Ranked choice voting

Proposal

In elections where there are three, four, or five candidates running for a single available seat and the intent is to elect a candidate with the support of a majority of the voters and write-ins are not counted, use the ranked choice voting method.

In elections where there are more than five candidates for a single seat and the intent is to elect a candidate with the support of a majority of the voters, use plurality voting in a primary round to identify up to five eligible candidates with write-ins, and then use ranked choice voting in an election without write-ins to determine the winner.

In referendums where there are three, four, or five choices and the intent is to select a choice with the support of a majority of the voters and write-ins are not counted, use the ranked choice voting method.

In referendums where there are more than five choices and the intent is to select a choice with the support of a majority of the voters, use plurality voting in a primary round to identify up to five choices with write-ins, and then use ranked choice voting in an election without write-ins to determine the winner.

For referendums using the ranked choice voting method, the number of available choices shall be considered as candidates in an election, and the number of answers to be submitted shall be considered as seats.

In local elections where ranked choice voting is applicable, use the ranked choice voting local method. In state and federal elections where ranked choice voting is applicable, use the ranked choice voting distributed method.

A “majority” or a “threshold” is determined based on the number of people who voted in the election. In elections with multiple rounds, a “majority” or a “threshold” is determined based on the number of people who voted in that round. A run-off election shall be considered a round of the preceding election.

When an election has more than one question on the ballot, each question shall be considered a separate election. The use of ranked choice voting shall be determined separately for each election.

Intent

Ensure elected officials in any branch of government have the most popular support. To empower people to express their first choice for any election without fear of “wasting” their vote on a candidate who is unlikely to win, via the ability to express their second choice, third choice, and possibly further choices if applicable.

When a ballot has a question for “who should be governor?” and a question for “who should be state treasurer?” each one of these is a separate election, and they are conducted concurrently using the same ballot. If a voter filled in a choice for governor, but did not fill in a choice for state treasurer, that voter participated in the election for governor but did not participate in the election for state treasurer. When determining the threshold for winning, that threshold is based on the number of people who participated in the election, not based on the size of the general population.

Discussion

Ranked choice voting is appropriate, useful, and worth the additional cost in situations where there are three to five candidates running for a single seat, where write-ins are not counted, and the intent is to elect the candidate with the most popular support, even if that candidate is not the first choice of a plurality of voters. In other situations, simpler voting methods are more appropriate. See the voting methods proposal for recommendations to when to use different voting methods.

This proposal for ranked-choice voting is an effective and easy-to-use way to ensure elected candidates have the most popular support. Ranked-choice voting allows voters to express their strong interest in any candidate while ensuring that their vote still counts if their preferred candidate does not win.

Ranked-choice voting can be used for single-seat elections, such as governor or mayor, and also multi-seat elections, such as electing a number of people to the city council or legislature. The procedure is essentially the same.

When only a single candidate is running for an office, the results are the same for a simple majority system, a plurality system, and a ranked choice voting system.

When two or more candidates are running for an office, the results may be different under each system. In a simple majority system, a candidate needs 50% + 1 of the votes to win. When there are only two candidates, it is simple. When there are three or more candidates, it may result in a run-off election. In a plurality system, a candidate just needs to get the most votes to win and that candidate may not have the support of the majority of the population. In a ranked choice voting system, a candidate in a race for a single seat needs 50% + 1 of the votes to win, like in a simple majority system, but unlike the simple majority system, a candidate may obtain the majority support via a combination of first-choice and second-choice votes, or even lower ranked choices, which means the candidate with the most popular support can be identified without having one or more run-off elections. This is why ranked choice voting is also sometimes called instant run-off voting.

For referendums or measures, the choices “Yes” and “No” are typically the only choices and in that case, either simple majority or plurality voting are more appropriate. However, if there are more than two choices, then ranked choice voting is a good alternative to plurality voting in order to identify the choice with the most popular support.

The proposal to only use ranked choice voting when there are three to five choices, and not more, is an arbitrary but practical proposal. Ranked choice voting is more complicated than simple majority or plurality voting, and its significant advantage in selecting a candidate with popular support is eroded when there are too many choices for people to rank. When there are too many choices for people to rank, they tend to leave empty spaces and when there are too many exhausted ballots the effectiveness of ranked choice voting in identifying the candidate with the most popular support is reduced, and therefore not worth the additional complexity. But with a small number of choices, ranked choice voting is more effective than the plurality method at identifying the candidate with the most popular support.

The voter experience (sticker option)

The sticker method helps to ensure that each candidate is mentioned only once in the rankings. However, disadvantages include accidentally tearing the stickers, and loose stickers resulting from a voter attempting to change their placement after their initial selection. If there is a variant of this idea that solves those issues it would be preferred over stickers — for example, doing something visually similar but using a voting machine instead of paper.

In a ranked choice voting election with five candidates, the ballot would look like this:

The candidates are A, B, C, D, and E. Relevant information about each candidate can be included, such as political party. The page with questions would be divided into two columns. The first column would have the questions. The second column would have the candidate stickers. Each candidate sticker has the candidate’s name and political party affiliation. If machine-counting is being used, the machine must use optical character recognition (OCR) to detect the candidate’s name. The stickers SHALL NOT contain any machine-readable code, because machine-readable codes are not readable to humans and we need to eliminate the possibility of error (intentional or unintentional) wherein the machine-readable code does not match the human-readable text.

1. Which of these candidates do you prefer? ___ (place sticker here)

If your preferred candidate is eliminated, your vote can be moved to support another candidate. The following questions are optional, but highly recommended to fill in as many spaces as you can.

