The Majority Rules, and When It Doesn’t, It Rolls

by John Sides on July 15, 2013 · 2 comments

in Legislative Politics

We welcome this guest post from Andrew Guess, a Ph.D. student at Columbia.


Since January, talk of the “Hastert rule” has gradually spilled over from the esoteric confines of the wonkosphere into mainstream political commentary. Shorthand for the norm that only bills with support from a majority of the majority should reach the House floor for a vote, the “rule” has generated interest insofar as it has been violated – three times during the current Congress so far. That may make a trend, but how common are such violations over the longer term? Over the course of three graphs, I will show that majority rolls of this kind have been occurring regularly (if infrequently) throughout a variety of different political contexts, despite any distinctive features of the current dynamic in Congress. That said, there are hints that any future such occurrences this term (perhaps for a vote on immigration reform, has some have speculated) will be restricted to cases in which the outcome is acceptable to most of the majority party’s members.

The share of votes for passage violating the Hastert rule has increased in the 113th Congress (for now).

The following figure shows majority party rolls starting in the 102nd Congress as a fraction of the total number of bills passed by the House that session (data from


Hastert rule violations continue to be very rare in an absolute sense, accounting for less than 2% of all votes for passage in any Congress up to last year. But in the current Congress, now back in session, that percentage has more than doubled. This reflects the temporarily smaller denominator (85 votes for passage in the House so far, as opposed to 304 in the whole 112th) but also explains why these incidents seem more common this year: It’s only July.

Various explanations have been offered as to whether violating the Hastert rule is becoming a favored tactic of Speaker John Boehner, and why this might be the case. Sarah Binder has argued that even if a majority of the majority opposes any given measure, it could benefit as a whole from an occasional “loosening” of the rule and give Boehner some leeway for allowing Democrats to carry a bill. Jonathan Bernstein has also emphasized the significance of following the Senate’s lead as Boehner’s approach to accommodating the various factions of his membership. Assuming this is accurate, I would simply add that bending the Hastert rule is a potential way out of numerous political binds that speakers might find themselves in. Boehner may face a divided Congress and a staunch conservative bloc whose support he can’t count on for national party priorities, but his counterparts in the past have used the same tactic to pass cross-cutting measures like NAFTA and the Balanced Budget Amendment (the former in a time of unified government).

For a majority roll to result in passage, near-unanimous support is necessary from the minority party (especially recently).

The next graph plots two trend lines, with the majority party illustrated by the blue and red shadings at the bottom. There are 36 votes for passage shown. In the time period covered, the greatest number of Hastert rule violations occurred in the 106th and 110th Congresses:

hastert2The top line illustrates the fraction of the minority party that voted for a bill brought to the floor without majority party support. For example, the last three points on the right, indicating votes on Hurricane Sandy aid, the Violence Against Women Act reauthorization, and “the acquisition and protection of nationally significant battlefields and associated sites of the Revolutionary War and the War of 1812,” passed without majority Republican support but with the votes of 96, 99, and 91 percent of House Democrats, respectively. Those same votes received 21, 38, and 44 percent support from House Republicans. As the top line suggests, minority parties seem to have become more cohesive over time, with less variation in the proportion supporting a given roll call vote and a tendency toward uniform support overall. This is consistent with historic and increasing levels of polarization in the House as well as possible selection effects due to agenda control.

Violations historically lead to cutpoints at the House median, but less so more recently.

Finally, to get a sense of the political consequences of temporarily suspending the Hastert Rule, I used Poole, Rosenthal et al.’s Optimal Classification (OC) algorithm to plot the “cutting points” of the same 36 votes in ideological space.  Here is how to think about OC scores and cutting points.

Suppose we have a group of legislators and a corresponding set of roll-call votes. The algorithm essentially orders the legislators (from “liberal” to “conservative”) such that we can predict who votes for any single bill by marking a cutting point. With perfect data, everyone to one side of the cutting point votes “yea,” and everyone to the other sides votes “nay.” Real life is more complicated, so the algorithm optimizes the ranking and cutpoints to maximize the percentage of individual voting decisions correctly predicted. It also computes OC “scores” for each legislator in ideological space. Usually this is all done in two dimensions to improve accuracy, but for simplicity I only graph first-dimension components below.


The red and blue lines show the mean first-dimension (left-right) OC scores in each Congress, with the Republican mean (the red line) on the conservative side and the Democratic mean (blue line) on the liberal side. The dotted line follows the chamber median, which jumps from one side of the midpoint to the other as party control shifts. And the dashes indicate estimated cutpoints for each of the votes for passage violating the Hastert rule since 1991 – that is, the thresholds separating those who voted “yea” and “nay.”

On average within a term, cutpoints tend to fall around the chamber median. This is not surprising: In allowing such votes to reach the floor, speakers temporarily transform Congress into the kind of institution explained by our classical models of legislative behavior.

But something different could arguably be going on in the past few Congresses. With one or two exceptions (the outlier near 0 in the 112th Congress is the fiscal cliff deal), cutpoints seem to be edging closer to the party means. This would seem to confirm the dominant narratives about Boehner’s strategic bind, that in order to pass even high-profile legislation favored by his own conference, he (paradoxically) needs to rely on at least a handful of Democratic votes.

What about immigration? As a quick exercise, I constructed a fictional roll-call vote for an immigration bill supported by all 201 Democrats and 19 Republicans whose seats are either contested for re-election or whose districts’ populations are at least 30% Hispanic – a minimal scenario for passage of comprehensive legislation. The result is that the hypothetical cutpoint (0.1) is significantly to the left of even the chamber median, illustrating that any major bill that could pass the House in this way would have to be a bitter pill to swallow for the great majority of Republicans. Since they are the ones empowering Boehner, a special set of circumstances would have to be in play for the historical pattern to change.

As the figures indicate, such an outcome isn’t unprecedented. It remains to be seen whether the trends from the past year continue, or if the Hastert rule is about to be shelved until another distinctive political situation demands it.


Comments on this entry are closed.

Previous post:

Next post: