The Arithmetic of the Wasted Vote — What the Spoiler Argument Proves, and What It Cannot Survive

A CONTEMPORARY CODA To the series Parties, Ideologies, and Systems Part 1 of 2

“I never submitted the whole system of my opinions to the creed of any party of men whatever in religion, in philosophy, in politics, or in anything else where I was capable of thinking for myself. Such an addiction is the last degradation of a free and moral agent. If I could not go to heaven but with a party, I would not go there at all.”— Thomas Jefferson to Francis Hopkinson, March 13, 1789¹

The Question Has an Answer

Victor Davis Hanson asked a question this month that he intended as rhetorical. Speaking with Jack Fowler about the gathering at Tucker Carlson’s home in Maine and the ten-point platform that followed it, he said that these were supposedly learned people, and that they should ask themselves when a third party ever got elected, and when a third party ever did anything other than harm the party most akin to it.²

The first half of that question has an answer. The answer is 1860.

The Republican Party was organized in 1854. Six years later it elected a president with 39.8 percent of the popular vote, and the party it displaced — the Whigs, who had elected two presidents and controlled the House — ceased to exist.³ Hanson knows this better than almost anyone alive. He did not forget it. He simply was not counting it, because the Republican Party of 1860 does not feel like a third party to us now. It feels like one of the two. That is the whole point, and it is the point this piece is about.

The second half of the question is harder, and it deserves a serious answer rather than a clever one. That is what follows.

I want to be clear about the terms. Victor Davis Hanson is among the most serious

historians this country has produced, and the classical scholarship that made his reputation will outlast every political argument he or I will ever make. Nothing here is a complaint about him. It is a disagreement with an argument — one he did not invent, one that circulates in every election cycle on both sides, and one that a very large number of Americans have accepted without ever seeing it examined. He gave the argument its clearest recent statement. That is why it is worth taking apart carefully, at full strength, before answering it.

Screenshot of the Victor Davis Hanson show with Jack Fowler (left) as the moderator

The Argument, Stated at Full Strength

Here is the case, assembled and improved.

Third parties do not win in the United States, and the reason is structural rather than accidental. Presidential elections are decided by the Electoral College, which allocates by state, which means a party can take a substantial national share and carry nothing. Ross Perot won 18.9 percent of the popular vote in 1992 and received zero electoral votes.⁴ Congressional elections use single-member districts decided by plurality, which means a party polling a consistent 25 percent across every district in the country wins no seats at all. Duverger’s law describes exactly this: plurality rules with single-member districts converge toward two parties, not as a cultural preference but as an equilibrium.⁵

Layered on top of the structure are ballot-access statutes, public-financing rules, and debate criteria that favor parties already established. Ballot access requires petition signatures gathered state by state under fifty different sets of deadlines and requirements. Public financing favors the incumbents twice over: primary matching funds require geographically distributed fundraising across at least twenty states, and a minor-party nominee generally must clear five percent of the national popular vote before qualifying for a general-election grant paid after the fact. Debate access has run on a fifteen percent polling threshold, set by a commission the two party chairmen created in 1987 and carried forward in modified form by the networks that hosted the 2024 debates. Each rule is neutral on its face and exclusionary in effect, and together they create a trap: a party cannot raise money without exposure, cannot get exposure without polling, and cannot poll without money.⁶

Given all that, what does a third-party vote actually accomplish? On Hanson’s account, one thing only. It moves the outcome away from the candidate the defector prefers and toward the candidate the defector likes least. He offers four cases. In 1912 Theodore Roosevelt bolted and delivered Woodrow Wilson eight years. In 1992 Perot took nineteen percent and denied George H. W. Bush a second term. In 1996 he repeated it and Bob Dole lost. In 2000 Ralph Nader ran as a Green and Al Gore lost Florida by a few thousand votes.

