Thinking about the US method of presidential selection

Thinking about the US method of presidential selection is something I do a lot, and have written about before (both at F&V and in academic works). This planting won’t have any new ideas on the topic. However, I want to call readers’ attention to a “symposium” at Balkinization on the topic, which began on 13 October. The first entry there makes some good criticisms against the current method that are less commonly articulated–for instance, that the electoral college is vulnerable to “stalking horse candidates” and to the whims of billionaires with egos as big as their asset portfolios.

The symposium is motivated by a couple of new books on the topic (see at the top of their post), and has had further installments posted in subsequent days.

Thanks to Alan for the tip.


Note: I found the various parts of the symposium hard to locate on the Balkinization blog when I went back in early February to re-read them. So, I am providing the links here. The one above takes you to Part 1. Then click these for Part 2, and Part 3. (The last one there is an idea from Edward Foley for state-by-state ranked-choice voting for presidential electors.)

34 thoughts on “Thinking about the US method of presidential selection

  1. From the third post. “Putting all that aside, however, my prediction is that the compact soon would prove unstable because of the “spoiler” effect associated with plurality-based elections. As soon as a third-party or independent candidate seriously threatened to deny a major-party candidate a victory achievable through an Electoral College majority fashioned from state-based wins, states in the compact would pull out in order to avoid this “spoiler” effect”

    This does not strike me as a particularly compelling reason for not persisting with the NPVIC. If states get into the habit of pulling out of the NPVIC for calculated partisan advantage, then it seems highly implausible that the Compact will get off the ground in the first place.

    • I’ve read Foley’s book and found most of his arguments persuasive, but this one was a shocker. I think there is a major problem with NPV because it sets up a national plurality system, but if states are going to bolt the NPV at election time there are multiple reasons they might do so.

      I do think his argument that states should award electors only to a candidate who wins more than half the votes actually accords with what the Framers of the XII Amendment thought they were achieving. At the time all states provided for an electoral or legislative runoff for electors and the framers do not seem to have envisaged the shift to plurality voting without runoffs. Andrew Jackson believed the majority rule had denied him the presidency in 1824 and within fairly short time all sates had moved to a plurality rule.

      The NPV is not as certain as either a constitutional amendment (which I admit is like climbing Mt Everest in the US) or state legislation to adopt a majority rule. Or maybe Mt Everest will be a lot more climbable in 20 days time.

      • It’s possible that the constitutional amendment requirement would cease to be quite so onerous if either the NPVIC becomes very well-established or the Electoral Collage puts a Democrat into office when a Republican wins the popular vote: after all, the gap between Republicans and Democrats on EC support really only ballooned after 2016.

        The more I consider Foley’s quote, the more confusing it seems. It would require a state dominated by a certain party not only to recognise that their candidate was being hurt by a spoiler (a difficult question to pick, the “was Nader a spoiler?” debate persists to this day) but to recognise that their candidate would be more hurt by that spoiler under the EC system than under the NPV, which seems very, very difficult to guess without actually having access to the results.

      • I’d certainly push ahead with NPV as well as the Foley proposal. If NPV actually goes into force there could then be bargaining for a constitutional amendment that provided for a national majority by instant or delayed runoff. In a similar way of if the Republicans were to lose their electoral college advantage they might be ready to talk seriously about constitutional amendment.

      • I rather like Foley’s proposal. I mean, for reasons of practicality and that even if it were the end of the process, it would be an improvement although an end point of actual direct election by two-round system* would be better.

        (* I am an RCV [AV] skeptic for direct national presidential elections, especially in the absence of a national electoral commission to actually administer them.)

      • This page at FairVote lists 12 states that use runoffs and two that had “recently” abandoned it, but the page (1) is undated, (2) is titled “archive”, and (3) includes 2006 and 2008 as “recent”, so.
        Of the 14 mentioned, one uses (and one used) a 40% threshold, one state said 35%, and Washington said the top two automatically went into the runoff. (MSS, is this California too now?)
        (Wikipedia isn’t much help).
        Am fairly certain I remember Lani Guinier mentioning one or another State – or possibly county or municipality – requiring 45% to avoid a runoff, but don’t have her book in front of me. Emmanuel Celler’s 1969 proposed direct-election amendment set the bar at 40%, supposedly so that Lincoln would still have qualified as a plurality winner on the first round.

