A well-thought proposal! Congratulations. It was well worth the wait!
I'll need to chew on your general system here. I totally agree that the Senate should have different virtues and different vices than the House, but I'm not sure whether these specifically would be the ones I'd choose. It'd be an improvement over the current system in isolation, but I want to chew on it coupled to your new House and see what blind spots you're leaving in Congress.
And, I share your great concern about writing "DW-NOMINATE" into the Constitution by reference... but you're right; sometimes the best system isn't a simple one.
But while I chew on that, several smaller issues that jump out at me:
* In your proposal, if nobody gets chosen by Section 2 consensus, Section 3 says the Senator must be one of several specific members of the nominating state legislature. I'm dubious about what this will do to the legislature. Worse, I can easily imagine a specific legislator who's somehow unable to move to DC (whether due to medical issues, family issues, or any number of other things) who messes up the political calculus of balloting. Perhaps the legislators in question could have the option of nominating someone else (by secret ballot so it's unknown who's nominating them), rather than themselves be automatically nominated? Or, simpler, decline the nomination and have the next-closest-scoring legislator be drafted?
* As a nitpick, Section 3 doesn't mention the redistribution of votes after a nominee is eliminated.
* I don't like suspending the Senatorial election when no eligible candidates present themselves under Section 3. I can see how it might be good to completely prevent party bosses from manipulating things thusly - but I think it'd be impractical for them to make all longserving state legislators resign! So I'd replace that with something like the three legislators nearest to qualifying.
* I absolutely do not want the House to be able to immediately change the Senatorial election method. At least, follow the Parliament Act of 1911 and make the House pass the same proposal twice over two years (which in our case will mean a general election intervenes).
* Plus, as another nitpick, Section 4 shouldn't say "Congress" but just "the House of Representatives."
* Section 7 will be the first time the word "justiciable" appears in the Constitution. I don't like the concept, and I don't want to implicitly approve it by leaving the implication some things aren't justiciable. And besides, this threatens Article 1 Section 5 Paragraph 1 where "each House shall be the Judge of the Elections, Returns and Qualifications of its own Members" - so I'd rather just restate that clause with the desired changes. So, what do you imagine justiciability looking like here?
1) I would add a topic or substantial similarity limitation in section 4, just in case
2) I disagree with footnote 3, I don't think people who disagree with the Senate should be married to the stuff beyond the text (I do however, think the clause entrenches itself, but restricting the Senate's powers is fair game)
And an unrelated comparative constitutional law/design question which I may or may not have asked already: what do you think about constitutional courts. They usually(!) a) are the primary body that interprets a constitution
b) decide on individual constitutional complaints against ordinary courts, in a collateral action similar to a habeas proceeding
c) are the sole body that can judicially review a statute, and they do it abstractly, vacating it for everyone. If an ordinary court (there usually exist a separate ordinary supreme court and other courts) thinks a statute is unconstitutional it has to refer it to the constitutional court. They also usually judge impeachments, for instance most recently in South Korea, but in lots of countries parliament and the Prime Minister are more powerful than the president.
For example, here in Croatia the constitutional court (see also powers section https://en.m.wikipedia.org/wiki/Constitutional_Court_of_Croatia ): a) decides on individual complaints, mostly in panels b) decides on the constitutionality of statutes, and the constitutionality and legality of administrative acts (administrative courts can also do the latter).
It can do b) sua sponte, on a request by some listed governmental body or on a proposal by any natural or legal person anywhere (yes, really). It also does some separation of powers and election-related stuff, see wikipedia.
However, ordinary courts can ignore legislation if it is contrary to treaties (which are considered self-executing and superior to statutes) or to EU law.
Interestingly in the Netherlands this is the only form of judicial review of statutes, the Dutch constitution bans judicial review of laws against the constitution itself, but they review them against international and EU law.
My questions are: 1) what do you think of this system, both in general or as maybe applied to America in an alternate world. I imagine if the US were to implement something similar it would be tacked onto SCOTUS instead of making a new court.
2) what do you think of moving impeachment judgments in the US to SCOTUS (but in most countries the impeachment itself takes 2/3 of the legislature, but I myself think majority is fine. It also usually takes 2/3 of voting judges to convict someone.)
3) what would you think of doing APA vacatur in the US, but for statutes, would that even be constitutional in the US. (Also, do you think the current APA vacatur is constitutional, seems at least Thomas doesn't, maybe you wrote about that?)
At the constitutional convention, the framers considered a Council of Revision that could veto legislation that it judged to be unconstitutional. (This was part of the Virginia Plan.) The Council would have been composed by the president and a number of federal judges. Ultimately, this proposal was narrowly defeated, and the veto was vested in the president.
In the U.S., the dispute over the role of the courts has been renewed by the debate over universal injunctions. The two basic views are judicial restraint (i.e., the role of the courts is to resolve the cases and controversies before it) and judicial activism (i.e., the role of the courts is to police the government for unconstitutional laws and actions).
Simply put, I lean towards judicial restraint because I do not believe the courts should be "the primary body that interprets" the Constitution. All three branches have an essential obligation to interpret the Constitution and ensure it is faithfully executed. Now, that is not to deny the role of judicial review, but used as only necessary to resolve said cases and controversies.
A process like APA vacatur for U.S. statutes would destroy the separation of powers. This would allow a judge to vacate any law he deigns is "arbitrary," "capricious," "an abuse of discretion," or "unsupported by substantial evidence." This is not the role of the courts. In fact, the Supreme Court recently struck down this sort of nonsense in U.S. v. Skrmetti.
Per Justice Thomas's concurrence in that case, "Taken together, this case serves as a useful reminder that the American people and their representatives are entitled to disagree with those who hold themselves out as experts, and that courts may not 'sit as a super-legislature to weigh the wisdom of legislation.' Day-Brite Lighting, Inc. v. Missouri, 342 U. S. 421, 423 (1952)." https://www.supremecourt.gov/opinions/24pdf/23-477_2cp3.pdf
Edit: To illustrate the terrors of "courts as a super-legislature," yesterday, a federal judge in the District of Massachusetts issued an ex parte TRO requiring the executive branch to fund Planned Parenthood, despite having just been defunded by Congress. Thankfully, the TRO itself is utterly unconstitutional (see the appropriations clause and OPM v. Richmond). But imagine the courts actually had such a "vacatur" power over duly passed laws. Insane.
Thanks for your thoughts! I was thinking of an APA vacatur remedy for statutes, but the standards for facial challenges would stay largely the same (the current ones are maybe a bit too strict with finding only one constitutional application saves the statute), just there would be a universal remedy at the end. I didn't think of importing the APA standards and concepts.
Apologies, I must not understand what you mean by "an APA vacatur remedy for statutes." Although I think your second comment might help clarify.
Do you mean nationwide relief? I.e., a judge striking down a regulation under the APA provides relief to non-parties to the case, whereas (post Trump v. CASA) a lower court striking down an unconstitutional statute only provides relief to parties.
In case it is helpful, I'll just note:
(1) The Supreme Court can already provide nationwide relief through its opinions and judgments. (See Justice Barrett's majority opinion in Trump v. CASA.)
(2) Congress might allow for nationwide relief via nationwide injunctions. (J. Thomas is skeptical this is constitutional. See his Trump v. CASA concurrence.)
(3) From 1937 to 1976, there were three-judge panels in the federal courts with jurisdiction to hear constitutionality challenges. Their decisions were directly appealable to the Supreme Court. https://sgp.fas.org/crs/misc/IF12746.pdf
1) I disagree, precedent isn't relief, there is no 1983 federal cause of action and Bivens is basically dead, so no damages, only prospective relief is possible against feds.
2) Yeah, I know. In my mind the vacatur works different mechanically than the injuction, it removes the act itself from the legal order, while injuctions operate on the parties being sued. That solution seems more elegant to me at least, tho I don't know if contempt is available. I suppose if the gov tries to use a vacated rule, they should be sanctioned for using non-existent authority. AFAIK there is also an APA provision for further relief in aid of an APA remedy, I wonder whether that allows nationwide injuctions post-CASA.
3) Interesting.
On CASA, long-term I don't see it doing anything, people will just say the magic words class-action. Especially after SCOTUS said in AARP that you don't actually need to do class certification analysis in a preliminary posture.
Re footnote 17: The best example of why not to use single-member IRV is perhaps ironically the one place (at least among more-or-less stable democracies) where it has lasted for over a century: Australia.
