On May 10, the government released its principles for electoral reform. In this series we are looking at those principles one by one.
Principle A:
“Effectiveness and legitimacy: that the proposed measure would increase public confidence among Canadians that their democratic will, as expressed by their votes, will be fairly translated and that the proposed measure reduces distortion and strengthens the link between voter intention and the election of representatives;”
Before we can look at how an electoral system can achieve this, we must understand what is meant by voter intention and the democratic will of Canadians as expressed by their votes, presumably the will of voters since the will of non-voters to not participate is already clear, and what is “voter intention” in the context of strengthening its link to the election of representatives.
The first thing to acknowledge is that different voters have different intentions, that voters can have a combination of intentions, and that this combination of intentions can change from election to election. Therefore strengthening the link between only one specific type of intention and the election of representatives is inappropriate, and it is particularly inappropriate if it decreases the legitimacy of another type of intention.
Intentions can include to elect a particular individual, or to dismiss a particular individual. Some are motivated by specific characteristics of candidates, whether age, sex, ethnicity, profession, or what have you, and want more members of that particular group in Parliament, or representing them. Some wish to have a personal and a community representative in Ottawa, a person who is solely responsible for speaking on their behalf. Some are loyal to a party and wish their party to take power. Some are opposed a party, often the party currently forming the government, and want it to lose power. Some wish to ensure that candidates of a party that is anathema to a large proportion of the population are kept out. Some are nearly equally satisfied with two candidates or with two parties, and want to help ensure that either one or the other is elected, and not a third. Some wish for stability and continuity, and some wish, beyond which party will be in power, to have the next government be a majority government or a minority government. Some are sensitive to social consensus and want to vote the way they think that their peers are voting to show solidarity with them. Some vote based on maximizing their self-interest, and some try to maximize the benefits to others. Particularly in the case of byelections, some wish to send a message of support or censure to a party without intending to change the government.
We must look at these intentions in the context of the other principles, such as Principle E. This means that the intent to elect a local representative, for instance, is of particular importance.
Some electoral systems are particularly good at satisfying multiple possible intentions at once, while some specialize in a smaller subset of the possible intentions. For instance, a closed party list single-district proportional representation system, like the ones used in Bulgaria or Israel, does well at representing the intentions of those who are loyal to a single party, but do not allow local representation nor votes for or against individuals. Even worse, some systems do not allow some of those intentions to be communicated, or force an interpretation of the vote that is inconsistent with some of the intentions, or endows one interpretation with legitimacy but not others.
Despite its flaws, the current system, FPTP or plurality voting or one-round single-member system, does have the benefit of allowing several different types of intentions, but is not always effective in translating the whole range of voter intentions into elected representatives. For instance, the wish to remove a party or an individual, or the wish for a majority or minority can be satisfied to some degree through the use of strategic voting. Strategic voting is a powerful tool to overcome the limitations of a ballot’s low-bandwidth communication of a voter’s more complex intentions, but it is an imperfect tool. In order for it to work, enough voters must evaluate information that is not easily available about the behaviour of others, and perform a difficult analysis. Two voters having identical intentions may end up voting differently, and because of it not achieve their common objective.
Some of the intentions that we have enumerated are difficult to reconcile with each other and with other principles. However there is one common situation that for all of these intentions has a high risk of distorting the democratic will. This is where in a riding, first-choice preferences are split for instance 33%, 33%, and 34% between three candidates and the one that wins with 34% is not the one that the majority would be most satisfied with. Going through the list of motivations, this result is inconsistent with almost all of them. Here, it is the fact that the ballot asks for a single choice which does not more clearly communicate the full range of preferences is what results in an outcome that is not the one most satisfactory to most voters. Not having a mandate from an absolute majority is where legitimacy of the outcome is called into question. Perhaps the majority would support that choice and perhaps not.
It is also where the reduction of distortion comes in. In a plurality system, it is possible for a change in a vote to have the opposite effect from what the voter intended. A voter may decide that while they marginally prefer candidate A over candidate B, their wish not to be represented by candidate C in a close race strongly overcomes that preference. They may vote for candidate B in order to optimize a combination of intentions. However there is a risk that switching from candidate A to candidate B in order to achieve those objectives has the opposite effect, contributing instead to a narrow victory of candidate C.
What about an interpretation of distortion as a difference from a proportionality measure? Remember, many members of our group believe that proportionality is a legitimate aspiration. However proportionality is not per se a voter intention when casting a ballot. It is a wish that involves the votes of other people, not just one’s own. It is a belief that the vote ought not to be counted on a local basis, and a wish that minority viewpoints as expressed by minority parties should have more importance, and a perception that first choices are paramount compared to second and third choices, but at the time of casting their ballot, their intention is to be represented by a member of their preferred party. It is a “distortion” relative to an expected political outcome that applies if all voters had that similar set of attitudes about the role of parties, and a definition of fairness including fairness to parties, but it is not a distortion of a wide range of voter intents. It is worthy of consideration, particularly when examining principle B, but it is difficult to reconcile with Principle E. We must also be conscious of the risk introducing new distortions. For instance, if a vote that was cast by a voter with the intention of electing an individual in one community is counted as though it were a vote for a party and is used to elect a different individual in another community, that would be introducing a distortion of that voter’s intention.
Effectiveness of voting is a separate issue. To what degree can one vote or a few votes affect an outcome of who is my MP or of who forms the government? We already discussed how it can have an unintended effect, but as a general rule the smaller the constituency the greater the effectiveness. In a constituency of one, my vote always determines the outcome for the MP, although I have little control over who forms the government. In a constituency of three, I must convince one other person in order to win every time. In fact, as the constituency grows, I need more information about the voting intention of others and I need to participate more in influencing other people in order to maximize my own effectiveness. When the constituency becomes too large I have no hope of convincing enough people to make a difference, unless I assume that the voting intentions of a large number of people are fixed and those voters support different candidates in nearly equal proportions, and I need then convince only a small number of undecided voters. In those cases, under plurality voting, there is an inverse relationship between effectiveness and legitimacy: With 5 candidates each expecting 20% of the vote, my vote is very effective but the resulting mandate lacks legitimacy. If one candidate has 70% support, my ability to make a difference is very limited but the mandate will have high legitimacy. If we switch to preferential voting, each MP will have the increased legitimacy of having over 50% support but without reducing my vote’s effectiveness. In fact, by allowing me to have a more expressive vote that spells out my priorities, I increase the effectiveness of my vote even if my first choice candidate is unlikely to gain consensus.
Going through the list of possible voter intentions, Preferential Voting or the closely related 2-round single-member voting will tend to count the vote in a way that is more consistent with the various intents of voters and in most cases will reduce the possibility that their vote may have the opposite effect from what they intended. Effectiveness is generally improved, except that it may reduce the probability that your own vote is more effective because it exploits a flaw that other voters suffer from.
PV does not reduce all distortions or perceived distortions, Arrow’s impossibility theorem proves that no system can do that, but it reduces most of them and does so in a way that respects the variety of intentions that voters may have. That allows those intentions to vary by person, geographically and temporally. Where voters were already compensating for those distortions through strategic voting, it provides similar result without requiring accurate information as to the intentions of others and without the risk of unintended outcomes.