2. If your preferred candidate is eliminated, which other candidate is your second choice? ____ (place sticker here)

3. If your second-choice candidate is eliminated, which other candidate is your third choice? ____ (place sticker here)

4. If your third-choice candidate is eliminated, which other candidate is your fourth choice? ____ (place sticker here)

5. If your fourth-choice candidate is eliminated, which other candidate is your fifth choice? ____ (place sticker here)

If there are fewer candidates, there would be fewer questions.

The voter experience (ink option)

In a ranked choice voting election with five candidates, the ballot would look like this:

The candidates are A, B, C, D, and E. Relevant information about each candidate can be included, such as political party.

1. Which of these candidates do you prefer? ___

If your preferred candidate is eliminated, your vote can be moved to support another candidate. The following questions are optional, but highly recommended to fill in as many spaces as you can.

2. If your preferred candidate is eliminated, which other candidate is your second choice? ____

3. If your second-choice candidate is eliminated, which other candidate is your third choice? ____

4. If your third-choice candidate is eliminated, which other candidate is your fourth choice? ____

5. If your fourth-choice candidate is eliminated, which other candidate is your fifth choice? ____

If there are fewer candidates, there would be fewer questions.

Ranked choice voting local method

The essential steps of ranked-choice voting are described here.

Before the election:

Step 1. The number of seats is determined. For example an election for a position that only has one elected official such as governor, secretary of state, mayor, etc. is an election for a single seat, whereas an election for representatives to the legislature or to a board may have multiple seats to fill such as two, three, or more seats.

Step 2. The threshold for election of any candidate is set based on the number of available seats. The threshold percentage is set to the inverse of one plus the number of available seats. For example The threshold for electing a candidate is set to one plus the inverse of one plus the number of available seats. For example if there is one seat the threshold percentage is 1/(1+1) or 0.5 expressed as 50%; if there are two seats the threshold percentage is 1/(1+2) or 0.33 expressed as 33%; if there are three seats the threshold percentage is 1/(1+3) or 0.25 expressed as 25%. The percentage is rounded down to the nearest one hundredth, expressed as the nearest integer in the percentage.

Step 3. All the candidates eligible for the election are listed on the ballot.

During the election:

Step 4. Each voter ranks the candidates in their order of preference. If the voter leaves an empty space, all selections after that space are moved up to eliminate the empty space. If the voter writes in a candidate where write-ins are not allowed, the write-in is treated as an empty space and all selections after that space are moved up to eliminate the empty space.

Counting the votes:

Step 5. All remaining voting cards are examined and any ineligible candidates are struck from each card. Ineligible candidates are candidates who are legally ineligible to be elected to the office (for example any write-in who is a non-resident or not even a real person), candidates who already won a seat in that election (only applies to second or subsequent rounds of tallies), and candidates who were removed from the list due to having the fewest votes in a previous found (only applies to second or subsequent rounds of tallies). In the first round, all “remaining” voting cards is all the voting cards since none have been counted yet. If any card has no eligible candidates remaining, that card is removed from the pile to be counted.

Step 6. The threshold for winning a seat is computed using the formula set in step 2 and using the number of cards remaining for the actual threshold calculation. The actual threshold is the number of ballots to be counted times the predetermined percentage threshold, rounded down to the nearest integer, plus one. If there are two or three or four candidates competing for a single seat (governor, secretary of state, mayor, etc.), the threshold is 50% + 1. If there are candidates competing for two seats (for example representatives to a legislature or to a board with two open seats), the threshold is 25% + 1. If there are candidates competing for three open seats, the threshold is 33% + 1. If the percentage calculation results in a non-integer, the percentage result is rounded down to the nearest integer. For example, if 1001 voter cards were submitted for a one-seat election, and 10 people wrote “Micky Mouse” as their only candidate with no second choice, then at this step there are 991 remaining voter cards and the threshold to win is 496 votes (one plus 50% of 990 rounded down, or 1 + 495).

Step 7. Each candidate’s votes are tallied from the voters’ highest available choice in their rankings. In the first round, this is generally everyone’s first choice unless someone used a write-in on their first choice and that write-in was not an eligible candidate.

Step 8. If any candidate received more than the required threshold of votes, that candidate wins a seat. The candidate is then removed from the list (having already won a seat), and all voting cards tallied for that candidate are then removed from the pile (ensuring the principle of one person, one vote is upheld).

Step 9. If there are no seats remaining, the process ends here. Otherwise, if there are any seats remaining: The candidate with the fewest votes is eliminated from the election. If there are two or more candidates tied for fewest votes, all the tied-for-last candidates are eliminated from the election. On each voting card that was counted for an eliminated candidate, the candidate is crossed out and the card is moved to the next highest ranked candidate on that card that has not yet been eliminated. That is, if a voter’s first choice was eliminated, that voter’s vote then moves to the voter’s second choice, and if that candidate is eliminated, that voter’s vote then moves to the voter’s third choice, and so on. The process then returns to step 5.

Ranked choice voting distributed method

For statewide or federal elections, or any election in which there are two or more levels of government involved, the distributed method enables distributed counting of votes and eliminate the possibility of fraud at higher levels of government where local observers cannot see the count.

Before the election:

Step 1. The number of seats is determined. Same as step 1 in the regular method.

Step 2. The threshold for election of any candidate is set based on the number of available seats. Same as step 2 in the regular method.

