Your state’s electors should match your state’s voters.

The Practicalities

Do states really have this power?

Yes. The Constitution says each state appoints electors “in such Manner as the Legislature thereof may direct.”⁷ In McPherson v. Blacker the Supreme Court upheld Michigan’s decision to choose electors by district and described that power as plenary.¹⁰ Winner-take-all is found in state law, not the Constitution, and Maine and Nebraska already depart from it. States can also require electors to honor their pledge, as the Court held in Chiafalo v. Washington, so Fair Slate can be made binding.¹¹

States have used that freedom widely. In the first elections, some legislatures picked the electors themselves, some states held a statewide vote, and others voted by district or mixed the approaches.¹⁰

Winner-take-all didn’t win out because anyone thought it was fairer. It spread defensively, as states hedged against each other. Jefferson would have won in 1796 if two of his strongest states hadn’t split their electors by district, so before 1800 Virginia switched to a statewide ticket. Jefferson admitted that “an election by districts would be best, if it could be general,” but with other states already consolidating their votes, he called it “folly & worse than folly” to hold out.¹² Massachusetts answered by having its legislature choose the electors itself, so every one would go to John Adams.¹²

Each state that consolidated made it costlier for the rest to hold out. By 1836, every state but South Carolina used winner-take-all, and South Carolina’s legislature kept choosing its electors until after the Civil War.¹² Today’s near-uniform system isn’t a requirement or even a design. It’s a standoff that states settled into at voters’ expense, and a fair compact is how they can step out of it together.

Ballot measures raise a narrower question, because Article II names the “Legislature.” Arizona State Legislature v. AIRC read that word in the Elections Clause to include a state’s voters, but the Court was interpreting a different clause.¹³ Expect a challenge to any measure passed that way.

What would a compact look like?

Probably a short, matching law in each state: proportional allocation that takes effect only when the named partner states have also enacted it. That trigger keeps a state that leans one way from disarming alone.

One practical wrinkle: 18 states and DC have enacted the NPVIC, including Colorado, New Jersey and Minnesota from the examples above. The compact stays dormant until it reaches 270,¹⁹ so any member could adopt proportional allocation now, but it would need to reconcile the two laws.

Would Congress have to approve it?

Maybe. The Constitution bars states from entering a compact “without the Consent of Congress.”¹⁷ But the Supreme Court reads that narrowly: consent is needed only for agreements that increase state power at the expense of the federal government (the “vertical” balance), and it has also asked whether an agreement harms states outside it (the “horizontal” balance).¹⁴ Whether the NPVIC crosses that line is hotly debated, and most of the arguments would apply to a Fair Slate compact too.¹⁸ A swing state switching on its own isn’t a compact at all, so the question only comes up for partnered states.

If a Fair Slate compact needs Congress’s approval, the NPVIC certainly does. But the reverse doesn’t follow, because a Fair Slate compact is a much smaller thing:

• Narrower. Two or three states, not a bloc holding enough electors to decide the presidency on its own. That answers the vertical concern.
• Fairer. Each partner gives up about what it gets back, so no group of states gains power over the rest. That answers the horizontal concern.
• Less disruptive. Each state still counts and certifies its own votes, other states’ electors still matter, and if no one reaches 270 the House still decides, as the Twelfth Amendment provides.⁸ Critics of the NPVIC point to exactly those two changes, sidelining non-member states and bypassing the House, as reasons it needs consent.¹⁸

And if approval is needed, a two-state deal that leaves both sides even is a far easier sell in Congress than the NPVIC. Consent comes as an ordinary bill: a simple majority in each chamber and the president’s signature. In the Senate a contested bill can still face a filibuster, but a compact that pairs a red state with a blue one gives neither party’s senators a power grab to object to.

Congress also doesn’t have to vote on every pairing. It can consent in advance to a whole class of compacts, as it did in 1934 for any states cooperating against crime.³¹ A single act could approve any Fair Slate compact that meets set terms, such as proportional allocation, a balanced trade and agreed rules for leaving.

