The Equal Protection Challenge to Winner Take All: A Legal Guide

15 min readSep 14, 2017


Today, the non-profit that I founded, EqualCitizens.US, announced it would crowdfund support for two lawsuits to challenge the way votes are allocated in the Electoral College. All but two states allocate their vote according to a winner-take-all system, in which the winner of the popular vote gets all the electoral votes for that state. We believe winner-take-all violates the Equal Protection Clause of the 14th Amendment, and we intend, through EqualVotes.US, to build a case (and a campaign) to get the Supreme Court to agree.

We’re not the first to have this thought. (I was brought to see it by Sam Issacharoff’s essay; last November, I published a related argument on Medium by Jerry Sims.) And we’re not so naive to believe that this fight will be easy. But we do believe that a credible and true argument can be made, and that there is real value in raising and pressing this argument now. In this post, I offer a guide to that legal argument, styled, for quick skimming, as an FAQ.

(1) What’s the practical problem with the winner-take-all system?

Winner-take-all skews the presidential selection process in two critical ways.

First, it focuses the attention of presidential campaigns — and hence, candidates — on no more than a dozen “battleground” states. Those are the swing states, the states that could conceivably go one way or the other. A rational campaign works hard to win those states by tailoring its message and platform to the interests of voters in those states. Thus, in 2016, two-thirds of campaign events happened in just 6 battleground states — Florida, North Carolina, Ohio, Pennsylvania, Virginia, and Michigan. Four battleground states — Florida, North Carolina, Ohio and Pennsylvania — saw 71% of campaign ad spending and 57% of candidate appearances. Altogether, the 14 battleground states saw 99% of ad spending and 95% of candidate campaign stops.

Yet these battleground states do not truly represent America. They are whiter, older, and more focused on 20th century industry than the rest of America. A presidential campaign aiming to please them is thus not a campaign for America.

Second, winner-take-all increases the probability that the loser of the popular vote will be selected as president. Two of the last 3 presidents were inaugurated after losing the popular vote. Some estimate that America’s changing demographics mean the probability of this will only increase over time. According to some, in close elections, there’s as much as a 40% chance that the popular vote loser will be selected as President. This again radically undermines the ability of the President to represent America as a whole.

The only way to solve these two problems perfectly would be to eliminate the Electoral College (by amending the Constitution), or to effectively (and very cleverly) evade it (as the National Popular Vote initiative would do, through a compact by states representing at least 270 electoral votes to pledge their electors to the winner of the national popular vote).

The first solution just will not happen. Amendments require 38 states to be ratified. There are easily 13 states that are ok with the current system.

The second solution could very much happen. But it will take an enormous political movement to convince the states to support the national popular vote compact. Even the most optimistic in that movement believe it will be years before they have enough votes committed.

Our aim is to make progress now. We want to build a movement to support the fight to get the courts to apply the principle of “one person, one vote” to our presidential elections, ban winner-take-all and thus force a change by 2020.

(2) The electoral college is part of the Constitution. How could a Court ever invalidate it?

Our lawsuit will not challenge the Electoral College. We will challenge the winner-take-all rules for allocating electoral college votes. Those rules are not part of the Constitution. They are the creation of state legislatures. It is the states that establish the procedures that erase the impact of any vote for any candidate other than the winning candidate. Those state procedures do not have the sanction of the framers’ Constitution. But they must conform to the standards of our amended Constitution—amended by the 14th Amendment.

(3) But doesn’t the Constitution give the states “plenary power” to choose their electors however they wish?

It does, but let’s be clear about what the Supreme Court has said “plenary power” means. As the Court wrote in Williams v. Rhodes (1968),

[T]he Constitution is filled with provisions that grant Congress or the States specific power to legislate in certain areas; these granted powers are always subject to the limitation that they may not be exercised in a way that violates other specific provisions of the Constitution. Nor can it be thought that the power to select Electors could be exercised in such a way as to violate express constitutional commands that specifically bar States from passing certain kinds of laws. Clearly, the Fifteenth and Nineteenth Amendments were intended to bar the Federal Government and the States from denying the right to vote on grounds of race and sex in presidential elections. And the Twenty-fourth Amendment clearly and literally bars any State from imposing a poll tax on the right to vote “for Electors for President or Vice President.” Obviously we must reject the notion that Art. II, § 1, gives the States power to impose burdens on the right to vote, where such burdens are expressly prohibited in other constitutional provisions. We therefore hold that no State can pass a law regulating elections that violates the Fourteenth Amendment’s command that “No State shall … deny to any person … the equal protection of the laws.”