Step 3. All the candidates eligible for the election are listed on the ballot. In addition, the number of permutations is computed in advance based on the number of candidates. That is, if there are 2 candidates there are 2 permutations with both candidates listed (“A,B” and “B,A”) and two permutations with only one candidate listed (“A” and “B”); and if there are 3 candidates there are 6 permutations with all candidates listed (“A,B,C”, “A,C,B”, “B,A,C”, “B,C,A”, “C,A,B”, and “C,B,A”) and 6 permutations with only two candidates listed (“A,B”, “A,C”, “B,A”, “B,C”, “C,A”, and “C,B”) and 3 permutations with only one candidate listed (“A”, “B”, and “C”). Each permutation is identified and assigned a unique number. For consistency, permutations shall be arranged shortest to longest, in alphabetical order (“A”, “B”, “C”, “A,B”, “A,C”, “B,A”, “B,C”, … “C,B,A”).

During the election:

Step 4. Each voter ranks the candidates in their order of preference. If the voter leaves an empty space, all selections after that space are moved up to eliminate the empty space. Write-ins are treated as empty spaces in the distributed method. Each polling station reports its results as a table showing all the available permutations and number of times each one occurred. In the subsequent steps, “voting cards” or “permutation cards” are the identified permutations and their counts representing the number of original ballots cast of each permutation, instead of the original ballots.

Counting the votes:

Step 5. All remaining voting cards are examined and any ineligible candidates are struck from each card. Ineligible candidates are candidates who are legally ineligible to be elected to the office (write-ins are removed and legitimate choices are moved up as if the write-in were a blank space), candidates who already won a seat in that election (only applies to second or subsequent rounds of tallies), and candidates who were removed from the list due to having the fewest votes in a previous found (only applies to second or subsequent rounds of tallies). In the first round, all “remaining” voting cards is all the voting cards since none have been counted yet. If any card has no eligible candidates remaining, that card is removed from the pile to be counted.

Step 6. The threshold for winning a seat is computed using the formula set in step 2 and using the number of cards remaining for the actual threshold calculation. The actual threshold is the number of ballots to be counted times the predetermined percentage threshold, rounded down to the nearest integer, plus one. If there are two or three or four candidates competing for a single seat (governor, secretary of state, mayor, etc.), the threshold is 50% + 1. If there are candidates competing for two seats (for example representatives to a legislature or to a board with two open seats), the threshold is 25% + 1. If there are candidates competing for three open seats, the threshold is 33% + 1. If the percentage calculation results in a non-integer, the percentage result is rounded down to the nearest integer. For example, if 1001 ballots were submitted for a one-seat election, and 10 people wrote “Micky Mouse” as their only candidate with no second choice, then at this step there are 991 remaining ballots and the threshold to win is 496 votes (one plus 50% of 990 rounded down, or 1 + 495). The number of ballots is computed as the sum of all the highest available choice counts on each permutation card.

Step 7. Each candidate’s votes are tallied from the voters’ highest available choice in their rankings. In the first round, this is generally everyone’s first choice. For example, if permutation 1 is candidate “A”, permutation 4 is candidates “A,B”, permutation 10 is “A,B,C”, and permutation 11 is “A,C,B”, then the number of permutations 1, 4, 10, and 11 from each county is tallied up and that total is the number of first-choice votes for candidate A.

Step 8. If any candidate received more than the required threshold of votes, that candidate wins a seat. The candidate is then removed from the list (having already won a seat), and all permutation cards tallied for that candidate are then removed from the pile (ensuring the principle of one person, one vote is upheld).

Step 9. If there are no seats remaining, the process ends here. Otherwise, if there are any seats remaining: The candidate with the fewest votes is eliminated from the election. If there are two or more candidates tied for fewest votes, all the tied-for-last candidates are eliminated from the election. On each permutation card that was counted for an eliminated candidate, the candidate is crossed out and the card is moved to the next highest ranked candidate on that card that has not yet been eliminated. Continuing the example from step 7, if candidate “A” is eliminated, then permutation 1 is removed because there are no remaining choices, permutation 4 strikes candidate A so that candidate B is the next choice, permutation 10 goes to candidate B also, and permutation 11 goes to candidate C. This means that everyone who ordered their candidates in the order A, B, C moves from A to B, and everyone who ordered their candidates as A, C, B moves from A to C. The result is the same as if the RCV process was operating on the original voting cards. The process then returns to step 5.

Write-ins

RCV local method can work with write-ins, but RCV distributed method does not work with write-ins because the available permutations must be predetermined and coordinated among all the local election offices.

On the national level, write-in candidates typically garner less than a percent of the popular vote and have never won a presidential election in the United States.

On the state level, some write-in candidates have been elected. However, in the United States, most states have rules regarding write-ins, such as requiring write-in candidates to submit a letter of intent before the election in order for their votes to be tallied, or prohibiting write-ins completely. Few states allow write-ins without restrictions.

The write-in restrictions make sense: It would be unjust to have a person campaigning for an office without ever registering as a candidate and having to follow election rules, and it places an additional burden on election officials to read people’s handwriting, determine if the write-in candidate is even eligible to be elected for that position (is it a real person? is that person eligible to serve in that office? age, residence, citizenship, criminal history, impeachment history, etc.).

For these reasons, this proposal allows write-ins for local elections but prohibits write-ins for statewide and federal elections where the distributed method is used. When tallying votes containing write-ins for a ballot question where write-ins are not allowed, the write-in should be ignored and all legitimate answers should be moved up as if the write-in was an empty space.

What is NOT in this method

Some descriptions of ranked-choice voting include the following provision, which is intentionally NOT included in this proposal:

“If a candidate receives votes in excess of the minimum number of votes required to be elected (the threshold amount), votes in excess of the minimum are transferred to lower-ranked active candidates.” (or some variation of this wording).

