Democracy's Missing Quarter

If a country banned a quarter of its citizens from voting, would you still call it a democracy?

I doubt many people would, and yet that is roughly the position every democracy on earth currently occupies. About a quarter of the population is under 18, and none of them can vote. What makes this stranger than the raw number suggests is that this same group will live with the consequences of today's elections longer than anyone casting a ballot. The decisions made now about schooling, debt, climate, immigration, retirement, and war will shape their lives for the next sixty or seventy years, and they have no formal say in any of it.

Stranger still is that we already let parents speak for their children in nearly every other domain we treat as serious. A parent can consent to surgery on a child's behalf, choose what school the child attends, what religion they are raised in, where they live, and which legal and financial decisions get made until the child is old enough to make them alone. Parental authority extends to the child by default almost everywhere. Voting is the one place where it stops, and the more I look at that exception the harder it gets to justify.

The good news is that there is already a worked-out proposal for fixing this. The bad news is that almost nobody has heard of it. This post is an attempt to change that.

Demeny Voting

The proposal is called Demeny voting, after the demographer Paul Demeny, who put it forward in a 1986 paper on fertility and political representation. The mechanism is straightforward: each minor child generates one proxy vote, cast by their legal guardians—half a vote each where two exist—and the vote expires when the child turns 18 and casts their own.

That is the whole thing. No existing voter loses a vote. No institution is replaced. The voting age is not lowered, the franchise is not means-tested, and the structure of elections does not change. What is being added is a single mechanism by which a child's stake in the outcome shows up in the count instead of disappearing entirely. In that sense it is the same kind of move as women's suffrage or lowering the voting age to 18: a franchise expansion, not a redesign.

The idea is not fringe. It has been discussed seriously in the German Bundestag, debated in Japan in the context of demographic decline, and shows up regularly in academic work on intergenerational justice. It just hasn't crossed over into general public awareness, which is why the question I want to spend most of this post on is not "what is the proposal?" but "why should you take it seriously?"

Why Parents Are Good Proxies

Whenever someone speaks on someone else's behalf, two things determine whether the arrangement is trustworthy. The first is whether the proxy is entangled with the principal—whether the proxy actually feels the consequences of decisions made on the principal's behalf. The second is whether their interests are aligned—whether voting in the proxy's own interest tends to serve the principal's interest as well.

These criteria are not unique to politics. They are the same criteria we already apply, implicitly, when we decide whether to let a parent make medical or educational decisions for a child. The interesting question is whether they hold up under the more general case of voting, and I think they do.

Entanglement

Start with entanglement. A good proxy bears the cost when things go wrong for the principal, and parents are entangled with their children in exactly this way. When schools fail, the parent watches their child struggle through them. When healthcare fails, the parent absorbs the bills. When the labor market fails, an adult child moves back home and the household's budget reflects it for years. The parent does not get to enjoy the savings of an underfunded school system while leaving the bad outcomes to someone else, because the bad outcomes arrive at their own door.

The escape hatch from this entanglement—forfeiting custody—is real, but the legal, financial, and social cost of taking it is so high that for almost all parents it is not a live option. They cannot exit cheaply, and that is precisely what makes the entanglement load-bearing. A proxy who can walk away when the costs come due is not really a proxy; a parent who cannot is.

Alignment

Alignment is the more contested criterion, and it is worth being honest about. Parents are not fully aligned with their children's interests. Full alignment would imply something close to total self-sacrifice—a willingness to subordinate every adult preference to the child's welfare—and most parents, very reasonably, do not behave that way. So if we required perfect alignment, we would have to admit parents are bad proxies.

But the alignment we actually need is weaker than that. We need the parent's interest to run in the right direction on the kinds of questions a vote can decide. And it does. A parent has a direct, durable interest in preventing their child's failure—in keeping them from becoming financially, physically, or emotionally dependent on the parent indefinitely. The child shares that interest. They both want the child to grow into an independent adult who is not crushed by debt, untreated illness, or a collapsing job market. The alignment is imperfect, but it runs the right way on exactly the questions a proxy would be voting on, and that is the bar a proxy needs to clear.