The system was introduced after a rural-focused centre-right party (the Country Party) emerged and split the centre-right vote in a safe seat, leading to the election of the Labor candidate. To avoid this happening again, the electoral system was changed from first-past-the-post to single-member instant-runoff.
The result of this has been, effectively, a two-party system masquerading as a multi-party system (though recent results have seen this break somewhat), where Labor is the main centre-left party and the centre-right is represented by a party focused on urban interests (called, at various times, Nationalist, United Australia, or currently Liberal) and a party focused on rural interests (originally Country when first formally organised, now National), which together are the Coalition. (Even with this, it has become increasingly rare for the Liberal and National parties to run candidates against each other.)
(This is actually true only in the two largest states, New South Wales and Victoria. The two centre-right parties merged in Queensland after disagreements between them led to the election of numerous Labor MPs in 2007; the additional candidates elected by the merged Liberal National Party account numerically for most of the shift away from Labor and toward the Coalition in 2010. In Western Australia and South Australia, the National Party exists but is not in formal coalition with the Liberal Party, and the one time a Western Australia National candidate was elected, in 2010, he strenuously objected to being counted as part of the Coalition's seat total by the media. The National Party does not exist in Tasmania or the Australian Capital Territory, while in the Northern Territory the centre-right party is the Country Liberal Party. Also, due to policy differences, the National Party split from the Liberal Party following the 2025 election and it is unclear whether a formal coalition will be in place in the next election.)
The result of this has been that even with a recent trend toward electing more candidates from outside the major factions, five or six in elections from 2010 to 2019, then sixteen in 2022 and thirteen in 2025, it is still the case that since the Coalition became more or less formal in 1934 there have been only two elections that have not seen one faction or the other win an outright majority, in 1940 and 2010. (Australia holds elections for its lower house roughly once every three years, sometimes more often, so this represents well over thirty elections in about ninety years! Which is less frequent than elections to most lower houses in the United States, to be sure, but more frequent than in most other parliamentary systems; even New Zealand's legislatures tend to actually last their full three years, and they've had thirty elections in that same time.)
As I've been given to understand it, an important factor in Australian elections is the two major parties trying to convince the plethora of minor parties to recommend that their supporters rank the one major-party candidate ahead of the other(s). This of course means making alliances with and concessions and promises to other factions before the election takes place, and Linz had something to say about that in "The Perils of Presidentialism".
Where Australia's multi-party system does shine through is its own Senate, where because each state has twelve Senators (each territory has two) they've used a proportional system (sort of) for elections to that body since 1948, and as a result neither major faction has been able to win a majority in it reliably since the 1960s or so (the exceptions being the Coalition in the late 1970s and in 2004).
I suspect that part of the reason instant-runoff has become popular is because no other system has ever really seen real-world use, at least in the political sphere. (As I've mentioned before, Borda, or at least partial versions thereof, gets used to select MVPs and other such awards in many sports leagues.) Australia uses it in a single-member context. Ireland uses it in a multi-member context (and tends to elect hung legislatures). Malta also uses it in a multi-member context but Malta is tiny, effectively a city-state as Valletta has expanded, and the two major parties (Labour and Nationalist) have very strong brands and there are almost always more candidates from each party in each district than there are seats in each district (each elects five). (For instance, in 2022, District 1 had eleven Labour candidates and seven Nationalist candidates, electing three and two respectively. Every district had at least five candidates from each major party.)
Nitpick: When listing off moderates from 2010, did you mean Blanche Lincoln (from Arkansas) or Ben Nelson (from Nebraska)?
Next: If you want to make your specification of DW-NOMINATE more proof against chicanery, you can say something like "as defined on such-and-such date". (More on this later in this comment.)
On banning bullet ballots (and requiring full ranking), to avoid the centre squeeze issue, why not incorporate Coombs' method? Maybe it has some failings that lead it to be less desirable than even Hare, but on paper it knocks out extremists who make a substantial chunk of the electorate vomit even as another chunk loves them.
Final suggestion: I said this before in the context of electing Popes, but just elect Senators in the way Venice elected its doge. Yes, many state legislatures are too small to use the full procedure, so instead specify the election of a representative assembly in a manner consistent with the method of election of the most populous branch of the state legislature and no smaller than it, and no smaller than, say, two hundred members. (For instance, using multi-member elections from state House districts that have to use a proportional method of some description, be that Australia's Senate method, or maybe Austria's method.) Then say that that body will elect the Senator out of all eligible citizens of the state. The operative clause here is that the method of election would be "the method of election of the Doge of Venice as of the year twelve hundred and sixty-nine" and if it looks weird to specify an exact time or year like that (I said I'd come back to this!) then I guess Section 18 of the Constitution Act, 1867 and especially Section 50(2) of the Constitution Act, 1996 (BC) must look weird! (But it makes sense there because while parliamentary privilege is really important, it's also extremely important that legislators not be allowed to just arbitrarily vote more privilege for themselves!)
Of course, this does have the problem that now Senators are being elected by random folks who might've run for the temporary assembly on a whim and gotten in effectively by luck, instead of by serving politicians who, whatever their faults, are at least generally well-versed on the issues of the day and familiar with the prospective candidates. (Unless every state can be convinced to have a lower house the size of New Hampshire's, anyway. Using state legislatures for this procedure would work there, I guess, but probably not in a state like Nebraska, Alaska or Hawaii. I guess you could specify that all serving state legislators would be part of the assembly and the random folks elected to top it up to the necessary minimum, but that also presents other problems.)
(This would still need a lot of work to make sure that the elections for those assemblies can't be too harshly gamed and whatnot but forcing multiple rounds of supermajority consensus where nobody even knows who will be voting in the next round should tend to knock out extremists, unless you get very unlucky.)
Combining three unrelated comments into one so as to be less spammy:
Section 1 of the comment: Enumerating the low-hanging bombs:
There is no topical limitation on section 3, so section 4 is effectively an unlimited power to amend the constitution.
DW-NOMINATE could be gamed by holding a lot of roll-calls on do-nothing motions. I think probably a majority party could even arrange the order of the role-call and thereby vote while already knowing all minority votes. If that flies and they do it often enough they can make the rankings of their own members whatever they want. Even if you ban that it is still easy for both parties to make a pact to statistically demoderate their respective moderates. The reason it hasn't been Goodharted yet is because it hasn't been made into a goal yet.
The amendment has a hard-coded assumption that recorded votes by name and roll-call votes are the same thing. I don't know if this is currently true as a matter of American ritual but it is certainly gameable. Just hold votes likely to produce the desired statistics by roll-call and others by electronic device, signed ballot, division, or whatever.
In the edge-case of perfect party-discipline DW-NOMINATE will collapse the spectrum to two points and the ranking will become undefined. With not-quite-but-near perfect party discipline it will be much easier to game or near random.
I don't know how easy it is for isolated members to force votes in the various legislatures, but if it isn't hard it probably can be made so. If roll-calls just happen very rarely, new members can be kept disqualified for a long time.
I don't think the states are required to make consent of all branches of the legislature necessary for legislation. Certainly various kinds of overridable upper-house vetos are common internationally. So a state constitution could just make the governor into a single-member and therefore least numerous branch of the legislature with the same veto power he now has outside of it. Or even have a "house of the senate majority leader".
If the presiding officer of the state senate doesn't like the expected section 2 winner, he can force section 3 by not waiting for nominations and just recessing the section 2 session immediately after opening it.
Section 2 of the comment: A comparative legal observation.
As far as I understand it, unconstitutional and federally preempted laws in the US are unenforceable but stay on the books (and maybe become enforceable later if the higher law changes). Section 5 would create a unique exception by potentially making a provision of a state constitution not just unenforceable but actually void. In German legal terms, you have reinvented Geltungsvorrang (validity precedence) while previously American law only new Anwendungsvorrang (application precedence).
Section 3 of the comment: How to elect the least hated member
Note that the least controversial hated will usually not be a Condorcet winner, because a Condorcet winner still can have a fairly large minority who hates him. But electing the least hated member is actually fairly easy: Use your favorite method of proportional representation (which should be STV but here it doesn't matter much) to elect an exclusion panel that has one less member than your state senate. The person not elected to the exclusion panel is your winner. In a two-party system, effectively the majority gets to exclude all of the minority, and the minority gets to exclude a lot of majority members in order of hate. If the majority is comfortably large, they also get to exclude some of their own.
You left out a key benefit of your proposal: it would create new shovel-ready jobs for Fortran programmers! (DW-NOMINATE is written in Fortran.) OK, fine, I’ll take my messaging hat off.
