What is the best way to elect the US President?

Our current Presidential election system is doubly broken. I will use EC for Electoral College, WTA for winner take all, and PV for a national vote.

  • WTAEC + FPTP: The status quo.
  • WTAEC + IRV: This is perhaps doable and seems to be gaining momentum. Still has problems with CA/TX having to centrally count a huge election.
  • Proportional EC + STV: Here, unlike usual STV, candidates can win multiple seats (electoral votes). This is unlikely to get used for the same reason that 96% of states choose to use winner-take-all.
  • WTAEC + Score or STAR: This could get used, and it would turn into BetterVotingAdvocacy’s concern.
  • Proportional EC + RRV or some other similar method: Wastes many of the properties of Score/Approval; unlikely to get used.
  • National single winner methods. All have the problem that they MAY require either a constitutional amendment or a change to the NPV compact. Otherwise you can have e.g. some states using Score, some using IRV, etc. However, suppose by some miracle all 50 states adopt the same method…
    • IRV: This is a HORRIBLE idea, but the Green Party is pushing for it. :frowning: The current momentum seems to be approaching either this, “state IRV” (each state computes IRV until one candidate gets >50%; final totals are then added up), which is even worse, or IRV for senate/governor and FPTP for president.
      Technically this is compatible with FPTP.
    • Score/Approval: Also compatible with FPTP. This could potentially work, and voters will pressure states with coarse scales to adopt finer ones until everyone is on 0-9.
    • STAR: This is not really compatible with Score even though the ballots are the same.
  • I think that apart from letting FPTP coexist with another method, and letting Approval and Score mix, compatibility rules just make everything more difficult. They have to be agreed on, and a state could still cause chaos by adopting a new method (e.g. Majority Judgement) that has no compatibility rules at all.

One of the issues here is that because the Electoral College forces separate elections, it will be very difficult to have a compromise candidate win, i.e. even proportional EC would mean at best, a majority-preferred candidate wins. That too, within only a two-party framework. Ultimately, that may be the compromise necessary to make; most Americans are unlikely to acquiesce to a parliamentary model, and a popular vote is far out of reach. Perhaps when voting reformers go to the statewide level, ballot measures could be used to unilaterally allow Approval/Score Voting to be counted alongside FPTP within the NPVIC?

There’s also Proportional EC + D’Hondt or Sainte-Laguë, which would not need to be counted centrally. (Not to suggest that they would be better overall.)

I think it is important to revisit the nature of single winner offices, and where possible introduce some kind of multi-member element that has meaningful and useful ongoing consequences for the single winner, then run the election on multi-member principles. This is my inclination for all single winner races, though in practice we should probably only go to the trouble for the most serious ones; governors of states are important enough and powerful enough, and probably mayors of big cities. Certainly the President is! If we can introduce some proportionality nuance into the execution of big important executive offices, the general checks and balances setting conditions on smaller less important single winners–Sheriff, etc–give people recourse absent when the single winner is at the top of the power pyramid.

So for some time now I have had the notion that we ought to have elections for the President in which people have a single vote for a single person, as now in state winner take all races for the state’s EC, and then have a new kind of Electoral College comprised of the top winning candidates exercising a proportional asset vote. The specific mechanism I have been thinking of for years is based on an arbitrary threshold of candidates to consider I set at 90 percent, but recent notions I have had just in the past few days suggest maybe another tack toward the basic body to be formed that might be less arbitrary.