That’s why, in the following year, in Moore v. Ogilvie (1969), the Court applied “one person, one vote” to Illinois’ system for petition gathering in the selection of presidential electors. And that’s why, more recently, in Bush v. Gore (2000), the Court affirmed the application of “one person, one vote” to “the Presidential selection process.” 531 U.S., at 107.

The reason, as Bush explained, is that once the state decides to give the vote for President to the people, that changes, fundamentally, the freedom the state has. As Bush described,

[w]hen the state legislature vests the right to vote for President in its people, the right to vote as the legislature has prescribed is fundamental; and one source of its fundamental nature lies in the equal weight accorded to each vote and the equal dignity owed to each voter. 531 U.S., at 107.

The legislature is certainly not compelled to give the people the right to vote for President. But once it does, its “plenary” authority has been fundamentally constrained. That is the meaning of the “one person, one vote” jurisprudence for almost 50 years.

(4) Ok, assuming, as Bush states, the “Presidential selection process” is constrained by “one person, one vote,” how does winner-take-all actually deny anyone an equal vote?

This is the critical step in the argument. Certainly, there is no decision by the Supreme Court that squarely addresses the issue. But the reasoning of the Court’s jurisprudence clearly points to the conclusion that winner-take-all violates “one person, one vote.”

To see why, however, we need to take a couple steps.

Winner-take-all systems are of two kinds. The most common—call it “final winner-take-all”—aggregates, and then discards, votes at the final stage of an election. Candidate X gets 1,000,000 votes; candidate Y gets 1,100,000 votes. Candidate Y thus wins. This is the way congressmen are elected. The winner of the popular vote at the final stage becomes the congressman. Those who voted for her opponent have their votes counted at the final stage. There are just not enough of them for the opponent to prevail.

But sometimes, winner-take-all systems aggregate (and discard) votes at both the final stage, and an intermediate stage (“intermediate winner-take-all”).

In Georgia, for example, until the early 1960s, the state primary for statewide office was conducted through a kind of mini-electoral college. The winner of the vote in a county got the delegates from that county. Those delegates were then aggregated across all counties to determine the nominee for the state.

Gray v. Sanders (1963) invalidated that procedure because the size of the counties was different, and hence the weight of votes between counties was different. Georgia tried to defend that inequality by pointing to the Electoral College. The essence of its argument was “we’re not as unequal as the Electoral College is.” The Supreme Court rejected that argument. As the Court held, the federal electoral college did not sanction parallels at the state level.

Yet though unremarked by the Court, just as troubling as the unequal size of the counties was the fact that the procedure threw away votes for a statewide office before they were effectively and finally counted. If you happened to vote for the candidate for governor who lost in your county, your vote had no effect at the final stage of the counting.

This is a distinct — but should be equally troubling — constitutional problem.

To see why, imagine a state had 4 districts of exactly the same population — 2.5 million each. Imagine each district had 1.5 million voters. And imagine a candidate for governor — call her Jones — lost 3 of the districts (A, B, and C) by just 10,000 votes, but won the 4th district (D) by 500,000 votes.

If the state counted every vote equally, Jones would win statewide by more than 450,000 votes. But if the state allocated delegates at the district level, like the Electoral College, and determined who got that delegate on a winner-take-all basis, Jones would win delegates from just one district, while losing in the other three. She would therefore lose the election, determined through district level winner-take-all reckoning, even though she would have plainly won the popular vote for the state — overwhelmingly.

Now there’s lots of ways our hypothetical state could achieve this same result more directly, without using the device of an electoral college. The problem for our hypothetical state is that each of these other ways has been clearly deemed illegal. If the state said, for example, that votes from districts A, B, and C were weighted at 20x a vote from district D, that would plainly violate the “one person, one vote” principle. Yet that would achieve the same result — Jones would lose the weighted count, even though she plainly wins the “one person, one vote” count.

Yet this is precisely what intermediate winner-take-all does. Votes are weighted contingently. If a vote is cast for a candidate who loses a district, that vote is weighted at zero. If the candidate wins the district, the vote is weighted at 1. That contingent weighting would just as effectively guarantee that Jones loses as the mini-electoral college would guarantee that Jones loses.