      • So the only examples of a sub-50% threshold are Kentucky, South Dakota, and South Carolina, the first one having abandoned, and the latter two potentially only using it for primaries.

    • The basic premise of this concern also has another issue. Is there any evidence that state governments would withdraw over the possibility of a three way race? In other words, would they feel compelled to protect…whatever it is they are protecting…because the winner may win 40-30-30? I don’t think there is a strong demand for the president to always have 50%+ of the popular vote. NPVIC is simply designed to ensure that whoever is first past the post wins.

      See for example the various proposals to implement a direct election via Amendment. How many of them provide for a runoff?

      • The last constitutional amendment proposal that won two thirds of the House (but failed in the Senate, where Biden was one of the no votes, incidentally) provided for a second round if the plurality was under 40%. The Costa Rican model of qualified plurality (also used in some US states).

      • Which US states use a 40% runoff? 0.o I do think I read one of the Dakotas in theory has a 33% runoff, which never comes into play in practice.
        40% runoff is certainly the model used in Israeli mayoral elections.

      • Miss

        I have counted more AV elections than I care to think about. There would be absolutely no difficulty in a federal EMB collating the results of an AV election conducted by state EMBs.

      • What Alan said, although I can count my own AV counts. But yeah, the National Elections Board would just tell the 50 State Board, “Okay, Perot/ Stockdale is third nationwide, so please now exclude that ticket and re-allocate each of their ballots to its higher preference (if any) as between Bush/ Quayle and Clinton/ Gore”, and this would be followed even in Maine and Utah, despite Perot/ Stockdale placing second in both those States.
        There is not reason all the ballots have to be in one place to count AV-STV, although with multiple seats and STV-PR it would make things more convenient, all else being equal, depending how long it would take to transport the ballot boxes there and whether one could fit 100,000,000 ballot-papers in one room. (Also, the States and municipalities would need to keep “long ballots” at home so they could also count Governor, Senator, Mayor, Commissioner of Dog-Catching, etc, further down the page.

    • Oh goodness, part 3 really goes off the rails, doesn’ it?

      How would ranked-choice voting work, when electors are voting for their states’ delegates to the Electoral College, rather than for a President directly (although the names of the two-person ticket appears on the current form of the ballot as an illusion)? Would Californians be faced with the names of 55 Republicans, 55 Democrats and various third-party and independent potential delegates on a tablecloth ballot paper? At that point, it would become very tempting for your average voter to ‘just vote 1’ and render the whole ranking method pointless.

      My father says he’s already mailed in his vote in Maine, but couldn’t tell me which offices were subject to preferential voting and which weren’t (he recalls punching a hole in the ballot paper more readily than writing down any numbers).

      The best solution is and always will be a form of PR for the allocation of ECVs, by whichever divisor method is acceptable to each state. If only that proposition in Colorado had passed in 2004…

      • Maine doesn’t elect individual electors by STV: rather, as this sample ballot shows, voters vote for a candidate and the candidate who wins a majority gets all the electors from the congressional district/state.

        Leaving aside the question of whether getting rid of the EC is a good idea or not, perhaps the fundamental problem with allocation of electors by PR is that it is almost never going to be in the political interest of any one state to adopt it. For safe states, like California or Louisiana, bringing in PR for electors basically means a free gift of 40%-ish of electors to the opposition party in that state. For swing states, on the other hand, PR for electors would dramatically reduce the attention candidates pay to a state, since the prize for a slim popular vote lead would go from a massive to a marginal one.

      • It would be the same as now: It looks like you are voting for president directly, but it’s for a slate of electors, only now you add a second (etc.) preference space on the ballot.

  2. Should Texas actually turn blue this year, the electoral college would no longer give Republicans a viable path to the White House. I’d think the road to Damascus would see a remarkable number of lightning strikes.