And there is a further claim underneath, which Hanson makes explicit. If the defectors already agree with the incumbent on most of what they want — his estimate was eight of ten points, and eighty to ninety percent of the agenda — then defection is not principle. It is temperament. It is the choice to burn the ninety percent because the ten percent was denied. He put it as a reductio: it is like saying one supports the transgender movement, the Green New Deal, diversity programs, and the open border, because at least it is not Trump. Incoherent, he said. Sad.

That is the argument. It is not a stupid argument. Every structural claim in it is true, and I know they are true because I wrote most of them myself.

Where I Have Made This Case Myself

In October 2025, in an article titled “Parties: The Coalitions That Seek Power,” I laid out the barriers to third-party viability at greater length than Hanson did and reached a conclusion close to his.⁷ The Electoral College punishes broad shallow support. Single-member districts convert consistent minority support into zero seats. Ballot access laws create a chicken-and-egg problem. Matching funds and debate thresholds exclude by rules that look neutral. Strategic voting is not irrationality but rational behavior under those rules. Media coverage reinforces non-viability by treating non-viability as the story.

I ended that section on a paradox: third parties offer ideological coherence and electoral irrelevance, major parties offer electoral viability and ideological incoherence, and voters must choose between influence and purity.

What I did not do in that article was follow the argument all the way down, because that article was about what parties are rather than about what a voter should do. Followed all the way down, the structural case supports a much narrower conclusion than the one it is routinely used to carry. It establishes that third parties rarely win offices. It does not establish that a vote for one is wasted, that defection is incoherent, or that a citizen owes a coalition his loyalty when the coalition has stopped delivering. Those are three separate claims, and only the first is proven.

The same article said something else, in its last paragraph, which is the standard this entire piece applies: work within parties when coalitional membership serves your goals, and work outside them when it does not.

What Is True in It

Before anything else, the concessions.

Nader probably did decide Florida. Gore lost the state by 537 votes. Nader received 97,488.⁸ It requires only a small fraction of those voters preferring Gore to Bush for the counterfactual to flip, and the political science work on the question generally finds that it does flip.

Judge Robert Rosenberg of the Broward County Canvassing Board uses a magnifying glass to examine a dimpled chad on a punch card ballot on November 24, 2000 during a vote recount in Fort Lauderdale, Fla. - Getty Images

The structural barriers are real, and they were built deliberately. The two parties that write the ballot access statutes are the two parties that benefit from them.

And no third party has won the presidency since 1860. That is 166 years.

Now the rest.

The Cases, Examined

1912

Hanson’s account is that Roosevelt created a third party, ran, and ensured Wilson eight years and Taft’s destruction. Every clause is accurate. The framing around it is not.

Wilson took 41.8 percent of the popular vote and 435 electoral votes. Roosevelt took 27.4 percent and 88. Taft — the sitting president of the United States, the nominee of the Republican Party — took 23.2 percent and 8, the fewest electoral votes any incumbent has ever received.⁹

Read that ordering again. The third party did not spoil the major party. The third party beat the major party, by more than four points and by eighty electoral votes. By Hanson’s own definition, in which the spoiler is the candidate who splits a natural majority and hands victory to the opposition, the spoiler in 1912 was William Howard Taft.

There is a further problem, and it is the one that matters most. 1912 was not a case of an outsider bolting a party that had rejected him at the ballot box. It was the first presidential cycle in which a large number of delegates were chosen by direct primary, and Roosevelt won nine of the twelve states that held them, including Ohio, which was Taft’s home state.¹⁰ Taft won one: Massachusetts. Roosevelt was the choice of Republican voters wherever Republican voters were actually asked.

He was not the choice of the machine. Taft’s delegate strength came from state conventions and from Southern delegations sustained by federal patronage in states the Republican Party had no prospect of carrying in November — those delegations backed Taft roughly five to one. When ninety-two contested delegate seats went before the Republican National Committee, the committee awarded every one of them to Taft.¹¹ Roosevelt walked his delegates out of the convention on June 22.

