Expand the House

Peter Baker, writing in the New York Times, picks up a theme that has been prevalent here for a while: the US House needs to be expanded, specifically to improve the reapportionment process and to restore one-person-one-vote.*

Not counting a two-election increase when Alaska and Hawaii were added** the House size has not changed since the 1912 election. Back then the US had about 95 million people, or around a third what it has today!

The House used to be expanded periodically to track population size (see graph at the second-linked item). Why not now? As the NYT notes, the US judicial system is about to be asked that question.

Some advocates of increased House size have suggested a House of over 1,000 Representatives. That’s ridiculous–and hardly helpful to the cause. The cube-root law (again, see second link) would suggest 620-660. But, really, even 600, or 550, would help restore Representativeness considerably.

Update: Nate Silver weighs in, but suggests focusing energies on expanding and reforming the Senate. I’m all for that, too.

* To one of the arms of the federal government, anyway.

** That is, those states came into the union between censuses, and a seat was added for each. With the subsequent reapportionment, those states’ Representatives came at the expense of voters in other states, in order to return the House at 435.

28 thoughts on “Expand the House

  1. One thing I never understood was why people insisted on one person, one vote in assemblies. In elections it is a good idea, of course, but really it’s not immediately obvious why a person who represents, say 31000 people, should have the same voting strength as one that represents 16000.

    This way, one could achieve proportionality/representatives without necessarily making the assembly huge.

  2. I do agree with increasing the size of the U.S house. I think increasing it to 669 members by using the Wyoming Rule is a bit too much. A modest to a 455 member House would be best after the 2010 Census, (which is 20 members more) and then allow the House to grow larger with each census.

    What the most numerous parliament in the world? How larger can a legislative body can be? I have heard that most legislative bodies grow no larger than 670 members. Can a parliament size limit be up to a 1000 members? How would such an assembly even function?

    Can increasing the size of the U.S senate be done by statue without Constitutional Amendment? Should all states have 3, 4, 5 members each?

    If states had 3 members, then we would have a 150 member Senate, and if states have 4 members we would have a 200 member senate, and states have 5 members, we would have a 250 member senate?

    I think it would be better if a state had three members, but that the third senator be appointed by the governor with the approval of the state legislature. What does everyone else think? That might be a good starting point to reform the procedure where the governor appoints a vacant senator seat without ratification by the state legislature albeit it is a strange rule for a country that requires members of cabinet and everything else to be approved by the legislative branch.

    Also while we are at it, would it be wise to look into electing House members using multi-member districts either using open party list PR, STV, or even MMP.

    Would it even be possible to amend the U.S constitution to set aside a nation wide tier of at large seats that consists no more than the size of 10% of the house? This is to enshrine nation wide proportionality and to end gerrymandering once and for all. I’m talking about leveling or adjustment seats by what the popular vote of a party gets.

  3. I like the idea of a sliding scale, like the cube rule. As Madison noted, no matter how small the population, the legislature should not be a tiny clique, and no matter how large, it should not be a confusing multitude.

    For Australia, I have long thought a formula like “divide [actual population plus 20 million] by 200,000” would give us a reasonably-sized House of Reps – averaging between (a) 1 per 100,000 and (b) around 200 MHRs.

    Eg, at present, with 21 million people, it would mean a House of 205. In the 1940s, with 7 million little Australians, a House of 135 (instead of the 75+/- that actually obtained from 1901 to 1948). With current projections of 35 million Australians in 2050, this would mean 275 MHRs. And if the population ever reached 50 million (the maximum, scientists say, this continent can sustain), there’d be 350.

    A similar formula could be adapted for the USA, albeit with bigger population numbers.

  4. The advantage of the Shugart number is that it has an underlying rationale. Why select 20 million as the add-on for the Round number? Why select 10% for the size of the Suazzaprodi national tier?