So the constitutional question comes to this — why is a system that contingently weights the votes in one district over another any better than a system that directly weights votes? Why is one that says “votes get weighted at 20x with certainty” any worse than one that says “votes get weighted at zero contingently”?

Our view is that there is no meaningful constitutional difference between a weighted vote system at an intermediate stage, and a contingently weighted vote system at an intermediate stage. Both systems violate the principle of one person, one vote. Both systems should therefore be forbidden by the 14th Amendment to the Constitution.

This is precisely the problem with winner-take-all in the Electoral College. The College is electing national officers — the President and Vice President of the United States. Citizens of the United States are voting for those national officers. But before their vote gets effectively counted, if they happen to vote for a candidate who loses their state, their vote is contingently discarded. If they vote for the loser, their vote is weighted at zero in the allocation of electoral college votes. It therefore has no power beyond their state, even if their candidate could win the election nationally. It thus violates the “one person, one vote” principle.

But notice the obvious point: Unlike the electoral college itself, which the Court in Gray signaled would receive a kind of 14th Amendment constitutional pass, intermediate winner-take-all is not part of the constitution. Even if Gray means that the unequally sized states within the electoral college don’t violate the Constitution, Gray says nothing about the rules for how electors get allocated by the states. Those rules don’t have constitutional sanction. Indeed, those rules, we believe violate the Constitution, as amended.

(5) Has the Court really never addressed this issue?

Never directly, indirectly once. And the authority of that one time has now been drawn into question by Bush v. Gore (2000).

The Court has never directly considered the question of whether winner-take-all violates “one person, one vote.” Delaware asked it to do so just over 50 years ago. The Court ducked the question then, and has never heard arguments on the issue since.

The Court did summarily affirm a lower court’s determination that winner-take-all does not violate “one person, one vote.” In Williams v. Virginia Board of Elections (1968), a 3 judge district court upheld Virginia’s winner-take-all rule against an Equal Protection challenge. As the lower court held, though winner-take-all did raise significant issues of fairness and equality, to violate the 14th Amendment, the plaintiff had to show that the rule was “invidious.” As the 3 judge panel wrote,

[I]n a democratic society the majority must rule, unless the discrimination is invidious. No such evil has been made manifest here. 288 F. Supp, at 627.

Six months later, the Supreme Court summarily affirmed, without argument or opinion, the lower court ruling. Williams v. Virginia Board of Elections (1969). And under the rules governing summary affirmance, even though 50 years old, this decision would preclude us from raising a challenge to winner-take-all again — unless “‘doctrinal developments’ illustrate that the Supreme Court no longer views a question as unsubstantial, regardless of whether the Court explicitly overrules the case.’ Bostic v. Schaefer, 760 F.3d 352, 373 (4th Cir. 2014) (quoting Hicks v. Miranda, 422 U.S. 322, 344 (1975)).”

We believe Bush v. Gore is that “doctrinal development.” For in that case, the Court found that a procedure for recounting votes did not meet the standards of “one person, one vote,” but without any suggestion of invidiousness. The question for the Court was not whether the standard for recounting established by the Florida Supreme Court was invidious. The question was whether it secured to every voter equal protection for their vote. Bush thus removes the invidiousness requirement for “one person, one vote” claims in the context of the “Presidential selection process.” It therefore permits us now to argue that winner-take-all violates the Equal Protection Clause.

(6) Yes, but didn’t the Supreme Court say that the rule of Bush v. Gore applied to that case only?

It is true that the Court did say that its “consideration is limited to the present circumstances.” 531 U.S., at 109. It’s also true that no one really knows what that bizarre statement means. The Court could not, honestly, be saying that it was crafting a 14th Amendment rule for that case alone. It could not have been saying that somehow the 14th Amendment protects Republicans differently from how it protects Democrats. As Justice Scalia said in a law review article,

[W]hen, in writing for the majority of the Court, I adopt a general rule, and say, ‘This is the basis of our decision,’ I not only constrain lower courts, I constrain myself as well. If the next case should have such different facts that my political or policy preferences regarding the outcome are quite the opposite, I will be unable to indulge those preferences.

