This morning I published a ranking of all fifty-six Speakers of the House, built on a rubric I stated in full: tenure, legislative output, institutional change, each graded zero to four, summing to twelve, emitted as a three-digit profile string so anyone can attack the arithmetic. In that table, Mike Johnson is twenty-seventh, with a profile of 220 and a score of four out of twelve.

I put the number in the paper because I believed it. I stand by the method. And I want to use this space to say plainly that the number is close to worthless as a statement about the man, that the reason is a specific defect in my own rubric, and that the defect has a shape I should have seen before I shipped it.

What twenty-seventh actually says

Begin with what the rank means arithmetically, because it is less flattering than it sounds. Twenty-seven of fifty-six is the median, almost exactly. He is tied with eight other Speakers on a score of four, and the band runs from rank twenty-four to rank thirty-two; the order within it is alphabetical, which is why he lands where he does. Sort that band by institutional change instead, and he drops to the bottom of it, at thirty-second.

His profile is 220. Two points of tenure, for the two years and eleven months he has held the gavel. Two points of legislation, for the two statutes I could verify as passing the House under his floor control and becoming law: the Laken Riley Act and the One Big Beautiful Bill Act. Zero points of institutional change, which in my scales means no permanent rule, precedent, or power of the office that he created or destroyed.

Hold the third digit in your mind, because it is the whole argument. I assigned it zero, and I will come back to why that was the mistake that matters.

A rubric that can only count production

The three components share a property I did not notice while I was writing them. Every one of them measures something a Speaker produced. Years served. Statutes passed. Rules written. A Speaker who does a great deal scores high and a Speaker who does very little scores low, and it is starkest at the bottom of my table, where the six men with a profile of 100 have nothing but the first digit to separate them, and Theodore Pomeroy, who served one day, sits below all of them on 000.

That is a real bias and I flagged it in the paper: the rubric “rewards the measurable,” and the Speakers who exercised power by preventing action “score low by construction.” I named Polk’s gag rule and Johnson’s recess as the examples.

But a limitation that says “this table cannot see obstruction” is not the same as an answer to it. I wrote the caveat and then let the number stand next to Johnson’s name anyway, where it does a different kind of work. A score of four out of twelve reads as mediocre, and mediocrity is a category of production. It is the score of a man who showed up and did a middling job. That is not what happened here, and my own table cannot say so.

The one component that should have caught him

Institutional change is the component I said was a judgment, argued in the open, with four points reserved for the Speakers who rebuilt the office: Reed’s 1890 rules ending the disappearing quorum, Cannon’s consolidation and the 1910 revolt that broke it, Clay’s conversion of the chair from referee into a national policy command post. I defined the scale as permanent rules and powers an incumbent “created or destroyed.”

Then I graded the freshman rules package of the 119th Congress at zero, and that was wrong.

On January 3, 2025, the House adopted H.Res. 5, and section 2(a) of that resolution rewrote rule IX on the motion to vacate. Under the old rule, any single member of the majority could move to remove the Speaker, which is how Kevin McCarthy lost the gavel in 2023. The new text is not ambiguous. A resolution declaring the office of Speaker vacant “shall not be privileged except if it is offered by a member of the majority party and has accumulated eight cosponsors from the majority party at the time it is offered.” One member became nine.

That is a change to the structural power of the Speaker’s own position, adopted at his conference’s initiative, and it is exactly the kind of thing my third component exists to capture. I missed it, or rather I discounted it, because I was picturing the Speakers who built the office and not the one who barricaded it. The scale as I wrote it had no room for a Speaker whose institutional achievement is making himself harder to fire.

The same package contains a second change I should have weighed. Section 2(g)(3) created a new rule I, clause 13, and it lets the Chair declare a “district work period” in which days stop counting as calendar or legislative days for the War Powers Resolution, for the rule XIII layover, and for the rule XV discharge clock. It is a mechanism for stopping the legislative clock from inside the chamber. That, too, is a precedent about what the office can do to time itself, and I recorded it as nothing.

What the record actually shows

Strip the arithmetic away and the case is concrete. It is a record of a man who used the office to prevent things from happening and to shield one person from consequences, and it is documented at primary source.

He tried to overturn the last election. In December 2020 he organized and led an amicus brief asking the Supreme Court to treat the presidential electors of four states as constitutionally invalid. The docket records the motion for leave, filed on behalf of Representative Mike Johnson and 125 other members. He recruited signers by telling colleagues the effort had Trump’s personal blessing and that Trump was “anxiously awaiting” the list, according to an email NBC News obtained. The Court denied leave on December 11, 2020, for lack of Article III standing. On January 6, 2021, he voted to sustain objections to both Arizona and Pennsylvania, which made him the most important architect of the objections: the New York Times found that about three-quarters of the members who objected relied on his arguments. He had presented colleagues with what he called a “third option,” faulting how states changed their voting procedures without endorsing the fraud claims. A year later he said he never bought those claims, that he “never egged on any of that,” and that he had been “a lone wolf crying in the wilderness.” Both things are on the record. They cannot both be true.