This “spill-over” mechanism is not included because there is no clear way to decide WHICH votes in excess of the minimum should go to their next choice. The outcome of the election should NOT be dependent on the order in which people cast their vote or on the order in which those votes are counted.

Instead of the spill-over mechanism, the threshold is recomputed on every round based on the number of remaining ballots. This ensures that every person’s vote is counted only once and only for their most preferred candidate who could have received their vote.

Comparison with a runoff election

Imagine an election with 3 candidates for a single seat, where no candidate got a majority of the vote (50% plus 1) in the election.

The candidate with the fewest votes is removed from the list and a runoff election takes place. The candidate with the most votes in the runoff election wins the seat.

Some people may choose not to vote in the runoff election. Some people may not vote in the first election but vote in the runoff election. Candidates may continue to campaign between the first election and the runoff election. Soldiers who are stationed overseas may get the ballot for the first election but may not have time to get the ballot for the runoff election and submit it in time, or the runoff election may only be held in polling locations and not by mail, depriving those deployed soldiers of the ability to vote in the runoff election.

In comparison to a runoff election, ranked choice voting has a similar effect but without the disadvantages: ranked choice voting upholds the principle of one person, one vote; there’s only one election so people don’t have to go back to the polls; soldiers deployed overseas can fully participate by mail and won’t be excluded from a runoff election.

Voter education

The following important notes about ranked choice voting must be presented to each voter at the time of voting and ideally also taught in public education:

Each voter gets one vote. When a voter’s vote is counted towards a candidate who wins a seat, that voting card is removed from the election and cannot be used to elect anyone else for the same position.

Each voter should select a candidate only once. Selecting the same candidate for multiple ranks doesn’t help the candidate and may cause the voter’s vote to be lost if their preferred candidate doesn’t win and they don’t have any remaining candidates on their ballot (voting card).

Voters should rank all candidates in preference order. Leaving any empty spaces on the voting card may mean the voter’s vote will be lost if their preferred candidate win and they don’t have any remaining candidates on their ballot (voting card).

Write-ins should be limited to eligible candidates. Writing in a candidate who is not eligible means that selection cannot be counted towards any winner, and it’s the same as simply not filling in the space.

Some of these issues are preventable with voting machines that can prevent voters from selecting the same candidate more than once and can warn voters if they leave empty spaces on their ballot or when they write-in a candidate.

Arguments in favor

Argument: RCV allows voters to elect the candidate with the most popular support when more than two candidates are running for the same seat while upholding the principle of one person, one vote. In contrast, under a most-votes-wins system if three candidates are running for the same seat and nobody gets a majority, that system could elect a candidate who does not have majority support.

Argument: RCV allows people to express their preference for a candidate who is unlikely to win the election without “throwing away” their vote. This eliminates the pressure of voting for “the lesser of two evils” and allows people to vote for who they really want and then list other candidates as a second or third choice. RCV allows people to vote for who they want and not “against” the candidate they don’t want.

Argument: RCV lessens political polarization because minor party candidates are no longer “spoilers” and therefore the two major political parties can avoid the expense of attacking minor party candidates. If their candidate has broad support as a first, second, or third choice they can still win the election. This puts the emphasis on appealing to more voters instead of attacking other candidates.

Argument: RCV can be done by hand with paper ballots or with a computer system.

Argument: RCV is auditable, meaning the ballots can be archived and the outcome of the election can be reproduced by recounting the ballots.

Argument: RCV together with other common-sense election reforms will end the Democrat-Republican duopoly and allow people more choice in electing their leaders, which will pressure all political parties to promote their strongest candidates who would appeal to the most voters even as a second choice. This will benefit voters by giving them more real choices.

Argument: RCV eliminates the problem of similar candidates splitting voters with similar values and allowing a third candidate who does not have majority to support to win the election. For example, 60% support the same values, split into 29% for candidate A and 31% for candidate B, and the remaining 40% for candidate C who has an opposing position. Candidate C wins in plurality voting even though the majority of the population prefers A or B. RCV allows the voters to designate their preference, so in this example 29% of the voters choose A then B, and another 31% choose B then A, and the remaining 40% choose C alone, C then B, or C then A. Candidate A is eliminated and those votes are moved to their next choice which is B, giving candidate B 60% of the vote. Candidate B wins, reflective of the majority support.

Arguments in opposition

Argument: RCV is confusing for voters. Citation of a 19-page instruction manual in Maine. Counter-argument: The RCV method can be presented and successfully used even by inexperienced voters with just one instruction “rank the candidates in order of preference”. The existence of a 19-page manual with details about how RCV works does not mean the method is confusing for voters to use. In a simple majority or a plurality system, the question is “which candidate do you prefer?”. In RCV, it’s the same question “which candidate do you prefer?” followed by, “if that candidate doesn’t win, which of the remaining candidates would you prefer?”, followed by the same question again until all the available spaces have been filled or the voter wouldn’t want any of the remaining candidates.

Argument: RCV is complicated for election workers. Counter-argument: The 9-step method outlined in this proposal is easy to follow. While the level of complexity may be subjective, government employees must routinely deal with far more complicated procedures than RCV, and every person who files their personal income tax return every year and reads the instructions on the form is dealing with a more complicated calculation than RCV.

Argument: RCV is more prone to errors than simple majority tally. Counter-argument: Election officials need to take care to follow the procedures correctly and tally the votes correctly regardless of which method is used to select winners. Errors do happen in elections and RCV is auditable and the results can be reproduced and errors detected on recount.