As you acknowledge, this proposal is extremely complicated. Unfortunately, the proposal is so complicated that I worry it would diminish the Senate’s legitimacy. The People may not directly elect their Senators, but they must at least understand the process by which they are elected!
Most people are not going to understand DW-NOMINATE, Condorcet winners, Bonham’s method, etc. I do not mean this judgmentally. I am an economist and even I quickly bounced off the Wikipedia article on DW-NOMINATE. It looks like they are assuming Gaussian utility functions and estimating the bliss point of each legislator based on a PCA of roll call votes?
The epicycles of the proposal stem from contingent elections as a method to break deadlocks. Why care about deadlocks? Previously, you mention that some Senate seats would remain open for months (sometimes years) because state senators could not agree on a candidate.
So what? If the Senate has a quorum to conduct its business, then let a state deprive itself of representation until it get its act together. Ultimately, if the state senators are unwilling to compromise, then they face accountability from voters whom (it seems) do not like deadlocks. Admittedly, this accountability would be undercut by the proposal’s use of the secret ballot.
Maybe I am just revealing my historical illiteracy, but why were deadlocks such a problem? In your prior piece, you mentioned occasional riots. That is a law enforcement problem not a constitutional problem. You also mentioned (per the deadlocks bill) that state senates were required to spend time on senate elections before moving to other business. OK, abolish that.
Let me assume, arguendo, that deadlocks are a problem that must be fixed. You discuss the merits of a conclave-like system, but you distrust individual state governments to enforce secrecy. Yet, Congress can regulate the election of senators (Art. I, Sec. 4). Congress could set forth regulations requiring a conclave-like system. Congress could also impose stiff penalties for states who fail to abide by the regulations, such as rejecting their senator-elects.
While Congress cannot regulate the place of choosing senators, I can imagine such regulation including safe harbor provision for states who conduct their conclaves in DC with federal oversight and assistance. E.g., Capitol Hill police stationed outside the conclave to prevent unauthorized access. The feds do a good job of keeping things secret when they want to!
Separately, I am brainstorming ways to game the proposed amendment. One idea: a highly-partisan state could pass a constitutional amendment creating a third legislative chamber with two members, each elected every six years by popular vote in a statewide election. This third chamber serves no legislative function. It is merely a proxy election for the state’s U.S. Senator.
Take Illinois as an example. A state constitutional amendment in IL requires a three-fifths vote of each house of the General Assembly, ratified by a majority of voters in a general election. Democrats could pass such an amendment through the General Assembly on a party-line vote. State-wide ratification would be harder but doable. (IL voted 54% Harris, 43% Trump in 2024.)
Such a system would operate similarly to the electoral college. Voters select the ticket for their preferred Senate candidate. In actuality, voters are electing candidates to this third legislative chamber whom (if elected) will choose their corresponding candidate for the open Senate seat. While Section 5 of your proposal prevents explicit “pledging,” parties can just choose loyalists.
(Nitpick: In Section 1, I would write “least numerous chamber” not “least numerous branch.”)
Your effort to come up with a Senate election proposal that was resistant to gaming by the political parties, was ingenious. I could not help thinking of a proposed state constitutional amendment proposal put in the Ohio ballot last year led by a citizen group, to eliminate gerrymandering in drawing state and Congressional district boundaries. it was fiendishly complex, involving a redistricting commission of non-politicians, nominated by a panel that itself was selected by retired state judges, all with a budget set in stone, with detailed line drawing criteria, hearing requirements, etc, etc.. The whole thing was a substantial fraction of the length of the entire Ohio constitution, which is not short! (Sadly, establishment pols ran ads that lied and deceived, and the measure went down to defeat).
All that to say that game-proofing a process against political capture is hard, and you deserve kudos for even making the attempt.
All that being said, I keep wanting there to be a simpler alternative to avoid the deadlock election than DW-NOMINATE. I'm half willing to tolerate deadlocks, forcing state pols to go unrepresented in the Senate if they can't come to a consensus.
The more I think about this, the more it makes sense to just let the single Senate seat from a given state go vacant, if the state senators deadlock. There will be enough states that are able to elect Senators (due to one-party domination if nothing else) that the "purple" states will have a powerful incentive not to be left out, and come to a consensus to elect someone.
I have a minor quibble to make about the phrasing of Section 3:
"Each member qualified for the U.S. Senate who has served in the body for at least two full years (even if non-contiguous) and who has cast at least one hundred roll-call votes shall be eligible for nomination." -- I would argue that the plainest reading of this would indicate that we are looking for people who were already in the US Senate, unless I am failing legalese...
Perhaps:
"Each member qualified for the U.S. Senate who has served in the State's body for at least two full years (even if non-contiguous) and who has cast at least one hundred roll-call votes shall be eligible for nomination."
Also, after writing that, I am curious if you are restricting this to current members of the State's body?
I think Section 3 is too complicated, even without the problems of defining DW-NOMINATE, etc. Instead, i think something simpler and somewhat random would serve. So:
If no nominee is elected within two calendar days, all nominees who received a majority of approval votes shall be listed in order of age. If the list is at least three nominees, the nominee with the median age is selected. If the are are an even number of nominees, the older (or younger?) of the two median nominees shall be selected. If fewer than three nominees received a majority of approval votes, the list shall be all members of the voting body who are eligible to serve in the Senate.
Problem 1: What if none of the 'moderate' 'state senators' actually WANT to be federal senators? Moving their family to DC sounds like a lot of work. Maybe they would prefer to just stay home. Or maybe they were bribed to say that.
Problem 2: DW-nominate doesn't care about what votes are ABOUT, and all votes are equal. A vote to adjourn the house for lunch counts just as much as a vote to implement single-party healthcare.
So, if you are a state senator who WANTS to someday be a federal senator, the winning tactic may be to always vote with the opposition on votes where the opposition is doomed, or where the vote doesn't matter. And it's easy to game this. The party in control of the house decrees that there will be three votes a day, on whether to break for breakfast, lunch, and dinner. 40% of the controlling party is ordered to either abstain or vote against the break. Bam, the 'designated moderates', who aren't really moderates, just increased their moderation score by voting with the opposition three times a day every session.
If the opposition catches on, and starts to vote against breaks, then you simply order the sergeant-at-arms to chain them to their desks and not give them any food until they vote in favor of breaking for food, so the 'moderate' can vote with them. Or, you switch tactics, and 51% of all house members, the ones you control, vote FOR breaking for food, and the 'designated moderate' votes AGAINST it, same as the opposition.
There are also many shenanigans available by, say, controlling who gets to close the vote, when. Maybe minority-party legislators have to cast their vote and stick with it, but majority-party legislators can switch their votes en masse at the last second, and then the vote-counter closes the vote right afterwards.
With that out of the way, now it's time to break the amendment every in way I can think of.
Section One:
So, the relevant body is whatever branch of the legislature has the least members. We can easily convert this back into a statewide senatorial scheme by erecting a new chamber with a single representative. This would allow, effectively, for popular election (but done 2+ years in advance), or, I suppose one could have rig a scheme where, while nominally popular, in practice this chamber will usually filled by appointments from the governor.
While we're here, I'll note that I did check whether any states had equally sized senate/house, as that would cause problems. There are not, currently.
Section Three:
DW-Nominate can be rigged. At the end of a session, just spend many votes on utterly trivial matters. Have the one faction vote but in lockstep, but without informing the other side how they will vote. Have the desired nominee vote to the contrary. If any partisans of one's own side are defecting, one can probe by telling them the opposite, and cut them out—one only needs the buy-in of a majority to keep forcing votes. The body of votes added by this process will, on average, make the desired candidate seem more opposed to the majority party than *any* member of the opposition. Since DW-Nominate is deterministic (I assume?), repeat until they will be projected to be the winner. I'm guessing that this could be iterated enough to make the average (arithmetic mean, I assume) extreme enough that they should be viable, even if they were extreme in all past legislative sessions.
I don't actually know the details of the algorithm, so I don't know if that works exactly, but I assume gamesmanship like that could be done.
As to a voting system, definitionally, a majority will always be able to push through a Condorcet winner. If you want to force moderation, you do not want an electoral system that always elects Condorcet winners. As is, this drops the threshold for guaranteed partisan victory from 2/3 to a 50.01% bloc.
Section Five:
After the fact punishments seem entirely permissible by parties here. What is to stop them from allocating funding for reelection as they wish? Of course, the ballots being secret makes this harder to do.
Section Seven:
Expect partisan courts. Was that ballot *really* secret?