So the sequence is this. Republican voters chose Roosevelt. Republican officials overrode them. Roosevelt ran anyway, finished ahead of the man the officials had installed, and the party lost. Hanson presents this as a cautionary tale about defection. It is at least as good a cautionary tale about what happens when a coalition refuses to be a coalition — when the mechanism for aggregating factions stops aggregating and starts excluding. The defection was the consequence. The exclusion was the cause.

Of the forty-eight states, only twelve produced a candidate with an actual majority of the popular vote.¹² Thirty-six states elected presidential electors with pluralities. The system did not produce a national majority in 1912. It produced a winner. Those are not the same thing, and the difference is the subject of this piece.

1992 and 1996

Hanson says Perot got twenty percent and ensured that George H. W. Bush would not be reelected.

The arithmetic does not work. Clinton took 43.0 percent, Bush 37.5, and Perot 18.9.¹³ The gap between Clinton and Bush was 5.5 points, or nearly six million votes. To overtake Clinton in a two-way race, Bush would have needed roughly two-thirds of everyone who voted for Perot.

He did not have two-thirds of them. In the Election Day exit poll, Perot voters asked for a second choice split evenly: 38 percent named Bush, 38 percent named Clinton.¹⁴ A later reallocation of the Voter Research Survey data by state found Perot voters leaning toward Clinton, 51 to 42.¹⁵ Both figures are far from what Bush required.

Here is where I will do something Hanson did not, and state the case against my own reading. There is serious work arguing the opposite. Alvarez and Nagler, analyzing the 1992 National Election Studies data with a multinomial probit model, concluded that Perot did draw more from Bush than from Clinton.¹⁶ Sean Trende has argued, persuasively, that the exit-poll answer is too simple — that Perot spent the spring departisanizing the critique of Bush, driving the incumbent down in a way a wounded Clinton could not have done, and that the honest answer is that we cannot know.¹⁷

Grant all of it. The arithmetic still holds. A directional lean is not two-thirds. The most that the strongest version of the counterargument establishes is that Perot’s presence affected the shape of the race, which is a claim about persuasion rather than about vote-splitting, and which is a description of what a political campaign is for.

1996 does not support Hanson’s claim at all. Clinton took 49.2 percent, Dole 40.7, Perot 8.4.¹⁸ Dole lost by 8.5 points while Perot received 8.4. Even a total transfer of every Perot vote leaves the outcome unchanged. Hanson raises 1996 not because it demonstrates

anything but because it is on the list, and lists acquire authority from length.

2000

Conceded above. Florida by 537 votes with Nader at 97,488 is the case the spoiler argument was built for, and it deserves to be named plainly rather than explained away.

It is also the cleanest of the four. The 1912 split may well have elected Wilson; the combined Taft and Roosevelt vote exceeded his. What does not follow is that Roosevelt rather than the convention is answerable for it.

The Arithmetic

The wasted-vote charge rests on a claim about consequence, and a claim about consequence can be measured. Somebody already has.

Andrew Gelman, Nate Silver, and Aaron Edlin computed the probability directly for the 2008 presidential election. A vote is decisive when the voter’s state proves necessary to an Electoral College majority and the vote within that state is tied. Multiplying those two probabilities across state-by-state forecasts, they found that in the states where a single vote mattered most — New Mexico, Virginia, New Hampshire, and Colorado — the chance of determining the national outcome was approximately one in ten million. Averaged across the country, it was one in sixty million.¹⁹

That figure sits badly against something conceded earlier in this piece, and the tension has to be settled before the figure can be used. If no vote decides an election, how did 97,488 votes decide Florida?

Both statements are true, and the distance between them is where the spoiler argument lives. Elections are decided by blocs. Ballots are cast by individuals. A bloc of ninety-seven thousand can be the difference in a state carried by 537 votes while no member of that bloc holds more than a sixty-millionth of the result. The first fact does not license the second.

The spoiler argument moves back and forth across that gap without marking the crossing. It establishes consequence at the level of the bloc and then assigns blame at the level of the person, as though the lever in one Broward County voter’s hand had been the only one wired to the result.

