I. Introduction

In the early 1960s, when the Supreme Court embarked on arguably the most consequential run of democracy-reinforcing judicial review in its history—the one-person-one-vote cases[1]—the Court did not linger on, and possibly did not even notice, some subtleties of the constitutional equality doctrine it was building. Was this new doctrine about equal representation: each representative representing an equal number of people? Or was it about some sort of voting-power equality, where the goal was for each district’s political contest to take place among the same number of eligible voters?[2] The Court used language open to both interpretations.[3] The looseness of the relationship between the nomenclature and the conceptual foundations of the doctrine was quite apparent in the fact that the original name of the whole enterprise—“one man one vote”—ignored the majority of the population, who were female.[4]

In practice, at first, these subtleties did not matter. The early redrawing of district lines would have looked roughly the same no matter the measure. (Indeed, it seems likely that even equalizing the number of men in each district, which nobody proposed or did, would have produced similar results!) The Supreme Court’s project was a radically democratic one: to completely reconfigure political power within jurisdictions across the country, undoing decades of profound malapportionment and even “nonapportionment” that favored rural white conservative bastions of political power in an increasingly baroque way.[5] The Court’s new doctrine succeeded spectacularly. Even if there were some lingering questions in theory about exactly what the Court was doing, in practice, one-person-one-vote simply worked. There did not seem to be an especially pressing need to be precise about what the Court was equalizing, in part because in many states, there is not much daylight between where the eligible voters are and where the people are, anyway. Either variable works as a reasonable proxy for the other.

But then there’s Texas. In Texas, a map of the state divided into districts of equal population looks significantly different from a map of the state divided into districts with equal numbers of eligible voters. That is because Texas contains very large areas that have far fewer eligible voters per total population than other areas of the state, due to a combination of age and citizenship status.[6] As Jowei Chen and Nick Stephanopoulos observe, while Texas is not unique, it empirically stands out from other states as the state in which both the racial and the partisan stakes of the gap between total population and citizen voting-age population (CVAP) are the starkest.[7] Despite their protestations, state lawmakers and political elites understand this dynamic.[8] That is why it is far from coincidental that Evenwel v. Abbott arose in Texas, as Professor Rachel Moran argues in her excellent 2023 Frankel Lecture, to which this Commentary is a response.[9]

It is laudable that Professor Moran used this platform to draw attention to the Evenwel issue and its centrality to the future of Hispanic political representation—in Texas and elsewhere. The issue has not received the attention it should, despite already reaching the Supreme Court once in Evenwel.[10] That case was a legal challenge in which a Republican political activist attempted to force Texas to switch from equalizing total population (the standard practice) to equalizing citizens of voting age.[11] If successful, it would have radically reconfigured political power in Texas. The Supreme Court unanimously rejected this lawsuit but did not fully resolve the question of what would happen in the future if Texas or some other jurisdiction decided to voluntarily draw districts based on citizens of voting age rather than total population.[12] The 2020 round of redistricting has not resolved the question because no jurisdiction (of which I am aware) chose to attempt to depart from the norm of using total population. This may be at least in part because of a lack of reliable citizenship data at a sufficiently granular geographic scale.[13]

In her lecture, Professor Moran generously engages with some of my work on this issue.[14] That work is informed by the unusual political dynamics of Texas. Before writing about this subject, I was aware that certain parts of Texas, particularly areas like the Rio Grande Valley, have strikingly few voters compared to their population.[15] I initially assumed this was a story primarily about turnout, and to the extent that it was about eligibility, I assumed the largest factor was citizenship status. Part of what caused me to focus on this topic in earnest was learning that when it comes to eligibility, the largest factor by a significant margin is simply age. When more of the population is under eighteen, there are fewer eligible voters, and this turns out to be a large part of why the Texas electorate has such a different racial profile from the Texas population.[16] In Hidalgo County, the large county in the valley that contains McAllen, over thirty percent of the population is under the age of eighteen.[17] The overwhelming majority of those under eighteen are citizens, even in cases where their parents are not.[18]

Overall, Professor Moran and I have no substantial disagreements. I take her argument as one that endorses and builds on the arguments I have made about virtual representation, with which she engages in the lecture.[19] Although there may appear to be some space between her account and mine, this is largely superficial: Moran emphasizes the problems and shortcomings of virtual representation—the ways it can go wrong.[20] I agree. Virtual representation has a variety of significant downsides in comparison to actual representation (where we actually vote for the officials who represent us). This is an underutilized but quite important and straightforward argument for expanding the electorate—giving more of the population the right to vote. Part I of this Commentary discusses what it would look like to do this: to meaningfully expand eligibility to vote in a way that would result in less virtual representation and more actual representation. Here, I defend something like the maximalist version of that project.[21]

However, enfranchisement is not a complete answer to the question of representation for those who cannot vote. You are never going to enfranchise everybody. This is not merely a statement of practical or political reality. There are compelling reasons, both in democratic theory and in actual democratic practice, why we should not enfranchise everyone. As I write this, my four-year-old is expressing his usual opinions—strongly held, not always internally coherent, but worthy of being listened to—but he is not ready to participate in democratic government. I do not think we ought to give him the vote. That is not because I worry he wouldn’t be the world’s most informed and careful voter. (He wouldn’t be, but that is the wrong question.) Instead—to draw the line in the right place—it is because I worry that he cannot do the thing asked at all. Giving him the vote would perhaps empower someone caring for him to fill out an extra ballot;[22] it is unlikely that his vote would be meaningfully his own. As I have argued elsewhere, this is also the best way to think about the question of mental competence and the vote among adults. Some adults are unable to manage complex financial affairs but, nonetheless, can coherently form a political intention to cast a ballot, and those people should be allowed to vote.[23] However, it does no favors to those who cannot meaningfully form such a political intention to hand them a ballot.[24]

In any democracy, not everyone will have the right to vote. At the very least, there will be children who are too young, immigrants who arrived too recently, and a few adults who lack the mental capacity to vote. This means that quite a lot of people—mostly children—will be represented by elected officials whom they did not have an opportunity to help select. This is virtual representation. It follows logically that those who cannot vote will be virtually represented.