  3. Oh dear, the old ‘eliminate the Electoral College’ solution. This could not realistically be achieved without opening up the Constitution. Talk about using a sledgehammer to crack a nut!

    If that came to pass, the U.S. would join Brazil, Mexico and Nigeria in being countries with directly elected executive Presidents by nationwide popular vote that nonetheless claim to be federations. At least France is honest about being a unitary republic.

    The real answer lies in allocating delegates by a proportional system in each state – something which already happens in many primaries. If this means neither candidate gets the magic 270 ECVs due to a few Greens elected in California and a couple of Libertarians in Texas and the House has to decide the Presidency, then so be it.

    If a cross-party “compact” is needed on anything, it’s uniform rules across all states on ballot access, polling place location, postal vote regulations, and method of enfranchisement and registration. Due to the precise way decentralisation works in the USA these are, at present, inconsistent even for federal elections, which is unusual as it is not the case in federations such as Australia, Canada and Germany.

    • Why is an indirectly elected executive a feature of federations? I see no such need for that, and I don’t even prefer a directly elected president. I’d prefer one electable, and recallable, by the House.

      What advantage does a proportional Electoral College provide that a direct election would not? Ensuring an opportunity for an also ran to win? Allowing an wholly undemocratic process in the House to pick a candidate when the people can’t agree?

      • Actually, that’s a fair point about the House. I had forgotten that in such a deadlock scenario, each state’s delegation is worth 1 vote. That is not a satisfactory outcome but again, it would require a constitutional amendment to alter it. It also raises the question of what would happen if there were a (proper) Congressman or Congresswoman for DC, which only needs ordinary legislation to be brought about. They wouldn’t constitute a state delegation, so that’s one less vote for the Dems.

        Federal parliamentary democracies have representative democracy by default. In some (Canada) a so-called “wrong-winner result” is more likely than in others (Germany). Federal republics with ceremonial presidencies have no need to elect the President directly, but there’s also no great harm if they choose to do so (like Austria, for example).

        A federation with an executive presidency is a different beast again. Only four sprang to mind and the USA turned out to be the outlier. Whilst I would normally caution against American exceptionalism, in this instance I think America is protecting the decentralisation of power far better than the other three.

        I guess I’m just accustomed to representative rather that direct systems, and greatly prefer the former. The Electoral College acts as a buffer preventing the crass nature of direct election, which makes watching election night coverage similar to following results under more familiar parliamentary systems. I just wish it were easier to do away with ‘winner take all’ – and in a better way than Maine and Nebraska’s pathetic little compromise.

      • I am not sure we should be basing systems on watching results from different areas of the country come in.

        More directly, I do not see why we need “representative elections” when voting for a single officer. It is naive to assume that the President of the United States is a creature of the states. He is a single person serving a fixed term. I cannot imagine anyone in 2020 designing a head of government who is neither directly elected nor selected from and removable by a parliament. There is nothing crass about asking the electorate who they want to fill a specific office. There is no need to filter their results through the formality of a body of mechanical idiots checking off a box as expected.

        As to DC, if it was given a seat in Congress as per the plan from a few years ago, it wouldn’t get a say in a contingent election. The current plan, however, is to admit it as a state. A rump area around the Capitol and White House would remain a federal district while the rest of the area is admitted to the Union with all the bells and whistles, including a vote in any contingent election

      • I personally struggle to see exactly what is so meaningfully ‘representative’ about a body which voters never see the members (who don’t campaign) of on their ballot papers, that meets for a matter of hours over the four-year term of a President, that never meets in a plenary session, and that we now know can be directed to vote in a certain way under threat of legal punishment. Perhaps to someone observing from a rather odd angle it may look like a parliamentary system, but it’s hard to find any meaningful difference.

      • Oliver R,
        Electing a president is still representative democracy. The electorate doesn’t run the executive branch directly; it elects a representative to do that.