Once blame is assigned that way, the selection becomes visible. Exit-poll estimates from Florida in 2000 put roughly 24,000 self-identified Democrats behind Nader and roughly 308,000 behind Bush. Roughly 191,000 self-identified liberals voted for Bush, against fewer than 34,000 who voted for Nader.²⁰ Every one of those blocs was large enough to have produced 537 votes many times over. Remove any of them, or redirect it, and the result changes. Only one of them is remembered as the spoiler.

The same holds for every other cause of that margin — the butterfly ballot in Palm Beach County, the purge lists, the recount standard, the weather in the panhandle. A margin of 537 votes out of nearly six million does not have a cause. It has several thousand of them. The one selected for blame is the one that had somewhere else to go, which is to say the one that can be disciplined by being blamed.

So the concession stands. Nader’s presence plausibly changed the outcome in Florida. What does not follow is that the Nader voter did something the Bush-voting Democrat did not, or that the lesson is that a citizen owes a party his ballot.

What a Vote Can Say

If individual decisiveness is not what makes a vote matter, something else has to be, and the honest answer is aggregation. Blocs move parties. A shift of two points in Pennsylvania is measured, studied, and answered, even though no individual in Pennsylvania moved it.

That is correct, and it is the ground on which the wasted-vote charge fails on its own terms, because once the question becomes what a bloc communicates, the two kinds of vote stop being equivalent.

A major-party vote carries an instruction and no explanation. It says this one rather than that one. It cannot separate the voter who wanted the candidate from the voter who feared the alternative, the convinced from the resigned, the person who endorsed the platform from the person who swallowed nine-tenths of it to stop something worse. The party receives a number, and a number of that kind gets read as approval, because approval is the flattering reading and nothing in the data argues against it. A voter who wanted lower deficits and no new wars, and who then watched the man he elected open a war and preside over a debt that crossed forty trillion dollars, has already spent the only instrument he had. He spent it saying yes.

A vote cast outside the two parties says something a coalition cannot round off: a person was here, in this precinct, who was available and was not obtained. That is information of a kind the machine is built to read. A coalition is not a creed. It is an apparatus for assembling winning combinations, and it revises itself according to the combinations it finds it can and cannot assemble.

Defection is the input. Loyalty is silence.

Two Different Measures

Hanson’s argument runs entirely on election results. It has to, because election results are the only evidence that supports the position. Set them beside a second measure and the picture inverts.

An election is a forced ordinal choice between two names on a specific Tuesday. Approval is an unforced evaluation of performance, asked repeatedly, with no alternative required. They are different instruments measuring different things, and they routinely disagree.

Start with the vote shares. Lincoln won in 1860 with 39.8 percent. Wilson won in 1912 with 41.8. Clinton won in 1992 with 43.0. George W. Bush won in 2000 with 47.9 percent while losing the popular vote outright.²¹ These are not obscure cases. They are four of the most consequential presidencies in American history, and none of the four men entered office with the support of a majority of the people who voted, let alone of the people eligible to.

Now the evaluations. Gallup asked the same approval question from the Truman administration until it retired the series in December 2025. The average across all presidents is 53 percent. John Kennedy holds the highest term average at 70.1 percent. Six presidents have averaged 50 percent or better; eight have not, including each of the last four.²²

And then the case that makes the point better than any argument could. George H. W. Bush averaged 61 percent approval across his single term — third highest in the entire Gallup series, behind only Kennedy and Eisenhower — and was defeated for reelection. The average is not the whole story, and the rest of it is the point. His ratings soared during and after the Gulf War and then collapsed; by October 1992 Gallup put him in the mid-thirties.²³ One man, one term, a swing of roughly fifty points, and a single Tuesday in November that recorded one instant of it. Hanson cites 1992 as proof that a third party overrode the public will. What the record shows is that the public’s evaluation of Bush moved further in eighteen months than the entire margin between the two parties in most elections, and that an election result captures none of that motion. Approval and electoral outcome are not the same measure. Bush is the proof, and he is Hanson’s own example.

