Skip to content

Forced vs. Voluntary Exhaustion — two exhausted ballots that deserve different names

Two very different things get lumped under the single word "exhausted." One is a voter's own choice and no cause for complaint. The other is the ballot design failing a voter who did everything right. They deserve distinct names — because collapsing them lets the whole problem be waved away as "well, that was the voter's choice."

This page is the fine-grained companion to Exhausted (Inactive) Ballots.

The two kinds

Voluntary exhaustion — self-truncation. You chose to rank fewer candidates than you were allowed (maybe only your favorite). If those candidates are eliminated, your ballot exhausts. That's the direct, foreseeable consequence of your own choice — a legitimate thing to do, and not a flaw in the method. (Standard term: voluntary truncation.)

Forced exhaustion — the ballot dropped out even though you ranked every candidate you were permitted to. You were as expressive as the rules let you be, and the count still discarded your ballot. This one isn't your choice — it's imposed by the ballot design. (Standard term: involuntary truncation. "Forced exhaustion" is this library's house label for it, because "involuntary truncation" buries the point — the voter didn't truncate anything; the ballot did.)

The key fact that makes "forced" the right word

In single-winner RCV-IRV, a fully-ranked ballot can never exhaust. If you rank every candidate, one of them is always still in the race until the final winner is chosen — so there's always somewhere for your vote to go.

Therefore forced exhaustion can only happen under a ranking cap — when the ballot lets you rank fewer candidates than are running (New York City: 5; Minneapolis: 3). You rank all 5 you're given; if all 5 are eliminated before the final round while 8 other candidates fight on, your ballot exhausts — not because you held back, but because the ballot wouldn't let you rank the rest.

The consequence is clean and important:

Forced exhaustion is a property of the ballot design (the ranking cap), not of the voter. Lift the cap — let voters rank the whole field — and forced exhaustion in single-winner IRV disappears entirely. Voluntary exhaustion is the only kind left, and that one is a choice.

(Caveat: this "fully-ranked can't exhaust" guarantee is specific to single-winner IRV. Multi-winner STV can exhaust even a fully-ranked ballot through quota/surplus transfers — a different mechanism.)

A worked micro-example

Six candidates — A B C D E F — but the ballot caps you at 3 rankings.

  • A voter's true order is C > E > A > B > D > F. They rank the top three they're allowed: C, E, A.
  • Rounds eliminate F, then D, then C, then E, then A — all three of the voter's ranked choices are gone before the final B vs. D… wait, D is already out; say the final is B vs. F-survivor. The point holds: once C, E, A are all eliminated, the ballot has no continuing candidate and exhausts — even though this voter would gladly have ranked B fourth and had a say in the final round.

Their neighbour who chose to bullet-vote C only also exhausts — but that was a choice. Same outcome on the tally sheet; completely different stories, and only one is the system's fault. That's why one word for both is misleading.

A runnable, BetterVoting-backed version of exactly this — where forced exhaustion discards more ballots than the winner receives — is the Forced Exhaustion Ceiling case (read as a ceiling, not a typical election).

Why the distinction matters

  • It changes who's responsible. "Exhaustion is just voters choosing not to rank everyone" is only true of the voluntary kind. Forced exhaustion is the method (plus a cap) discarding a maximally-expressive ballot — and it's fixable by removing the cap.
  • The cap makes it common where it's least visible. In a large field the cap bites hardest. New York City's 2021 mayoral primary ran 13 candidates against a 5-rank cap; a large share of the exhaustion there was forced, not chosen.
  • It's uneven. Voters who rank an underdog or a minor candidate first burn their limited slots on candidates who get eliminated early — so the cap force-exhausts their ballots more than mainstream voters', quietly tilting the result.

Methods that don't force-exhaust

  • Ranked Robin (Condorcet) reads every ballot in every pairwise contest — no eliminations, so nothing exhausts, forced or voluntary.
  • Scored methods (STAR, Score, Approval) never drop a ballot: a blank reads as 0 and still counts. (A different trade-off — a 0 is a real lowest score, not a neutral abstention — but the ballot never leaves the count.) See scores vs. ranks.