        You may have an argument for Mexico, but Brazil and Nigeria are very decentralised. Reading David Samuels’ book on Brazil demonstrates how much influence governors have over the federal government, and how federal deputies actually advance their careers in the states, rather than continued service in the federal govt. If anything, maybe it’s arguably less than federal because it is in many ways more a de facto confederation – but then you have to ignore the president, who the one clearly federal actor. (Of course, if developments over the past 15 years have disturbed this equilibrium, I’d love to hear more about it.) Now Nigeria is also clearly decentralised based on policy. Its northern states have sharia courts! That’s quite a lot of autonomy. And from the little I’ve followed of Nigerian news, news about governors and other state-level politics gets a lot of attention, which I would also think of as indicative.

        Now of course, I should have started off by offering a definition. To me, a polity is federal if it has 1)two or more levels of government, each of which enjoys entrenched (aka ‘constitutionally protected’) autonomy and 2) each of those governments is independent of the other, which I would operationalise based on whether or not a level of govt has a directly-elected veto player. Based on this definition, all three aforementioned countries are federal. Now, one might argue an ostensibly federal country based on this definition is in fact so centralised that it is “de facto” unitary, which seems to me what you are ultimately saying (correct me if I’m wrong). Again, I contend that maybe one can say that for Mexico. Not for Nigeria or Brazil.

      • jd

        Mexico was, of course, a presidential autocracy during the long period of PRI rule. One book I read in the early 90s used the phrase sistema tlatoani, incorporating the Nahuatl word for an Aztec emperor. However, that was written when Salinas was president and my understanding is that degree of centralisation has now largely vanished and state autonomy is now much more genuine.

  4. If the electoral college is abolished, is the federal government going to count and tabulate all the votes regardless if plurality, 2 round, or ranked choice voting is used? Territories can be added to voted for the President which would be a good thing.

    How do other Federal States like Canada, Australia, Germany, Austria, Switzerland, Brazil, Argentina and etc handle this? Only Austria has a directly elected President (2 rounds). Most Federal States rarely have directly elected leaders from the people Brazil, Argentina, and Austria are the exceptions.

    Even if Ranked Choice Voting is used, there is no guarantee there isn’t vote spoilers and exhausted ballots unless there is a requirement to rank all candidates.

    • Indonesia, Mexico and Nigeria have directly elected presidents. The US is the only democracy, and the only federation, that elects an executive president by electoral college.

      More seriously, the US could maintain its decentralised electoral system so long as congress or some other federal authority added up the numbers reported by each state. Coordinating preference distribution would be fiddly, but no more so than Australia conducting STV elections for the senate when the ballots are still physically dispersed over huge distances. Under the Foley proposal for RCV to choose the electoral college the problem would not even arise.

      • The Federal Government may require the states that it will count and tabulate the votes for President, it need not print the ballot paper for the President. Unless each state gives the Federal government its vote total and the Federal government will tabulate it. Some states will not like the Federal government interfering with their election administration. A lot of details would have to be worked out. How do other Federations handle this?

  5. I am a little startled to discover* that in 1787 Connecticut, Delaware, Maryland, New Jersey, New York, North Carolina, Rhode Island, Vermont**, and Virginia all allowed noncitizens to vote. Georgia enfranchised noncitizens in 1868, New Hampshire in 1792, Pennsylvania in 1790, and South Carolina in 1790. The last state to disfranchise noncitizens was Arkansas in 1926.

    Typically a noncitizen had to live in the state for 6 months and swear allegiance. Now there’s an exciting electoral reform for Democrats to consider.

    *I cannot tell a lie, I did it with my little Netflix.

    **Vermont did not join the Union until 1791.

  6. There are three reasons the polls should close everywhere simultaneously.

    There is the LBJ argument. He typically arranged for the West Texas vote to come in late in case the numbers need to be adjusted after early counting in the rest of the state.

    In a country that occupies multiple timezones late voting regions (Alaska, Hawaii, Western Australia) can feel left out.

    Where, as in the US, there is no authority that formally declares the result, a candidate can claim that early trends make them president, a particularly dangerous situation when an incumbent president seeks re-election.

    Voting should therefore run for 24 hours and close when it is 8pm in the most populous timezone. If we can do it with supermarkets we can do it with polling booths.