I think a common theme of some of the other comments made is this: this amendment assumes that political actors will continue to abide by the current conventions, but the optimal partisan political outcomes are achieved by changing the conventions, and thus the conventions will change. The only way to avoid this is for voters to be willing, en masse, to punish politicians who try to game the system, which is dubious on multiple grounds.
This is, in effect, the same problem faced by the people who want to use the current electoral system for President to bootstrap a nationwide popular vote via an interstate compact: political actors who dislike that can simply start playing by different rules. (For instance, state legislatures might remember that they can just directly appoint electors for President, and not bother to have any sort of statewide vote on the matter.)
EDIT: To elaborate somewhat, the point of this amendment seems to be that the Senate would become fifty clones of Sen. Moderate Milquetoast giving sober second thought (a much better analogy, in my opinion, and one actually supported by historical evidence as having been made) to the bills passed by the 218 clones of Faction A's Rep. Ragey McRageFace over the strenuous objections of the 217 clones of Faction B's Rep. Ragey McRageFace in the House.
The problem is this: what happens when Faction A games the system to return 25 clones of Sen. Ragey McRageFace but Faction B respects the spirit of the system to return 25 clones of Sen. Moderate Milquetoast?
To make this system stick in spirit, you need a certain critical mass of voters who would be willing to overlook any and all ideological differences with Faction B (you may ascribe to them whatever policies you personally find abhorrent for the purposes of this thought experiment and assume that even their Moderate Milquetoast clones more or less support them) to make sure that Faction A, as well as any future Factions C, D or E that swear up and down that they are not just Faction A with a different haircut, is driven out of politics forever.
Could you realistically expect to find such a critical mass, especially if Faction A makes it known that, yes, they are going to be sending their Ragey McRageFace clones who are popular with the base while Faction B makes it known that they are going to be sending their Moderate Milquetoast clones whom the base hates? State Senates (even if you allow them to be malapportioned again) are still full of plenty of clones of Ragey McRageFace because they are popularly elected, and as noted in other comments you can just hold meaningless votes to make one particular State Sen. McRageFace look like a moderate. (And that's to say nothing of the notion that some have brought up of just creating a third house that has only one or two members that only serves as a proxy for the Senate election; that might perhaps be mitigated by placing a minimum size of 15 on the legislative body that would select federal Senators, since Alaska's is smallest at 20.) If you can't find a critical mass willing to put democratic norms and constitutional conventions (or even constitutional text; as you've argued, everyone who voted for the current occupant of the Oval Office in November 2024 was violating the Constitution by casting a vote for a candidate who did not enjoy passive suffrage) above policy preferences, democratic systems collapse.
(You could, perhaps, implement a system like that found in the Basic Law of Germany whereby parties that sufficiently threaten the democratic underpinnings of the system are banished from politics, with strict monitoring to ensure that they don't crop up again having done little more than changing the colour of their shirts, which has seen the Sozialistische Reichspartei Deutschlands banned in 1952 and the Kommunistische Partei Deutschlands banned in 1956, and may yet see the Alternative für Deutschland banned following a recent report that they may meet the criteria for such. But then you are placing a lot of faith in the continued independence of that adjudicative process.)
Why not just make it illegal for party organizations to endorse Senatorial candidates? You could even have a broad, vague legal understanding of what exactly is a 'party organization' established through caselaw, to disincentivize attempting to work around the prohibition.
Something like, "Any private organization that runs, nominates, or endorses candidate for election to public office may not endorse, nominate or otherwise select a candidate for any election to the United States Senate".
Penalties that are onerous enough would get parties and affiliated partisan organizations walking on eggshells around the possibility of even being perceived as having endorsed. While more distantly related organizations may feel safe to do so, removing the central coordination point will make it difficult for single-issue advocacy groups, to unite around individual candidates capable of winning on their own, decreasing the salience of the Federal Senate elections to the choice of state senator (because no one party can credibly promise to elect a specific Federal Senator if they win majority control of the state Senate; even if private organizations have pledged state senators to vote for a particular Federal candidate, they're not necessarily THE SAME candidate across state senators).
By making it difficult for external organizations to express a universal preference, you make it easier for state senators to make their own decisions without having to fear directly defying powerful outside forces.
This isn't that major a new step in American law, either: we already do this in the tax code, where certain non-profits are banned from endorsing candidates. That would give you an existing body of law to build on.
This might be a crude solution, but what if we simply make it so that, after a certain number of rounds of voting, if the legislature can't agree on a candidate, they must all resign, and are ineligible from ever holding any political office ever again? That would almost certainly outweigh any pressure from party bosses.
You know, one of the easier filters would just be to ban people who have experience with first-past-the-post elections from being senators.
Try this:
Rule 1:
No person who has previously been a member of any house in the state legislature, or the Governor or lt. Governor in the State executive branch, or been a member of the US Federal House of Representatives, or held any other office which is normally determined by means of a statewide election, or who has been President or Vice President of the United States, may ever be a US federal senator under any circumstances, ever.
Each state may determine, using it's own laws, whether or not Town mayors, town councilmen, country sheriffs and other county-wide elected offices may be permitted or denied from holding the office of US Federal Senator, according to the preferences of each state.
And while we're at at it:
Rule 2: No person who has served in the Federal Judiciary may ever again run for, or hold office in, or otherwise be employed by, any part of the federal executive or legislative branches. But prior service in those branches is not disqualifying from being a member of the federal judiciary.
No person who has served in, run for office in, or been employed by the Legislative branch may ever again serve in the executive branch in any capacity, but prior service in the executive branch is not disqualifying from being a legislator.
Each state may choose whether or not to extend these rules so that prior service in that state's executive branch may or may not be considered equivalent to prior service in the federal executive branch for purposes of this rule, and shall also determine whether or not the state legislative branch is likewise equivalent to the federal legislative branch, and shall also determine whether or not the state judiciary branch likewise be considered equivalent to the federal judicial branch.
(E.G., Federal Justices can NEVER run for POTUS. States have the right to determine whether or not justices from that state may ever run for POTUS. Note that this is based on whether or not you ever served as a justice in that state, NOT based on which state you're living in when you run for office. If you were once a state district judge in Florida, and Florida says their judges count as federal judges for purposes of this rule, that means you can't ever run for, say, House Representative for the First District of Georgia, because your prior service in the Florida Judiciary blocks you from ALL future federal executive and legislative employment, not just the roles specific to Florida. )
Why not just start using Single Transferable Vote with a 2/3rds threshold, instead of going straight to complex algorithms that most people can't check on their own? I feel like once you get to the point where a computer scientist is involved with selecting candidates there's a sudden lack of transparency in the process.
God help us if there are incorrect entries in the databases used to calculate the candidates. And it wouldn't just be a "Well, the state made an error, but the state chose their senator so the senator has to be legitimate." It's in the constitution to use the algorithm correctly.
What do you mean by "Single Transferable Vote with a 2/3rds threshold"?
If by threshold you mean the number usually called the quota, and given the degenerate case of electing a single winner (that degenerate case also often being called AV or IRV), that would leave you with three sub-possibilities:
1) you insist on every voter numbering the whole ballot. Then you will ritually get a 2/3 majority by counting people as having "voted" for their last "choice", but the winner will always be the exact same person who would win with unmodified STV/IRV. If the ritual 2/3 is what you want, you might as well make it a ritual unanimity, which is also the same system.
2) you could do the same thing usually done with the usual quota and say the quota on every count is 2/3 (rather than the usual floor(1/2)+1) of the currently unexhausted ballots. This will also always elect the same candidate as the standard algorithm, in this case by eliminating more candidates to create more exhausted ballots. You again get something you can call a 2/3 majority in a very technical sense but it will be 2/3 of the wrong base group.
3) Or you really insist on a 2/3 of votes cast while simultaneously allowing truncated ballots. Then the system will often fail to elect anybody and you are in the same place as if you required the 2/3 without an STV wrapper.
Hey, @everyone! Lots of great, in-depth comments on this one, which is wonderful. (Please keep them coming for as long as the Spirit moves you!)
As I sat down to start replying to all of them, I realized that doing so would take many hours, and that, since several of you raised similar points, it would be duplicative as well. So I'm now planning to do a "highlights from the comments" follow-up to this post in a few weeks (after I take my time off and then post the next part of If They'd Made Me Pope).
A well-thought proposal! Congratulations. It was well worth the wait!
I'll need to chew on your general system here. I totally agree that the Senate should have different virtues and different vices than the House, but I'm not sure whether these specifically would be the ones I'd choose. It'd be an improvement over the current system in isolation, but I want to chew on it coupled to your new House and see what blind spots you're leaving in Congress.