What stops a state from backing out?

A compact isn’t a handshake. It’s a binding contract between states, and the Supreme Court has said a member can’t simply nullify it or decide for itself what it means: “A State cannot be its own ultimate judge in a controversy with a sister State.”¹⁵ Disputes between states go straight to the Supreme Court, which can order a state to comply and has allowed damages for a broken compact.¹⁷¹⁶

Being held to a compact means being held to its terms, including the terms for leaving, so those matter. The NPVIC only blocks withdrawals in the last six months of a president’s term.²¹ A Fair Slate compact can ask for more: for example, a withdrawal that takes effect only after the next presidential election, so no state can switch back once a race is underway.

The trigger is a second safeguard. If a partner does leave, the other state’s law switches back to winner-take-all too, so neither side is left giving up electors alone. Enforcement is firmest when Congress approves a compact, which makes it federal law.¹⁷ A narrow compact may not need that approval, but partners could seek it anyway.

How are the electors calculated?

We use the largest-remainder method. Each candidate gets the whole-number part of their share of the electors, and leftover electors go to the largest fractions. For example, Arizona in 2024 had 11 electors: Trump’s share came to 5.7 and Harris’s to 5.1, so each gets 5 and the last goes to Trump. Other formulas exist and would change a few close results. We count votes for named candidates, not scattered write-ins or Nevada’s “None of these candidates.”

Why not award electors by district, like Maine and Nebraska?

Districts can be gerrymandered, and most are safe for one party. A statewide count has no lines to draw. Every vote in the state counts the same, wherever it’s cast.

Why ranked ballots are a wise addition

Proportional allocation lets more voters win some representation, which is exactly why a few third-party electors appear in the tables above. Ranked ballots keep the benefit and reduce the risk. A voter can rank a third-party candidate first and still have their next choice count once that candidate is eliminated, so fewer electors end up stranded and a House contest becomes less likely. Maine already uses ranked ballots for president.²⁶

We can’t re-run 2016 with ranked ballots, because nobody’s second choices were recorded. As a rough bound, if we count only votes for the two major candidates, 2016 comes out 270 Clinton, 268 Trump. Any real result would depend on how those voters ranked their choices.

How a larger House could sharpen the result

Electors are whole people, so each state’s result gets rounded. The fewer electors a state has, the coarser that rounding is. With 3 electors, each one stands for about 33% of the state, so Wyoming’s 72% vote for Trump in 2024 becomes 2 of 3, or 67%. Pennsylvania’s 19 electors move in steps of about 5%.

Even that coarse rounding is far more accurate than winner-take-all. We measured how far the leading candidate’s share of electors landed from their share of the vote in all 153 state results from the last three elections. Winner-take-all typically missed by 43 points, and by 40 or more in 100 of them. Fair Slate typically missed by 3. Even in states with 5 or fewer electors, the typical miss was 6 points and the worst was 16. Wisconsin in 2024 shows the gap: winner-take-all missed its vote by 50 points, and Fair Slate by less than 1.¹²³

Every state’s elector count is its House seats plus two, so adding House seats gives every state more increments to work with. That raises the resolution of the result everywhere, not just in small states, and lets each state represent fractional vote shares more closely. It also shrinks the share of a small state’s electors that comes from its two Senate seats, which is the root of the Wyoming-versus-California complaint.

It’s the same rounding problem the House already has. Each state’s House seats are its share of the national population, rounded to whole seats, so some districts hold far more people than others. A state’s electors inherit that rounding, and then its result gets rounded again. A larger House shrinks both: seats track population more closely, which makes the House more representative, and every state gets finer steps for its electors, which makes the Electoral College more representative.

The two reforms work on different levels. The House has been capped at 435 by federal statute, so changing it takes Congress, not an amendment.²⁷ Fair Slate works at today’s size, and whenever the House grows, it gets more accurate automatically.