Another key idea here is that beyond the Presidential election, the necessary candidates for the President they elect to win are encouraged and enabled to stick around the Executive branch–definitely not with any of them becoming VP, that should remain a candidate/party choice to prevent perverse incentives, but perhaps with special rights to assume other executive offices as on the Cabinet, and to tie the ongoing operation of the President to their ongoing advice and consent, such as linking veto power to these also-rans, and other quantifiable Presidential powers. Say we formed an Elector Council of the candidates with proportional shares of EV adding up to 9, so that 5 votes are needed to win, and the plurality candidate and the runner-up each have 4 EV with a third candidate having the balance of 1. (This is what would happen in an election replicating the distribution of PV between Bill Clinton, GHW Bush, and Ross Perot in 1992). So say “Clinton” and “Perot” agree to make Clinton President, so Perot has special rights to serve in the Clinton administration. And the veto is dependent. I long ago decided we had better not make the veto simply the ratio of EV held by the Executive Council members (here just Clinton and Perot) to total–here that would be 5/9, or 55.555etc percent, obviously much lower than 2/3! But suppose Clinton and Perot can agree to invite Bush in as well, ought that mean that the three of them can veto anything with no override at all? I’d say no, after all the three of them put together do represent almost all popular votes for President that year–but not quite all, and that much power is dangerous. So I figured, they count up to the level needed to get a majority, and after that, additional EV count as only half their number–so getting Bush to join with them would mean that all 4 of Bush’s EV are superfluous, so they only count as 2–for a combined ratio of 7/9. Actually, it would run Clinton’s 4 and one of Bush’s 4 count in full, and the rest of Bush’s 3 become 1.5 to add to Perot’s 0.5 but it comes out the same collectively of course. 7/9 is a heck of a strong veto to try to override, but if the three of them were to do something seriously egregious a sufficiently outraged Congress could still do it, and getting all three on the same page probably does merit executive power that is almost but not quite total.

Similarly take say judicial appointments. With fixed small integer EV, a fair balance between Clinton and Perot would be for Clinton to name 4 justices, but then after he has named 4, Perot is now entitled to throw in another one of his choosing. In practice this might mean Perot lets Clinton name say 20, and then cashes in his chips saying he’s got a couple to name, which resets the score of those Clinton has over Perot’s share back to 4.

These special powers shared out among the junior Executive Council members ought to form a basis for an ongoing culture of mutual consultation of the ExCoun members, all of whom after all received a substantial share of the votes of the American people. This combined with junior EC members also serving specified discretionary duties presumably at Cabinet level means that there is ongoing skin in the game for the also-rans who choose to back a leading candidate to become President. (I thought it would go without saying, maybe I should mention I assume the rule is the coalition forming a majority of EV must elect the one among them who has the most popular votes; Bill Clinton would not have the option of handing Perot the supreme office nor could Perot expect that as a legal option. Thus if to bypass Perot Clinton and Bush agreed to team up, freezing Perot out instead of Bush, Bush must accept that Clinton cannot offer to let Bush continue as President–a lot of people favored Bush, but more favored Clinton so it has to be that way. Bush and Clinton would not have done this–but if Perot had been Hitler instead, it is an option. Mind, if it were Hitler instead of Perot, but a somewhat less hotblooded version of Hitler able to patiently endure being in a junior position, I would seriously wonder whether Bush would team up with him to remain President. Or if Clinton just might too).

Assuming none of the strong candidates are in fact as bad as Hitler, the upshot is a much wider popular base of support for the current administration. A duopoly team up in the case above would give the Clinton-Bush team a veto around 70+ percent–note that is not too much higher than 2/3.

To complete the picture I think we need a new office, one I call “Tribune.” A Tribune is an Elector who does not participate in the Executive council and does not serve the administration. The exact nature of the powers a Tribune should have remain poorly defined in my mind, but mainly involve powers of investigation. The point is to give candidates winning enough PV to make a significant share of the total, who dissent from the ruling coalition, some official standing and role such that a threat of an Executive Council member to resign is more credible, and to make a decision by a President to leave an Elector out of coalition have some cost to it, and give the opposition a similar role that keeps them hanging around haunting the administration.

With a package of reforms along these lines, voters could simply cast a single vote for their most preferred honest candidate, and let the numbers resulting guide how their chosen candidate is involved in sorting the outcome out. A much broader coalition than those devoted to the person of the one who won the supreme office exists connecting other voters to the administration via their own favorite if they wind up in the Executive Council, and to opposition with some defined powers to dog the administration’s heels should they not.

That could create “center squeeze”. It’s not uncommon in, say, 3 candidate elections, for the best compromise option to be the one with the fewest first choice votes. This is one reason why IRV is so unpopular here.