What Demeny Voting Would Actually Change

The legitimacy case is one half of the argument. The other half is what the franchise expansion would do if it were adopted, because the strongest case for taking Demeny voting seriously is not just that the current arrangement is unfair—it is that the current arrangement produces predictably bad policy.

Democracies, as currently constituted, systematically discount the future. Voters are mortal, and the median voter in most rich democracies is now in their late forties or older. Decisions that pay off thirty or fifty years from now are decisions whose benefits accrue to people who either cannot vote (children) or are not yet born. The political class responds to this incentive in the obvious way, by under-investing in education, infrastructure, climate adaptation, and long-horizon research, and over-investing in transfers to current voters.

Demeny voting changes the math. It pulls the median voter younger by making children's stake count, and it does so by routing that stake through the people most invested in those children's long-term outcomes. You do not have to assume parents are saintly to see why this would shift policy: you only have to assume that on questions about the world thirty years out, parents care more about the answer than non-parents do, because their children will be living in the answer.

This is not a guess. It is the same logic that makes us comfortable with parents being primary decision-makers in every other domain.

If Proxies Work, Why Not Epistocracy?

A natural follow-up is: if we are willing to let one group vote on behalf of another, why stop at parents and children? Jason Brennan has argued that the most knowledgeable citizens should hold disproportionate political power on the grounds that they will make better decisions for everyone. If the proxy logic is good enough for parents, why not for experts?

I think this is a useful test of the proxy framework, and the answer is that the same two criteria distinguish the cases cleanly. Parents pass them both; experts fail both, in opposite directions.

Entanglement

Experts are not entangled with the people they would govern in anything like the way parents are entangled with their children. A Nobel laureate's life does not depend on what happens to high-school dropouts. Bad economic policy can devastate the lives of those at the bottom of the labor market while leaving those at the top almost untouched, and the same is true for nearly every other domain in which expertise operates. There is no built-in feedback loop forcing experts to bear the cost of their own decisions, which means the central thing that makes a proxy reliable is missing.

Alignment

Alignment, in the epistocratic case, runs in the opposite direction—and this is the part that I think is least appreciated in the standard debate. Credentialed self-interest does not run toward the broader population; it runs against it. Medical associations cap how many doctors may practice. Bar associations restrict who may give legal advice. Tenured academics gatekeep the next generation of academics. Every credentialed group has the same incentive: keep itself small, keep itself prestigious, and keep outsiders outside.

This is structurally the inverse of the parental case. Parental selfishness happens to coincide with the child's interest, because what is good for the child is generally good for the household. Expert selfishness, by contrast, is sustained precisely by harming those outside the credential. Education does not solve this problem; if anything, it makes the defense of the arrangement more articulate, because the people defending it are better at constructing reasons.

So the proxy logic does not generalize to experts. The criteria that make parents trustworthy proxies are exactly the criteria that fail in the epistocratic case.

Objections

A few objections come up almost every time Demeny voting is discussed. The two that matter most:

"The proxy votes on every issue, not just family policy."

This is true, and it is supposed to sound damning, but it shouldn't. No voter's franchise is restricted to issues they have direct stakes in. Childless adults vote on schools. City dwellers vote on rural policy. Retirees vote on student loans. The proxy follows the same rule everyone else already follows; restricting it to "family-relevant" votes would be a stricter standard than we apply to ordinary voters, and it is not obvious why the proxy alone should have to meet it.

"Why not just lower the voting age?"

Lowering the voting age is compatible with Demeny voting, not competitive with it. If the age were lowered to, say, 14, the proxy would simply cover ages 0–13 instead of 0–17, and the underlying logic would be identical. The two only come into real conflict if someone proposes to enfranchise infants directly—at which point the rationale for having any voting age at all has dissolved, and we are no longer arguing about proxies.

The Asymmetry

The thing I keep coming back to is the asymmetry between voting and every other domain in which children are represented. We already entrust parents with medical, legal, financial, and educational decisions on a child's behalf. We do not pretend the parent is a perfect representative; we accept that the alternative—pretending the child has no stake until they turn 18—is much worse. Voting is the only place where we still pretend.

A democracy that excludes the group with the longest stake in its outcomes has not finished extending its franchise. It has simply stopped one step short, and there is already a name for the step that finishes it.