He changed the rules so he could not be removed. The vacate threshold went from one member to nine, in the first week of the Congress.

He shut the government down for forty-three days, the longest lapse in American history. The House Historian’s own table records it: funding ended September 30, 2025, was restored November 12, and the gap ran forty-three days, longer than the thirty-four-day shutdown of 2018 and 2019. The House was out of session for most of it.

He stonewalled the Epstein records until the floor forced his hand. He kept the chamber closed for nearly two months, and he declined to seat Representative Adelita Grijalva, who had won a special election on September 23, 2025, until November 12, which was the day she signed the discharge petition and delivered its two hundred eighteenth signature. When the vote finally came on November 18, it passed 427 to 1, and he described it on the floor as “a raw and obvious political exercise” while voting for it anyway.

He ducked a war-powers-adjacent vote by going home. On September 16, 2026, he cancelled the last day of votes before the midterm recess, sending members home a day early and avoiding a vote on a privileged resolution to impeach the Secretary of Defense. The House then stayed out until after the election while more than one hundred Democrats asked him to stay and legislate on artificial intelligence. His answer to that was that staying would cost the country the AI race: “If Congress just races in and does some sort of emergency session to try to regulate AI, we will lose the race to China, and that is a threat to every single American.”

He said Congress could abolish a federal district court. Speaking to reporters in March 2025, he floated congressional intervention in the judiciary: “We do have the authority over the federal courts, as you know. We can eliminate an entire district court.” He later said the remark was meant to illustrate Congress’s constitutional powers rather than to threaten a specific court, and the clarification belongs in the record with the original.

I am not going to pretend this is a balanced ledger. It is not. On the other side there is an appropriations process that eventually funded the government, three defense authorization acts, the stablecoin law, and the Laken Riley Act, and a Speaker who did, in April 2024, move a Ukraine supplemental that a majority of his own conference opposed and pass it with 210 Democratic votes. That was a genuine act of institutional nerve and he paid for it with a motion to vacate. It is the one entry on this side of the ledger that costs him something.

But that is a thin defense of a speakership, and it does not weigh against organizing a brief to void an election.

Why “mascot” is the wrong word, and contempt is not

The temptation is to call him a mascot, a front man, a hollow figurehead inflated by a movement that needed a face. I understand the impulse and I think the word is wrong in a way that flatters him.

A mascot does not file the brief. A mascot does not personally recruit 125 colleagues by telling them the president is waiting to see the list. A mascot does not rewrite the motion to vacate so that nine names are required instead of one. A mascot does not hold the floor calendar closed to protect an impeachment target. Each of those is an exercise of the office by the man who holds it, aimed at a specific outcome, and each worked at least partly. He is not a passenger. He is the instrument, and he has been a willing and effective one.

So I will say what I actually think, in my own voice, which is what this space is for: I hold Mike Johnson in contempt. Not because he is a fool, which the record does not support. Because he is a lawyer who spent twenty years arguing constitutional cases and then tried to nullify a presidential election, changed the rules so he could not be held to account for it, and has spent the rest of his tenure using the Speaker’s calendar as a device for preventing the House from voting on things he would rather not lose. His contempt for the institution is not a charge I am inventing. It is legible in the rules he wrote.

That is a judgment, and judgments are not the same as measurements. A ranking is a claim about a record that can be checked and argued with. A judgment is a claim about what the record means. I can defend both, but they are different kinds of statement, and I should not have let the first one stand as though it settled the second.

What the number is still for

Here is the honest use of the ranking, and where I think it earns its keep.

It shows that the office’s own history does not hand you a verdict on the man who holds it, because the rubric that grades the office counts tenure and statutes, and the worst things a Speaker can do to the institution are not tenure and statutes. It shows that Johnson’s 220, a middling score, sits in a table next to Reed’s 234 and Cannon’s 334, and that those profiles describe Speakers who built things while the middle of the table describes a Speaker who mostly subtracted. The profile string was designed to show the difference between a 340 and a 223. It cannot yet show the difference between a Speaker who did a little and a Speaker who did damage, and that is the next thing the method needs, if anyone wants to build it.

I would grade the 119th rules package as a one, not a zero, if I were doing it again: a minor institutional change by the standard I set, but real. That single point would move him from a score of four to a score of five, and from a band spanning ranks twenty-four to thirty-two into the band at seventeen to twenty-three. One digit, and he goes from the median to the upper third of the table, past William Bankhead and Nathaniel Macon and James K. Polk. That is the fragility the number has always had. The arithmetic is coarse enough that a marginal call about one rule is worth several places, which is another reason not to read the placement as a verdict.

A ranking can tell you where a man stands. It cannot tell you what he did to the ground he is standing on. For that you have to write the sentence yourself.

Sources

PRH | huffmanwrites.org | © Philip Huffman