Preferential Balloting - A More Democratic System?

by Richard Hennick

The present winner-take-all system for the election of candidates has several drawbacks. It encourages the marketing of image and charisma, and leads to polarisation on issues (if issues are dealt with at all), to domination of elections by major parties with large amounts of money, and to a lack of media and public attention to "minor" candidates. It takes no account of either the widespread public recognition of the need for compromise and the "middle way", or the order of preference which voters may feel, especially for those minor candidates.

Preferential balloting can provide a way to overcome or at least ameliorate these and other problems. The results that are obtained with PB are, I believe, closer to representing a consensus of the opinions of voters. In PB, instead of marking the ballot with a cross for the candidate(s) of choice, the voter must rank the candidates in order of preference. After the polls have closed, the resulting "scores" for each candidate are tallied, and various manipulations can then be carried out.

For purposes of illustration, the following is an idealised and simplified version of PB. I don't think it is an exact description of any system currently in use (actually, Australia's is the only system that I have learned anything at all about).

If the voters' ranking is done in the common fashion, i.e. a 1 for first choice, 2 for second choice etc., then the candidate with the lowest overall "score" is the winner.

The winner will not necessarily be the candidate with the largest number of first-choice votes (except in an election with only two candidates, in which case the results would be the same regardless of the balloting system used). Instead, it turns out that preferential balloting produces as a winner the candidate who is the "least unacceptable" to the greatest number of people, and who may have been almost nobody's first choice! This is because highly controversial, polarised candidates tend to eliminate each other in PB - their first- choice votes are in effect cancelled out by the large number of last- choice votes that they also collect.

A couple of examples will probably be helpful at this point.

First, let's look at a three-candidate election in a very small riding; it only has 11 voters! Candidates A and B represent major parties; candidate C represents a minor party and would normally be considered an "also-ran".

Under the WTA system, the results look like this:

Voter >  1   2   3   4   5   6   7   8   9   10  11  total

Cand. A X X X X X 5
Cand. B X X X X 4
Cand. C X X 2

Sure enough, candidate C got left way behind, and candidate A is elected with 45.5% of the vote.

Now let's hold the election again, using PB.

Voter >  1   2   3   4   5   6   7   8   9   10  11  total

Cand. A 1 1 1 1 1 3 3 3 3 2 3 22
Cand. B 3 3 3 3 3 1 1 1 1 3 2 24
Cand. C 2 2 2 2 2 2 2 2 2 1 1 20

The polarisation between A and B has done them no good at all. All of A's and B's supporters would prefer that candidate C get elected rather than their respective sworn ideological enemy, and they get their wish - candidate C has won! Note that the two voters whose first choice was for C had split their preferences for second and third choice between the other two, but this was not actually necessary for C's victory.

Let's look at a more complicated example. Same riding, same first three candidates, but now they are joined by two others, with D representing a "lunatic fringe" party and E representing an interesting genuine alternative to all of the others (I'll leave the names of the parties to your imagination). D and E are both outsiders in conventional WTA elections, and ignored more or less equally by the media pundits.

Voter >  1   2   3   4   5   6   7   8   9   10  11  total

Cand. A 1 1 1 1 5 5 5 4 2 5 4 34
Cand. B 5 5 5 5 1 1 1 3 4 2 3 35
Cand. C 3 4 2 3 3 2 4 1 1 4 2 29
Cand. D 2 3 4 4 4 4 2 5 5 1 5 39
Cand. E 4 2 3 2 2 3 3 2 3 3 1 28

Running this one under WTA, the results would be A - 36.4%, B - 27.3%, C - 18.2%, D and E tied for last place, 9.1% each (obviously there are limits to what you can demonstrate with only 11 voters). The preferential balloting tells a very different story. Again, the hard- line polarisation between A and B has spoiled both their chances. Now E is first, C second, followed by A, B, and D (I said that D was lunatic fringe). Obviously most voters saw E as representing some kind of credible alternative - notice that both E and C have no last- choice votes at all. Notice also that E can no longer be regarded as an outsider on an equal footing with D, undeserving of attention from the media or the voters!

Preferential balloting can also be used when there is more than one candidate to be elected, such as municipally, where its effect would be to drastically reduce the impact of "slate" voting. In addition, there do not have to be people's names and parties on the ballot; PB might also provide a way to get around some of the drawbacks to referenda, by allowing the introduction of "multiple-choice" questions and the concept of compromise, and greatly reducing the potential for simplistic solutions and hard-line polarisation.

PB also provides a real incentive for voters to educate themselves about all of the candidates in an election, or all of the options in a referendum, because it would quickly become obvious that one's second and third etc., choices are just as important to the final result as one's first pick.

PB does not represent a change in our constitutional form of government, unlike Proportional Representation and other modifications. It is therefore quite easy to introduce PB at any level, by amending the appropriate sections of Federal or Provincial Elections Acts or the Municipal Act (for municipalities in BC). Since no constitutional changes are required, PB could be an option, with some provinces or municipalities using it and others staying with WTA.

Now for the disadvantages. There are actually only two that I am aware of, but I'm not an expert on PB so I have to assume that there are more. Neither of them are insuperable.

First is voter education. Having worked in elections, I know how hard it is for some people to even grasp the idea of putting an X in a particular space on a ballot paper. Since there has been so much effort spent over the years to get people to do this, and since PB requires that EACH candidate MUST receive a ranking for a ballot to be valid, the number of spoiled ballots (incomplete, or marked with X's) is bound to increase and to stay high if a manual system were used.

The second problem is the much greater time required, firstly for voters to complete their ballots and secondly for the points to be counted accurately and checked.

I think there must be a way to address the first problem in the tallying system; and I believe that both these problems can be largely overcome through technology, that is, the use of some form of voting machine and computer tallying.

What do you think? Does Preferential Balloting seem more democratic? Does it more accurately reflect the perceptions and wishes of voters? Does it elect "consensus candidates"? Is it an alternative or perhaps an asset to Proportional Representation? Draw up some charts, imagine some candidates, run some paper elections or computer simulations of your own!

Originally written & presented to the Green Party May 1987