National Civic
Review
Instant Runoffs: A Cheaper, Fairer, Better
Way to Conduct Elections
By Rob Richie and Caleb Kleppner, Center for Voting
and Democracy, and Terrill Bouricius, Vermont state legislator
Introduction
In 1997, Democrats in New York City faced an uphill challenge in
seeking to oust incumbent mayor Rudolph Giuliani. Their leading
contender, Manhattan Borough President Ruth Messinger, finished a
clear first in the September primary, but fell short of a popular
majority. The open question was whether she won less than 40% of the
vote, which would trigger a citywide runoff election against Rev. Al
Sharpton, the second-place finisher.
Messinger's vote total was so close to 40% that delays in
tabulating ballots forced the city and the candidates to prepare for
the possible runoff. The city spent some $2 million, while Messinger
was delayed in turning her attention to uniting Democrats for the
general election. She ultimately avoided the runoff, but the
uncertainty contributed to her general election defeat, in which
voter turnout among black voters -- Sharpton's political base -- was
particularly low.
Meanwhile, in San Leandro (Calif.), Sheila Young in 1996 won a
four-person city council race with only 39% of the vote, far less
than a majority. Two years later, also in a one-round, non-partisan
election, she was elected mayor with 44%, again less than a
majority. In the same election, Glenda Nardine won a city council
seat with 37%. These and other plurality wins led to calls for a
majority requirement with a second-round runoff. The idea of
runoffs, however, prompted concern about making taxpayers pay for
and participate in two elections. Skeptics also pointed to
remarkably long campaign season associated with many local runoff
elections in California. In 2000, many jurisdictions will have the
first round election in early March and a runoff fully eight months
later if no candidate wins a majority.
Despite the political costs created by the runoff law in its 1997
mayoral primary, New York City Democrats have good reason to ensure
nominees can obtain a reasonable threshold of support -- a primary
winner with less than 25%, as happened in a key congressional race
in 1998, might well be a weak standard-bearer. In San Leandro,
opponents of runoffs have reason to be wary of the downsides of
two-round runoff elections.
Fortunately, there is a simple reform that resolves these
dilemmas and promises to improve political participation,
representation and governance. Instant runoff voting lowers the
campaign costs for candidates, saves taxpayers money, boosts voter
turnout, reduces negative campaigning and ensures majority rule.
Because of these advantages, voters in Vancouver (Wash.) and Santa
Clara County (Calif.) recently passed charter amendments to allow
the use of instant runoffs. Legislation to enact the instant runoff
is under consideration in a rapidly growing number of states and
cities.
The Return of the Voter?
Many Americans bemoan low voter turnout, especially in local
elections. It is disturbing enough that most eligible Americans
abstained from voting in the 1996 presidential election, and turnout
in local elections is usually much lower. Turnout among registered
voters fell well below 10% in major Texas cities like Dallas and El
Paso in May 1999 mayoral elections.
No single reform will bring most non-voters back to the polls,
but certainly we must address their frustration and growing cynicism
with campaigns that too often skirt issues and devolve into personal
attacks. Attitudes such as, �All elected officials are liars� and
�What�s the point of voting?� are disturbingly common today.
Reformers propose many remedies to low turnout: same-day voter
registration, vote-by-mail, internet voting, weekend or holiday
voting, increased access for candidates to the media, spending and
contribution limits, public financing of campaigns, codes of conduct
for candidates, tax credits for voting and compulsory voting. At
least some of these proposals deserve consideration, but none
address an important underlying cause of our dwindling
participation: the voting system.
By "voting system", we mean the method of translating votes for
candidates into winners. Most American elections use a plurality
voting system. This means that the candidate with the most votes
wins, even if the candidate receives less than a majority (50% plus
one) of votes cast. Another widely used voting system, particularly
in party primaries and municipal government, is the two-round
runoff. In a runoff election, the two highest vote-getters advance
to a second round runoff if no candidate receives a majority (or
established threshold, such as 40%) in the first round.
Plurality voting systems and runoff elections both contribute to
such problems as costly elections, low voter turnout, negative
campaigning and simplistic debate. The instant runoff is gaining
support across the country because it addresses such problems and
complements other proposed reforms.
Defects in Plurality Voting
When the United States was founded, there was little experience
with alternative voting methods. The only recognized alternative to
plurality voting's failure to ensure majority rule was holding
repeated elections, in hopes that a natural culling of candidates or
modification of voter sentiments would lead to a majority
winner.
