I have already intimated to you the danger of parties in the State, with particular reference to the founding of them on geographical discriminations. Let me now take a more comprehensive view, and warn you in the most solemn manner against the baneful effects of the spirit of party generally.
…
the common and continual mischiefs of the spirit of party are sufficient to make it the interest and duty of a wise people to discourage and restrain it.
...
It serves always to distract the public councils and enfeeble the public administration. It agitates the community with ill-founded jealousies and false alarms, kindles the animosity of one part against another, foments occasionally riot and insurrection. It opens the door to foreign influence and corruption, which finds a facilitated access to the government itself through the channels of party passions.
...
In offering to you, my countrymen, these counsels of an old and affectionate friend, I dare not hope they will make the strong and lasting impression I could wish; that they will control the usual current of the passions, or prevent our nation from running the course which has hitherto marked the destiny of nations. But, if I may even flatter myself that they may be productive of some partial benefit, some occasional good; that they may now and then recur to moderate the fury of party spirit, to warn against the mischiefs of foreign intrigue, to guard against the impostures of pretended patriotism; this hope will be a full recompense for the solicitude for your welfare, by which they have been dictated.
- George Washington, Farewell Address (1796)
Voting Methods for Single-Winner Elections
Partisan extremism is unhealthy for democracy, but is fueled in the United States by a two-party system in which single-party primary elections are usually more important than the subsequent inclusive general elections. What is worse is that these single-party primary elections are funded by taxpayers, many of whom are not members of a major political party and in some states are not allowed to vote at all in the partisan primary elections.
The manner of electing public officials should not be controlled by political parties. Instead, all public elections should be neutral with respect to political parties. All voters should be entitled to vote and to have their votes counted equally regardless of the party affiliation of the voter or the candidates they vote for. All candidates should be treated equally and face the entire electorate in any public election.
For more information on open primaries, see https://www.openprimaries.org/ .
Several states have reformed elections so that candidates for public office must face all voters in both primary and general elections, but most of these reforms are flawed due to antiquated election methods. In particular, Top-2 primary elections with plurality voting can end up nominating two relatively unpopular candidates due to vote-splitting among a large number of more popular candidates. California, Washington, and Louisiana use Top-2 elections followed by a runoff. Alaska recently established a Top-4 primary system with successive elimination (called “instant runoff voting”) in the general election. This is an improvement over plurality voting, but is not optimal. In the August 2022 special election for Alaska's vacant house seat, the third place candidate, Nick Begich, was preferred by 52.5% of voters to the “winner” Mary Peltola. This is clearly an undesirable result. The successive elimination method is so complicated that a programming error in California's Alameda County was not discovered until 50 days after the election, by which time the winning candidates had already been certified (https://www.sfchronicle.com/bayarea/article/Alameda-County-admits-tallying-error-in-17682520.php).
This document presents a comparison of commonly proposed election methods: Comparison of Election Methods
Here is a comparison of election methods applied to Alaska's August 2022 Special House Election: Election Methods re 2022 AK Special Election
The supposed purpose of “ranked-choice voting” is to obtain majority support for the winning candidate. But the commonly-used successive elimination method is not a reliable way to achieve that result. The only way to ensure majority support for the winning candidate is simply to elect the candidate who has majority support against each of the other candidates. And if by chance no candidate has majority support against each of the others, elect the one who comes closest. In other words, use ballot rankings to perform a round robin of instant runoffs between each pair of candidates, and elect the one who defeats, or comes closest to defeating, every other candidate. This Instant Round Robin is called “(Condorcet) Minimax” or “Simpson-Kramer” in academic literature.
The Instant Round Robin method is simple, transparent, and yields objectively better electoral results that other currently used or proposed methods. The basic idea is that a candidate (called a Condorcet winner) who is preferred by voters to each of the other individual candidates should be elected. Voters may rank candidates in order of preference, or give each candidate a favorability “score”. In an election with four candidates, each precinct could report results for all six two-way runoffs, or make a concise 4x4 table showing all of the named candidates in each row and column (the report could be expanded to show top write-in candidates or add “other” to denote all other vote recipients combined). A number in each table cell shows how many voters preferred the row “candidate” to the column “opponent”. Summing the results from each precinct yields all the information needed for anyone to determine the winner. Just find the worst runoff result (strongest opponent) for each candidate. The winner is the candidate with the best runoff result against their strongest opponent. That candidate is either a Condorcet winner to comes closest to being one.
This method guarantees that any losing candidate has the weakest possible claim that they should have won instead. It comes as close as possible to guaranteeing majority support for the winning candidate. This paper by Richard Darlington of Cornell University summarizes the current status: https://arxiv.org/pdf/1807.01366. Academics at Princeton University also make a strong case for Condorcet methods: https://www.princeton.edu/~cuff/voting/theory.html. Various Condorcet methods differ only in how they deal with the rare instances when there is no Condorcet winner.
Interested parties can try out different election methods at the Condorcet Internet Voting Service (CIVS). You can also explore different voting methods at the Smart Voting Simulator.
For Top-4 or Top-5 open primary elections, I recommend simple plurality voting. It is unlikely that vote splitting would eliminate a highly competitive candidate, and primary voters should not be expected to carefully evaluate a large field of candidates.
Multi-Winner Elections
Ideally, multi-winner elections should elect a slate of candidates with proportional representation, meaning that each chosen candidate is elected with votes (or fractions of votes) from disparate groups of voters. This prevents a single plurality voting block from electing all of the winners and leaving everyone else unrepresented. A method for achieving this is called “Multi-Winner” or “Proportional” Ranked-Choice Voting (P-RCV). This method is explained here. The basic idea is that each voter gets a single vote (which might be split between candidates during tabulation), and any candidate who exceeds a threshold number of votes is elected. For “N” voters filling “k” positions, the threshold is N/(k+1). For example, to be elected to one of three positions, a candidate must receive more than 1/4th of the total number of votes (it is only possible for three candidates to exceed this threshold). If a candidate exceeds the threshold by more than one vote, then the excess votes are mathematically attributed to all ballots counted toward the winning candidate, and assigned to the next choice on those ballots.
There are two problems with P-RCV. First, it is quite complicated to implement, requiring all ballot information to be readily available for counting, and mathematically splitting votes into fractions when a candidate's number of votes exceeds the threshold. Second, if too few candidates reach the threshold, then someone will be elected whose public support does not meet the threshold.
A much simpler method that approximates proportional representation is a multi-winner version of Bucklin voting:
Any candidate exceeding the threshold with only 1st-choice rankings is elected.
If seats are still open, 2nd-choice rankings are added and then candidates with the most votes exceeding the threshold are elected in order until either all positions are filled or no other candidates exceed the threshold.
Additional rankings are added until all positions are filled.
If all positions are filled using 1st-choice rankings, then the result is the same as proportional voting. Minority voting blocs can still elect candidates. If the number of rankings counted reaches the number of positions to be filled, then the result is similar to “vote for many”, in which a majority voting block could pick all the candidates. Fewer voters would be represented than with proportional voting, but every candidate elected would have support from more than the threshold number of voters.
Note that if the number of positions is reduced to one, then the threshold is 50%, or a majority. The problem with single-winner Bucklin voting is that ranking more than one candidate can reduce the chances of the favorite winning. That is much less of a concern when filling multiple positions, and voters would be well-advised to rank at least as many candidates as there are positions being filled.
Created: April 22, 2021; Last updated: June 17, 2024.
Copyright © 2021-2024 Robert A. Close