And, I share your great concern about writing "DW-NOMINATE" into the Constitution by reference... but you're right; sometimes the best system isn't a simple one.
But while I chew on that, several smaller issues that jump out at me:
* In your proposal, if nobody gets chosen by Section 2 consensus, Section 3 says the Senator must be one of several specific members of the nominating state legislature. I'm dubious about what this will do to the legislature. Worse, I can easily imagine a specific legislator who's somehow unable to move to DC (whether due to medical issues, family issues, or any number of other things) who messes up the political calculus of balloting. Perhaps the legislators in question could have the option of nominating someone else (by secret ballot so it's unknown who's nominating them), rather than themselves be automatically nominated? Or, simpler, decline the nomination and have the next-closest-scoring legislator be drafted?
* As a nitpick, Section 3 doesn't mention the redistribution of votes after a nominee is eliminated.
* I don't like suspending the Senatorial election when no eligible candidates present themselves under Section 3. I can see how it might be good to completely prevent party bosses from manipulating things thusly - but I think it'd be impractical for them to make all longserving state legislators resign! So I'd replace that with something like the three legislators nearest to qualifying.
* I absolutely do not want the House to be able to immediately change the Senatorial election method. At least, follow the Parliament Act of 1911 and make the House pass the same proposal twice over two years (which in our case will mean a general election intervenes).
* Plus, as another nitpick, Section 4 shouldn't say "Congress" but just "the House of Representatives."
* Section 7 will be the first time the word "justiciable" appears in the Constitution. I don't like the concept, and I don't want to implicitly approve it by leaving the implication some things aren't justiciable. And besides, this threatens Article 1 Section 5 Paragraph 1 where "each House shall be the Judge of the Elections, Returns and Qualifications of its own Members" - so I'd rather just restate that clause with the desired changes. So, what do you imagine justiciability looking like here?
Some observations from a non-American:
1) I would add a topic or substantial similarity limitation in section 4, just in case
2) I disagree with footnote 3, I don't think people who disagree with the Senate should be married to the stuff beyond the text (I do however, think the clause entrenches itself, but restricting the Senate's powers is fair game)
And an unrelated comparative constitutional law/design question which I may or may not have asked already: what do you think about constitutional courts. They usually(!) a) are the primary body that interprets a constitution
b) decide on individual constitutional complaints against ordinary courts, in a collateral action similar to a habeas proceeding
c) are the sole body that can judicially review a statute, and they do it abstractly, vacating it for everyone. If an ordinary court (there usually exist a separate ordinary supreme court and other courts) thinks a statute is unconstitutional it has to refer it to the constitutional court. They also usually judge impeachments, for instance most recently in South Korea, but in lots of countries parliament and the Prime Minister are more powerful than the president.
For example, here in Croatia the constitutional court (see also powers section https://en.m.wikipedia.org/wiki/Constitutional_Court_of_Croatia ): a) decides on individual complaints, mostly in panels b) decides on the constitutionality of statutes, and the constitutionality and legality of administrative acts (administrative courts can also do the latter).
It can do b) sua sponte, on a request by some listed governmental body or on a proposal by any natural or legal person anywhere (yes, really). It also does some separation of powers and election-related stuff, see wikipedia.
However, ordinary courts can ignore legislation if it is contrary to treaties (which are considered self-executing and superior to statutes) or to EU law.
Interestingly in the Netherlands this is the only form of judicial review of statutes, the Dutch constitution bans judicial review of laws against the constitution itself, but they review them against international and EU law.
If you wanna read here are wiki links to the French, German and Italian constitutional courts: https://en.m.wikipedia.org/wiki/Constitutional_Council_(France)
https://en.m.wikipedia.org/wiki/Federal_Constitutional_Court
https://en.m.wikipedia.org/wiki/Constitutional_Court_of_Italy
My questions are: 1) what do you think of this system, both in general or as maybe applied to America in an alternate world. I imagine if the US were to implement something similar it would be tacked onto SCOTUS instead of making a new court.
2) what do you think of moving impeachment judgments in the US to SCOTUS (but in most countries the impeachment itself takes 2/3 of the legislature, but I myself think majority is fine. It also usually takes 2/3 of voting judges to convict someone.)
3) what would you think of doing APA vacatur in the US, but for statutes, would that even be constitutional in the US. (Also, do you think the current APA vacatur is constitutional, seems at least Thomas doesn't, maybe you wrote about that?)
At the constitutional convention, the framers considered a Council of Revision that could veto legislation that it judged to be unconstitutional. (This was part of the Virginia Plan.) The Council would have been composed by the president and a number of federal judges. Ultimately, this proposal was narrowly defeated, and the veto was vested in the president.
In the U.S., the dispute over the role of the courts has been renewed by the debate over universal injunctions. The two basic views are judicial restraint (i.e., the role of the courts is to resolve the cases and controversies before it) and judicial activism (i.e., the role of the courts is to police the government for unconstitutional laws and actions).
Simply put, I lean towards judicial restraint because I do not believe the courts should be "the primary body that interprets" the Constitution. All three branches have an essential obligation to interpret the Constitution and ensure it is faithfully executed. Now, that is not to deny the role of judicial review, but used as only necessary to resolve said cases and controversies.
A process like APA vacatur for U.S. statutes would destroy the separation of powers. This would allow a judge to vacate any law he deigns is "arbitrary," "capricious," "an abuse of discretion," or "unsupported by substantial evidence." This is not the role of the courts. In fact, the Supreme Court recently struck down this sort of nonsense in U.S. v. Skrmetti.
Per Justice Thomas's concurrence in that case, "Taken together, this case serves as a useful reminder that the American people and their representatives are entitled to disagree with those who hold themselves out as experts, and that courts may not 'sit as a super-legislature to weigh the wisdom of legislation.' Day-Brite Lighting, Inc. v. Missouri, 342 U. S. 421, 423 (1952)." https://www.supremecourt.gov/opinions/24pdf/23-477_2cp3.pdf
Edit: To illustrate the terrors of "courts as a super-legislature," yesterday, a federal judge in the District of Massachusetts issued an ex parte TRO requiring the executive branch to fund Planned Parenthood, despite having just been defunded by Congress. Thankfully, the TRO itself is utterly unconstitutional (see the appropriations clause and OPM v. Richmond). But imagine the courts actually had such a "vacatur" power over duly passed laws. Insane.
Thanks for your thoughts! I was thinking of an APA vacatur remedy for statutes, but the standards for facial challenges would stay largely the same (the current ones are maybe a bit too strict with finding only one constitutional application saves the statute), just there would be a universal remedy at the end. I didn't think of importing the APA standards and concepts.
Apologies, I must not understand what you mean by "an APA vacatur remedy for statutes." Although I think your second comment might help clarify.
Do you mean nationwide relief? I.e., a judge striking down a regulation under the APA provides relief to non-parties to the case, whereas (post Trump v. CASA) a lower court striking down an unconstitutional statute only provides relief to parties.
In case it is helpful, I'll just note:
(1) The Supreme Court can already provide nationwide relief through its opinions and judgments. (See Justice Barrett's majority opinion in Trump v. CASA.)
(2) Congress might allow for nationwide relief via nationwide injunctions. (J. Thomas is skeptical this is constitutional. See his Trump v. CASA concurrence.)
(3) From 1937 to 1976, there were three-judge panels in the federal courts with jurisdiction to hear constitutionality challenges. Their decisions were directly appealable to the Supreme Court. https://sgp.fas.org/crs/misc/IF12746.pdf
1) I disagree, precedent isn't relief, there is no 1983 federal cause of action and Bivens is basically dead, so no damages, only prospective relief is possible against feds.
2) Yeah, I know. In my mind the vacatur works different mechanically than the injuction, it removes the act itself from the legal order, while injuctions operate on the parties being sued. That solution seems more elegant to me at least, tho I don't know if contempt is available. I suppose if the gov tries to use a vacated rule, they should be sanctioned for using non-existent authority. AFAIK there is also an APA provision for further relief in aid of an APA remedy, I wonder whether that allows nationwide injuctions post-CASA.
3) Interesting.
On CASA, long-term I don't see it doing anything, people will just say the magic words class-action. Especially after SCOTUS said in AARP that you don't actually need to do class certification analysis in a preliminary posture.
Re footnote 17: The best example of why not to use single-member IRV is perhaps ironically the one place (at least among more-or-less stable democracies) where it has lasted for over a century: Australia.