Eventually most states settled on one of the two available
pragmatic "solutions" -- one-round elections with plurality rules,
often combined with laws making it difficult for candidates to gain
access to the ballot, or two-round elections with a runoff between
the two top vote-getters.
That we already use more than one voting system suggests that no
system is perfect. In fact, economist Kenneth Arrow won a Nobel
Prize for proving that there can never be a perfect voting system --
they all inevitably have a weakness, and not all jurisdictions have
the same goals for elections. The instant runoff is gaining support
because it fixes important defects in our most popular voting
systems - plurality elections and two-round runoff elections.
Plurality voting suffers from three principle defects:
� It does not provide for
majority rule, allowing candidates to win with a narrow band of
support.
� It can create
opportunities for "spoiler" candidacies.
� It promotes negative
campaigning.
Violates majority rule: By definition, a candidate in a
plurality election can be elected with less than a majority of the
vote. As a result, it is quite possible that most voters dislike the
winner who "represents" them. At the very least, minority-vote
winners have a weaker mandate to govern.
In the 1992 presidential election, for example, the winner of all
the electoral votes in 49 out of 50 states was opposed by a majority
of that state's voters. George Bush failed to win a majority of the
popular vote in any state, while Bill Clinton only won a majority in
Arkansas. Senate Republicans pointed to Clinton's 43% of the popular
vote as a rationale for their frequent filibusters with 43% of
Senate seats.
In fact, more than a fifth of our current governors have been
elected by a plurality vote in one of their primaries or general
elections, including several who won with less than 40%. In 1997,
Jim Baca was elected mayor of Albuquerque (New Mexico) with 29% --
meaning 71% of voters supported a different candidate.
An observer of one-round, non-partisan elections for the troubled
Washington, D.C. school board noted that some winners simply have
found the angriest group in their district, pitched their message to
them and won with some 25% to 35% of the vote -- barely even
acknowledging the vast majority of their future constituents in
their campaign. If winning instead required 50%, a successful
candidate would have to reach out and win the support of more
constituencies -- a particularly valuable demand in racially and
ethnically diverse districts.
Winning office with less than majority support is not necessarily
wrong for all elections -- to elect legislative bodies, we support
proportional representation, which ensures both fair representation
to the majority and the minority -- but it is unjust for elections
where the majority can lose the right to decide, let alone be
completely shut out.
Permits "spoiler" candidacies: Plurality elections also
suffer from the �spoiler� phenomenon in races with more than two
candidates. Someone with no chance of getting elected can gain
enough votes to swing the race between the two leading
candidates.
By gaining the power to determine the winner, unscrupulous
candidates can gain leverage over major candidates. At the same
time, many minor candidates genuinely seek to raise important
issues. Their supporters must make a tough decision: to vote for
their favorite candidate, knowing that the candidate won�t win and
might even throw the race to the supporters� least preferred
candidate, or to settle on a less preferred candidate who has a
chance to win. In other words, voters must accurately judge not only
which candidate they prefer, but whether that candidate has a chance
of winning.
This need for calculation illustrates a profoundly undemocratic
feature of an outwardly democratic and simple system. Yes, it is
simple for everyone to cast one vote, and have the candidate with
the most votes win. But it is hardly simple or democratic for some
worthy candidates to have a disincentive to participate and for
their supporters to have a disincentive to vote for them. With the
rise in support for third party and independent candidates today and
with bigger fields in many non-partisan elections, plurality
voting's anti-participation bias is an increasing problem.
Promotes negative campaigning: Plurality elections lead to
negative campaigns that, with ever-increasing sophistication about
how to win office in today's multi-media world, drive far more
voters away from politics than bring them to it. Knowing that voters
will ignore minor candidates because they can't win, the two leading
candidates can focus their attention on each other. Turning a voter
away from an opponent is just as effective -- and often considerably
easier and cheaper -- than persuading an undecided voter to support
you. Plurality voting creates direct incentives for negative
campaigning and personal attacks; candidates generally will only
eschew negative campaigning if they think they can win without
it.