Argument: RCV means not every ballot counts. Counter-argument: This would only happen if a person does not fill in all the spaces for rankings, which is the same effect as if there were an election and no clear winner and the voter decided not to vote in the run-off election, or if a person decides to write-in an ineligible candidate. It is not factually correct that the RCV method throws ballots in the trash.

Argument: RCV forces people to vote for candidates they don’t want. Counter-argument: This is not factually correct. A voter can choose to only rank the candidates they want and not rank any additional candidates. If all their preferred candidates are eliminated and their ballot is discarded due to no remaining eligible candidates on it, it’s the same as voting for the “throw away” minor party candidate in the current system or writing in a non-eligible candidate or choosing not to participate in a run-off election where their preferred candidate didn’t make the cut to the run-off.

Argument: RCV doesn’t make sense because the initial winner loses and the initial loser wins. Counter-argument: This is not factually correct, and it’s not possible to predict the outcome of any election in this manner. In RCV, the candidate with the most popular support wins. It’s misleading to say that the candidate with the most first-choice votes is the initial winner. The winner is not known until the final round. When the winner has fewer first-choice votes than other candidates, it indicates that the winner is a compromise candidate. That means fewer people ranked this candidate first, but more people are satisfied with this candidate as the winner than any other candidate, even if it wasn’t their first choice.

Argument: RCV is more prone to fraud than simple tallies. Counter-argument: This is factually incorrect. Assuming there are no procedural or counting errors, the vulnerability to fraud is the same between RCV and simple majority tally whether it is hand-counted or machine-counted.

Argument: RCV suppresses the vote by 3% to 5%, according to research by San Francisco State University (Reference). Counter-argument: There’s more than one way to present RCV on a ballot so obviously choosing a less complicated way to present it would encourage people to complete the ballot. It’s also not clear how much voter education was done to teach people how RCV works and what to do. Furthermore, the research conflated unrelated issues such as not showing political party affiliation on the ballot. Nothing in this RCV proposal prohibits showing the candidate’s political party next to their name on the ballot. Furthermore, when politicians and influencers tell people that their vote won’t count in RCV instead of teaching them how to vote in RCV, they might be contributing to the lower vote turnout and suppressing the vote themselves.

Argument: RCV is inaccurate. Citation of a California election in which a data entry error went undetected and the wrong winner was certified. Counter-argument: Election officials are responsible for presenting true ballots to voters and data entry errors are possible in any system, for example Florida’s famous “hanging chad” problem in the 2000 presidential election. Human error or machine error is always possible and nothing about the RCV method in this proposal would inherently cause the results to be inaccurate.

Argument: RCV exempts certain offices. Counter-argument: Voters don’t get to cite the reason they voted “no” on a ballot so if voters reject a limited RCV proposal because doesn’t cover all offices, politicians can misunderstand that as voters don’t support RCV instead of voters want RCV to cover more or all offices. If an RCV measure on a ballot (to approve the use of RCV itself) applies only to certain offices, and voters want it to apply to more or even all offices, then voters should approve RCV and then propose future ballot measures to expand the use of RCV to additional offices.

Argument: RCV costs more to implement. Counter-argument: The benefits of RCV are worth the additional cost to implement and to educate voters.

Argument: RCV delays election results, RCV is slow, or RCV is time-consuming. It would takes weeks to determine the winner. Citation of Alaska not even starting to count ballots until 15 days after election day. Counter-argument: Machine-counting with RCV is fast and does not have a noticeable delay. Hand-counting with RCV takes longer when two or more rounds are needed, but each round is faster and faster due to fewer ballots being counted in each round. It would not delay the results significantly. Nothing about this RCV proposal would cause a delay of 15 days before counting even starts. While there is more than one way of organizing an RCV count, at least one way is possible which is only marginally slower than a simple majority count due to the movement of second choice votes of eliminated candidates, but does not require recounting all the votes multiple times.

Argument: RCV destroys the concept of “one person, one vote” by counting some people’s votes more than once while not counting other people’s votes the same number of times. Counter-argument: This is factually incorrect. In an RCV election, every voter votes exactly one time and submits one ballot. Every voter has the same opportunity to rank all the candidates on their ballot. Voters are not required to fill in all the spaces. This is the equivalent to voting in run-off elections. Nobody says that run-off elections destroy the concept of “one person, one vote” just because the number of people who vote in the run-off could be higher or lower than the number of people who voted in the first round.

Argument: RCV forces people to vote for candidates they don’t want. Counter-argument: This is factually incorrect. Voters are not required to fill in all the spaces. A voter should only rank the candidates they want to win, in their order of preference. If a voter does not want a certain candidate to win at all, the voter should not list that candidate on their rankings.

Argument: In RCV, a candidate can collect the votes of defeated candidates without the permission of voters. Counter-argument: This is factually incorrect. In RCV, a candidates can only receive votes from voters who voted for that candidate by placing that candidate in their rankings. In RCV, voters explicitly give permission to move their vote from their first choice to their second choice when their first choice is eliminated, and from the second choice to their third choice when their second choice is eliminated, and so on. That is what the rankings mean. If a voter does not want their vote to transfer to a particular candidate, the voter should not mention that candidate in their rankings.

Argument: RCV has unpredictable outcomes. Counter-argument: A voting method should not be selected based on how easy it is for a pollster to predict the outcome, but in any case pollsters can figure out how to adjust their polls to understand their levels of support for different candidates and continue predicting elections. Even under the plurality method that many elections use, pollsters sometimes predict the outcome incorrectly.

Argument: RCV lacks transparency in tabulating results. Counter-argument: The process for counting votes in RCV is described in this proposal. If people distrust the outcome, an audit can be performed and the ballots can be hand-counted by a different crew to confirm the outcome. In any case, other than a few observers the general public does not get to see the votes being counted in any election, so in that sense the lack of transparency applies to all voting methods currently in use in general elections.