The system was introduced after a rural-focused centre-right party (the Country Party) emerged and split the centre-right vote in a safe seat, leading to the election of the Labor candidate. To avoid this happening again, the electoral system was changed from first-past-the-post to single-member instant-runoff.
The result of this has been, effectively, a two-party system masquerading as a multi-party system (though recent results have seen this break somewhat), where Labor is the main centre-left party and the centre-right is represented by a party focused on urban interests (called, at various times, Nationalist, United Australia, or currently Liberal) and a party focused on rural interests (originally Country when first formally organised, now National), which together are the Coalition. (Even with this, it has become increasingly rare for the Liberal and National parties to run candidates against each other.)
(This is actually true only in the two largest states, New South Wales and Victoria. The two centre-right parties merged in Queensland after disagreements between them led to the election of numerous Labor MPs in 2007; the additional candidates elected by the merged Liberal National Party account numerically for most of the shift away from Labor and toward the Coalition in 2010. In Western Australia and South Australia, the National Party exists but is not in formal coalition with the Liberal Party, and the one time a Western Australia National candidate was elected, in 2010, he strenuously objected to being counted as part of the Coalition's seat total by the media. The National Party does not exist in Tasmania or the Australian Capital Territory, while in the Northern Territory the centre-right party is the Country Liberal Party. Also, due to policy differences, the National Party split from the Liberal Party following the 2025 election and it is unclear whether a formal coalition will be in place in the next election.)
The result of this has been that even with a recent trend toward electing more candidates from outside the major factions, five or six in elections from 2010 to 2019, then sixteen in 2022 and thirteen in 2025, it is still the case that since the Coalition became more or less formal in 1934 there have been only two elections that have not seen one faction or the other win an outright majority, in 1940 and 2010. (Australia holds elections for its lower house roughly once every three years, sometimes more often, so this represents well over thirty elections in about ninety years! Which is less frequent than elections to most lower houses in the United States, to be sure, but more frequent than in most other parliamentary systems; even New Zealand's legislatures tend to actually last their full three years, and they've had thirty elections in that same time.)
As I've been given to understand it, an important factor in Australian elections is the two major parties trying to convince the plethora of minor parties to recommend that their supporters rank the one major-party candidate ahead of the other(s). This of course means making alliances with and concessions and promises to other factions before the election takes place, and Linz had something to say about that in "The Perils of Presidentialism".
Where Australia's multi-party system does shine through is its own Senate, where because each state has twelve Senators (each territory has two) they've used a proportional system (sort of) for elections to that body since 1948, and as a result neither major faction has been able to win a majority in it reliably since the 1960s or so (the exceptions being the Coalition in the late 1970s and in 2004).
I suspect that part of the reason instant-runoff has become popular is because no other system has ever really seen real-world use, at least in the political sphere. (As I've mentioned before, Borda, or at least partial versions thereof, gets used to select MVPs and other such awards in many sports leagues.) Australia uses it in a single-member context. Ireland uses it in a multi-member context (and tends to elect hung legislatures). Malta also uses it in a multi-member context but Malta is tiny, effectively a city-state as Valletta has expanded, and the two major parties (Labour and Nationalist) have very strong brands and there are almost always more candidates from each party in each district than there are seats in each district (each elects five). (For instance, in 2022, District 1 had eleven Labour candidates and seven Nationalist candidates, electing three and two respectively. Every district had at least five candidates from each major party.)
Nitpick: When listing off moderates from 2010, did you mean Blanche Lincoln (from Arkansas) or Ben Nelson (from Nebraska)?
Next: If you want to make your specification of DW-NOMINATE more proof against chicanery, you can say something like "as defined on such-and-such date". (More on this later in this comment.)
On banning bullet ballots (and requiring full ranking), to avoid the centre squeeze issue, why not incorporate Coombs' method? Maybe it has some failings that lead it to be less desirable than even Hare, but on paper it knocks out extremists who make a substantial chunk of the electorate vomit even as another chunk loves them.
Final suggestion: I said this before in the context of electing Popes, but just elect Senators in the way Venice elected its doge. Yes, many state legislatures are too small to use the full procedure, so instead specify the election of a representative assembly in a manner consistent with the method of election of the most populous branch of the state legislature and no smaller than it, and no smaller than, say, two hundred members. (For instance, using multi-member elections from state House districts that have to use a proportional method of some description, be that Australia's Senate method, or maybe Austria's method.) Then say that that body will elect the Senator out of all eligible citizens of the state. The operative clause here is that the method of election would be "the method of election of the Doge of Venice as of the year twelve hundred and sixty-nine" and if it looks weird to specify an exact time or year like that (I said I'd come back to this!) then I guess Section 18 of the Constitution Act, 1867 and especially Section 50(2) of the Constitution Act, 1996 (BC) must look weird! (But it makes sense there because while parliamentary privilege is really important, it's also extremely important that legislators not be allowed to just arbitrarily vote more privilege for themselves!)
Of course, this does have the problem that now Senators are being elected by random folks who might've run for the temporary assembly on a whim and gotten in effectively by luck, instead of by serving politicians who, whatever their faults, are at least generally well-versed on the issues of the day and familiar with the prospective candidates. (Unless every state can be convinced to have a lower house the size of New Hampshire's, anyway. Using state legislatures for this procedure would work there, I guess, but probably not in a state like Nebraska, Alaska or Hawaii. I guess you could specify that all serving state legislators would be part of the assembly and the random folks elected to top it up to the necessary minimum, but that also presents other problems.)
(This would still need a lot of work to make sure that the elections for those assemblies can't be too harshly gamed and whatnot but forcing multiple rounds of supermajority consensus where nobody even knows who will be voting in the next round should tend to knock out extremists, unless you get very unlucky.)
Combining three unrelated comments into one so as to be less spammy:
Section 1 of the comment: Enumerating the low-hanging bombs:
There is no topical limitation on section 3, so section 4 is effectively an unlimited power to amend the constitution.
DW-NOMINATE could be gamed by holding a lot of roll-calls on do-nothing motions. I think probably a majority party could even arrange the order of the role-call and thereby vote while already knowing all minority votes. If that flies and they do it often enough they can make the rankings of their own members whatever they want. Even if you ban that it is still easy for both parties to make a pact to statistically demoderate their respective moderates. The reason it hasn't been Goodharted yet is because it hasn't been made into a goal yet.
The amendment has a hard-coded assumption that recorded votes by name and roll-call votes are the same thing. I don't know if this is currently true as a matter of American ritual but it is certainly gameable. Just hold votes likely to produce the desired statistics by roll-call and others by electronic device, signed ballot, division, or whatever.
In the edge-case of perfect party-discipline DW-NOMINATE will collapse the spectrum to two points and the ranking will become undefined. With not-quite-but-near perfect party discipline it will be much easier to game or near random.
I don't know how easy it is for isolated members to force votes in the various legislatures, but if it isn't hard it probably can be made so. If roll-calls just happen very rarely, new members can be kept disqualified for a long time.
I don't think the states are required to make consent of all branches of the legislature necessary for legislation. Certainly various kinds of overridable upper-house vetos are common internationally. So a state constitution could just make the governor into a single-member and therefore least numerous branch of the legislature with the same veto power he now has outside of it. Or even have a "house of the senate majority leader".
If the presiding officer of the state senate doesn't like the expected section 2 winner, he can force section 3 by not waiting for nominations and just recessing the section 2 session immediately after opening it.
Section 2 of the comment: A comparative legal observation.
As far as I understand it, unconstitutional and federally preempted laws in the US are unenforceable but stay on the books (and maybe become enforceable later if the higher law changes). Section 5 would create a unique exception by potentially making a provision of a state constitution not just unenforceable but actually void. In German legal terms, you have reinvented Geltungsvorrang (validity precedence) while previously American law only new Anwendungsvorrang (application precedence).
Section 3 of the comment: How to elect the least hated member
Note that the least controversial hated will usually not be a Condorcet winner, because a Condorcet winner still can have a fairly large minority who hates him. But electing the least hated member is actually fairly easy: Use your favorite method of proportional representation (which should be STV but here it doesn't matter much) to elect an exclusion panel that has one less member than your state senate. The person not elected to the exclusion panel is your winner. In a two-party system, effectively the majority gets to exclude all of the minority, and the minority gets to exclude a lot of majority members in order of hate. If the majority is comfortably large, they also get to exclude some of their own.
You left out a key benefit of your proposal: it would create new shovel-ready jobs for Fortran programmers! (DW-NOMINATE is written in Fortran.) OK, fine, I’ll take my messaging hat off.