One does not have to like virtual representation. One does not have to endorse it as a good idea. But one ought to notice that it is, in fact, always happening. Therefore, we have some choices to make: There are legal and policy questions about how to set up the system so that it does a better job of virtual representation rather than a worse job. That is the subject in Part III.

II. The Maximalist Case for Expanding the Franchise

The United States disenfranchises far more people than necessary. This results in considerably more virtual representation (less actual representation) than necessary. Consider just a few policy levers that we might pull to enfranchise significantly more people to help choose their own elected representatives.

First, consider immigration law. The process of naturalization in the United States is unconscionably expensive and slow. Helping people become citizens more quickly is a straightforward policy goal, and there is no shortage of models from around the world to show us how we might do this better, more quickly, and cheaper. Beyond naturalization, we also need broader immigration reforms that would give more of the people who are de facto members of our communities the de jure status that would enable them to begin the naturalization process.

Further, we need to end felon disenfranchisement. The best way to understand this is not as a nice thing we ought to do for felons. Instead, it is about getting clear about what voting is. When we disenfranchise people because they committed crimes, we are telling them, and ourselves, that voting is a privilege for law-abiding people.[25] We are collectively saying that, like recess in elementary school, voting is a privilege you can lose if you do not follow the social compact.[26] That is not the best way to understand what voting is. Voting is a right that helps constitute citizenship itself—a right that helps give citizenship in a democracy a clear meaning. Thus, to revoke the right to vote because a citizen has committed a crime is not like revoking the driver’s license of a person guilty of drunk driving. Driving is a privilege contingent on safe and law-abiding behavior, not a defining component of democratic citizenship. When we make voting conditional on law-abiding behavior, we create two classes of citizens, and in that way sap citizenship itself of meaning.

Voting is also a duty. It is something we do to contribute to a public good. The public good in this case is a functioning representative democracy. Sometimes when people commit crimes, we require them to do community service. In effect, we tell them that they need to use some of their time and energy to contribute in various specific ways to the public good. We ought to view felon voting in exactly this way. Instead of prohibiting people from contributing to this particular public good, we should understand the duty to vote as part of ex-felons’ rehabilitation as citizens. One positive effect of this would be more actual, and less virtual, representation.

In numerical terms, the largest category of people who cannot vote is people who are simply too young.[27] In 1971, the United States lowered the national voting age to eighteen through an unusual combination of judicial action and a quickly enacted constitutional amendment.[28] This Vietnam-era shift is an underappreciated and important building block of modern American democracy.

There is a strong case for pressing further on this lever and moving the voting age younger than eighteen. This would have several important effects. It would promote actual, rather than virtual, representation. More teenagers would get to participate in democratic government, and many households would obtain some votes for the first time. This would also improve the quality of the virtual representation of people in mixed-status households who still cannot vote.

Since 2013, a handful of localities in Maryland, plus one in Vermont, have enfranchised sixteen- and seventeen-year-olds to vote in all local elections; in addition, Berkeley and Oakland, California have enfranchised the same age group for school board elections.[29] States have the power to do this for their own elections and at the very least can ensure that state law empowers localities to make this choice for local elections.[30]

Because any age we choose is somewhat arbitrary, the sticking power of eighteen since 1971 is an interesting phenomenon. It may reflect the fact that eighteen feels less arbitrary than most ages because it is the legal age of majority in a variety of contexts.[31] Twenty-one was at least the drinking age; and sixteen is at least a moderately significant age (in many states, the driver’s license age).[32]

But from the perspective of transforming virtual representation into actual representation, none of these unrelated semi-arbitrary age cutoffs is as important as the two institutions that people under the age of eighteen are generally embedded within, from which people often move on starting right around age eighteen: the families in which they grow up, and their high schools. There are significant and underappreciated benefits to creating actual representation for citizens who are still embedded firmly in these two institutions. These benefits strongly support moving the voting age to sixteen, and indeed provide good reasons to consider moving it even lower, specifically to fourteen.

The obvious objection to lowering the voting age is that teenagers do not know enough to vote intelligently. This line of attack rests on some widely held intuitions. It resonates with centuries of efforts, both by conservatives and by progressive good-government reform types, to limit the franchise in ways based broadly on some conception of competence or knowledge—for example, to limit voting rights to those who could read English.[33] Such reforms were often framed in high-minded ways, around the idea that screening voters for competence or knowledge would produce better election outcomes. But in fact, their main effect, and often their purpose, was to screen out radical ideas and/or to blunt the political power of the poor, racial minorities, and immigrants.[34]

Such efforts to limit the franchise to the supposedly most competent were, and are, fundamentally antidemocratic. Perhaps inevitably, they are seeing a contemporary revival in this age of minority rule—both in political philosophy and in our politics.[35] The civic knowledge and competence of teenagers relative to other voters is a complex empirical question discussed below. But arguments of this general form—arguments for limiting the franchise based on generalizations about voters’ knowledge or competence—merit extreme skepticism. We let adult voters with little knowledge and competence vote all the time and for a very good set of reasons. Limiting the franchise to more “competent” voters predictably shifts political power away from communities that are already disproportionately disempowered; it creates enormous room for political mischief in the specification and application of the limits; and fundamentally, what it leads to is less actual representation, which is a poor result. As I have argued in the context of mental disability and voting, the best place to draw lines of this kind is not the place that maximizes or even satisfices voter knowledge or competence according to some enlightened expert view of what knowledgeable, competent voting looks like. Instead, the goal should be to draw lines that best approximate an answer to the question of who is capable of voting at all.

