Chamber size and party ‘strength’

What do folks think the correct answer to this question is: How does the size of an assembly affect the strength of political parties?

By strength, I mean the relative freedom of the individual member to cultivate constituency ties and to dissent from party leadership on votes on legislation. I also mean, holding other factors constant.

Suppose a country’s assembly is significantly smaller than its expected size, per the cube-root law. If nothing else changes, how would raising the size be expected to affect the strength of parties?

Obviously, I am thinking about potentially expanding the US House, so a starting point of non-hierarchical parties, and only two of them (and presidentialism, etc.). But I am interested in the question more broadly, and whether features of US party and legislative politics, aside from the small House size, change the impact of increased size on party strength in a manner that might be different from how it would play out in other contexts.

I ask because I genuinely do not know. I could see it going either way. A larger house, for a given population, means each member represents fewer voters, obviously. This could make personal-vote and constituency-service strategies more viable, thereby in some sense making parties “weaker”. On the other hand, a larger assembly (here, independent of population) makes internal collective action more challenging. This could result in members delegating (or simply losing) more authority to internal party leadership, making parties “stronger.” Note that these possible directions of change are closely connected to the two factors that go into the cube root law itself–this is a logical model that is based on balancing (and minimizing overall) two types of “communication channels”: those between legislators and constituents, and those among legislators themselves.

It possible both directions of change can happen at the same time, implying parties get weaker in some ways and stronger in others. That is, more constituency-oriented behavior, but also more party leadership control over votes and especially over speaking time. I am not sure what that means for overall strength. Maybe that isn’t even the right way to frame the question; skepticism over my own question framing is why I use the inverted commas in the title of this post.

Finally, theoretically and all else equal, a larger assembly means more parties should be represented (per the Seat Product Model). I have my doubts that this would be realized in the US, however, given all the other barriers to third-party representation. Unless the House were truly huge, I do not expect much impact there as long as it is elected in single-seat districts, and with primaries (or with “top two”/”top four” rules). However, parties’ internal strength could be affected. But which way?

11 thoughts on “Chamber size and party ‘strength’

  1. “It possible both directions of change can happen at the same time, implying parties get weaker in some ways and stronger in others.”

    Yes, and it is also possible that the effect could be in one direction in presidential systems and the opposite direction parliamentary systems, or opposite directions for different (pre-existing) effective numbers of parties. Or that the effect could be limited to some such circumstances and nil in other such circumstances. I don’t have any hypotheses about these interaction effects, other than a vague sense that interaction effects are somewhat likely here.

  2. Comparing Australia to the UK I’d say party discipline is much stronger in Australia. When ministers and parliamentary secretaries are taken into account, there’s 30 plus out of a party room that is never much more than 100 that are part of the executive. Most of the rest think they have a shot at it. If there were a party room of a few hundred, then most would accept that their career is as a backbencher and a representative of their constituents or local party branch primarily.

    • Yes, this is a theory I’ve heard as well: the ratio of cabinet members to total legislature members is meant to be significant because cabinet members are subject to stricter discipline, and potential cabinet members will toe the party line to improve their chances of promotion. I’m not sure how this dynamic (if it is indeed important) would translate to presidential systems. A very rough analogy could be made between cabinet membership and membership on important committees: in a smaller legislature, I suppose a higher proportion of legislators could be on committees, thus increasing party discipline? I don’t know that that makes sense, though: there’s no principle of committee solidarity comparable to cabinet solidarity, and I don’t know that committee size is necessarily based on the same principles as cabinet size.

  3. The US should abandon the 5-minute rule for speeches, particularly in public committee hearings. You cannot do anything in 5 minutes except repeat slogans. At 76 members the Australian senate is roughly comparable with the US senate in size, but the normal speech limit is 15 minutes, reduced from 20 in 2020.

    And then there is the annual estimates process, where departments are represented typically by the minister, the secretary (confusingly for Americans officials like the secretary of the treasury are civil servants) and advisers. There is no limit on speeches, numbers of questions, or total length of the hearing. If the minister does not have a document an estimates committee wants, it is usually produced by the end of the day, when the hearing is still live. Former Senator John Faulkner once suggested the process was so gruelling it should be referred to the human rights commission, but it is a major accountability mechanism where running down the clock is not an option.

    Allocation of time would be a major concern in a larger assembly, but there are ways round that.

    The quorum for the Athenian assembly was 6000. The Syracusan assembly was even larger. (It’s a myth that Athens was the first, largest or most radical Greek democracy)

    Athenians, Syracusans, Thebans and Argives managed to get by without clocks, electronic tabulation, or formal party leaders. A large assembly with short speeches and leadership control of who spoke would strengthen party leadership. The same assembly with different mechanism would not. Even the Athenian council of 500, which handled administration and prepared assembly business, was larger than either US chamber. On the late Roman republic the senate was fixed at 900 members. The Venetian great council had 2476 members and a quorum of 600 from 1527 to 1797.

