't win in their riding and they really want to stop Party C from winning. Others may want A's candidate as their local representative, but prefer to see Party B take power. Still others might loathe Candidate X but struggle to express through a single vote their view that Candidates Y and Z are equally sound.

A value-neutral way to describe the desire for a new voting system is to say that Canadians believe the current system doesn’t explicitly express their wishes – that they prefer a system which chooses a government based on more and better information about what voters want.

Many commentators have suggested that each party, based on its political interests, prefers a particular system. Liberals, we’re told, like the "preferential ballot" option (which we'll call IRV for "instant runoff voting") because many voters on both the left and right would pick the Liberals as their second choice. Many assume the Conservatives would like to stick with first-past-the-post (FPP) because their faithful base – with modest additional support and a few advantageous vote-splits among their rivals – can win them a plurality in a majority of ridings. Finally, the NDP and Greens tend to favour "proportional representation" (PR), which is assumed to offer them more seats.

These assumptions may or may not be true, but there are reasons to at least expect disagreements along party lines – and even within the parties themselves. In a Commons vote on December 3, 2014 the Liberal caucus was split – 16 in favour and 15 against – on a vote on "a form (unspecified) of mixed-member proportional representation." And there are many more options to consider. So Parliament probably will have a hard time reaching a multi-party consensus.

And what happens if Parliament can't reach consensus? Many are calling for a referendum to decide the issue. But what if, as seems likely, there is multi-party consensus only for a ballot question that offers voters a choice between three or more options? How could we design a referendum that would respect the idea that Canadians want to give more and truer information about their political preferences?

Michigan Congressman John Dingell once said, "If you let me write the procedure and I let you write the substance, I'll (beat) you every time." A referendum could, by design or accident, unduly favour one outcome or another. So let’s think about how to design a voting system for the referendum itself that would allow citizens to better articulate their opinions.

Suppose, for argument's sake, there are three proposals to decide between: IRV, PR and FPP.

Consider how FPP supporters might word the referendum. If they thought that their opponents were badly split between PR and IRV, they might frame the question like so – "Do you support changing to IRV?" – on the assumption that supporters of the status quo and and fans of PR would combine to defeat IRV. Or they might hold a three-way "first-past-the-post" referendum between PR, IRV and the status quo – again, to take advantage of vote-splitting.

Supporters of PR – if they thought that they outnumbered IRV-supporters, and that a majority of voters wanted some kind of change – might opt to split the referendum into two questions: "Should we change the voting system?" and "Assuming it changes, should it change to PR or IRV?"

Supporters of IRV, meanwhile, might set up the referendum as an IRV ballot, asking voters to rank the three options – if they thought that the second choice of both PR and FPP supporters would be IRV, or if they thought that the format of the referendum itself would help get voters comfortable with IRV.

You see the problem. The format of the referendum itself could unduly influence the process of finding out what Canadians want. That shouldn’t happen.

If we decide that a referendum is necessary, it should be structured through "score voting" – or through its far simpler cousin, "approval voting." Parliament would provide, say, three to five options for voting systems. For each option, voters would give a numbered score indicating their level of support. Under score voting, voters usually pick a numbered score from a range of values (one to nine is typical). Approval voting is binary – a straight yea/nay vote on each option, although voters are allowed to approve of more than one option.

Approval is the simplest and most studied form of score voting; its disadvantage is that it sacrifices some precision. But both forms of score voting have significant advantages:

In fact, although we’re only talking here about how to structure a referendum on voting reform, score/approval voting could also be a good approach to elections themselves. There are ways to hybridize it with PR, if necessary. Instead of voters being asked to pick just one option, voters would be asked for their opinions on each option.

Which gets to the heart of the voting reform debate – how to make the system better reflect what Canadians think and want. If the goal of an election is to legitimize the choice of government, what better way to improve legitimacy than to ask voters for more information?

Ted Hsu was the Liberal MP for Kingston and the Islands from 2011 to 2015; he did not run in the 2015 election. Warren D. Smith is the former president of The Center for Election Science and has a PhD in mathematics.

The views, opinions and positions expressed by all iPolitics columnists and contributors are the author's alone. They do not inherently or expressly reflect the views, opinions and/or positions of iPolitics.

Return to main page