It is nonetheless worth noting that there are various ways to study the empirical claim that those younger than eighteen lack the competence to vote. And as intuitive as that claim may be to many people, there is substantial evidence that it is false. A variety of lines of social science research using different methodologies have found that sixteen-year-olds in particular are about as knowledgeable and about as politically competent as voters a few years older, with fourteen-year-olds close behind.[36] Many of the mental tools used when deciding how to vote seem to decline, not improve, with age.[37] Obviously, at some point this argument runs out. Sufficiently young children lack the capacity to form political intentions to vote at all. But high school students are not remotely in that category.[38] Debates about their relative competence are fundamentally asking the wrong question—and then assuming what substantial evidence suggests is the wrong answer to that wrong question.

High school students also have certain advantages that eighteen- and nineteen-year-olds lack. Their greatest advantage is that they are in high school. Especially for the one-third or so of them who will not go on to any further education, this is their last chance for civic education in any formal educational setting.[39] Enfranchising sixteen-year-olds (or even better, fourteen-year-olds) has the potential to turn high schools into significant sites of civic engagement and political organizing, the way colleges often are now for those who attend. Local candidates would aim to speak or debate in front of high school audiences if those audiences were full of eligible voters. High school teachers could give assignments to students relevant to building political competencies, such as writing an essay from a point of view with which they disagree. Teachers, of course, sometimes do this already. But imagine how much more relevant such civic education would feel if high school was also a student’s first opportunity to vote as an equal citizen with her peers.

High school students are also embedded in families, and here we can begin to see the enfranchisement of sixteen- or fourteen-year-olds differently: as a means of creating the first eligible voters some families will ever have. Obviously, many eighteen- and twenty-year-olds also live with their parents—and every family situation is different—but for quite a few families in the United States, the first time anyone in the family will vote is when a child who is a citizen of the United States reaches voting age.[40]

III. The Inevitability of Virtual Representation

The argument above is something like a maximalist approach to enfranchisement. Suppose it were possible today to adopt this maximalist approach: quick naturalization and broad immigration reforms, no felon disenfranchisement, and high school students enfranchised. This would mean millions more eligible voters. But it would still leave an enormous number of people who cannot vote. (The vast majority of these people are still too young to vote.[41]) These people will not have actual representation. What they will have instead is virtual representation.

This raises a question that is uncomfortable, but also inevitable: What kind of virtual representation will it be? Virtual representation is always imperfect. But there are big differences in the kinds of virtual representation that different voting rules and maps produce.

In the United States, the way we provide representation is generally through geography-based districts: lines on a map. This system has received a lot of criticism over the years, and for good reason.[42] It is at the root of the fine American art of partisan and racial (and conjoined partisan-racial) gerrymandering. It leaves many voters with a sense that politicians are choosing them rather than the other way around.[43] But here, I am reminded of Churchill’s famous line about democracy: Geographic districts may be the worst form of virtual representation, except for all the others.

Professor Moran points out that in any system with virtual representation, some voters will feel that they have to take into account the interests of nonvoting relatives or others they care about, and this is a burden on the voters.[44] For example, consider Puerto Rican voters in the mainland United States, many of whom feel obligated to vote on behalf of the interests of their families on the island. This is not an argument against virtual representation. Instead, it illustrates the phenomenon itself: virtual representation is occurring.

The hard question for any scheme of virtual representation is how the voters relate to the nonvoters with whom they are grouped in a district or otherwise. It’s easy enough to identify edge cases of a very good or very bad relationship between the voters and the nonvoters. First, you do not want the two groups to be too distant. That was the situation that first gave virtual representation a bad name in this country: before the American Revolution, the British said that the American colonists’ interests were “virtually represented” in the British Parliament.[45] As Alexander Hamilton put it (in the musical): “Why should a tiny island across the sea regulate the price of tea?”[46]

On the other hand, you don’t want the voters and the nonvoters to be close but antagonistic. Suppose there is a population of farmworkers, all of whom are noncitizens, and then in the same district, a population of landowners, all of whom are citizens who vote. This is not going to go well. There may be occasional areas of interest convergence, but it is not going to feel to the farmworkers that they get much in the way of representation, virtual or otherwise.

Add a racial edge to this stylized example, and you get the very situation that led the Supreme Court to construct the one-person-one-vote doctrine in the first place. Before one-person-one-vote, many rural Southern jurisdictions had very large, heavily disenfranchised Black populations, living alongside a white planter class whose central political goal was to keep them disenfranchised. Malapportionment gave those white planters extremely outsized political power, which they used to preserve the racial status quo. In this way, as J. Douglas Smith explains, “malapportionment served as a cornerstone of white supremacy, ensuring the overrepresentation of the most ardent segregationists . . . .”[47] This was not a recipe for adequate virtual representation for the disenfranchised Black citizens.

And yet, overall, geography-based representation seems to do a lot better than that. A major reason for this is that today, so many people live in families and households—and often, extended families live geographically near one another—that include both voters and nonvoters.[48] Families include both children and adults; many families include both citizens and noncitizens. Often, there is a considerable degree of alignment of interests within families and/or within households, which translates into voting.[49] To the extent that this is true, if some of the members of households and families have the right to vote, the preferences and interests of those with the right to vote will relatively align with the preferences and interests of those without. And, to the extent that people living near one another have politically relevant things in common—an extent augmented by various forms of geographic “sorting,” but one that begins with extended families living near one another—this logic then extends beyond the household to the community. The preferences of the eligible voters in a community will tend, as a result, to be relatively similar to the preferences of the nonvoting members of a community.