The current numbers are where the structural argument lands. In the Reuters and Ipsos survey conducted August 21 through August 24, 2026, 33 percent of American adults approved of President Trump’s job performance and 65 percent disapproved. Support for the war with Iran stood at 31 percent, down from 37 in March, with 83 percent expecting the fighting to continue for an extended period.²⁴ Congressional approval in Gallup’s last readings before the series ended sat at 14 percent against 80 percent disapproval.²⁵ Asked how they would vote in a congressional election held today, independents favored Democrats over Republicans by 33 percent to 19 — which is to say that a majority of independents named neither.²⁶

Trump entered his first term with a 45 percent approval rating, the lowest first reading for any incoming president in the series, and finished that term with a four-year average of 41 percent, the lowest full-term average on record.²⁷ Biden’s term average was 42.2, the second lowest.²⁸ The two lowest-rated presidencies in the eighty-year history of the measurement are the two most recent completed ones, back to back, produced by the same binary.

A president approved of by a third of the country. A Congress approved of by fewer than two in ten. A war approved of by three in ten, six months after it began. Two consecutive record-low presidencies. And a mechanism whose defenders insist that the correct response is to choose between the two organizations that produced all of it, and to regard any other choice as a failure of seriousness.

An election result tells you which of two names got more marks. It tells you nothing about whether the person who made the mark wanted to. When a mechanism reliably produces officeholders whom large supermajorities disapprove of, the failure is not located in the voters who declined to participate in the choice. It is located in the choice.

Coalitions Are Instruments, and Instruments Get Discarded

The deepest problem with Hanson’s question is the assumption buried in it — that the two parties currently on the ballot are permanent features of the landscape, and that a citizen’s options are therefore to pick one or to opt out of consequence.

Parties are not permanent, and the historical record on this is not ambiguous. A party is a coalition of factions that align because none of them can win alone, and coalitions dissolve when their members conclude they would do better elsewhere. The Federalists are gone. The Democratic-Republicans are gone. The Whigs are gone, and they went in a single decade, because the slavery question cut across their coalition rather than along it and the coalition could not metabolize it.

Nor is this an American peculiarity. The British Liberal Party was one of the two governing parties of the later nineteenth century and was displaced as the principal party of the left by Labour within roughly a quarter-century of Labour’s founding. The Progressive Conservative Party of Canada held a majority government in 1988 and won two seats in 1993 — two, out of 295. In France, the presidency belonged for decades to the Socialists and to the Gaullist and center-right formations that eventually became Les Republicains; in 2022 their nominees took 1.7 and 4.8 percent, and the runoff was reached by a movement founded in 2016.²⁹ Coalition mechanics are not a quirk of the American constitution. They are what happens to coalitions.

So when Hanson asks when a third party ever got elected, the accurate answer is that the question misdescribes the process. Third parties do not usually get elected. They get absorbed, or they trigger a realignment in which one of the two incumbents is replaced and a new duopoly forms around the new division. That is 1854 to 1860 exactly: Free Soilers and anti-slavery Whigs and anti-slavery Democrats, a party organized in 1854, a second-place finish in 1856, the presidency in 1860, and the Whigs erased.

The forms of victory that follow are the ones Hanson does not count. The Populists lost every presidential election they entered, and within a generation several of the demands in their 1892 platform — a graduated income tax, direct election of senators, stronger federal control of the railroads — had been adopted or carried a long way forward by the parties that beat them.³⁰ Counting only offices won is like measuring a strike by whether the picketers got hired.

Hanson’s own subject matter says this more forcefully than I can. He has spent a career explaining that institutions which stop performing their function do not persist out of respect for their history. Applied to the American party system, the observation cuts the way it cuts everywhere else.

Defense and Counterargument

The strongest objection to everything above is not one Hanson raised, so I will raise it.

Duverger’s law cuts against me, and 1860 is not the precedent I want. The Republican Party did not add a third pole to American politics. It replaced one and restored a two-party equilibrium within a decade. If the structural logic is as strong as I have said, then any successful new party is not a break from the duopoly — it is the next duopoly, with the same incentives, the same coalition arithmetic, and the same eventual disappointment for whichever faction stops being electorally valuable. A libertarian who imagines that a breakthrough would produce a durable multiparty system in which his faction is permanently represented has misread the mechanism he is invoking.

