Electoral reform needs better PR: Childhood addendum

This is a brief addendum to the previous discussion about electoral reform. As I noted in Part III, s 3 of the Charter grants a right to vote to “[e]very citizen”, but not every citizen can vote. Most notably, citizens under 18 cannot vote. I’m going to discuss two less infringing alternatives than giving children no vote: give a proxy to their parents or enhance the vote of their future self.

These alternatives haven’t been discussed in the case law, so far as I know, but the age restriction has been challenged by some 16 year olds in Alberta. In Fitzgerald v Alberta, 2002 ABQB 1086, Lefsrud J accepted that their s 3 rights had been infringed, but found that infringement was demonstrably justified per s 1. To reach that conclusion, Lefsrud J accepted that the 18 year old age cutoff was minimally impairing because an age-based cutoff was appropriate (to prevent babies from voting)¹, an age-based cutoff was necessary (because there is no rightful substantive test for individual capacity to vote)², and 18 years of age was not an arbitrary choice³.

Later cases may make a direct attack of that decision possible⁴, but that’s not what I want to address. Instead, I’d like to ask whether the choice is really between under-18s having a vote and having no vote.

Oakes identified four hurdles to justifying an infringement under s 1:⁵

the objective of the infringing measure must be “pressing and substantial”;
the measure must be “rationally connected to the objective”;
the measure must “impair “as little as possible” the right”
the measures’ effects must be proportional to the objective.

I’m going to focus on the third hurdle: minimal impairment. In its most-government friendly form, it invalidates a measure only when there is an alternative that is as good on every pressing and substantial objective and is less infringing.⁶ If the alternative fulfills some pressing and substantial objectives but raises other policy problems, then we have to consider proportionality.

I see two possible middle grounds between fully impairing the right to vote and fully granting the right to vote. These create alternatives that are less impairing than providing no-vote but fulfill the pressing and substantial objective identified in Fitzgerald that “those eligible to vote are mature enough to make rational and informed decisions about who should represent them in government”⁷.

The first alternative is to allow parents of children to vote on the latter’s behalf. Children, still, would be unable to vote; but they would likely be more effectively represented (to use the Boundaries Reference language) if someone whom otherwise is expected to act in their best interests votes, than if they have no vote at all. Children’s rights would still be infringed, but they’d be infringed less.

Various objections are possible. First, one might object that not all parents will vote in their children’s best interests and so not contribute to effective representation. This is true for many aspects of parenthood, and yet parents can make decisions for their children generally; overall, our social system presumes that parents are better at representing their children’s interests than others are. Today, we essentially make the choice that children having no vote is okay by dividing it among all other citizens. Are parents better voters for their children than society at large? Probably.

Second, one might object that this would simply give parents an extra vote, unfairly to non-parents. This objection is stronger the more that one thinks that parents will vote in their own interests rather than in their child’s. The more you hold that cynical view, the more parental proxies are offensive. At the same time, the more you see this as a child’s vote being exercised by the parent, the less offensive parental proxies are. Non-children have previously been unfairly advantaged by being able to infringe the rights of children. That an advantaged group feels that reducing unfairness is unfair is common but not something courts should care about.

A more prosaic, but non-trivial, objection is that it’s very hard to make a fair and secret parental vote for children. I’m going to assume first that only custodial parents can exercise their child’s right to vote. I’m also going to assume that the elections authority has a register of parental relationships with children, and knows who the custodial parents are. I see four options for parental proxy voting: “joint voting”, “individual voting”, “several voting”, and “joint-and-several voting”. All have problems.

In joint voting, all custodial parents must agree on a ballot in order to cast it. This model seems unlikely to work if parents live in separate ridings but share custodial rights; at very least, they’d have to agree on which riding to vote in, and would presumably have to attend together in person to get the ballot and fill it out. This seems likely to cause problems, especially for divorced custodial parents.

In individual voting, whichever custodial parent attends first to the ballot box can cast a vote for the child. This has multiple problems. First, if the parents have separate polling stations, there is the potential for double-voting by the child, if parents attend at their polling stations at close to the same time. The child might have been given two votes by the time the polling stations record the child’s vote and transmit it to the other station. This problem could be fixed with a delay and some technology, but it could cause some friction. Second, and more fundamentally, it creates a race to vote, especially for parents who disagree.

In several voting, each custodial parent can exercise 1/n of their child’s right to vote, where n is the number of custodial parents. This solution may mean that a child doesn’t fully vote. It also may reveal the vote of a parent when custody is shared between more than two parents (e.g., if there are few enough three-parent families, then a 1/3 vote may be unique at a polling station and so those parents’ votes for their children may not be kept properly secret).⁸

In joint-and-several voting, each custodial parent exercises 1/n of their child’s right to vote, where n is the number of custodial parents who cast a ballot for the child. This solution maps the closest onto the notion of “shared custody”, because each parent can act fully on behalf of their child, if the other one does not. Implementing this system with paper ballots, however, would require making a special ballot for each child so that it’s possible to track how many ballots are cast for the child.⁹ Such a special ballot would reveal how each parent exercised the vote for their child both to each other and possibly also to the state.

