Whoever Writes the Agenda Wins the Vote. Until 329 MPs Said No.

In a body that decides by simple majority, the member who writes the agenda usually beats the member in the middle. I tested this claim against three written rules. Each rule gives one actor the gate. Each rule also gives the floor majority a key to that gate. My result: the agenda holder predicts outcomes better than the median voter in normal times. But its power is a loan from the floor majority, and the loan has a written repayment clause. On 25 March 2019 the House of Commons called in that loan by 329 votes to 302 [9].

Question

The usual model of a majority vote looks at the voter in the middle. If you line up 435 members from left to right, member 218 decides every pairwise vote. The agenda-control model gives a different answer. The vote only takes place on proposals that somebody allowed onto the floor. So the actor who decides what comes to a vote also decides which median-voter outcomes can happen at all.

My question is narrow. In three bodies with public rules, does the written agenda holder predict outcomes better than the median voter? And under which written condition does that prediction fail?

Data and where it came from

I used three primary texts and the research that tests them.

Rule in force (US House): The Committee on Rules reports special rules that set the terms of debate for each bill. CRS states that this committee "generally operates as an arm of the majority party leadership," with nine majority and four minority members [2]. The House adopts a special rule by simple majority. The other main route to the floor, suspension of the rules, needs two-thirds of those voting and permits no amendments [2]. The bypass is the discharge petition: 218 signatures, a majority of the whole House.

Rule in force (UK Commons): Standing Order 14(1): "Save as provided in this order, government business shall have precedence at every sitting." The same order sets aside 20 opposition days (17 for the Leader of the Opposition, 3 for the second largest opposition party), 35 backbench days (at least 27 of them in the chamber) and 13 Fridays for private members' bills [3]. Arthur Balfour moved the original version on 30 January 1902. He called it a codification of existing practice [4].

Rule in force (corporate board): I used Sony's board charter, as filed with the SEC in a 2010 exhibit [5]. Article 9: "the agenda of the Board of Directors shall be determined by the Chairman of the Board, in accordance with proposals made by Directors and Corporate Executive Officers, or upon his/her own decision as the Chairman of the Board." Article 10: the quorum is a majority of directors who may vote, and a resolution passes by a majority of directors present. Article 8 contains the bypass. If a director submits a written agenda and asks for a meeting, and the chairman sends no notice within five days, that director may convene the meeting [5].

I read the annex first, as usual. In this charter the most important sentence is in Article 8, not Article 9. I return to it below.

For outcomes, I used four published sources. Cox and McCubbins measured House final-passage votes from the 45th to the 105th Congress (1877 to 1998) [1]. CRS counted discharge petitions from 1931 to 2002 [6]. Patterson and Schwartz challenged the roll-rate evidence [7]. A public list counts Commons government defeats since 1945 [8].

Method

For each body I recorded three items:

1. Who holds the written agenda power, and under which article.
2. What vote the floor needs to bypass that holder.
3. How often the floor used the bypass, and with what result.

The agenda model predicts that bypasses are rare and that items on the floor move policy toward the agenda holder. The median model predicts that the agenda holder's position adds no information once you know the median. I computed all ratios below by hand from the cited counts, without the Lab. A reader can check each one with the stated inputs.

Result

US House: the gate holds, and the key is rarely used

Cox and McCubbins report that after Reed's rules became permanent in the 1890s, "over 80% of the bills allowed to reach the final passage stage in the typical Congress proposed to move policy toward the majority-party median" [1]. The share rose "abruptly, substantially, and permanently" when the House adopted those rules [1]. The median voter did not move in 1890. The agenda rule did. In my reading this is the best single piece of evidence for the agenda model, because the change follows a rule change, not a change in members.

The bypass data point the same way. CRS reports that from 1931 to 2002, members filed 563 discharge petitions. Of these, 47 obtained the required signatures, the House voted to discharge 26 times, it passed 19 of the measures, and 2 became law [6]. The ratios:

$$\frac{47}{563} = 8.3\%, \qquad \frac{19}{563} = 3.4\%, \qquad \frac{2}{563} = 0.36\%$$

So in 72 years, a petition against the agenda holder produced a law about once in 280 attempts. The median-voter model has trouble here. Some of these 563 petitions surely had majority support on the merits. A signed petition is a public act against your own party's gate, and most members of the majority do not sign it.

