NYT endorses a larger House, with STV

Something I never thought I would see: The editorial board of one of the most important newspapers in the United States has published two separate editorials, one endorsing an increase in the size of the House of Representatives (suggesting 593 seats) and another endorsing the single transferable vote (STV) form of proportional representation for the House.

It is very exciting that the New York Times has printed these editorials promoting significant institutional reforms that would vastly improve the representativeness of the US House of Representatives.

The first is an idea originally proposed around 50 years ago by my graduate mentor and frequent coauthor, Rein Taagepera, based on his scientific research that resulted in the cube root law of assembly size. The NYT applies this rather oddly to both chambers, then subtracts 100 from the cube root result. But this is not something I will quibble with. Even an increase to 550 or 500 would be well worth doing, while going to almost 700 is likely too much, the cube root notwithstanding.

The second idea goes back to the 19th century (see Thomas Hare and Henry R. Droop) but is as fresh and valid an idea today as it was then. The NYT refers to it as “ranked choice voting in multimember districts” and I have no problem whatsoever with that branding. In fact, I think it is smart.

Both ideas could be adopted separately, but reinforce each other if done jointly.

They are not radical reforms, and they are not partisan reforms (even though we all know that one party will resist them tooth and nail and the other isn’t exactly going to jump on them any time soon). They are sensible reforms that would bring US democracy into the 21st century, or at least into the 20th.

And, yes, we need to reform the Senate and presidential elections, too. But those are other conversations…

45 thoughts on “NYT endorses a larger House, with STV

  1. Good on the Gr[e]y L[a]dy, but as the examples of A Blair and J Trudeau demonstrate, there’s still something of an implementation gap between “left-liberal pundits conclude that FPTP is outmoded and incompatible with left-liberal democratic ideals” and “actually existing left-liberal politicians take steps to replace FPTP.”
    Keep up the pressure and don’t let Beto quietly shelve the report from his 2021 Presidential Commission of Inquiry Into Voting Systems.


  2. I’m a big, big fan of increasing the size of the House of Representatives (and many other legislative bodies in the U.S.). But I’m afraid that linking an STV proposal to a proposal for a larger legislature is a good way to torpedo STV. Too many rank and file voters believe that a good way to cure what ails the body politic is to have fewer politicians, not more of them. I despair of being able to overcome this misconception.


    • There was a bill, mentioned in the editorial, in 2017 implementing STV and prohibiting gerrymandering, while specifically not expanding the House. It even allowed states to run primaries if they wanted to choose X candidates for a district. Hopefully the House can give that one meaningful debate.

      Though I do believe the House should be bigger. And I am a bit concerned that the bill, as submitted, technically could still have allowed a determined state legislature to tullymander their House delegation.


      • I think that section was a sop to the establishment. It essentially is a long list of processes to preserve partisan primaries, where they exist. Louisiana wouldn’t be obligated to construct one. I’m not sure how Top 2 Primaries would be effected. A top-10 primary?


    • On the other hand Malaysia (and Singapore, too) might give certain U.S. gerrymanderers ideas if they hear about what happens there/


      • Interesting that Malaysia is a Federation that allows seat boundaries to be drawn across state lines, so is Malaysia a Federation masquerading as a Unitary state? Is Malaysia, Mexico and perhaps Russia the only Federations to have districts to be drawn across state lines?

        “First is Article 46, which allows the parliament to arbitrarily allocates seats across states and federal territories, and which has allowed inter-state malapportionment to worsen since 1974. Second, the demand by Section 2(c) of the Thirteenth Schedule for constituencies within the same state to be “approximately equal” is compromised by the vaguely-worded “area weightage” and the 1974 removal of a cap on maximum deviation from state averages.”

        Is it a requirement for a Federation to have a requirement that the lower house of parliament requires minimum representation for each state an example would be 1 seat in the US or 8 seats in Brazil? Does this not make a country a Federation if boundaries are drawn across state lines?

        What is the purpose of a mandate minimum representation for a state, providence, and other regions? Some countries like Brazil have maximum representation of 70 members.