I think that objection is correct, and I do not have a rebuttal to it. What I have is a narrower claim, which is the one I am actually making: replacement is still change, and change is still the only thing that has ever moved a party off a position it had grown comfortable with. The Whigs being replaced by Republicans did not end coalition politics. It ended the political viability of the Slave Power’s national veto, which was the point.

The spoiler risk is real and 2000 proves it. Also correct. A voter in a genuinely close state, weighing two candidates with materially different consequences, is making a real trade and should not be told the trade is imaginary. What he should be told is the size of it — on the 2008 map, roughly one in ten million in the most pivotal states, varying by election and by state — and then be left to weigh it himself rather than be shamed out of the calculation.

The honest resolution is structural, not moral. The tension between expressing a preference and affecting an outcome is manufactured by plurality voting, and it dissolves under approval voting, ranked-choice voting, proportional allocation of electors, or fusion balloting, all of which permit a voter to state a first choice without abandoning a hedge. Anyone who genuinely believes the spoiler problem is serious should be working on the counting rule. That almost none of the people who deploy the spoiler argument have any interest in changing the counting rule is worth noticing. The argument is not usually offered as a problem to be solved. It is offered as a reason to stop asking.

And the last objection, which is Hanson’s real one. If the defectors agree with the incumbent on eighty or ninety percent, defection is disproportionate. That claim is empirical, it is testable against a published document, and it is the subject of the companion to this article (Part 2).

The Liberty Test

In the Self-Evident series, every episode ends against a single question: did this leave Americans more free, or more governed? It is worth applying here, with the honest limit stated first — the test does not resolve close calls, and it does not tell a voter in a swing state what to do on a Tuesday.

What it does resolve is the framing. A rule that tells a citizen his only legitimate options are the two coalitions currently holding ballot access, and that any third choice is a moral failure that makes him responsible for the outcome, is not a description of how elections work. It is a claim on his consent that he never granted. He is more governed for accepting it, and no freer for the acceptance.

What Part 2 Takes Up

Hanson made a second argument, longer than the first, and it was not about elections at all. It was about motive. Carlson broke with Trump because he was cut off from Air Force One and lost the president’s cell phone number. Greene broke because she read Georgia polling. Massie broke because he was primaried. Kent broke because his appointment was in jeopardy. On this reading, none of the four has a grievance. All four have wounded vanity, and the platform is the rationalization.

Part 2 tests that claim the only way it can be tested — against the ten points Carlson actually published on August 5, 2026, and against the record of what this administration has actually done on debt, on transparency, on war, and on the relationship between the state and private industry. Hanson’s estimate was eight of ten. The document is public. We can count.

Self-Reflection Prompts

1. Before reading this piece, what did you take the phrase “wasted vote” to mean? Having seen the decisiveness figures, does the phrase still describe something real — and if so, what?

2. In 1912, Republican voters chose Roosevelt in nine of twelve primaries and Republican officials nominated Taft. Which of those two groups bears responsibility for what followed in November? Does your answer change depending on which party the example involves?

3. George H. W. Bush averaged 61 percent job approval and was defeated for reelection. What does that pairing tell you about what an election result actually measures?

4. If you have ever voted for a candidate you did not want in order to prevent an outcome you feared, how did the party you voted for learn that this was what happened? What mechanism existed for you to tell them?

5. The Republican Party replaced the Whigs and then became half of a new two-party system. If a new party displaced one of the current two, how long do you expect it would take before it disappointed you in the same way, and does that expectation change what you should do now?

6. Did the forced choice between two coalitions leave you more free, or more governed?

Sources and Further Reading

Andrew Gelman, Nate Silver, and Aaron Edlin, “What Is the Probability Your Vote Will Make a Difference?” Economic Inquiry 50, no. 2 (April 2012): 321–326.