Argument: RCV does not guarantee a majority winner. Someone who did not get the most votes can win the election. It violates the principle of majority rule. Counter-argument: This is factually incorrect and conflates the first choice votes of the first round with the final tally which includes votes that were transferred from eliminated candidates. The plurality voting method uses just one round and whoever has the most votes wins, even if it’s not the majority. In the United States, in federal elections, the electoral college overrides the popular vote and it has happened that a president was elected who did not win the majority of the popular vote, even when using a supposedly simple majority voting system. The outcome of an RCV election is the candidate with the most popular support, and RCV delivers this result in a timeframe comparable to a regular election, in contrast to running multiple run-off elections that could take months to complete and come up with the same result.

Argument: RCV tilts the odds to favor the party in power. Counter-argument: This is factually incorrect. Nothing about this RCV proposal gives an advantage to a party already in power. If a party is going to cheat in an election and defraud voters, it can do that with or without RCV.

Argument: RCV opens new paths for cheaters. Counter-argument: This is factually incorrect. If a party is going to cheat in an election and defraud voters, it can do that with or without RCV. The RCV method in this proposal is fully auditable. If the argument is that cheating can happen at any step and more steps means more opportunities to cheat, even a simple most-votes-wins method can be implemented in a convoluted way to introduce opportunities to cheat. See also the argument about statewide elections.

Argument: RCV for statewide elections will be counted at the state level not at the county level, and this is bad because statewide officials can’t be trusted with counting the votes accurately and honestly. Counter-argument: The RCV distributed method addresses this concern and eliminates the possibility of fraud at higher levels of government. Each election district tallies up the votes and publishes a report of each permutation and its count, to be added up by the next higher level of government (state officials). Since the individual county tallies are public, any observer can reproduce the statewide results and confirm that they are correct. Similarly, if states are reporting their tallies for a federal election, any observer can reproduce the federal results and confirm that they are correct.

Argument: RCV encourages fringe candidates. Counter-argument: The most-votes-wins voting system encourages a duopoly and the majority-to-win voting system encourages a duopoly and requires run-off elections when no candidate gets a majority. In contrast, RCV allows more choice while ensuring that the elected officials have majority support. Election officials determine the eligible candidates according to established procedures, and that is a separate issue from the way that a winner is determined. If the procedures require a certain level of popular support (usually indicated by signatures on a petition) and a candidate meets the requirements to be on the ballot, it doesn’t matter if the candidate is fringe or not. If the candidate is fringe, by definition the candidate won’t appeal to a majority of voters and won’t win the election, and there’s nothing to fear. In fact, RCV is a better way to deal with fringe candidates than most-votes-wins because a fringe candidate won’t get a majority of votes so votes for eliminated candidates will be redistributed to the voters’ next choice and the candidate with the most support will get elected, which is unlikely to be the fringe candidate. Stated another way, if an ostensibly fringe candidate wins an RCV election, that candidate was not actually a fringe candidate.

Argument: RCV encourages radical splinter parties. Counter-argument: Elections are about letting the people choose who they want to represent them or work for them in government. Splinter parties occur when there are differences in values, beliefs, or opinions that cannot be reconciled and people feel they would be represented better by another party. RCV encourages choice and ensures that the candidate with the most majority support wins the election. This encourages candidates to appeal to more voters. If the “radical” splinter party appeals to more voters than the party it came from, then it will win. This would be democracy working as designed. To say that it’s a splinter party is undesirable merely because it’s new, without any relevant information about what is different and why it separated from the main party, is to accept a conservative bias that all change is bad merely because it’s a change without considering whether the change itself is good or bad.

Argument: RCV relies on technology that many voters don’t trust. Counter-argument: The most-votes-wins method and the majority-wins method can be counted by hand or by machine, just like RCV can be counted by hand or by machine. Some people don’t trust the machine counts regardless of the voting method they implement. Some people don’t trust hand counts regardless of the voting method they implement. Some people don’t trust the government in general, and there’s nothing you can say or do to change their mind. Some people question the integrity of elections on purpose even when they don’t have any evidence of fraud, because they think they will benefit in some way if people distrust the election results — typically this would be a foreign government or an organization that wants to overthrow the government and replace the democracy with an authoritarian form of government such as a dictatorship or monarchy. For everyone else who supports democracy and trusts their local government to count votes correctly (or wants to participate as an observer), especially if they distrust higher levels of government, RCV is a better choice.

Argument: The final winner in an RCV election is decided by computer software, potentially AI, and people cannot trust it. Counter-argument: RCV has a simple algorithm for determining the winner, which can be done by hand or via software. Both methods can be used, for example using software to quickly determine the result, and following up with a hand-count to verify the results. AI is not required to implement RCV and it would be extremely inappropriate to determine the outcome of an election via AI.

Argument: RCV allows voters to disenfranchise themselves accidentally by over-voting. Over-voting is when a voter marks two or more candidates for the same rank. Counter-argument: This can happen on paper ballots. However, it’s also possible to over-vote on simple majority or plurality paper ballots by filling in more than one circle. In all methods, voters need clear instructions on how to fill out the ballot. Ranked choice voting on paper can be done with a matrix, a list on which the voter numbers each item according to preference, or two adjacent lists where the voter draws lines between the rank and the candidate. Over-voting is possible with all of these. However, using voting machines, it is possible to prevent over-voting in all systems by allowing the voter to change their selection but not to select more than one candidate in simple majority or plurality systems and not to select more than one candidate for the same rank in RCV.