        I would believe that the basis of a Federation is that the Central Government can’t abolish, merge, split, and/or redraw it’s component states without the consent of the states and/or the people of that state by referendum.


  3. I think they removed 100 to account for the senate as from the article “This is because the United States Senate is a more significant lawmaking body than the smaller chambers of other countries”


      • I think in this case, it is more the power the upper house has rather then the size. (I know down here in Aus we have 76)


      • The question of relative sizes of first and second chambers is something my coauthors and I looked at briefly in A Different Democracy (2014, p. 210 & 214). We note that the UK is the only one (of our comparison set, at least) with a second chamber (“upper house”) that is larger than its first chamber (“lower house”). We then say:

        The United States thus has one of the smaller second chambers relative to its first chamber (100/435=.23), but not as small as Germany’s (69/603=.11)… The mean ratio for federal systems is about .34, whereas for unitary systems it is .53. The difference is statistically significant.

        [Error in German ratio corrected, as called out by Alan in a later comment; fortunately, it is correct in the book, and I simply made a transcription error here.]

        The NYT proposal would reduce the Senate to only around 0.17 times the size of the House, which might be too small (though I have no idea what the actual implications might be).

        Perhaps this is an argument for increasing the size of the Senate, which is something I actually have proposed in print (as a way to introduce PR).

        We could do something like 150 seats in the Senate and 500 in the House. These sizes sum to something a bit less than the NYT’s suggested 693, but we also need to maintain a number for the Senate that is divisible be 50 (I am assuming the current states, and the need to stick to equal representation thereof). So this combo is close. We get 150/500=.30 not far from the current relationship.

        If I could get a 150-seat Senate (for 50 states), I’d be pretty happy, provided all (at least in any given state) were elected at the same time by a non-majoritarian electoral rule.

        I’d be quite happy with 200/500, too!

        And, no, it is not something I expect to see. But it is an intriguing thought!


        • I think you may have typed the wrong number for the German ratio. I will take any chance I get to call the kettle black.


      • Would a 150 member Senate actually be better? There’d almost certainly be 50 from the GOP and 50 from the Dems elected first and second in each state (assuming STV or, honestly, almost any proportional system), but the 3rd senator from each state would go the state’s majority. Am I wrong in thinking that would make it fairly easy for the GOP to cement a near certain majority? A second Democrat winning against the standard split of the state would be a hard sell in most places.


      • The Philippine senate has 24 members,which must surely be among the smallest legislative chambers around, apart from the small island states.


      • Mark, with a senate of 150 seats, even if all states went 2-1 for their current plurality party, I’d still see this as an improvement. There surely are several states where that third seat would be closely contested, and thus the plurality party had to fight for it. In very safe states, at least the second party is likely to get a seat (even if there would not be much competition, except perhaps intra-party in general elections). It would be quite a change in the character of the Senate, even if the majority did not shift.

        And, of course, we are so far leaving out the increased probability that a third party occasionally might have a chance in some states.

        I agree that more than 3 per state would be better, but 4 has its own issues (as later comments discuss) and I can’t quite imagine a 250-seat senate.


  4. Even if the US senate went to 150 with 1 senator elected in every state every 2 years by a majoritarian (or pluralitarian) method, I’d think you would get a more responsive senate because it would smooth out cases like this election where the senate map was extremely favourable to one party and that party could therefore campaign only in those states actually voting. To get to a PR senate I’d guess you\d have to go to at least 3 senators elected in every state by PR every 2 years, but a senate of 300 would perhaps make heads explode.


    • Every two years, elect all three senators from each of 16 or 17 states, adding up to 50 states and 150 senators in each six year cycle.


      • Your proposal still has only part of the country voting at each election, which means that one party has the potential for massive advantage at a particular election.


      • With PR, is the regional advantage thing really such a problem for schemes where not the whole country is up for election each time?

        Bob Richard almost literally described how Argentina does things.


      • JD

        You could well be right. One of the interesting stories (to me anyway) about the the US midterms is that the Democrats managed to keep their senate losses to 2 despite having such an adverse map.