Maurice Duverger, Political Parties: Their Organization and Activity in the Modern State (1954).

R. Michael Alvarez and Jonathan Nagler, “Economics, Issues, and the Perot Candidacy: Voter Choice in the 1992 Presidential Election,” American Journal of Political Science (1995).

Gallup, Presidential Job Approval Center (series retired December 2025).

Federal Election Commission, “Understanding Public Funding of Presidential Elections.”

American Presidency Project, election returns and Gallup approval series.

Jeff Kellick, “Parties: The Coalitions That Seek Power,” Consequential Actions, October 21, 2025.

Endnotes

1. Thomas Jefferson to Francis Hopkinson, March 13, 1789.

2. Victor Davis Hanson, in conversation with Jack Fowler during the podcast “Victor Davis Hanson: The FBI’s Secret Trump Probe, Birthright Citizenship, and the WNBA’s Race War,” August 11, 2026, 01:05:01

3. Abraham Lincoln received 39.8 percent of the popular vote in 1860. On the Republican Party’s founding in 1854 and the collapse of the Whig coalition, see Kellick, “Parties: The Coalitions That Seek Power,” October 21, 2025.

4. Perot’s 1992 share and zero electoral votes.

5. Maurice Duverger, Political Parties: Their Organization and Activity in the Modern State (1954). Duverger described a strong tendency rather than an exceptionless law; the distinction matters, since India and Canada sustain multiparty competition under plurality rules.

6. On ballot access generally, see Kellick, “Parties: The Coalitions That Seek Power,” October 21, 2025. On public financing, see the Federal Election Commission: primary matching funds require more than $5,000 in matchable individual contributions in each of at least twenty states, while a minor-party nominee qualifies for a proportional general-election grant by receiving at least five percent of the popular vote. The five percent figure applies to the general-election grant, not to primary matching funds; the description in the October 2025 article conflates the two and should be corrected there. No major-party nominee has accepted the general-election grant since 2008. The fifteen percent debate threshold was a rule of the Commission on Presidential Debates, created in 1987 by the chairmen of the Republican and Democratic national committees, and was carried forward in modified form by CNN and ABC for the 2024 debates; ABC also required ballot access sufficient to reach 270 electoral votes.

7. Kellick, “Parties: The Coalitions That Seek Power,” Consequential Actions, October 21, 2025, section titled “Third Parties: Ideological Purity and Electoral Futility.”

8. Florida 2000 certified margin of 537 votes; Nader’s Florida total of 97,488.

9. 1912 general election, American Presidency Project returns: Wilson 6,294,327 (41.8 percent, 435 electoral votes); Roosevelt 4,120,207 (27.4 percent, 88); Taft 3,486,343 (23.2 percent, 8); Debs 900,370 (6.0 percent, 0). Totals vary across datasets by minor-party and state-return conventions; this article uses the American Presidency Project throughout. Taft’s eight electoral votes are the fewest ever received by an incumbent president.

10. Twelve states held Republican presidential primaries in 1912. Roosevelt carried nine, including Ohio, Taft’s home state; Taft carried Massachusetts; La Follette carried North Dakota and Wisconsin.

11. The Republican National Committee’s credentials process awarded all ninety-two contested delegate seats to Taft. Delegations from the former Confederate states supported Taft by approximately five to one. Roosevelt directed his delegates to withdraw from the convention on June 22, 1912.

12. Of the forty-eight states voting in 1912, only twelve produced a candidate winning an outright majority of that state’s popular vote.

13. 1992 general election: Clinton 43.0 percent, Bush 37.5 percent, Perot 18.9 percent.

14. Election Day exit polling on Perot voters’ second choice: 38 percent Bush, 38 percent Clinton.

15. A 2023 state-by-state reallocation of the 1992 Voter Research Survey exit-poll data, published by Split Ticket, found Perot voters’ second choices at 51 percent Clinton and 42 percent Bush, with the remainder naming another candidate or no vote. This is a later secondary analysis of the VRS dataset, not a contemporaneous finding by the original pollsters.