To be sure, there are several ways this alignment can fail to materialize or can break down. In a deeply divided community, if one side of the community has substantially more enfranchised voters than the other side, it may be impossible for the less enfranchised side to win any offices. This is a classic gerrymandering problem and indeed something of a classic gerrymandering technique. Professor Moran’s examples of Oaxacan immigrants in different Southern California jurisdictions raises the question of whether this might occur—particularly in San Diego. The problem in that example is not that the immigrant community lacks any citizens who can vote. Instead, the problem is that the immigrant community’s small number of voters are submerged in, and outnumbered by, a larger political community with whom they are in a relatively antagonistic relationship.[50] Overall, however, this situation is the exception. In general, most of the members of any community have connections with, and at least some politically relevant things in common with, the enfranchised members of the community. To the extent that this is the case, our system does a surprisingly tolerable job of ensuring that each community has reasonably good virtual representation.

However, even when communities offer relatively good virtual representation across most issues, there will be specific issues where this breaks down. Climate change is the unusual issue that seems to have quite a significant age skew.[51] (The age skew can look bigger or smaller depending on the specific question but it is certainly there.)[52] In effect, telling young people “don’t worry that you can’t vote right now—you’ll be able to in a few years, and in the meantime you’ll be represented by the people your parents elect,” doesn’t work as well if the young people disagree with their parents about what they view as their central and most important political issue. Young people in that situation are perhaps the best argument of all for lowering the voting age. And it’s not surprising that absent a lowering of the voting age, some are finding means of taking political action of various kinds despite being too young to vote.[53]

It might seem that the law is too distant, or perhaps too blunt an instrument, to meaningfully improve the quality of virtual representation. But it can and does. As Professor Moran argues correctly, the choice to use CVAP rather than total population—the choice raised in Evenwel—is itself subject to Section 2 analysis.[54] It’s very possible that in Texas, as she suggests, the choice to switch from total population to CVAP would illegally dilute Hispanic voting power.[55]

Even short of that, when drawing lines, it is possible for the law to consider what kind of representation a district provides to the people who live there, including those who cannot vote. In the 2006 Texas redistricting case LULAC v. Perry, the Court struck down a map that contained a new gerrymandered district that stretched from the Rio Grande Valley to south Austin.[56] Republicans drew this district with a Hispanic majority basically to make up for deleting a different Hispanic-majority district.[57] Justice Kennedy wrote for the majority that the new district didn’t count as compact—not primarily because of its rather long and snakelike shape, but instead because the wildly disparate populations that were grouped together at the North and South ends of the district simply didn’t have much in common—other than being Hispanic. They did not form a coherent community.[58]

In this analysis, one can see the germ of a Voting Rights Act jurisprudence that would look at the community that is being grouped into a district and ask how much that entire community of people has in common. Is this a group of people with enough in common to be capable of having a “candidate of their choice”? This is a question that parties in litigation under the Voting Rights Act already confront.[59] Perhaps they ought to confront it not only regarding the voters in a community but also the nonvoters, such as noncitizens and children.

Drawing district lines is an exercise in grouping people together with other people. We decide whether to create districts composed of people with antagonistic interests, where one group can vote and the other cannot. In that situation, the group that can’t get actual representation is not going to get much virtual representation either. Alternatively, we can choose to group together sets of people who are relatively more similar in their interests and views, along whichever are the central political axes of our time.

Various districts in the Rio Grande Valley in South Texas—and also in heavily Hispanic areas throughout the state—simply do not have many voters.[60] Turnout is low among those who are eligible.[61] Far fewer people are eligible than elsewhere in the state—primarily (and surprisingly) because of age.[62] To the Evenwel plaintiffs, this was a sign that these districts should simply have less political power. Those plaintiffs argued, essentially, that power should flow to the places where the voters live to give each voter more of an equal chance to decide the election.[63]

For the moment, that is not our system in the United States, and there is a good reason why. Whether people vote or not—and indeed, whether they are eligible or not—they make essentially the same kinds of demands on government. They make most of those demands not individually, but in groups.[64] Politics—especially local politics—is an exercise in group mobilization, coalition-building, and collective action. It is not a story, at the end of the day, about one voter and her minuscule power to change the outcome of an election.

That individualistic story fundamentally fails to explain why anybody votes.[65] Anyone’s chance of being the decisive voter in an election is amazingly small, not only because there are so many voters but also (and more importantly) because most elections in most places are not especially close.[66] In a district with a solid partisan lean, it’s nearly certain that nobody will provide the one decisive vote. As the lopsidedness of a district increases, and the already minuscule chance of casting a decisive vote curves sharply (if asymptotically) toward zero, so-called “rational” voter models predict that because voters are rational, self-interested individuals, they will not vote, even if the costs of voting are relatively low. These models are wrong. Indeed, people often vote even when the chances of casting a decisive vote are ridiculously low and the costs of voting are substantial, as exemplified by long lines at the polls.[67]

Voting is not a fundamentally selfish act. We vote in part because we (correctly) feel a duty to do so. It is a way of contributing to the public good of representative government. And yet that is not the whole story. We also vote because it is something we wish to do together with others, to elect people who represent our communities.

I am not entitled to have any particular representative represent me. But we, as a group—if there are enough of us—are entitled to have a representative who represents our community. By leaving behind the excessively individualistic approach of the rational voter models, we can begin to see not only why people vote but also who is entitled to representation: not individuals, but groups and communities. Those groups and communities invariably include both voters and nonvoters.