  5. Why select any number of seats for a legislative house? Why Switzerland’s 200 or Greece’s 300 or South Africa’s 400 – or the UK’s (roughly) 1 per 100,000? There’s a huge amount of line-drawing. The closest I know to a “rational” size is the US Senate (50 States x one Senator by majority vote per election) – it could only be made more “rational” if the US switched to 3 Senators per State or else changed from three rotation classes to two.

    The problem with a global (in both senses) formula like the cube root is that a country doesn’t just fix the size of its lower house in isolation. It seems to be affected by whether there is any upper house at all: whether the upper house has directly elected members; the country’s geographic area; and whether there are powerful (and bicameral?) regional and local councils as well. Trying to come up with a global, one-size-fits-all formula that can apply to all countries (and believe me, many envelopes have been covered with many numbers in my attempts) seems to collapse into too many variables. The cube root is a good predictive device, but where it departs too far from societal expectations (as MSS notes above that it would in the US), it’s the cube rule that gives way. Whereas the Round Number (thank you) is one that I could imagine, in the right circumstances, a majority of Australian voters in four States accepting at a referendum to replace the current nexus ratio, which bears no relationship to either population or House functionality.

  6. I have this pathetic belief in the rationality of the electorate. I doubt the Round number would get the numbers at a referendum. The yes case would argue ‘We picked the Round number because it feels nice’. The no case would scream ‘More politicians!’.

    The electorate, at least the hypercynical Australian electorate, would determine, quite rationally, that something smelt bad about the referendum and vote no.

    The 1999 republic referendum is a precedent where the republicans managed to get themselves manoeuvred into a corner where their principal argument was that the republic would make no difference.

    There is a fixed number of MPs because there needs to be. You cannot have an assembly where the number of members alters hour by hour or day by day. There is then a question about what number is best in terms of which values.

  7. Alan, we seem to have crossed our wires somewhere because I am not arguing for a constantly changing number of MPs. Rather, what I am saying is that every (say) five or seven years, when the boundaries get majorly re-drawn, the number of MPs gets revised (as in Britain and Ireland – and, indeed, Australia, although only by 1-2 seats depending how the remainders round off) via a transparent and fairly simple formula that gives 50% weight to population size and 50% weight to having a House of Reps that will, in light of the likely population of this particular polity, not deviate hugely from 200 MHRs.

    If you think this is too complex to sell at a referendum, why would you think the cube rule would be any easier?

  8. Tom at #5, so when a feather floats on an air current, does gravity “give way”?

    The cube rule is the one equation that covers all national assemblies, within a strikingly narrow range of deviation, and in spite of the fact that few actual setters of legislative size are likely ever to have known of its existence.

    I would imagine that a model for subnational assemblies and councils could be created within the same parameters (i.e. number of communication channels), but I do not know if it has been tried. Taagepera already has shown the way for taking into account second chambers and supranational assemblies.

  9. Tom at #7, I do not think anyone has ever proposed that the cube rule be used as an actual formula for a specific country’s assembly size, although it would be an interesting exercise to try and propose it. (We accept economists’ models for some aspects of economic policy, so why not a political scientist’s model for some aspects of political-institutional policy?)

    The biggest advantages the cube rule has are its mathematical elegance and its solid logical foundation. Still, it is a multi-country generalization of an observed pattern, and not a prescriptive model.

  10. Tom, I was answering your rhetorical point about why particular numbers are used for particular assemblies. There has to be some number and various places fix it in various ways. I agree Australia has one of the silliest both in its stated reasons (the senate has to be at least half the house or it will tend to lose influence) and its practical effects. I would guess the nexus was designed with joint sessions after double dissolutions in mind, but that could easily be dealt with by a proportional formula

    I also agree that the number of representatives should be related to population and should be regularly revised. In Australia that happens frequently in particular states, where population shifts cause the transfer of MHRs from shrinking states to growing states. I notice New South Wales is in th process of losing a seat to Queensland at the moment with the consequent redistribution’s in both states. Sadly what doesn’t get revised is the overall number of MHRs.