Defects in Two-Round Runoffs
Two-round runoff elections were developed in response to
non-majority victories in plurality voting. In addition, they allow
voters to scrutinize the final two candidates more thoroughly during
the period between the two rounds of election. Runoffs first came
into wide use at the turn of the century, with the advent of party
primaries, because primary elections in races without an incumbent
frequently had more than just two candidates vying for a particular
office. Seventeen states have used runoffs in federal primaries at
some time. A dozen states, mostly in the south, use runoffs today
for primaries.
Runoffs also are frequently used in non-partisan elections
because, like primaries, they tend to draw more than two candidates.
A 1986 survey of American cities with populations above 25,000
revealed that most of the 946 cities contacted had a runoff
provision.
Although successful in guaranteeing that winners receive a
majority of the second round vote, runoff elections create new
problems and fail to correct all of plurality voting's defects.
Problems with runoff elections include:
� Candidates must raise money for a
second campaign, often with little time to raise funds.
� Taxpayers must pay for
administration of two elections.
� Voter turnout can be
significantly lower in one election round because voters must pay
attention to two campaigns and go to the polls twice.
� Runoffs do not correct
plurality voting's defects; they still can thwart majority rule,
create opportunities for �spoiler� candidacies and promote
negative campaigning.
Exacerbates problems with money in politics: Campaign costs
at all levels of election continue to mount. Runoffs create a
particular burden for candidates, who usually have to raise money
for a runoff election on a very short timeline. Candidates in the
first round often spend nearly all their campaign funds in seeking
to avoid a runoff or to fight their way into one. They then can have
as little as two weeks to mount a second campaign.
Runoff elections thus increase opportunities for donors and
potential donors to influence candidates. At the same time, many
campaign contributors grow weary of the increased demands created by
runoffs.
Adds costs to taxpayers: Elections are not cheap. Setting up
polling places, training and paying poll workers, printing ballots
and voter guides and counting ballots all cost tax dollars. Election
administration typically costs $1-to-$2 per resident and can cost
much more in small and special elections. A citywide election in San
Francisco costs some $1 million. The next runoff in New York City,
likely to occur in the 2001 citywide elections, may cost more than
$10 million.
In addition, more jurisdictions are having trouble finding
pollworkers. Doubling their burden with runoffs makes this task all
the harder. In 1996 in Texas, some jurisdictions had to open polls
eight times: for city elections with runoffs, for federal primaries
with runoffs, for school board elections with runoffs and for three
U.S. House elections that required general election runoffs.
Discrepancies in voter turnout : Voter turnout often varies
widely between the two elections, particularly when either the first
or second round coincides with elections that draw more interest. In
the 1992 U.S. Senate race in Georgia, for example, incumbent Wyche
Fowler won 1,108,416 votes in the November election, but fell just
short of a majority. In the December runoff, without the draw of the
presidential race, Paul Coverdell beat Fowler with 635,114 votes --
a majority of the runoff vote, but nearly half a million votes
fewer than Fowler's vote total a month before.
In California, many jurisdictions will have a first-round in
March 2000, when turnout likely will be relatively low. A top
finisher with 49.9% will have eight more months of campaigning and
face a runoff in the November election with much higher turnout. A
candidate with 50.1% in the March round will be elected, although
must wait nearly a year to take office.
In both cases, a decisive election can take place when far fewer
voters participate than when the decisive election might have
occurred. The whole concept of runoffs producing majority rule is
brought into question by such contradictions.
Finally, the more times that citizens are asked to vote, the less
likely they are to vote in any given election. Many political
scientists attribute low voter turnout in the United States to our
unusually high number of elections. Runoffs only increase that
burden.
Fails to correct plurality voting's defects: At the same
time, runoff elections fail to correct any of plurality voting's
major defects. Runoffs still can thwart majority rule, create
opportunities for �spoiler� candidacies and promote negative
campaigning.
Voter turnout discrepancies raise serious questions about
majority rule when a candidate can win in the second round with far
fewer votes than the loser won in the first round. Furthermore, if
there are several candidates and no clear frontrunner -- as is often
the case when no incumbent is running -- the combined vote total of
the two candidates who advance to the runoff can be less than a
majority of the first-round vote. In Salt Lake City's mayoral race
in 1999, the two candidates making the runoff gained only 42% of the
first-round vote -- the second-place finisher made the runoff with
19.5%, while the third and fourth place finishers were eliminated
with 19.4% and 18%, respectively. Runoff participants had a combined
vote under 50% in two of five city council races in Vancouver,
Washington in 1999 -- results which contributed to the city's
passage of an amendment authorizing instant runoffs in November.
