A unanimous Court let a state fine an elector for how he cast a vote in a federal proceeding. Put on the same two-axis rack as Trump v. Anderson, Chiafalo fails both: it’s directionally wrong (it crosses the manner/substance line that vindicates Anderson, and kills the institution’s only reason to exist) and its reasoning is unreasonable (liquidation borrowed from the wrong practice). The keystone of the trilogy — the case that created the orphaned-elector problem Anderson tripped over.
Links: Originalism and the Limiting Document (the keystone — this is a construction-zone failure), Trump v. Anderson (the manner/substance line, here violated), Wickard v. Filburn (the third sibling in the same outcome-engineering pattern), Registration and Scope Creep, The Gödel Governance Problem
Chiafalo v. Washington, 591 U.S. ___ (2020), Justice Kagan, unanimous; Thomas concurring in the judgment (Gorsuch joining in part). Companion: Colorado Dept. of State v. Baca (same day) — Colorado had removed and replaced a faithless elector; the 10th Circuit ruled for the elector and was reversed.
2016: three Washington “Hamilton electors,” pledged to Clinton (the state’s popular winner), voted Colin Powell instead, hoping to spark cross-party defections. Washington fined each $1,000. Held: the state may do this.
The whole federal election-law tradition protects the act of voting (Yarbrough, Classic, Reynolds, Burdick). It doesn’t transfer here because the Court treats the elector as a delegated public office, not a citizen exercising a personal right: he casts the state’s electoral votes as the state’s instrument, and an officeholder can be sanctioned for not performing the office. That instrumentality theory is the load-bearing wall — and it’s the same wall that will collapse below.
Correction to a common overstatement: Chiafalo blessed a civil fine and (in Baca) removal/replacement. It did not bless imprisonment — Kagan expressly reserved whether harsher or criminal sanctions raise other constitutional problems. Jailing a citizen for a ballot probably would collide with the franchise cases; that question is open.
The Anderson analysis is vindicated by a clean rule: a state controls the manner/instrument (who is appointed, the ballot mechanics) but never the substance of the federal function. Anderson is directionally right because it enforces that line — Colorado may not make the substantive federal eligibility call.
Apply the same line to Chiafalo and it condemns the holding. Coercing how an elector votes is not manner-control; it is substance-control of a federal act. Once appointed, the elector casts the 12th-Amendment vote that is transmitted and counted by Congress — a federal constitutional proceeding. Penalizing the content of that vote is the state reaching into federal substance: the very thing Anderson struck Colorado down for. The appointment power is spent at appointment; it does not include a remote control over the federal act that follows.
So the two cases do not reconcile on the tidy line — the line that saves Anderson convicts Chiafalo. And measured against purpose, the direction is wrong twice: the Electoral College’s only distinctive reason to exist is interposed judgment; bind the electors and you keep the shell while killing the function.
Liquidation is a legitimate tool, but it requires the specific practice to be long, open, and unchallenged. Watch which practice actually has the pedigree:
Kagan borrowed the legitimacy of the old voluntary-fidelity tradition to license a new coercive-power conclusion with no comparable lineage. Liquidation establishes “electors usually choose to follow the party”; it does not establish “states may punish those who don’t.” The argument switches objects mid-stride.
Thomas’s 10th-Amendment route has its own hole: the Amendment reserves powers “not delegated to the United States” — but conducting the presidential vote is delegated to the federal sphere (it’s a federal proceeding states merely participate in by appointing). It isn’t a reserved domain.
Tempting to reach for original intent — but the Anti-Federalists cut the other way. George Mason attacked the EC on practical grounds (ill-informed electors; he predicted the House would decide “nineteen times out of twenty”); Federal Farmer accepted it; Anti-Federalist No. 72 wanted more popular control, less elector independence. The robust deliberative-elector vision is essentially Hamilton’s alone (Federalist 68). So the Court couldn’t even claim original intent for discretion — and resting on the papers is a trap in either direction. The discretion case is stronger from the text than from anyone’s commentary. (The Federalist/Anti-Federalist papers are argument and expectation, not operative law — a discipline worth keeping against people who quote them as gospel.)
This is an argument a textualist should accept — and one did.
Ray v. Blair (1952) held only that a state party may require a pledge as a condition of running for elector, and expressly reserved whether the pledge is enforceable (Reed assumed it might be “legally unenforceable” and upheld the requirement anyway). It decided the easy half and ducked the hard half. Chiafalo leapt the reserved question — pledge-requirement-OK became coercion-OK across a 68-year gap, with no intervening authority closing it.
The dissent in Ray is the canonical statement of the other side — Justice Robert Jackson (joined by Douglas):
“No one faithful to our history can deny that the plan originally contemplated, what is implicit in its text, that electors would be free agents, to exercise an independent and nonpartisan judgment as to the men best qualified…”
The irony seals it: Jackson, the author of Wickard, was the textualist about elector independence. The man who bent “commerce” past breaking refused to bend “elector.”
The pledge, properly read, is a promise to discharge the duty of the office — and the duty is to vote. A faithless elector votes; he discharges the duty. What he declines is to obey the state’s preference about the content — which was never the duty; it is the discretion the office is made of. Executing the state’s will was finished at appointment.
So Kagan’s “appoint includes condition and enforce” equivocates between:
The exact historical proof: pre-17th-Amendment U.S. Senators were appointed by state legislatures, which could instruct them — but the doctrine of instruction was political, never legally coercive. Defiant senators faced the ballot box, not a fine. Same structure (state appoints a federal officeholder), and the Founders’ answer was: you may pick and pressure, you may not fine the federal vote. Chiafalo grants states a power over electors they never had over identically-appointed Senators.
| Direction | Reasoning | |
|---|---|---|
| Wickard | wrong (guts enumerated powers) | unreasonable (contradicts clear text) |
| Anderson | right (structurally sound) | unreasonable (§5 invented) |
| Chiafalo | wrong (substance-control; kills the EC’s purpose) | unreasonable (liquidation bait-and-switch) |
Chiafalo lacks Anderson’s saving grace: Anderson reached the constitutionally correct place by a bad road; Chiafalo took a bad road to the wrong place — a holding floating above its own foundations (Ray reserved the very question, the founding-opinion record doesn’t supply the rationale, the text supplies the opposite, the structure forbids the coercion).
The one honest caveat — it isn’t Wickard-brazen. The operative text doesn’t guarantee discretion in so many words, so this is resolving a real ambiguity (wrongly) rather than contradicting clear text. Which means it bottoms out at a philosophy-of-interpretation fork:
All three cases run the same move: the result is fixed first and the text is bent to reach it — expanding power (Wickard, to save the New Deal), contracting a constitutional bar (Anderson’s §5 holding, to avoid electoral chaos), or coercing a federal function (Chiafalo, to avoid 538 loose cannons). Outcome-engineering is direction-agnostic. And Chiafalo is the keystone: by converting the elector from deliberative principal into pure relay, it is the case that orphaned the eligibility-screening function — the orphan Colorado then tried to adopt in Anderson. The honest discipline the whole pattern lacks is the same one each time: when the structure no longer serves, amend it; don’t gut it and keep the shell.