    Our difference is that it seems to me that if you go in search of a non-arbitrary number for the lower house you can do a lot better than adding an arbitrary number to the population and dividing that by a quota. Why not simply divide the population by a smaller number? Or better argue for the Shugart number which has serious empirical evidence to support it.

  11. The National People’s Congress of China has 2,987 members, and the Supreme Soviet of the USSR had 1,500 members. I’ve read arguments that the large size of these bodies ensured that individual members had very little power, but there probably were other factors in play. In 1989, when the Supreme Soviet briefly acted as a real legislature, its membership was cut down to 542.

    To limit the examples to democracies, the British House of Commons has generally had over 600 members in the twentieth century, and exceeded 700 at one point. Again, I’ve seen arguements that its easier for the party whips to control such a large body, because individual members just don’t have that much influence.

    On the other extreme, the 550 elected members of the Lok Sabha represent an average of 2.2 million constituents. The US House would probably come in second, the average district size is probably currently about 700,000.

    If the House of Representatives expanded to 642 members, the House of Commons level, that would cut the average district population down to about 470,000, about where it was in the 1970s. This would mean increasing the size of the House by 50%. I’m not sure if this is really worth it. The problem here is really that the US has too large a population for there to be a close relationship between members of the legislature and their constiuents.

    To my knowledge, the US is the only country where legislators draw their own district lines. Consider a metropolitan area of a small city, with a population of around 700,000. If this area is mostly covered by one House district, as it most likely would be with a neutral boundary commission, then all the local media would cover its one Congressperson. The Congressperson could well be the former mayor of the city, or a former country commissioner. Now consider the usual situation in the US, where parts of the metropolitan area are put in three different districts, all of which include other parts of the state in other media markets, and one of which has its population center in another metropolitan area. I suspect that voters would know much more about their Congressman in situation A than situation B.

    In other words, the large size of Lok Sabha districts probably does not prevent them from being more representive than the US House districts, because India uses a neutral boundary commission. There is a long list of other reforms that would strengthen democracy in the US that should be tried before expanding the size of the House.

  12. The Roman Senate had a traditional membership of 900 for a considerably smaller population than the US. The Greek cities had even larger assemblies. Far be it from me to argue for the superiority of classical assemblies, but larger bodies than the US house have functioned quite successfully. The numbers in ceremonial legislatures like the Chinese NPC and the pre-perestroika Supreme Soviet do not really matter. Zero divided by any number is still zero.

  13. The Roman Senate was not elected and at least in composition was more similar to the House of Lords than to any current legislative assembly. The House of Lords had over a thousand members before the hereditary peers were removed. I imagine, as with the House of Lords, that a relatively small proportion of Roman Senators attended regularly.

    Other classical assemblies tended to be for voting than for deliberation. Sometimes the “assembly” was essentially the male citizens of the community. Modern representative legislatures function quite differently.

  14. The Chinese National Peoples Congress is not and the former Supreme Soviet was not elected and at least in composition they resemble the House of Lords more than any current legislative assembly. It is therefore hard to understand why they count in the numbers debate but the Roman senate does not.

    The Greek assemblies have not been well-studied. The documentation is scanty anyway and early American thinkers like Madison and Jefferson did a quite inaccurate hatchet-job on Greek democracy because they wanted to drape themselves in togas and build lots of domes and pillared facades.

    We do know the Spartan assembly was thought to have very limited powers because it functioned in the way Ed describes because Aristotle says so in his Politics. We do know that at least at Athens the assembly was fully deliberative with an important distinction between standing laws and temporary decrees, bills, amendments, rules of procedure, etc etc. The quorum was 6000.

  15. The EU Parliament comes close to the US Congress in the contest for the second-highest number of constituents per member, it’s currently at ~680 000. It could exceed the American number in the next few years, depending on how quickly various countries are allowed to join and whether the Treaty of Lisbon is ever implemented.