This way of understanding voting and representation expands the circle of who is represented. It takes in the nonvoters, the children, and even the noncitizens who are part of our communities, and recognizes the ways they, too, are represented in government. In that way, it explains why it makes sense that we assign districts by total population, for reasons that run deeper than the inadequacies of citizenship data.

We ought to draw districts that group people into communities in such a way that those who vote and those who do not will be adequately represented. From this perspective, drawing district lines based on total population is even more important in Texas than it would be elsewhere, largely because of the Hispanic population and its relative youth.[68]

IV. Conclusion

In this Commentary, I’ve made two related points. First, I argued for lowering the voting age and for a variety of other strategies that would enfranchise people who are capable of voting but are currently disenfranchised. I argued that this would have a variety of benefits, one of which is that by increasing actual representation, these changes would decrease the amount of, and need for, virtual representation. Second, I argued that even if we pursue a maximalist version of these strategies, our system will still include lots of virtual representation. Because virtual representation is inevitable, I argued that we ought to do our best to provide better virtual representation. This means taking account of the representation not only of the voters but also of the nonvoters: they, too, are part of the groups and communities that our representatives represent.

These two strategies are complements, not substitutes. They work together. Enfranchising more people in a given community will improve the quality of virtual representation for others who still cannot vote. In particular, we need the concept of virtual representation to understand an important part of the benefit of lowering the voting age: Lowering the voting age gives mixed-status families their first eligible voters sooner. Those voters will then have the benefit, and inevitably the burden, of being the ones able to make sure the entire family gets some representation. Virtual representation is no one’s first choice. But, particularly in a state like Texas, quite a lot turns on our doing it well.

  1. See generally Baker v. Carr, 369 U.S. 186 (1962) (establishing that problems of voter dilution are justiciable by courts) and the many cases that have followed from it.

  2. In other work, I have explored why the voting power story does not make sense on its own terms, even conceptually. Joseph Fishkin, Weightless Votes, 121 Yale L.J. 1888, 1893–98 (2012). The chance that an individual vote will affect an election outcome depends primarily on how close an election is, not on whether there are somewhat more or fewer eligible voters. Id. at 1894–95. In landslides, no one vote has any chance of affecting the outcome. Id. at 1896–97. And, in any event, affecting outcomes through one’s individual vote is not a good description of why people vote. See e.g., id. at 1898; Joshua Harder & Jon A. Krosnick, Why Do People Vote? A Psychological Analysis of the Causes of Voter Turnout, 64 J. Soc. Issues 525, 534–39 (2008).

  3. See Evenwel v. Abbott, 578 U.S. 54, 71 (2016) (Ginsburg, J., majority opinion) (“For every sentence appellants quote from the Court’s opinions,” suggesting that one-person-one-vote could mean some version of equal voting power, “one could respond with a line casting the one-person, one-vote guarantee in terms of equality of representation . . . .”).

  4. Google Books Ngram suggests that “one person one vote” did not decisively overtake “one man one vote” in books published in English until after 1974. See Google Books Ngram Viewer, https://books.google.com/ngrams/graph?content=One+man+one+vote%2C+one+person+one+vote&year_start=1800&year_end=2019&corpus=en-2019&smoothing=3 [https://perma.cc/7WGK-D4WP] (last visited Jan. 19, 2024).

  5. See Pamela S. Karlan, Reapportionment, Nonapportionment, and Recovering Some Lost History of One Person, One Vote, 59 Wm. & Mary L. Rev. 1921, 1940–41, 1944, 1955 (2018) (describing the failure to reapportion certain state legislatures at all, thereby locking in rural power as the population moved to urban areas, in the early decades of the 20th century).

  6. See infra text accompanying notes 16–18.

  7. Jowei Chen & Nicholas O. Stephanopoulos, Democracy’s Denominator, 109 Calif. L. Rev. 1019, 1042–45 (2021). The authors’ experimental approach to this question is unusual; it is a helpful way to see differences.

  8. See, e.g., Alexa Ura, Republicans Say Texas’ New Political Maps Are “Race Blind.” To Some Voters of Color, That Translates as Political Invisibility, Tex. Trib. (Oct. 20, 2021, 5:00 AM), https://www.texastribune.org/2021/10/20/texas-redistricting-race-discrimination/ [https://perma.cc/7RQ8-84DD] (describing lawmakers’ ability to draw districts that seem to take advantage of racial differentials in eligibility and turnout even without making direct use of racial demographic data); Rachel F. Moran, The Perennial Eclipse: Race, Immigration, and How Latinx Count in American Politics, 61 Hous. L. Rev. 719, 726 (2024).

  9. Moran, supra note 8.

  10. See Evenwel v. Abbott, 578 U.S. 54, 63–64 (2016).

  11. See id. at 62–64.

  12. Id. at 55, 64, 74–75 (“[W]e need not and do not resolve whether, as Texas now argues, States may draw districts to equalize voter-eligible population rather than total population.”).

  13. See Moran, supra note 8, at 723–26; Nathaniel Persily, The Law of the Census: How to Count, What to Count, Whom to Count, and Where to Count Them, 32 Cardozo L. Rev. 755, 773–74, 776 (2011).

  14. See Moran, supra note 8, at 726, 729–33, 737, 748.

  15. This reality is stark enough that Texans hear about it in local media. See, e.g., Pablo De La Rosa, The Rio Grande Valley Is a 2022 Political Hotspot, but Voter Turnout Has Been Historically Low, Tex. Pub. Radio (Jan. 16, 2022, 12:24 AM), https://www.tpr.org/border-immigration/2022-01-14/the-rio-grande-valley-is-a-2022-political-hotspot-but-voter-turnout-has-been-historically-low [https://perma.cc/MNX8-N399].