Runoff elections also still can suffer from �spoilers.� There are
no spoilers in the head-to-head runoff election, but in the first
round, voters still must decide whether to support their favorite
candidate or a candidate with a better chance of getting into the
runoff.
Finally, just as with plurality elections, runoff elections
promote negative campaigning, particularly in the runoff itself. In
a head-to-head runoff, campaigning can become extremely negative
because it is a zero-sum game -- "I win when you lose." In a
statewide runoff in Louisiana in November 1999 for insurance
commissioner, the incumbent overcame a recent indictment with a
slashing attack against his opponent's integrity; only 27% of
registered voters ultimately participated in the runoff. Having more
than two candidates can moderate negative campaigning in the first
round, at least, but it still is an "all-or-nothing" vote.
Candidates know that they won't benefit from reaching out to voters
who are sure to support other candidates.
How the Instant Runoff Works
The instant runoff corrects the defects found in plurality and
runoff elections. Although new to many Americans, it is recommended
by Robert's Rules of Order and is a proven system that has been used
for more than a century in this country and elsewhere.
The instant runoff works much like a traditional runoff, but with
significant improvements. In a traditional runoff, each voter casts
one vote in the first round. If a candidate gains a majority, that
candidate wins. If not, the top two candidates advance to the second
round. In that second round, those who voted for one of the
advancing candidates in the first round likely will continue to
support that candidate, but the remaining voters must choose between
the two advancing candidates.
The instant runoff simulates a traditional runoff exactly, with
the added refinement of eliminating only one candidate at a time
rather than all but the top two candidates. (As detailed in Robert's
Rules of Order, this gradual elimination avoids the problem of the
top two candidates having less than a majority of the first-round
vote.) The instant runoff can occur in a single election because its
inventor had a simple realization: that voters can indicate which
candidate they would support in a runoff without having to return to
the polls.
Rather than vote for just one person, then, voters in the instant
runoff can rank candidates in order of choice: a first choice for
their favorite candidate, a second choice for their next favorite
and so on. Ranking candidates indicates which candidates voters
would support if their top choice were defeated. Doing so is
literally as easy as "1, 2, 3," and voters have shown a quick
capacity to use the instant runoff wherever it has been adopted.
Ballots are counted like a series of runoffs. If a candidate wins
a majority of first choices, that candidate wins. If there is no
initial majority winner, the last-place finisher is eliminated, and
a second round of counting takes place. In each round, ballots count
as one vote for the top-ranked candidate who has not been
eliminated. In other words, your vote counts for your first choice
candidate as long as that candidate remains in the running. If your
first choice candidate is eliminated, then your vote counts for your
next-choice candidate, and so on. The election is over when one
candidate wins by gaining a majority of votes.
Ranking more than one candidate is optional. Voters set in their
ways or unfamiliar with more than one candidate are free to vote for
a single candidate just as they do in plurality voting. In fact,
most voters usually will rank a candidate first who is among the
final two candidates, meaning that their ballot will count for their
top choice throughout the ballot-count.
Even so, a voter has every incentive to rank more than one
candidate because ranking additional candidates never harms a
voter's top-ranked candidate. If there is any chance that a voter's
top choice candidate will be eliminated in the course of the count
and if that voter has any preference among the remaining candidates,
that voter should rank his or her next favorite candidate -- just as
they might still return to the polls for the second round of a
runoff election after their preferred choice lost in the first
round. Ranking other candidates increases the electoral chances of
voters' next-choice candidates in case their first choice loses.
Note that voters usually are allowed to rank all candidates in a
race, but to simplify ballot design and ease the administration of
the election, the number of permitted rankings can be limited with
minimal impact on the performance of the system.
How Instant Runoffs Fix the Defects of Plurality and Runoff
Elections
The instant runoff addresses each of the defects found in
plurality voting and traditional runoffs. In contrast to two-round
runoffs, the instant runoff only requires one election to produce a
majority winner. Doing so cuts the cost of campaigns nearly in half
and saves taxpayers the cost of administering a second election. By
scheduling the election when voter turnout is highest or when other
state or local elections take place, turnout is maximized for the
decisive election.