Of these options, several or individual voting appear the best to me, but obviously all are somewhat flawed. They also all suffer by neglecting children under public guardianship, or the children of parents who are themselves not old enough to vote. Implementing parental proxies without providing any voting method for those (plainly disadvantaged) groups of children would be discriminatory. And, as discussed, the notion that parents are acting in their child’s best interest is not always true. What’s the alternative? Giving the child’s right to vote to their future selves.

Here’s how it would work. In each election where a child is not permitted to vote, a credit is added to that child’s vote bank by the elections authority. In the first election that the child can vote, their vote bank is emptied and they get an additional ballot for each credit in the bank.

What does this do? Well, it means that parties will have an interest in appealing to teenagers, who will soon have the right to vote. Teenagers would also have deep reason to become engaged with politics because they’ll be super-voters for their first election.¹⁰ How super? In the 2019 federal election, 27.4M voters were eligible to vote.¹¹ Of those, about 1.8M were people who had just aged into their first election where they had the right to vote.¹² If all of those voters were citizens at birth, then they’d have been alive for 5 elections where their right to vote was removed. Their potential vote would thus rise from 1.8M (6.6% of the total vote) to 10.8M (30% of the total possible vote). Elections would radically change, as parties aim to attract the 18-22 year olds who hold the power of all their previous disenfranchisement.

Would 18-22 year olds holding such power be less infringing than the present? I think so. On the Boundaries Reference ideal of “effective representation”, we can say that children’s interests are going to be more effectively represented if their future just-adult selves are supervoters than if they’re ordinary voters. Parties then have much more reasons to try to appeal to the future-18 year olds, and that will shape policies they make that affect non-voting children.

On the Figueroa ideal of playing a meaningful role in the electoral process, one way of seeing the role of children with vote banking is as them casting a long shadow over the electoral process. Although they don’t get to vote, the promise of their future vote may shape political decisions. This is not a very meaningful role, but it is more meaningful than playing absolutely no role. And if we instead used the dignity-based vision of meaningfulness that I suggested in Part III, vote banking communicates to children that their perspectives are important and valued, and will be taken into account when they are of an age to reflect properly. The state is still acting parentalistically, but vote banking builds into it more respect for the wishes of children than granting them no vote at all.

I do not mean to suggest vote banking is a full solution, merely a scheme that is less infringing than present. Clearly 18 year old supervoters will be imperfect proxy voters for their past selves in many respects. And parties will surely not be fully responsive, aiming to do what is best for an 8 year old in the hope that a decade later the then-18 year old will remember. Extra power in a later election doesn’t make up for no power in an earlier one, especially when decisions are made that affect the future. Nonetheless, the bar here is very low: what is better than nothing?

Vote banking and parental proxies also work quite well together. One might reasonably impose an age-cutoff for parental proxies that roughly corresponds to puberty / the teenage years (e.g., ~13 years old). Above that age, children are beginning to make significant life choices for themselves and exercise autonomy from their parents. Children may also feel strongly then that their parents do not share their views, political or otherwise — and for the expressive purpose of voting (as opposed to the legislative purpose), such feelings are core. And, if children are right in that, granting a parent proxy for those ages will be more like creating an additional parental vote than allowing another to exercise the child’s vote. Just as we can expect parents will be worse proxies for teenagers than for prepubescent children, we can also expect that the teenage years will be fresh in an 18 year olds’ memory in the way that infancy and childhood will not. Their future adult self is a better proxy voter for a teenage child than for an infant child.

Relying on their future self also helps cure certain defects in relying on parental proxies alone. In the parental proxy system, parents are supposed to act essentially as agents for their children; sometimes, they won’t. We can’t do anything about this if parents vote wrongly (because that would require deciding what a rightful vote is), but we can if parents don’t vote. If parents don’t vote, then they clearly have not acted properly as an agent for the child. In such cases, this agency deficiency can be remedied by having the child’s vote added to their vote bank. Similarly, if the “several” proxy system is used and all parents don’t vote, then the remaining partial votes can be added to the bank. And, most importantly, if parents are prohibited from exercising their child’s right to vote (e.g., the parents are too young themselves or are non-citizens) or the child is in public guardianship, again, the vote can be added to the bank. All of these defects are partially, if not fully, cured by transferring the child’s right to vote to their future self.

In sum: if we take “minimal impairment” seriously, there’s a less impairing alternative to removing the right to vote from children entirely. Courts shouldn’t accept that a ban on children voting minimally impairs their right to vote.