16. R. Michael Alvarez and Jonathan Nagler, “Economics, Issues, and the Perot Candidacy: Voter Choice in the 1992 Presidential Election,” American Journal of Political Science 39, no. 3 (August 1995): 714–744, finding that Perot drew more votes from Bush than from Clinton.

17. Sean Trende, “We Don’t Know Whether Perot Cost Bush in 1992,” RealClearPolitics, July 10, 2019.

18. 1996 general election: Clinton 49.2 percent, Dole 40.7 percent, Perot 8.4 percent. On the reported totals, Clinton received 47,401,185 votes while Dole and Perot together received 47,282,763, leaving Clinton ahead by roughly 118,000 even on a complete transfer.

19. Andrew Gelman, Nate Silver, and Aaron Edlin, “What Is the Probability Your Vote Will Make a Difference?” Economic Inquiry 50, no. 2 (April 2012): 321–326.

20. Exit-poll estimates for the Florida 2000 vote: approximately 24,000 self-identified Democrats voted for Nader; approximately 308,000 voted for Bush; approximately 191,000 self-identified liberals voted for Bush against fewer than 34,000 for Nader. Jim Hightower, “How Florida Democrats Torpedoed Gore,” Salon, November 28, 2000, attributing the calculations to Tim Wise. Hightower is a partisan columnist and the underlying computation is not shown; these are survey estimates of party identification, not administrative counts of voter registration.

21. Popular vote shares: Lincoln 39.8 percent (1860); Wilson 41.8 percent (1912); Clinton 43.0 percent (1992); George W. Bush 47.9 percent (2000), trailing Gore in the national popular vote.

22. Gallup: average job approval across all presidents in the series is 53 percent; Kennedy holds the highest term average at 70.1 percent; six presidents averaged 50 percent or higher and eight below.

23. Gallup: George H. W. Bush’s term average of approximately 61 percent, third highest in the series. Contemporaneous readings show the collapse: 33 percent approval against 58 percent disapproval on October 1–3, 1992, and 34 against 56 on October 13–15. See the American Presidency Project’s Gallup series for Bush.

24. Reuters/Ipsos national survey of 1,215 U.S. adults, conducted August 21–24, 2026, margin of error 3 percentage points: 33 percent approval and 65 percent disapproval of President Trump; 31 percent support for U.S. military action against Iran, down from 34 percent earlier in August and 37 percent in March, with 69 percent of Republican respondents in support against 77 percent in March; 83 percent expecting the war to continue for an extended period. Reported by Jason Lange, Reuters, August 24, 2026. U.S. and Israeli strikes on Iran began February 28, 2026.

25. Gallup congressional job approval at 14 percent against 80 percent disapproval.

26. Reuters/Ipsos, August 21–24, 2026: asked how they would vote in a congressional election held today, independent respondents favored Democrats over Republicans, 33 percent to 19. The midterm elections fall on November 3, 2026.

27. Gallup: Trump’s initial job approval of 45 percent in January 2017, the lowest first reading for any incoming president in the series; four-year first-term average of 41 percent, the lowest full-term average in Gallup’s history; final first-term reading of 34 percent.

28. Gallup: Biden’s full-term average of 42.2 percent, the second lowest in the series.

29. The Liberal Party of Great Britain was formally constituted in 1859 and was displaced as the principal party of the left by Labour, founded in 1900. The Progressive Conservative Party of Canada fell from a majority government in 1988 to two seats of 295 in 1993. In the 2022 French presidential first round the Socialist and Republicain nominees received 1.7 and 4.8 percent respectively; La Republique En Marche was founded in 2016.

30. On the Populist platform of 1892 and its subsequent adoption, see Kellick, “Parties: The Coalitions That Seek Power,” October 21, 2025. Note the chronology: federal railroad regulation began with the Interstate Commerce Act of 1887, before the Populist platform, and the Sixteenth and Seventeenth Amendments were ratified in 1913. The claim is one of adoption and advancement by the major parties, not of exclusive Populist causation.