UK Commons: the gate holds until the majority loses its majority

Standing Order 14 gives the government the floor at almost every sitting [3]. The defeat record follows the size of the government's majority more than anything in the text. The public list records 0 defeats for the Eden and Macmillan governments (1955 to 1963) and 4 for the Thatcher governments (1979 to 1990). It records 34 for the Callaghan minority government (1976 to 1979) and 33 for the May minority government (2017 to 2019) [8].

The 2019 case is the decisive one for my question. On 25 March 2019 the House approved the Letwin amendment by 329 to 302 [9]. The amendment set aside Standing Order 14 for one day, so that backbenchers could hold indicative votes on Brexit options. The margin was 27 votes. Three ministers resigned to vote for it [9]. Note the vote rule. A one-day suspension of the government's agenda power needed only a simple majority of members voting. No supermajority protected the gate.

Corporate board: the gate is wide, and the evidence is thin

Sony's chairman sets the agenda under Article 9 [5]. But the quorum and vote rules make the floor very small. Take a board of 12 directors as an illustration (my number, not Sony's). Under Article 10, a quorum is 7. A majority of 7 present is 4. So 4 of 12 directors, one third of the board, can adopt a resolution that is on the agenda. Whoever controls the agenda therefore controls which items 4 directors can pass at a thin meeting.

Article 8 closes that loophole in part, and I admire the drafting. A director who is refused a meeting for five days may call it herself, with her own agenda [5]. That is a discharge petition with a quorum of one. I found no study that measures how often directors use such a clause, or whether chair agenda control predicts board decisions. Boards rarely publish vote counts. I mark the board case as a rule reading only, with no outcome data behind it.

Summary table

Body	Agenda holder (rule)	Bypass rule	Observed bypass
US House	Rules Committee, 9 to 4 majority [2]	Discharge, 218 signatures	47 of 563 petitions qualified, 2 laws, 1931 to 2002 [6]
UK Commons	Government, SO 14(1) [3]	Motion to set aside SO 14, simple majority	329 to 302 on 25 March 2019 [9]
Board (Sony)	Chairman, Art. 9 [5]	Self-convened meeting after 5 days, Art. 8	No public data

My estimate: in the two legislatures, the written agenda holder predicts what reaches a final vote better than the median voter does. I put that at 0.7 for the House over 1890 to 2002 and 0.65 for the Commons under governments with a working majority. For minority governments I drop it to about 0.4, because the 2017 to 2019 record shows the floor taking the agenda back. For boards I hold no number, because I have no outcome data.

Sensitivity: which assumption moves the result most

The assumption that moves my result most is this: low majority roll rates show agenda control. A "roll" is a final vote that passes against a majority of the majority party. Cox and McCubbins read tiny majority roll rates as proof that the majority keeps hostile bills off the floor [1].

Patterson and Schwartz give the strongest case against that reading. They show that "the observed roll rates are approximately what would result from chance alone or from chance constrained in several natural ways" [7]. They also argue that some rolls may be the intended result of the majority's own agenda choices [7]. If they are right, the roll-rate evidence cannot tell the agenda model from the median model. My House estimate would then rest on the Reed-rule break and the discharge counts alone. I would lower it from 0.7 to about 0.55.

I do not think their point reaches the Reed result. A chance model has to explain why the share of bills that move toward the majority median jumped at a rule change and stayed high [1]. Chance does not read the rule book. But I concede that the roll rate on its own is a weak test, and I will stop citing it as one.

The second assumption is that the bypass vote is the true limit on agenda power. Tsebelis models each actor whose agreement a law needs as a veto player. In fifteen West European countries from 1981 to 1991, more veto players and wider distance between them meant fewer significant laws [10]. That finding supports my standing view that written veto points predict delay. But it also warns me. An agenda holder with a simple-majority bypass behind it is a veto player only while its own side stays with it. The Commons in 2019 did not change its rules. It used a rule that was already there.

This connects to procedure in other posts on this site. The earlier post on the European credit-score case reports an appeal that took seven years. I extend it with one question: which body set the docket in those seven years, and under which article? A seven-year delay is usually an agenda fact before it is a legal fact.

What would change my mind

Two findings would move me. First, a study that shows bills with majority support in the House reaching the floor at the same rate under strong and weak Rules Committee control. Second, a Commons record in which a government with a working majority loses control of its agenda as often as a minority government does. Either result would mean that the median voter, not the gate, does the work.

Motion: that the reader, for any vote in any body, first ask who wrote the order paper, and then ask what vote the floor needs to rewrite it. If the second answer is "a simple majority," the agenda holder has power only on loan.