    • Supermarkets are not ballot boxes and there is no election in the United States where Alaskan and Hawaiian voters can flip who wins in New York or Florida.

      Closing polls early and simultaneously sounds like a bad way to get around potential fraud and fraudulent actions. Actually combating potential fraud (but not in the “traditional” American way of simply disenfranchising people) and not letting candidates declare themselves winners sound like much better ideas.

      Elections are not entertainment. The broadest possible path towards getting people to vote should be used. Generous hours for early voting and ballot drop off, postage free mailing of ballots until polls close and expedited delivery, and late closings on Election Day itself.

      • Dixville Notch has just voted 5-0 for Biden-Harris. I trust that even Justice Kavanaugh (who is admittedly not my go-to expert on the nuances of consent) would regard this as a sufficiently decisive snapshot of “the consent of the people”, and would want more votes to come in.

      • This seems to be a similar mindset, which appeals to conservative (or at least right-wing) temperaments – that a simple and quick answer is preferable even if it is of dubious accuracy and arguable injustice.
        A similar mindset came out in the AV referendum in UK 2011. It seemed to offend a surprisingly large number of Britons that, in Australia, the candidate with most primary votes at 10 PM on election night was not necessarily home and hosed. Criticisms of this typically confused two separate respects in which Australia differs from the UK – counting second and later preferences, and allowing a two-week deadline for postal votes to arrive.
        This mindset manifests in other areas as well, among conservative and/or libertarian thinkers. So, eg, fans of Nozick or Hayek argue that (i) their guru’s ultra-simplified rules of acquisition will give results that overlap 90% with what an omniscient god or state would allocate on the basis of hard work and just deserts, and (ii) any attempt to eliminate that final 10% gap, where a Martin Shkreli or a Paris Hilton can become a millionaire without hard work or just deserts, would be futile since it would require the rules to become massively more complex. See “Simple Rules for a Complex World” (1992) by Richard A Epstein, a leading libertarian jurist, making precisely this point.

      • It is ironic that soi-disant originalists would argue for “quick and dirty” as a constitutionally mandated, or at least constitutionally permissible, inroad on the right to have one’s vote counted, since a minutes’ perusal of Article I Section 1 of the US Constitution, as revised by the Twelfth Amendment, leaves some… textual doubt that the Founding Fathers were hell-bent keen to see an immediate – even if rough and ready – Presidential election result right away…
        PS: I thought the Supreme Court in 2000 said Bush v Gore was not supposed to be cited as a precedent?

      • Polling booths and counting centres must be guarded and secured for some days before and after an election anyway. Th logistics of 24-hour voting were not nearly as difficulty as you suggest. Adopting one electoral reform does not increase or decrease the difficulty of adopting any other electoral reform. Quite minor electoral reforms can make a huge difference. Thabo Mbeki instead that Zimbabwean returning officers post a local count on the doors of the polling booth as well as sending the numbers into a central tally room. That’s why the opposition won the 2008 parliamentary election.

        There are too many ‘No-one could have imagined’ stories. No-one could have imagined a strategy of suborning the postal service, but a US district court judge is, as we speak, intimating the will consider criminal contempt charges against ISPS leadership for defying a court corder to deliver ballots on time.

        As Judge Sullivan wrote:


        blockquote>Sullivan contrasted the chaotic mishmash of Election Day rules with the relative simplicity of the federal income-tax deadline: “Think about it. Every year everyone knows to file taxes by April 15th. It’s seamless. If you don’t file, there’s penalties. But everyone knows — that’s a given.”
        By contrast, state vote-by-mail deadlines present a spaghetti-like tangle for the Postal Service and voters to navigate.
        “Postmarks matter, postmarks don’t matter. … Delivery matters, delivery after a date doesn’t matter. Why can’t there be one set of rules?” Sullivan said, concluding, “Someone needs to be tinkering with the system to make sure it works seamlessly and better for the American voters.”

        24 hour voting is not directly related to the bungled mishmash of deadlines for mail voting the US. However the idea of a single standard voting time would make electoral education and campaigning far, far easier that it is now.

Leave a Reply

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

You are commenting using your 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.