  5. Could the staggered nature of the US Senate be abolished outright and just have all the Senators elected at the same time? Could an at large form of representation be added to the Senate, but then that would dilute the no state shall be denied it’s equal suffrage requirement? What about Dual Member PR system just requiring the 2nd Senator comes from that State?


    • A few years ago we discussed PR for the US Senate. A long shot, but 4 seats per State, elected simultaneously by PR-STV (no rotation) for 4 years, with a national-popular-majority-elected Veep having a casting vote (total, 200.5 seats) would mitigate the grossest minoritarian bias, introduce (sorry, Wilf) reasonable proportionality, and be achievable by ordinary constitutional amendment (ie 38 States instead of all 50), since every State would gain more Senate seats, ad every State would keep its current share (indeed, inasmuch as the Veep’s casting vote counts as 0.5 of a vote, each State would have 2% of a 200.5-vote body instead of 2% of a 100.5-vote body, which is a – very – small increase).


      • Are there any states where one party routinely gets less than 40% of the vote? With 4 seats, it would almost always be a 2-2 split between left and right.

        That would not be a bad thing per se, as it would effectively make the Senate into a revising chamber. Assuming party discipline, only bills that the president would likely veto anyway would fail to pass.


      • I could picture the Utah delegation going 3-1 Red and the NY delegation going 3-1 Blue. I suppose the danger is if these lopsided States didn’t cancel out. With the chamber being designed to go 100.5 to 100 on most issues, even a couple of extra States one way or the other might tip the balance.
        With larger district magnitudes, Australian Senate results have sometimes given a majority even with even numbers, although these tended to be landslides with 10 seats (6-4 for the Coalition in a couple of States in 1975). The Coalition’s 4-2 win in Queensland in 2004 was accomplished by running two separate tickets under a functionally largest-remainder system. The Qld conservatives have since corrected this error and merged into a single party so won’t be running two separate Senate tickets in the foreseeable future.


      • Even magnitudes are not optimal with STV. The ACT Proportional Representation (Hare-Clark) Entrenchment Act 1994, Section 4(1) sets out a list of defining features and the first 2 are:

        (a) at a general election, an odd number of members of the Legislative Assembly shall be elected from each electorate;

        (b) at a general election, at least 5 members of the Legislative Assembly shall be elected from each electorate;

        In a district of magnitude 4 the quota will be 1/5 rounded up and securing a majority of seats will require just over 3/5 of the votes. A district of magnitude 3 will have a quota of 1/4 rounded up and wining a majority of seats will require just over 1/2.


      • The idea district magnitude under STV would be odd numbers for every electorate, true. Second-best, however, would be either uniform even numbers or else a maximum of one even number per State/ Province/ Territory to make the numbers fit (eg, 5, 6, 7 for a State/ Province/ Territory with 18 seats in the national lower house). Worst of all is allowing mixing of several even magnitudes with several odd numbers ad lib, which opens the way to “Tullymandering”.
        ie, with 30 Representatives for your State, yes, six 5-seaters would be ideal but five 6-seaters would be better than one 7-seater, four 6-seaters and one 5-seater. In other words, having fewer even magnitudes simpliciter is not semper et ubique preferable.


  6. The purpose of an Upper House is to stop the Lower house from doing things. USA, Italy, you name it.

    All the Canadian provinces that originally had upper houses have abolished them. New Zealand has abolished its.

    Germany has a different kind of Upper House; it’s part of their federalism structure. As far as I know it never blocks measures that have been adopted by the Bundestag.

    What democrat needs an Upper House anyway?