  16. See, e.g., Ashley Lopez, Latinos Are the Biggest Ethnic Group in Texas, but Their Political Power Lags Behind, npr (Aug. 15, 2023, 5:01 AM), https://www.npr.org/2023/08/15/1193747771/latino-political-power-texas [https://perma.cc/B62X-S87Y].

  17. See QuickFacts: Hidalgo County, Texas, U.S. Census Bureau, https://www.census.gov/quickfacts/fact/table/hidalgocountytexas/PST045222 [https://perma.cc/7KVT-32RX] (last visited Jan. 2, 2024). This is more than five percentage points higher than the state as a whole. See QuickFacts: Texas; United States, U.S. Census Bureau, https://www.census.gov/quickfacts/fact/table/TX,US/PST045223 [https://perma.cc/7SBC-W8U9] (last visited Jan. 11, 2024).

  18. This is because children born in the United States are citizens regardless of their parents’ immigration status. Cf. Hidalgo County, TX, Data USA, https://datausa.io/profile/geo/hidalgo-county-tx# (last visited Feb. 12, 2024) (noting that 81.8% of county residents are citizens—that is the figure for residents of all ages; for those under 18, the figure would be even higher).

    Moran cites a different statistic in her lecture: Only seven percent of the undocumented population in Hidalgo County is under the age of sixteen. See Moran, supra note 8, at 737–38. But that figure is lower for a straightforward reason: birthright citizenship. Because all children born in the United States are citizens, quite a lot of families contain undocumented parents and citizen children; indeed, among families with undocumented parents, this is the norm. See, e.g., National Academies of Sciences, Engineering, and Medicine. The Integration of Immigrants into American Society 121–22 (2015) (noting that out of the 5.2 million children in the United States who “resided with at least one undocumented immigrant parent,” 4.5 million or 86% of the children were U.S.-born citizens). Thus, the statistic Moran cites is unsurprising because one would expect the undocumented population to be older than the general population.

  19. See generally, Joseph Fishkin, Taking Virtual Representation Seriously, 59 Wm. & Mary L. Rev. 1681 (2018).

  20. See Moran, supra note 8, at 729–31.

  21. Of course, what I characterize as a maximalist argument—colored, inevitably, by what is politically imaginable to me today—might strike a reader someday in the future as something far short of maximalist. I hope that may be the case. At the same time, as I suggest just below, there are reasons to think there are real limits on how far enfranchisement ever will (or should) go.

  22. Giving parents extra votes on account of their children would be consistent with certain arguments about how to achieve good virtual representation, but would violate more fundamental and obvious principles of democratic equality (as between parents and nonparents, who are equal citizens). Cf. Robert W. Bennett, Should Parents Be Given Extra Votes on Account of Their Children?: Toward a Conversational Understanding of American Democracy, 94 Nw. Univ. L. Rev. 503 (2000) (using the utter lack of support for giving parents extra votes as an opening for a broader inquiry into the underlying norms of American democracy).

  23. See Joseph Fishkin, Equal Citizenship and the Individual Right to Vote, 86 Ind. L.J. 1289, 1353–54 (2011) (arguing for limiting the mental competency restriction on suffrage: “the key line to be drawn is whether an individual is able to form a conscious intention to vote”).

  24. Id. This argument suggests that there may be good reasons for setting some minimum age to vote. However, it is far from clear that this minimum age should be as high as eighteen. See discussion infra, Part II.

  25. See Pamela S. Karlan, Convictions and Doubts: Retribution, Representation, and the Debate over Felon Disenfranchisement, 56 Stan. L. Rev. 1147, 1150–55 (2004) (explaining how the reconceptualization of voting as a right rather than a state-granted privilege undercuts the case for felon disenfranchisement).

  26. See Note, The Disenfranchisement of Ex-Felons: Citizenship, Criminality, and “The Purity of the Ballot Box,” 102 Harv. L. Rev. 1300, 1304–07 (1989) (exploring the inadequacies of social-contractarian arguments for felon disenfranchisement); but see Green v. Bd. of Elections of N.Y.C., 380 F.2d 445, 451 (2d Cir. 1967) (defending felon disenfranchisement on explicitly Lockean contractarian grounds).

  27. Of all the people living in the United States who are legally ineligible to vote for any reason, roughly three-quarters of them cannot vote because they are too young. See Sex by Age by Nativity and Citizenship Status, U.S. Census Bureau, https://data.census.gov/table/ACSDT1Y2022.B05003?q=citizenship [https://perma.cc/H7VX-DHBB] (last visited Jan. 5, 2024) (showing that the United States currently contains about 21 million noncitizens, compared to about 72 million people under age eighteen). Those who cannot vote because of felony convictions currently number about 4.6 million. Christopher Uggen et al., Locked Out 2022: Estimates of People Denied Voting Rights Due to a Felony Conviction 6 (2022), https://www.sentencingproject.org/app/uploads/2024/02/Locked-Out-2022-Estimates-of-People-Denied-Voting.pdf [https://perma.cc/WRR7-W6HY]. This is a politically significant number, especially because this group of disenfranchised people is heavily concentrated in certain states. See id. at 7. Estimates of those disenfranchised by reason of mental incapacity are harder to come by, but this is a smaller number. Nationwide numbers appear likely to be in the hundreds of thousands. See, e.g., Matt Vasilogambros, Thousands Lose Right to Vote Under ‘Incompetence’ Laws, Stateline (Mar. 21, 2018, 12:00 AM), https://stateline.org/2018/03/21/thousands-lose-right-to-vote-under-incompetence-laws/ [https://perma.cc/8QKW-BKJP] (citing an estimate that 32,000 people had lost the right to vote for this reason in California in a decade—and noting that most, but not all, states have similar restrictions). There is a small amount of overlap among these categories (e.g., someone who is both a noncitizen and younger than eighteen). After accounting approximately for such overlap, it appears that there are roughly 95 million people living in the United States who cannot legally vote, and approximately three-quarters of them cannot legally vote because they are too young.