  16. This isn’t really relevant, but the People’s National Congress definitely is elected, and the Chinese government is even allowing more than one candidate to run for a few of the seats.

  17. Communist countries have long been noteworthy for having exceptionally large assemblies (discussed in Taagepera and Shugart, 1989).

    If people’s democracy is good, more people’s democracy is better, right?

  18. The NPC is formally, although indirectly, elected.Roughly each tier of government has a peoples congress that elects delegates to the next highest level. Provincial peoples congresses are limited to 110 candidates for every hundred seats. That there are only slightly more candidates than seats rather nugates the concept that the election is in any way meaningful.

    I accept the NPC has been showing some slight signs of life since the early 90s. In calling it unelected I was merely following the ancient Confucian practice of the rectification of names.

  19. A strong committee system (especially with a different selection method), combined with two developments since Madison’s time, would prevent a very large House from being a mob, or from being too gridlocked to do anything. First, limited debate time rather than the unlimited time practiced by the Senate, which was also the practice of the early House and, unless I’m mistaken, by every assembly the Founders had been familiar with. When you consider how the Constitutional Convention went in circles for several months, of course Madison wouldn’t have wanted to see several hundred members, let alone thousands of them — but the procedures are much more formal now, and floor debate itself is largely ceremonial, with the actual deliberation happening elsewhere. Second, the internet and modern communication technology in general means there’s no reason they have to meet in person. If we want to keep the (already largely ceremonial) live floor debates and votes, amendments from the floor can be replaced by online consideration by the Committee of the Whole, and the quorum can be defined by members able to participate by registering their votes, not by those physically in the chamber (which is already the de facto definition).

    A much larger House has benefits well beyond the fact that the more members there are the less inequality there will be in the apportionment process. Although that’s true, it distracts people to the Senate; even though the Senate is indeed apportioned even more unequally, or not “apportioned” in the usual sense at all, talking about fixing it is a waste of time because of the entrenchment clause. The House, on the other hand, can be expanded by statute.

    Campaigns can be run less expensively, members will be more like their constituents, and a local activist is more likely to seriously impact a race. Every state can have enough Representatives to implement proportional representation.

    More, the House will have enough manpower to republicanize the regulatory state. Regulatory rulemaking is too many steps away from any kind of direct popular accountability, and the accountability when it does exist is through the incredibly blunt instrument of a Presidential election; the President is a single winner whose election is essentially unrelated to any given particular regulatory issue. The public comment period is grossly inadequate, more like a calculated insult: the sovereign people is invited to make comments and then the actual decision makers will take it into consideration. To a large extent the Congress passes the buck to the executive precisely in order to prevent democratic accountability.

    But there’s another reason, one which I think Madison himself expected. In a little essay called “Consolidation” (498-500 in the Library of America collection) he predicted that the sheer volume of business to be handled by a unitary state (and the same principle must apply in proportion as the federal government grows more powerful, because the volume of business grows proportionately with the responsibilities assumed) would force a transfer of decision making from the legislature to the executive. Because the legislative process given by the Constitution cannot address the number of things now governed by the federal government, and a traditional legislative process of any kind cannot cover it because of the coverage and detail and adjustability now called for in the rules, Congress delegates its non-delegable legislative responsibility to executive-branch agencies.

    The more this happens, more more aggrandized the President becomes (which was Madison’s concern), and even worse, the more we come to assume that making decisions is for credentialed experts, not the people, which is directly anti-republican.

    The solution is to transfer regulatory rulemaking to the House committees. The House will need many more voting members, preferably enough that every member can belong to no more than one committee, and that some can belong to none. The composition of committees should also change, so that members are elected through some proportional method by the House as a whole instead of by appointed based on seniority. Policymaking will be vested in the people’s representatives, not bureaucrats, as the Founders intended.