    • A democrat who wants effective parliamentary review of secondary legislation for a start:

      With this controversy in the background, the Senate, following the general election of 1931, resolved to incorporate in the standing orders a requirement that a Standing Committee on Regulations and Ordinances be appointed at the commencement of each session of Parliament.9 Only the House of Lords, when it created a committee in 1925 to examine regulations requiring an affirmative resolution to become law, had previously acted in this field. Eventually many houses of parliaments followed a similar course of establishing a committee to oversee statutory instruments, but one which has not done so is the Australian House of Representatives. Thus responsibility in the Commonwealth for active and systematic scrutiny of this extensive field of legislation falls upon the Senate. Maurice Blackburn, later a Labor member of the House of Representatives, had explicitly contended in 1930 that:

      the House of Representatives is not likely to do that work well, or, in fact, to do it at all. Upon its vote turns the fate of the ministry. The regulation is made by the ministry, and a proposal for its disallowance would certainly be treated as a vote of want of confidence, and would be tested on party lines. No ministry depends on the vote of the Senate and it is quite likely that in that chamber a regulation would be considered on its merits…. 10

      Secondary legislation now exceeds primary legislation by volume in most democracies. All Australian states now have independent officers of parliament (electoral commissioner, ombudsman, auditor-general, independent commissioner against corruption) whose appointments are subject to parliamentary veto. The Blackburn argument for bicameral veto of secondary legislation applies to these appointments in exactly the same way. Especially in a parliamentary system, a unicameral assembly is limited in its ability to check the executive.


      • I concur with Alan, and would add that committees of upper houses also serve as a useful mechanism for investigating government actions. Even under the most proportional legislature, governments are reluctant to appoint committees to seriously investigate themselves.


      • I. And in a presidential system, bicameralism may actually help governability if the more “executive-adjacent” functions of legislative chambers (budgeting, confirming appointments) can be exercised by the executive plus either chamber.

        II. I’ve often noted the following two-step among what I call the “flog-Mandela crowd” on the British Right wing (you know the type – their two articles of faith are “(1) FPTP is great because it gives the country a responsive two-party system, excludes minor-party cranks from power, and empowers the voters to sack governments who don’t act for the public good”, and (2) “Both Labour and the Tories have conspired for the past decades to ignore the British People’s will regarding the EU, which happens to coincide exactly with UKIP’s stance on the EU”):

        Odd-numbered days: “There is no ‘best’ or ‘perfect’ voting system. Different electoral methods have differing goals, maximise incommensurable criteria, and cannot possibly be judged by a common yardstick of ‘representing the popular will’ since no such thing exists. [Second premise in syllogism illegible due to white-out.] Therefore we should use first-past-the-post voting, obviously, because it alone ensures majority rule.”

        Even-numbered days: “What on earth is the point of having an elective upper house? Only two outcomes are possible. Either the Government will have a majority, in which case the upper house will be uselessly supine and involve a colossal waste of public funds; or else the Opposition will have a majority, in which case the upper house will be irreconcilably obstructive and will render it impossible for the government to function, or even pass a budget.”



      • Oddly enough in Alaska the executive functions of the legislature (confirmations, veto overrides, executive order overrides, etc etc) are vested in a joint session of both houses. Even more oddly impeachments are brought by the senate and tried by the house.


    • Wilf:
      “The purpose of an Upper House is to stop the Lower house from doing things. USA, Italy, you name it.”
      Seeing so many upper houses today have only delay powers, this is clearly exaggerated.

      “All the Canadian provinces that originally had upper houses have abolished them. New Zealand has abolished its”
      Makes a lot of sense for units of a federation. There are a number of good reasons to want to keep elections and accountability simple at that level. NZ, on the other hand, isn’t a federation and the upper house it abolished was at the national level. Given that they still had FPTP at the time, that change made the system even more majoritarian. They would have been much better served by reforming it to be elected by PR.

      “Germany has a different kind of Upper House; it’s part of their federalism structure.”
      Yes, federalism is probably one of the best reasons to have bicameralism, even if many (most, even) federations have upper houses that don’t enhance federalism one bit (I’m looking at you, Canada).

      “As far as I know it never blocks measures that have been adopted by the Bundestag.”
      This is patently untrue. Roughly 50% of all legislation is subject to the absolute veto of the Bundesrat; German govts routinely have to negotiate and compromise with the state govts in the Bundesrat, even when it has a friendly majority there, and sometimes this means that the federal govt’s aims are simply frustrated. This is not least because in Germany, the state govts have primary responsibility for enforcement of federal law.