As you acknowledge, this proposal is extremely complicated. Unfortunately, the proposal is so complicated that I worry it would diminish the Senate’s legitimacy. The People may not directly elect their Senators, but they must at least understand the process by which they are elected!
Most people are not going to understand DW-NOMINATE, Condorcet winners, Bonham’s method, etc. I do not mean this judgmentally. I am an economist and even I quickly bounced off the Wikipedia article on DW-NOMINATE. It looks like they are assuming Gaussian utility functions and estimating the bliss point of each legislator based on a PCA of roll call votes?
The epicycles of the proposal stem from contingent elections as a method to break deadlocks. Why care about deadlocks? Previously, you mention that some Senate seats would remain open for months (sometimes years) because state senators could not agree on a candidate.
So what? If the Senate has a quorum to conduct its business, then let a state deprive itself of representation until it get its act together. Ultimately, if the state senators are unwilling to compromise, then they face accountability from voters whom (it seems) do not like deadlocks. Admittedly, this accountability would be undercut by the proposal’s use of the secret ballot.
Maybe I am just revealing my historical illiteracy, but why were deadlocks such a problem? In your prior piece, you mentioned occasional riots. That is a law enforcement problem not a constitutional problem. You also mentioned (per the deadlocks bill) that state senates were required to spend time on senate elections before moving to other business. OK, abolish that.
Let me assume, arguendo, that deadlocks are a problem that must be fixed. You discuss the merits of a conclave-like system, but you distrust individual state governments to enforce secrecy. Yet, Congress can regulate the election of senators (Art. I, Sec. 4). Congress could set forth regulations requiring a conclave-like system. Congress could also impose stiff penalties for states who fail to abide by the regulations, such as rejecting their senator-elects.
While Congress cannot regulate the place of choosing senators, I can imagine such regulation including safe harbor provision for states who conduct their conclaves in DC with federal oversight and assistance. E.g., Capitol Hill police stationed outside the conclave to prevent unauthorized access. The feds do a good job of keeping things secret when they want to!
Separately, I am brainstorming ways to game the proposed amendment. One idea: a highly-partisan state could pass a constitutional amendment creating a third legislative chamber with two members, each elected every six years by popular vote in a statewide election. This third chamber serves no legislative function. It is merely a proxy election for the state’s U.S. Senator.
Take Illinois as an example. A state constitutional amendment in IL requires a three-fifths vote of each house of the General Assembly, ratified by a majority of voters in a general election. Democrats could pass such an amendment through the General Assembly on a party-line vote. State-wide ratification would be harder but doable. (IL voted 54% Harris, 43% Trump in 2024.)
Such a system would operate similarly to the electoral college. Voters select the ticket for their preferred Senate candidate. In actuality, voters are electing candidates to this third legislative chamber whom (if elected) will choose their corresponding candidate for the open Senate seat. While Section 5 of your proposal prevents explicit “pledging,” parties can just choose loyalists.
(Nitpick: In Section 1, I would write “least numerous chamber” not “least numerous branch.”)
Thank you for the shout out in footnote 8!
James, I will say you did not disappoint.
Your effort to come up with a Senate election proposal that was resistant to gaming by the political parties, was ingenious. I could not help thinking of a proposed state constitutional amendment proposal put in the Ohio ballot last year led by a citizen group, to eliminate gerrymandering in drawing state and Congressional district boundaries. it was fiendishly complex, involving a redistricting commission of non-politicians, nominated by a panel that itself was selected by retired state judges, all with a budget set in stone, with detailed line drawing criteria, hearing requirements, etc, etc.. The whole thing was a substantial fraction of the length of the entire Ohio constitution, which is not short! (Sadly, establishment pols ran ads that lied and deceived, and the measure went down to defeat).
All that to say that game-proofing a process against political capture is hard, and you deserve kudos for even making the attempt.
All that being said, I keep wanting there to be a simpler alternative to avoid the deadlock election than DW-NOMINATE. I'm half willing to tolerate deadlocks, forcing state pols to go unrepresented in the Senate if they can't come to a consensus.
The more I think about this, the more it makes sense to just let the single Senate seat from a given state go vacant, if the state senators deadlock. There will be enough states that are able to elect Senators (due to one-party domination if nothing else) that the "purple" states will have a powerful incentive not to be left out, and come to a consensus to elect someone.
I have a minor quibble to make about the phrasing of Section 3:
"Each member qualified for the U.S. Senate who has served in the body for at least two full years (even if non-contiguous) and who has cast at least one hundred roll-call votes shall be eligible for nomination." -- I would argue that the plainest reading of this would indicate that we are looking for people who were already in the US Senate, unless I am failing legalese...
Perhaps:
"Each member qualified for the U.S. Senate who has served in the State's body for at least two full years (even if non-contiguous) and who has cast at least one hundred roll-call votes shall be eligible for nomination."
Also, after writing that, I am curious if you are restricting this to current members of the State's body?
I think Section 3 is too complicated, even without the problems of defining DW-NOMINATE, etc. Instead, i think something simpler and somewhat random would serve. So:
If no nominee is elected within two calendar days, all nominees who received a majority of approval votes shall be listed in order of age. If the list is at least three nominees, the nominee with the median age is selected. If the are are an even number of nominees, the older (or younger?) of the two median nominees shall be selected. If fewer than three nominees received a majority of approval votes, the list shall be all members of the voting body who are eligible to serve in the Senate.
Pretty sure I can break this system.
Problem 1: What if none of the 'moderate' 'state senators' actually WANT to be federal senators? Moving their family to DC sounds like a lot of work. Maybe they would prefer to just stay home. Or maybe they were bribed to say that.
Problem 2: DW-nominate doesn't care about what votes are ABOUT, and all votes are equal. A vote to adjourn the house for lunch counts just as much as a vote to implement single-party healthcare.
So, if you are a state senator who WANTS to someday be a federal senator, the winning tactic may be to always vote with the opposition on votes where the opposition is doomed, or where the vote doesn't matter. And it's easy to game this. The party in control of the house decrees that there will be three votes a day, on whether to break for breakfast, lunch, and dinner. 40% of the controlling party is ordered to either abstain or vote against the break. Bam, the 'designated moderates', who aren't really moderates, just increased their moderation score by voting with the opposition three times a day every session.
If the opposition catches on, and starts to vote against breaks, then you simply order the sergeant-at-arms to chain them to their desks and not give them any food until they vote in favor of breaking for food, so the 'moderate' can vote with them. Or, you switch tactics, and 51% of all house members, the ones you control, vote FOR breaking for food, and the 'designated moderate' votes AGAINST it, same as the opposition.
There are also many shenanigans available by, say, controlling who gets to close the vote, when. Maybe minority-party legislators have to cast their vote and stick with it, but majority-party legislators can switch their votes en masse at the last second, and then the vote-counter closes the vote right afterwards.
*Kyrsten Sinema
With that out of the way, now it's time to break the amendment every in way I can think of.
Section One:
So, the relevant body is whatever branch of the legislature has the least members. We can easily convert this back into a statewide senatorial scheme by erecting a new chamber with a single representative. This would allow, effectively, for popular election (but done 2+ years in advance), or, I suppose one could have rig a scheme where, while nominally popular, in practice this chamber will usually filled by appointments from the governor.
While we're here, I'll note that I did check whether any states had equally sized senate/house, as that would cause problems. There are not, currently.
Section Three:
DW-Nominate can be rigged. At the end of a session, just spend many votes on utterly trivial matters. Have the one faction vote but in lockstep, but without informing the other side how they will vote. Have the desired nominee vote to the contrary. If any partisans of one's own side are defecting, one can probe by telling them the opposite, and cut them out—one only needs the buy-in of a majority to keep forcing votes. The body of votes added by this process will, on average, make the desired candidate seem more opposed to the majority party than *any* member of the opposition. Since DW-Nominate is deterministic (I assume?), repeat until they will be projected to be the winner. I'm guessing that this could be iterated enough to make the average (arithmetic mean, I assume) extreme enough that they should be viable, even if they were extreme in all past legislative sessions.
I don't actually know the details of the algorithm, so I don't know if that works exactly, but I assume gamesmanship like that could be done.
As to a voting system, definitionally, a majority will always be able to push through a Condorcet winner. If you want to force moderation, you do not want an electoral system that always elects Condorcet winners. As is, this drops the threshold for guaranteed partisan victory from 2/3 to a 50.01% bloc.
Section Five:
After the fact punishments seem entirely permissible by parties here. What is to stop them from allocating funding for reelection as they wish? Of course, the ballots being secret makes this harder to do.