Yet if the Court was really saying, “here’s the 14th Amendment we’re applying here, so as to guarantee a Republican gets elected President, but in every other case, please apply a different 14th Amendment” then we shouldn’t let it get away with such lawlessness so easily. Our presumption should be as Scalia declared it — that the “basis of” the decision constrains the lower courts and the Supreme Court. And the basis of the decision in Bush v. Gore is that invidiousness is no longer a requirement.

(7) If you win, won’t the states just allocate electors at the congressional district level?

They may try, but if we win, they won’t be allowed to. The basis upon which we would prevail would also foreclose allocation by congressional district.

The reason again is the “one person, one vote” principle.

If we win, the default remedy would be either national popular vote, or proportional allocation of electoral votes at the state level. From the perspective of equality, congressional allocation would be worse than either of these two solutions because it would less certainly assure a result that would be consistent with the “one person, one vote” results. Once we show that fact, the state would then bear the burden of justifying this less equal scheme. There is no justification for that added inequality. Electors have no ongoing connection to any congressional district. There is no other justification for accepting the less equal result. “One person, one vote” thus shows why the rule must be proportional allocation of electors, either at the state level, or nationally — at least so long as the states vest a vote for President in the People.

(8) If you win, won’t the system become unstable. Won’t third party candidates throw the election into the House?

Under the Constitution, if no candidate for President gets 270 votes in the College, then the House of Representatives must determine who the next President will be. The House votes by state delegation — each state gets one vote. Currently, 31 states are controlled by Republicans. Thus, if the election were decided by the House, it is most likely that the Republican candidate would win, regardless of who won the “one person, one vote” election.

Third party candidates increase the probability that no candidate would achieve a majority in the College. They therefore increase the chance that an election for President would be thrown into the House. But the consequence of that result, though constitutionally sanctioned, is an even greater probability that the selected President would not be the candidate who won the “one person, one vote” election.

This fact justifies the traditional state remedy for third party candidates. States have long had the power to limit third party candidates. In this case, the possible election in the House would justify the states limiting any allocation of their electors to any third party candidate who could not win in the college. Such an allocation rule by the states would increase the probability that the candidate who was ultimately selected was the candidate who won the “one person, one vote” vote. That fact would, at the very least, justify state rules that gave voters a ranked choice for President, or that excluded allocations to candidates who could not win in the College.

To say this is not to denigrate the Constitution’s scheme for resolving elections in which no candidate achieves a majority. There needs to be a system for addressing that problem, when that problem arises. But even at the founding, election by the House was not the preferred method for selecting the President. As Hamilton explained in The Federalist Papers, it was “desirable that the sense of the people should operate in the choice of the person to whom so important a trust was to be confided.” That “sense” is expressed today through elections. A state rule that aims to preserve that sense is plainly justified — and especially so under the amended Constitution.

So yes, the Constitution gives the House the power to determined the President when no majority exists in the College. But a Constitution complemented by the “one person, one vote” principle certainly permits the states to adopt rules aimed at increasing the probability that the College would determine the President and Vice President consistent with the “one person, one vote” election.


Thus an outline of a legal argument for the conclusion that “one person, one vote” renders winner-take-all in the states unconstitutional. I’m happy to extend the analysis in response to questions raised in the comments. We’re eager for feedback and criticism. Note, we don’t consider “you’re not going to win” criticism. In our view, the question is (1) do we have a correct legal claim, and (2) will bringing that claim help build the movement necessary to bring about reform, whether through this litigation or not.

I’ve tried to answer (1) in this post. I will say more about (2) later. But this should be enough to establish the core belief that is necessary to justify bringing a legal action in good faith: Under the law as it is right now, given the Constitution as it has been amended, states should not be allocating their electoral college votes according to the winner-take-all principle.

I have long believed that the courts alone will not solve the problems with our democracy. I remain committed to that view. But EqualCitizens.US aims to rally as many Americans as possible to the understanding that our current democracy does not respect all citizens equally. This case will make that point clear in the context of the presidential selection system. In other fights, we will show why it is also true because of the private funding of public campaigns, because of gerrymandering, and because of the way we are not equally free to vote. All of these inequalities destroy the possibility of America being a “representative democracy.” Yet a “representative democracy” (what the framers would have called, “a Republic”) is what we were promised.

Help us fight to get that promise back. Join us at EqualVotes.US.