  28. See Oregon v. Mitchell, 400 U.S. 112, 130 (1970) (holding that Congress had the power to lower the voting age to eighteen—as it had done in the Voting Rights Act Amendments of 1970—but only in federal elections); U.S. Const. amend. XXVI, § 1 (“The right of citizens of the United States, who are eighteen years of age or older, to vote shall not be denied or abridged by the United States or by any State on account of age.”). The 26th Amendment, ratified on July 1, 1971, helped put state and local elections back in alignment with federal ones. See id.; Mitchell, 400 U.S. at 130.

  29. See Voting Age Status Report, Nat’l Youth Rts. Ass’n, https://www.youthrights.org/issues/voting-age/voting-age-status-report/ [https://perma.cc/W2X2-NYDY] (last visited Nov. 30, 2023); It’s Official. Greenbelt, MD Lowers Its Local Voting Age to 16, Vote16USA, https://vote16usa.org/its-official-greenbelt-md-lowers-its-local-voting-age-to-16/ [https://perma.cc/V4U4-JLBR] (last visited Jan. 30, 2024).

  30. See State Legislatures Consider Bill to Lower Voting Age to 16 in 2021, Vote16USA, https://vote16usa.org/state-legislatures-consider-bills-to-lower-voting-age-to-16-in-2021/ [https://perma.cc/UR4M-T2US] (last visited Jan. 9, 2023). A state allowing voters younger than eighteen to vote in federal elections would be inadvisable because the nature of federal elections requires a certain degree of similar treatment of voters across states and litigation over this issue could have unfortunate consequences.

  31. Now You Are 18!, Hous. Bar Ass’n, https://www.hba.org/?pg=Now-You-Are-18 [https://perma.cc/72LX-TRCC] (last visited Jan. 9, 2024).

  32. 21 Is the Legal Drinking Age, Fed. Trade Comm’n Consumer Advice (Sept. 2013), https://consumer.ftc.gov/articles/0386-21-legal-drinking-age [https://perma.cc/3TQP-2TLG]; see Kara McGinley, How Old Do You Have to Be to Drive?, PolicyGenius (July 2, 2021), https://www.policygenius.com/auto-insurance/driving-age-by-state/ [https://perma.cc/GS7E-66JM].

  33. See Alexander Keyssar, The Right to Vote: The Contested History of Democracy in the United States 53, 115–18 (2000).

  34. See id. at 97–98.

  35. See, e.g., Jason Brennan, Against Democracy 138–39 (2016); Meg Kinnard, Ramaswamy Proposes Raising Voting Age to 25, Unless People Serve in Military or Pass a Test, Associated Press (May 11, 2023, 8:04 PM), https://apnews.com/article/vivek-ramaswamy-voting-age-2024-president-ea1429836e8f809fbf301b7b027f4ab9 [https://perma.cc/77TX-AJ8D] (describing a presidential candidate’s unconstitutional, but ostensibly serious, proposal to raise the voting age to twenty-five and allow younger citizens to vote if they pass a knowledge test used in naturalization).

  36. E.g., Daniel Hart & Robert Atkins, American Sixteen- and Seventeen-Year-Olds Are Ready to Vote, 633 Annals Am. Acad. Pol. & Soc. Sci. 201, 209–13 (2011) (using a survey methodology and finding that sixteen-year-olds are the same as voters in their twenties on a variety of measures from civic knowledge and interest to tolerance and “political skill”; this is also true for fourteen-year-olds except that they have less “tolerance” than slightly older voters do); Valérie-Anne Mahéo & Éric Bélanger, Lowering the Voting Age to 16? A Comparative Study on the Political Competence and Engagement of Underage and Adult Youth, 53 Canadian J. Pol. Sci. 596, 604–06 (2020) (finding no differences between sixteen-year-olds and slightly older voters in political knowledge and other variables; sixteen-year-olds were slightly more likely to intend to vote and slightly lower in subjective self-efficacy). See generally Mark N. Franklin, Consequences of Lowering the Voting Age to 16: Lessons from Comparative Research, in Lowering the Voting Age to 16: Learning from Real Experiences Worldwide (Jan Eichhorn & Johannes Bergh eds., 2020) (collecting comparative studies).

  37. See, e.g., Richard R. Lau & David P. Redlawsk, Older but Wiser? Effects of Age on Political Cognition, 70 J. Pols. 168, 175 (2008).

  38. The empirical literature on this point is basically about sixteen-year-olds. See, e.g., Dieter Stiers et al., Are 16-Year-Olds Able to Cast a Congruent Vote? Evidence from a “Voting at 16” Initiative in the City of Ghent (Belgium), Electoral Stud., Nov. 2019, at 2, 9. Congruence here refers to voters’ abilities to successfully cast a vote that is congruent with their political preferences. Id. at 2.

  39. News Release, College Enrollment and Work Activity of High School and College Graduates — 2022, Bureau of Labor Statistics (Apr. 26, 2023), available at https://www.bls.gov/news.release/pdf/hsgec.pdf [https://perma.cc/Q3NV-D6UP].