Section Seven:
Expect partisan courts. Was that ballot *really* secret?
The even sections seem difficult to game.
I think a common theme of some of the other comments made is this: this amendment assumes that political actors will continue to abide by the current conventions, but the optimal partisan political outcomes are achieved by changing the conventions, and thus the conventions will change. The only way to avoid this is for voters to be willing, en masse, to punish politicians who try to game the system, which is dubious on multiple grounds.
This is, in effect, the same problem faced by the people who want to use the current electoral system for President to bootstrap a nationwide popular vote via an interstate compact: political actors who dislike that can simply start playing by different rules. (For instance, state legislatures might remember that they can just directly appoint electors for President, and not bother to have any sort of statewide vote on the matter.)
EDIT: To elaborate somewhat, the point of this amendment seems to be that the Senate would become fifty clones of Sen. Moderate Milquetoast giving sober second thought (a much better analogy, in my opinion, and one actually supported by historical evidence as having been made) to the bills passed by the 218 clones of Faction A's Rep. Ragey McRageFace over the strenuous objections of the 217 clones of Faction B's Rep. Ragey McRageFace in the House.
The problem is this: what happens when Faction A games the system to return 25 clones of Sen. Ragey McRageFace but Faction B respects the spirit of the system to return 25 clones of Sen. Moderate Milquetoast?
To make this system stick in spirit, you need a certain critical mass of voters who would be willing to overlook any and all ideological differences with Faction B (you may ascribe to them whatever policies you personally find abhorrent for the purposes of this thought experiment and assume that even their Moderate Milquetoast clones more or less support them) to make sure that Faction A, as well as any future Factions C, D or E that swear up and down that they are not just Faction A with a different haircut, is driven out of politics forever.
Could you realistically expect to find such a critical mass, especially if Faction A makes it known that, yes, they are going to be sending their Ragey McRageFace clones who are popular with the base while Faction B makes it known that they are going to be sending their Moderate Milquetoast clones whom the base hates? State Senates (even if you allow them to be malapportioned again) are still full of plenty of clones of Ragey McRageFace because they are popularly elected, and as noted in other comments you can just hold meaningless votes to make one particular State Sen. McRageFace look like a moderate. (And that's to say nothing of the notion that some have brought up of just creating a third house that has only one or two members that only serves as a proxy for the Senate election; that might perhaps be mitigated by placing a minimum size of 15 on the legislative body that would select federal Senators, since Alaska's is smallest at 20.) If you can't find a critical mass willing to put democratic norms and constitutional conventions (or even constitutional text; as you've argued, everyone who voted for the current occupant of the Oval Office in November 2024 was violating the Constitution by casting a vote for a candidate who did not enjoy passive suffrage) above policy preferences, democratic systems collapse.
(You could, perhaps, implement a system like that found in the Basic Law of Germany whereby parties that sufficiently threaten the democratic underpinnings of the system are banished from politics, with strict monitoring to ensure that they don't crop up again having done little more than changing the colour of their shirts, which has seen the Sozialistische Reichspartei Deutschlands banned in 1952 and the Kommunistische Partei Deutschlands banned in 1956, and may yet see the Alternative für Deutschland banned following a recent report that they may meet the criteria for such. But then you are placing a lot of faith in the continued independence of that adjudicative process.)
Why not just make it illegal for party organizations to endorse Senatorial candidates? You could even have a broad, vague legal understanding of what exactly is a 'party organization' established through caselaw, to disincentivize attempting to work around the prohibition.
Something like, "Any private organization that runs, nominates, or endorses candidate for election to public office may not endorse, nominate or otherwise select a candidate for any election to the United States Senate".
Penalties that are onerous enough would get parties and affiliated partisan organizations walking on eggshells around the possibility of even being perceived as having endorsed. While more distantly related organizations may feel safe to do so, removing the central coordination point will make it difficult for single-issue advocacy groups, to unite around individual candidates capable of winning on their own, decreasing the salience of the Federal Senate elections to the choice of state senator (because no one party can credibly promise to elect a specific Federal Senator if they win majority control of the state Senate; even if private organizations have pledged state senators to vote for a particular Federal candidate, they're not necessarily THE SAME candidate across state senators).
By making it difficult for external organizations to express a universal preference, you make it easier for state senators to make their own decisions without having to fear directly defying powerful outside forces.
This isn't that major a new step in American law, either: we already do this in the tax code, where certain non-profits are banned from endorsing candidates. That would give you an existing body of law to build on.
This might be a crude solution, but what if we simply make it so that, after a certain number of rounds of voting, if the legislature can't agree on a candidate, they must all resign, and are ineligible from ever holding any political office ever again? That would almost certainly outweigh any pressure from party bosses.
Probably an unjustified intrusion on internal state affairs by the federal government.
You know, one of the easier filters would just be to ban people who have experience with first-past-the-post elections from being senators.
Try this:
Rule 1:
No person who has previously been a member of any house in the state legislature, or the Governor or lt. Governor in the State executive branch, or been a member of the US Federal House of Representatives, or held any other office which is normally determined by means of a statewide election, or who has been President or Vice President of the United States, may ever be a US federal senator under any circumstances, ever.
Each state may determine, using it's own laws, whether or not Town mayors, town councilmen, country sheriffs and other county-wide elected offices may be permitted or denied from holding the office of US Federal Senator, according to the preferences of each state.
And while we're at at it:
Rule 2: No person who has served in the Federal Judiciary may ever again run for, or hold office in, or otherwise be employed by, any part of the federal executive or legislative branches. But prior service in those branches is not disqualifying from being a member of the federal judiciary.
No person who has served in, run for office in, or been employed by the Legislative branch may ever again serve in the executive branch in any capacity, but prior service in the executive branch is not disqualifying from being a legislator.
Each state may choose whether or not to extend these rules so that prior service in that state's executive branch may or may not be considered equivalent to prior service in the federal executive branch for purposes of this rule, and shall also determine whether or not the state legislative branch is likewise equivalent to the federal legislative branch, and shall also determine whether or not the state judiciary branch likewise be considered equivalent to the federal judicial branch.
(E.G., Federal Justices can NEVER run for POTUS. States have the right to determine whether or not justices from that state may ever run for POTUS. Note that this is based on whether or not you ever served as a justice in that state, NOT based on which state you're living in when you run for office. If you were once a state district judge in Florida, and Florida says their judges count as federal judges for purposes of this rule, that means you can't ever run for, say, House Representative for the First District of Georgia, because your prior service in the Florida Judiciary blocks you from ALL future federal executive and legislative employment, not just the roles specific to Florida. )
Why not just start using Single Transferable Vote with a 2/3rds threshold, instead of going straight to complex algorithms that most people can't check on their own? I feel like once you get to the point where a computer scientist is involved with selecting candidates there's a sudden lack of transparency in the process.
God help us if there are incorrect entries in the databases used to calculate the candidates. And it wouldn't just be a "Well, the state made an error, but the state chose their senator so the senator has to be legitimate." It's in the constitution to use the algorithm correctly.
What do you mean by "Single Transferable Vote with a 2/3rds threshold"?
If by threshold you mean the number usually called the quota, and given the degenerate case of electing a single winner (that degenerate case also often being called AV or IRV), that would leave you with three sub-possibilities:
1) you insist on every voter numbering the whole ballot. Then you will ritually get a 2/3 majority by counting people as having "voted" for their last "choice", but the winner will always be the exact same person who would win with unmodified STV/IRV. If the ritual 2/3 is what you want, you might as well make it a ritual unanimity, which is also the same system.
2) you could do the same thing usually done with the usual quota and say the quota on every count is 2/3 (rather than the usual floor(1/2)+1) of the currently unexhausted ballots. This will also always elect the same candidate as the standard algorithm, in this case by eliminating more candidates to create more exhausted ballots. You again get something you can call a 2/3 majority in a very technical sense but it will be 2/3 of the wrong base group.
3) Or you really insist on a 2/3 of votes cast while simultaneously allowing truncated ballots. Then the system will often fail to elect anybody and you are in the same place as if you required the 2/3 without an STV wrapper.
Hey, @everyone! Lots of great, in-depth comments on this one, which is wonderful. (Please keep them coming for as long as the Spirit moves you!)
As I sat down to start replying to all of them, I realized that doing so would take many hours, and that, since several of you raised similar points, it would be duplicative as well. So I'm now planning to do a "highlights from the comments" follow-up to this post in a few weeks (after I take my time off and then post the next part of If They'd Made Me Pope).
Until then, I'll be reading attentively!