Argument: RCV allows voters to disenfranchise themselves accidentally by under-voting. Under-voting is when a voter skips some ranks, for example selecting a candidate for first choice and another candidate for fourth choice and not selecting any candidate for second or third choice. Counter-argument: This can be easily addressed when counting votes by ignoring the empty spaces and moving candidates up to the first available space. Continuing the example, the candidate listed in the fourth choice would be treated as if it were the second choice, because there’s nothing in between.

Argument: RCV allows voters to disenfranchise themselves accidentally by exhausted ballots. Exhausted ballots are ones where all of the voter’s choices are eliminated and the voter’s ballot is not included in the final tally. Counter-argument: This is a factually incorrect. This is the same situation as in a simple majority system where a voter decides not to participate in the run-off election because it doesn’t include their preferred candidate, and it’s the same situation as in a plurality system where the voter chooses a candidate who does not receive the most votes — just because a voter’s vote didn’t count towards the winner doesn’t mean they were disenfranchised. If a voter only wants a single candidate and not the others, and the voter’s single candidate got the fewest votes and is eliminated from contention, that voter was not disenfranchised. They voted, their vote was counted, and their candidate didn’t make the next round. That is very much “enfranchised”. Not everybody wins. To be disenfranchised is to be denied the right to vote, which is clearly not what happens when a person votes and their vote is tallied and their candidate doesn’t win because the candidate got the fewest number of votes and was eliminated from contention. In fact, ranked choice voting provides a way for voters to still have their vote count even if their first choice doesn’t make it. It’s one of the advantages of ranked choice voting.

Argument: RCV demands that voters have a large amount of information about candidates’ differing views. Counter-argument: Voters should know something about candidates’ differing views under any system discussed here. If voters only know about some candidates, probably the ones most interesting to them, they can choose to rank only the ones they know and not fill in the rest of the spaces. Leaving empty spaces is better than choosing remaining candidates at random because it shows support for the candidates they know and not fake support for candidates they don’t know. This is the same principle as voting or not voting in an election where there’s only a single candidate. If you know the candidate and support them, you fill in the circle. If you don’t know them or don’t support them, you don’t fill in the circle.

Argument: RCV doesn’t work because of an information deficit. It is well documented that American voters often lack basic information about candidates’ policy positions. It seems unlikely that they could confidently rank multiple candidates based on a sound assessment of their platforms. Counter-argument: This information deficit already exists, so it’s true that if a candidate doesn’t have enough information to answer the question “which candidate do you prefer?” in a simple majority or plurality system, they also wouldn’t have enough information to answer the follow-up question “if that candidate is eliminated, which of the remaining candidates do you prefer?”. However, the duopoly and the trend towards increased political polarization may be contributing to the information deficit because voters don’t need to bother learning about candidates if they are only going to select candidates based on the name of the party next to their name, and where voters are consistently being told that a vote for any minor party candidate is a wasted vote. RCV may get people interested in learning about candidates again. And if it doesn’t, they can still vote for a single candidate based on the party next to their name.

Argument: In RCV, less knowledgeable voters are more likely to rank fewer candidates, potentially denying them influence over the election outcome. Counter-argument: RCV does not deny less knowledgeable voters influence over the election outcome. This argument is equivalent to saying that in a plurality voting system, voting for a candidate who doesn’t win denies the voter influence over the election outcome. It is a voter’s choice whether they rank just one candidate or multiple candidates. Election officials can choose to provide a voter’s guide that lists the candidates and their positions and statements, or not, and voters can choose to learn about the candidates prior to the election, or not, and this is the case in any of the voting systems discussed here. Furthermore, in a simple majority or plurality system, a voter can still decide not to vote for any candidate running for a position they don’t care about or where they don’t know the candidates.

Argument: In the Democratic primary election for candidates for governor in 2018, there were seven candidates on the ballot and more than seven percent of the ballots were exhausted by the end of the fourth round of tabulation. Counter-argument: This means that seven percent of the primary voters selected four candidates who had the least amount of support. In a simple majority system, none of those four candidates would have been anywhere near a majority. In a plurality system, showing each of the candidates and the number of votes they got in decreasing order, the same four candidates would have been the long tail of very low numbers in the last four positions. Either way, those candidates weren’t going to win. Voters with exhausted ballots are voters who chose only losers, which is the same as voting for losers in simple majority or plurality elections. If those voters didn’t care about the other candidates, their outcome is the same either way. However, RCV gives voters the opportunity to move their vote to another candidate if their preferred candidate is eliminated, which is an advantage over simple majority and plurality voting.

Argument: In the 2011 mayoral race in Portland, where ranked-choice voting was used, 15 candidates appeared on the ballot, voters had 15 choices, and almost 18 percent of the votes were exhausted before a winner was determined. Counter-argument: This means that 18 percent of the voters selected only candidates who were not going to win anyway in a plurality system and didn’t know or care about the remaining candidates. For those 18 percent of voters, the outcome would have been the same in a simple majority or plurality system, which is their preferred candidates didn’t win. That being said, having 15 choices to fill in order of preference is a lot, and in a simple majority system it’s likely that there would be a run-off election with many of the candidates eliminated, which would be equivalent to voters learning about and including more candidates on their ballot, whereas in plurality voting with 15 choices it’s likely that a candidate wins the election without having the support of the majority of the voters. While RCV is better in this situation, election officials could decide to limit the number of candidates who appear on the ballot under any of the systems discussed here. For example, candidates can gather signatures on their petition to be included on the ballot, and for example only the top five candidates with the most signatures by the due date even appear on the ballot. Alternatively, a plurality voting election is held first, like a public primary, and the top five or six candidates then compete in an RCV election. To be clear, this is not part of the RCV proposal and for a primary convention, where all the votes are tallied locally, having 15 candidates with any number of spaces works just fine and it would even work with write-ins.