  40. Richard Fry et al., A Majority of Young Adults in the U.S. Live with Their Parents for the First Time Since the Great Depression, Pew Rsch. Ctr. (Sept. 4, 2020), https://www.pewresearch.org/short-reads/2020/09/04/a-majority-of-young-adults-in-the-u-s-live-with-their-parents-for-the-first-time-since-the-great-depression/ [https://perma.cc/M2DR-VQN8]. Most children of immigrants are citizens, but over 7 million such children currently have only noncitizen parents. See Part of Us: A Data-Driven Look at Children of Immigrants, Urb. Inst. (Mar. 14, 2019), https://www.urban.org/features/part-us-data-driven-look-children-immigrants [https://perma.cc/FB5L-TYAE].

  41. See supra note 27 (estimating that those too young to vote account for about three-quarters of all those who cannot vote for any reason).

  42. For instance, Lani Guinier has argued that forms of proportional representation are more effective than single member districts at representing minority interests and avoid many gerrymandering-related pathologies. See Lani Guinier, The Representation of Minority Interests: The Question of Single-Member Districts, 14 Cardozo L. Rev. 1135, 1156, 1159–60 (1993) [hereinafter Guinier, The Representation of Minority Interests]; Lani Guinier, The Triumph of Tokenism: The Voting Rights Act and the Theory of Black Electoral Success, 89 Mich. L. Rev. 1077, 1137, 1150 (1991).

  43. See Guinier, The Representation of Minority Interests, supra note 42, at 1156–58.

  44. Moran, supra note 8, at 731–32.

  45. As Justice Breyer put it during oral argument in Evenwel, virtual representation “sounds an awful lot [like] what they had in 1750 or something, where the British Parliament said, well, don’t worry, America, you’re represented by the people in England because after all, they represent everybody in the British Empire.” Transcript of Oral Argument at 33, Evenwel v. Abbott, 578 U.S. 54 (2016) (No. 14-940); see Fishkin, supra note 19, at 1685, 1690–92.

  46. Lin-Manuel Miranda et al., Farmer Refuted, on Hamilton: An American Musical (Original Broadway Cast Recording), at 01:33 (Atl. Recs. 2015).

  47. J. Douglas Smith, The Supreme Court Could Send America Back a Century, TIME (Dec. 8, 2015, 10:14 AM), https://time.com/4139409/one-person-one-vote/ [https://perma.cc/AB6H-ERCL]; see Pamela S. Karlan, Reapportionment, Nonapportionment, and Recovering Some Lost History of One-Person, One Vote, 59 Wm. & Mary L. Rev. 1921, 1937, 1955 (2018) (explaining how the failure to reapportion by population led to a white rural lock on political power so profound that when the Supreme Court finally intervened in the one-person-one vote cases, “everyone understood [the] racial implications”). For a terrific student note on the racial dimensions of Baker v. Carr, see generally Robert M. Crea, Note, Racial Discrimination and Baker v. Carr, 30 J. Legis. 289 (2004).

  48. Fishkin, supra note 19, at 1692–94.

  49. Id. at 1693.

  50. See Moran, supra note 8, at 731–32.

  51. See Matthew Ballew et al., Global Warming’s Six Americas Across Age, Race/Ethnicity, and Gender, Yale Program on Climate Change Commc’n (Apr. 5, 2023), https://climatecommunication.yale.edu/publications/global-warmings-six-americas-age-race-ethnicity-gender/ [https://perma.cc/7ZVB-KZ5P].

  52. Id.

  53. Megan Carnegie, Gen Z: How Young People Are Changing Activism, BBC (Aug. 8, 2022), https://www.bbc.com/worklife/article/20220803-gen-z-how-young-people-are-changingactivism [https://perma.cc/D3PM-8G6T].

  54. Moran, supra note 8, at 746.

  55. Id. at 746, 750–52.

  56. League of United Latin Am. Citizens v. Perry, 548 U.S. 399, 423, 428, 442 435 (2006).

  57. Id. at 413, 423–24.

  58. Id. at 424, 434–35.

  59. This standard comes from the text of the Voting Rights Act. See 52 U.S.C. § 10301 (Voting Rights Act of 1965) (guaranteeing any racial group an equal opportunity “to participate in the political process and to elect representatives of their choice”).

  60. See supra notes 1617–18 and accompanying text.

  61. See, e.g., Graysen Golter, Texas Counties with the Highest and Lowest Election Turnouts – and Why, KXAN (Dec. 10, 2020, 5:54 PM), https://www.kxan.com/news/us-politics/election/texas-counties-with-the-highest-and-lowest-election-turnouts-and-why/ [https://perma.cc/PCF7-PUZU] (illustrating that the lowest-turnout counties, including Hidalgo, tend to have extremely high percentages of racial minorities).

  62. See supra notes 16–17 and accompanying text.

  63. See Fishkin, supra note 2, at 1893, 1899 (explaining why equalizing CVAP would not in fact give all the voters an equal chance of deciding the election).

  64. Id. at 1902–03; Fishkin, supra note 19, at 1688.

  65. See, e.g., Ilya Somin, Empowering Hispanics to Vote with Their Feet, 61 Hous. L. Rev. 777, 781 (2024) (arguing that people have little individual incentive to vote because they have very little chance of affecting the outcome). This common way of modeling the decision to vote—framing it as being heavily about one’s chance of affecting the outcome—creates a significant blind spot for many economists, political scientists, and others who approach the question this way.

  66. Fishkin, supra note 2, at 1893–94.

  67. The costs of voting—for example, the time it takes to vote—are markedly higher in some communities than in others. But many people do vote even in places where voting is more difficult. See, e.g., Keith Chen et al., Racial Disparities in Voting Wait Times: Evidence from Smartphone Data, 104 Rev. Econ. & Stat. 1341 (2022) (using phone location data to find that voters were 74% more likely to spend more than half an hour at the polling place in Black neighborhoods than in white ones).

  68. Cf. Moran, supra note 8, at 752–53 (noting that it might be only in Texas that switching to CVAP would violate the VRA).