  20. Doesn’t Italy use a system somewhat like what Aaron’s outlined? except that Parliamentary Committees can actually enact and repeal statutes, not merely regulations. A specified minority of the Committee’s MPs can demand the matter be referred back to the plenary Chambers but Aaron’s proposal to make it two-tiered is constitutionally more elegant.

  21. I. I’ve always thought what is called delegated legislation elsewhere (because it remains subject to legislative review) should be called surrendered legislation in the US. In addition the president seems to have acquired large legislative powers as a function of the commander-in-chief role.

    2. The Australian House has a a parallel legislative stream called the Main Committee. All MHRs belong to it and the deputy speaker chairs it. It can only pass an item unanimously or send it back to the House as an unresolved item.

  22. My understanding is that the US Supreme Court, in a (to me) misguided enthusiasm for purist “separation of powers” over “checks and balances”, has invalidated the legislative veto and refuses to let Congress reserve itself a power to disallow not only one-off administrative decisions, but even delegated legislation, authorised by an enabling statute.

    I like AA’s plan. It’s not as if Congressional Committee chairs don’t already exercise a lot of power. (Vividly illustrated many years back when a US remake of a BBC show – I forget which one, but it was a spy thriller – converted the character of the British Cabinet Minister into, not a US Cabinet Secretary, but the Senator who chaired the intelligence committee).

  23. I agree with Tom’s enthusiasm for the plan, although I would quibble with some details, like restricting representatives to a single committee. I would hope the committee chairs were also elected by some accountable process.

  24. I wasn’t sure how people would take the idea, so I’m glad at least a couple people like it. A few hundred million to go…

    Tom Round;

    I looked up the website of the Italian Parliament but I didn’t see anything like that, but maybe I just missed it or the English version doesn’t have everything. I did see that their Constitution requires that (ordinary) legislation be sent to committee before going before the full chamber, which is interesting. I would assume Government bills are approved in committee as a matter of course. It also requires that each section of a bill pass separately before the whole bill is voted on; I wonder if this has different effects from just allowing amendments that strike sections of bills.

    I also looked up the Supreme Court case you mentioned; it is Immigration and Naturalization Service v. Chadha. The grounds the Court relied on were bicameralism and the Presentment Clause. I need hardly point out that this contradicts the Court’s repeated finding that delegated rulemaking authority is permissible, since regulations are not passed by both chambers, and are not presented to the President (and even they were, they are not presented by Congress, otherwise it would be ordinary statute-making legislation, not the kind of regulation we’re talking about). Justice Stevens is the only member of that Court to belong to the current one. Unfortunately he was on the wrong side. Justices White and Rehnquist (who was not Chief Justice yet) dissented and Justice Powell concurred on a narrower ground. I have no idea if there’s any particular reason why the two dissenters were conservative Justices, or if that would carry over to the present.


    My idea is that members would have more time to devote to some particular area of policy. Maybe a low maximum, but more than one, would work just as well.

    On chairmen: of course. Either they should be elected by the committee from their own members, or else chosen by the House along with the rest of the committee.

  25. Chadha reads really strangely to an Australian because the equivalent language in our constitution is almost identical to the US constitution, and where it varies it is more restricted than the US clause. The High COurt of Australia reached a precisely opposite to conclusion on more restrictive language. Naturally neither the queen nor the governor-general ever exercise their veto powers except on ministerial advice and no bill ever passes without ministerial approval because the cabinet have a majority in the house.

    Regulations are only ever rejected by the Senate because they are made either by the governor-general on ministerial advice or by ministers themselves, so a vote on a regulation in the House is always a question of confidence.

    Equally the commander-in-chief language is almost identical but that title, which belongs to the governor-general, has been held tone am independent source of authority as it is in the US, indeed one former governor-general derided the whole notion of the commander-in-chief as man on horseback in a speech to our military academy.

  26. Given the size of the US, and the concerns about oligarchy expressed during the debate on ratification, I think a house of thousands is worth thinking about.

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.