    At risk of diverting the thread, which as you all know I would never do, the Athenians managed secret ballots rather effectively. Each citizen was handed two balls that’s served as ballots, one black and one white. They then walked between two urns and dropped one ballot in each urn. The urns were designed so that each ballot clanked audibly against a brass plate and then fell into the body of the urn. To vote yes you dropped the white ballot into the voting urn and the black ball into the dummy urn. The ballots were then poured into a decimal grid which made sorting and counting easy and fast. For minor issues a random subset of an odd number of members was selected which voted on behalf of the full assembly.

    Judging by the glacial pace at which both US chambers vote, the Athenians could probably hold substantive votes faster in their assembly than either US chamber can manage.

  4. The fifty states vary substantially with respect to the size of their legislatures, both in absolute terms and relative to the cube root rule. That’s easy to measure. Is there a database with a good measure of adherence to party discipline? The average percentage of members voting with their own and with the other party, across many roll call votes, would seem to be a good independent variable for this analysis.

    Even if feasible, this research wouldn’t directly answer the question here, which is what happens when you change the size of the legislature. But it would be relevant.

    • In this time of extreme polarisation, why would the percentage of legislators voting in agreement with other legislators be a good measure of party discipline? It might be a good measure of party unity, but voluntary unity by virtue of agreement does not signify discipline, it signifies cohesion. Discipline is obedience to orders given by the party, usually backed by implicit sticks and carrots of future advancement, demotion, and maintenance of access to the party label. Discipline says something about party organistion and, therefore, its strength; this is not necessarily the case with cohesion, which is not necessarily the result of organisation or design. America’s current high party cohesion is certainly not by any design.

  5. Thanks for the comments (so far). Lots of good points to chew on. I agree that the direction of the main effect likely varies with executive format, with existing number of parties (even though we know the latter is shaped by assembly size itself), and other factors. Likely some complex interactions and feedbacks here!

    On the question Bob raises about state legislative size, this is a really good point. I know of a couple of scholars working on some aspect of this question. I need to follow up and see if they have papers yet.

    A couple of caveats, though: We do not know if the cube root law should apply to state legislatures. I can think of reasons why it should, and also reasons why not. (Basically, federal systems may imply all chambers “should be” smaller than in unitary, for given population, because the different levels of legislative sovereignty reduce both types of communication channel for any given assembly, especially perhaps the channels with constituents.)

    The other caveat is the one JD mentions on “discipline”, which we should be careful never to confuse with voting unity. Chris Kam’s 2009 (2010?) book on Westminster systems has several tests of discipline, things like members’ career trajectories stalling from what would otherwise be expected. It is tricky to measure; I think his book is one of the best efforts. Also a journal article from some years ago by Benedetto and Hix gets at this (on UK).

    The point about impact of cabinet size, relative to assembly, is also an important one. Surely relevant only to parliamentary systems. There’s certainly work about this on New Zealand, where in the smaller parliament of the FPTP era the cabinet often would consist of a majority of the majority caucus, but the larger parliament (and smaller leading party seat share) has changed that dynamic in the MMP era. Obviously also a stark contrast to the UK, with its huge assembly. I think Kam also discusses this (although he covers only the FPTP era in NZ; he also has Australia in his study).

    • I’m not sure if cabinet size is only relevant to parliamentary systems. I know Virginia better than most, so I’ll use it as an example: it has just three directly-elected statewide officers, while the secretaries are all picked by the governor and confirmed by the General Assembly. (Judges and commissioners are picked by the GA as well.) While nationally many US senators may pass on a 4-year cabinet post as they have accumulated much legislative power, it’s a common career trajectory in Virginia to serve in the part-time legislature for a few decades and then do at least 3 or 5 years as a full-time secretary, commissioner or judge. The reason for this is that the defined-benefit retirement system is based on your best 3-or 5-year period (it changed to 5 years last decade), while number of years includes all the time spent earning $20k a year in the GA (and paying very little in premiums). It is a not uncommon tactic for a governor to appoint someone of the opposing party, or in their party but out of step with the leadership, to a position, especially if that person is out of step with their district and they hope to flip a seat. Otherwise, they tend to reward long serving members. GA leaders have also done this with judgeships and commissionerships. (Becoming a university administrator or professor at a state school is common as well.) The spoils are strong, and people need to stick around for a while to collect them, either sucking up to leadership or creating their own impenetrable base. Part-time legislatures mean advancing into the executive or judiciary is very profitable, often moreso than becoming a lobbyist.

  6. Matthew, you say “We do not know if the cube root law should apply to state legislatures. . . . Basically, federal systems may imply all chambers “should be” smaller than in unitary, for given population, because the different levels of legislative sovereignty reduce both types of communication channel for any given assembly, especially perhaps the channels with constituents.”

    In Canada, federalism for the last seventy-five years or so has meant shared jurisdiction, joint programs, and many other complications. A lot of an MP’s time is spent saying “that’s provincial” and an MLA’s time saying “that’s federal.” If they’re lucky. Reduced communication time? Sadly, I think they would all chuckle at that.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.