Argument: African Americans, Latinos, voters with less education, and those whose first language is not English are more likely to be disenfranchised with a ranked-choice voting system because of using the ballot incorrectly or not making all available choices and having their ballots exhausted. Counter-argument: African Americans and Latinos are perfectly capable of filling out an RCV ballot, so like any other sphere of life, if there’s evidence that a particular group has a disadvantage in an area where other groups are succeeding, efforts should be made to help. In the case of voters whose first language is not English, the voter guide can include a link to instructions and candidate positions and statements in the most popular languages. Either way, making a single choice and not filling out all the spaces is up to the voter, and if the voter had the opportunity to vote, then the voter was not disenfranchised.

Argument: A candidate in Maine in 2018 prevailed in a ranked-choice election without receiving a true majority of the votes cast. Citation of the 2nd congressional district election between candidates Tiffany Bond, Jared Golden, William Hoar, and Bruce Pollquin. Counter-argument: This is equivalent to a simple majority vote where none of the candidates get a majority, the two candidates with the fewest votes are eliminated, and a run-off election is held between the two candidates who got the most votes in the first round. Some voters may not participate in the run-off election and in that case, using the same numbers as the result of the Maine election cited, it can also be said that the winner of the run-off election didn’t get a true majority when the people who voted in the first round but not the run-off are included in the calculation. In a plurality voting system, the winner isn’t required to have a true majority. The exhausted ballots in RCV are equivalent to votes cast for people who either don’t make the run-off in simple majority voting, or have the least support in plurality voting and wouldn’t win anyway.

Argument: The primary-runoff system is better because voters aren’t disenfranchised by exhausted ballots. In the primary-runoff system, each voter casts one vote for their preferred candidate, and then the candidates with the top two totals compete in a run-off election. Counter-argument: A voter who wants one of the candidates to win, but whose candidate didn’t get the most or second-most vote totals in the first round of a simple majority or plurality system, is then forced to either vote in the run-off for one of two candidates the voter doesn’t want at all, or opt-out of voting in the run-off. These situations are not better than exhausted ballots. It’s not reasonable to expect RCV to achieve a “true majority” of all the original voters in the first round, while accepting that simple majority voting with run-off may also select winners who don’t have a “true majority” in the first round, and that plurality voting usually does not result in a winner who got the majority of votes. A candidate who wins a simple majority vote may still not have the support of a “true majority” of the population because not everyone participates in the election.

Argument: RCV exhibits non-monotonicity, meaning that more first-place votes can hurt, rather than help, a candidate. An example that has been used is an election with 3 candidates for 1 seat, where 100 ballots are cast and there are no exhausted ballots, so 51 votes are required to win. If ballots are 39 for ABC, 35 for BCA, and 26 for CAB, then candidate C is eliminated and those 26 votes are moved to candidate A, who then wins the election with 39+26=65 votes. However, if in an alternate universe 10 of the BCA votes are instead ABC votes, then the ballots are 49 for ABC, 25 for BCA, and 26 for CAB. In this scenario, candidate B is eliminated and those 25 votes move to candidate C, who then wins the election with 26+25=51 votes. This is non-monotonic because 10 more first-choice votes for A, who was already leading, caused A to lose the election. Counter-argument: This scenario is presented as if the move of 10 votes to A caused A to lose the election, but this is misleading. The loss of 10 votes for B caused B to be eliminated, so A faced off against C in the final round, and C had more popular support with 26 first choice votes and 25 second choice votes. The result would have been the same if the 10 votes from B didn’t move to A at all but just decided not to vote. With 90 votes, the threshold would be 46, and the results would be 39 for A, and 51 for C. This makes it clear that the 10 votes moving to A didn’t cause A to lose, because the result was the same when 10 B voters stayed home. What actually happened is that B lost support, and because of that B got eliminated, causing A to face a different candidate in the final round, and C had more popular support than A. In fact, the B voters in the scenario who moved to A actually had A as their last choice, which doesn’t make sense. If the B voters were BAC instead of BCA, meaning if A had more popular support as a second choice among people who preferred B, then when B got eliminated candidate A would have received those 25 second-choice votes and won the election by a landslide with 49+25=74 votes (if gaining the 10 first choice votes) or 39+25=64 votes (if the 10 voters stayed home). Even without RCV, the same phenomenon is present in party primaries. A candidate can be very popular within their party and get the nomination but lose the election if they face off against a candidate with broader appeal to the general population. People call this concept “electability”, which means a candidate’s ability to win 51% of the vote in the election is more important than their ability to win 90% of the vote in their party’s primary.

Argument: RCV promotes the duopoly by taking away a third party’s influence on the two major parties. This is because when people can vote their preferred candidate as first choice, and they know their preferred candidate will lose, they will put one of the major party candidates as their second choice. Confident of getting those second-choice votes, a major party candidate doesn’t need to cater to people who would be third party voters. Counter-argument: If the major party candidate wants to receive those second-choice votes, the candidate still needs to appeal to those voters. Otherwise they may put the other major party candidate as their second-choice vote, or not fill in a second choice vote, and the major party candidate who doesn’t bother with them may lose because of that.

• This topic was modified 3 weeks, 4 days ago by Jonathan Buhacoff.
• This topic was modified 3 weeks, 4 days ago by Jonathan Buhacoff.