      “What democrat needs an Upper House anyway?”
      If democracy is only about majority rule, then I suppose I’m no democrat. If democracy is also about deliberation, checks and balances, compromise, oversight, pluralism and so forth, then bicameralism is often a logical auxiliary institution.

      That’s an interesting scheme for presidentialism, but I think strong, elective upper houses are a bad idea for presidential systems, as they add another veto gate while making accountability and elections even more complicated for voters. If you have to have presidentialism (and please avoid it if at all possible), the best kind of bicameralism is one where the upper house is appointed or indirectly elected in some way that makes it relatively independent from the lower house and very independent of the president, and its powers are relatively limited (e.g. very brief delay for money bills, joint-sitting override otherwise). Ironically, what actually exists in the world is the opposite: the strongest forms of bicameralism are most common in presidential systems, while unicameral legislatures are disproportionally parliamentary.

      The even-numbered day position you describe is exactly equivalent to the statement often attributed to Abbe Sieyes: “if an upper house dissents from the lower house, it is dangerous; if an upper house agrees with the lower, it is superfluous.” Possibly my least favourite quote on any topic, let alone constitutional design.


      • Indeed, JD, I was thinking of Abbe Sieyes but as filtered through the Right wing of the Anglosphere which holds that (a) all electoral systems bar FPTP produce hopelessly deadlocked hung parliaments (sheesh, I mean, they even persuaded 68% of British voters that this was true of Alternative Vote), and (b) any elections at all for the upper chamber must of necessity produce a government or opposition majority.,
        Whereas it is possible that an electoral system other than FPTP might produce a hung upper house but, being an upper house, this might perhaps be a good thing? Maybe…? Just possibly…?
        As noted, this conjunction of sound-bites is common on the Anglosphere Right. It does not seem very well thought-out. But I repeat myself…


      • Just for interest I ran through the Wikipedia list of legislatures and compared the size of upper houses from largest to smallest. I excluded nondemocratic parliaments and parliaments with upper houses smaller than the US..

        Britain 808
        France 348
        Italy 315
        Spain 266
        India 245
        Japan 242
        Romania 137
        Indonesia 132
        Mexico 128
        Canada 105
        Pakistan 104
        Colombia 102
        Poland 100
        United States 100

        Britain is an outlier, but so is the US when you compare its population with the upper houses around 100.

        The ratio of population to upper house member in the US and the 2 countries on that list with larger populations than the US are instructive.

        India 4,847,900
        Indonesia 1,798,669
        United States 3,143,467

        The highest and lowest populations per seat from that list is:

        India 4,847,900
        Britain 81,388


      • A US senate with 1/3 of the states electing 5 senators by STV every 6 years would not be unreasonably large by international standards and would approach the level of proportionality of the Australian senate while maintaining the equal representation of the states. It is very likely not all senators would be Republicans or Democrats. The transitional arrangements would be a horror show, if you could get it passed.


      • As positive as electoral reform of the Senate might be, I think reducing its powers to be more important, even if it would be harder. Abolishing direct elections and turning to something more federal might also be better.


      • Turkey’s former Senate had 150+ members (like Italy and Chile, former presidents got seats for life) so, had it not been abolished, it would be about halfway down that Top Ten list.
        I think the ratios in Sweden before 1974 were something like 150+ upper house, 230+ lower house so unicameralism actually meant a slight reduction.


    • The NYT definitely based its calculations off the cube root law. I think it even included a link to Taagepera’s original paper on the topic.

      It is good to have papers on both coasts mention it! Now, we need some from the “heartland” to join in.


  7. There are very few “eminent Caliornia political scientists” who have a spare couple of million dollars lying around. That’s what it costs to get a proposal on the statewide ballot. Heck, there are very few Californians, period, with the resources to do this.

    I’m sorry that I misremembered the New York Times article and appreciate the correction.


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 )

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.