In contrast to plurality voting, the instant runoff provides that
winners have majority support and eliminates fears of "spoiling."
Even if a majority of voters split votes between two or more
candidates, the candidate with the most overall support will win in
an instant runoff.
Eliminating fears of "spoiling" and maximizing the number of
people who influence the outcome of elections could significantly
improve our electoral process.
� Voters would probably have more
choices, but more importantly, they would be more likely to like
their choices on election day.
� With the need for winners to
appeal to other candidates' supporters, there would likely be more
acknowledgment of other people's concerns, more coalition-building
and more positive, issue-oriented campaigns.
� Opportunities for good
governance would increase. Winners would take office after more
inclusive, positive campaigns. They also would often have a
clearer mandate due to receiving majority support and learning
more about the issue priorities of supporters. Such a mandate
could establish greater accountability.
The combination of better choices, less money in politics,
clearer mandates and less negative campaigning could lead to higher
voter turnout and increased overall participation in
politics.
Why Aren't Instant Runoffs More Widely Used?
The instant runoff in fact is widely used. For decades, millions
of people in Australia and the Republic of Ireland have used instant
runoffs for their most important elections, and Malta has used a
similar candidate-ranking system in multi-seat districts.
Instructively, Malta and Australia have had the highest voter
turnout in the word in the 1990's.
More nations are moving to the instant runoff. It was used in
Fiji for parliamentary elections this year and recently was adopted
for Bosnia's next presidential election; in both cases, promoting
coalitions among ethnic groups was seen as crucial. The instant
runoff also will be used in London's first-ever election for mayor
in May 2000 because of interest in producing a majority winner in
one election.
In the United States, the American Political Science Association
is among many private American organizations using the instant
runoff. No governmental jurisdictions currently use it, but it has
been used in the past. Starting in 1912, four states -- Florida,
Indiana, Maryland and Minnesota -- adopted versions of the instant
runoff for some party primaries. Seven other states used another
candidate-ranking system known as "the Bucklin system" for important
offices, including gubernatorial races. In Bucklin, voters ranked
only two candidates. If the first-count did not produce a majority
winner, all second-choice ballots were counted
simultaneously. Bucklin was found to be defective as it sometimes
resulted in a voter�s second-choice vote helping to defeat that
voter�s first-choice. As a result, most voters learned to refrain
from indicating second choices, which thwarted the goal of
discovering which candidate was favored by a majority of voters.
By the 1930's all of these ranked-ballot systems had been
eclipsed, generally by two-round runoffs that were seen as easier to
administer. In the years since, some cities used the system,
including Hopkins (Minn.) for mayoral elections and New York City
for electing Staten Island's city councilor at the same time that
the rest of the city used a proportional system also based on
voters' ranking city council candidates.
The most recent use of the instant runoff was in Ann Arbor
(Michigan) in its 1975 mayoral race. A third party, the Human Rights
Party, had created lively, three-way elections that caused concerns
about splintering the vote. After adopting the instant runoff in a
1974 initiative, it was used for the first time a few months later.
The Democratic nominee, the city's first-ever black mayor, won a
cliffhanger on the strength of being the second choice of nearly all
supporters of the Human Rights Party candidate; he trailed the
Republican incumbent 49% to 40% after the first count, but
ultimately won a majority.
Republicans, who had benefitted from split liberal votes under
plurality rules, sought to eliminate the system. A legal challenge
failed as the U.S. Circuit Court upheld the constitutionality of
instant runoffs, but a repeal succeeded in a low turnout special
election in 1976. Although tinged with racial undertones, the repeal
effort focused on the difficulty of counting ballots by hand -- and
indeed there had been problems given the short time for election
administrators to prepare.
Election administration almost certainly is the major reason why
instant runoffs haven�t been more widely used in the United States.
Most jurisdictions for years have used automatic voting equipment
that was incompatible with instant runoffs. Only recently, in 1997,
did Cambridge, Massachusetts become the first city to use modern
ballot-scanning voting machines with the type of rank-order ballots
used in instant runoff elections. Implementing instant runoffs in
the past would have required jurisdictions to switch to counting
ballots by hand, which few jurisdictions were willing to do. In
contrast, nations like Ireland and Australia that use instant
runoffs have always counted ballots by hand.
Plurality and runoff elections also generally escape scrutiny.
Most people aren�t aware that alternative voting systems even exist;
they rarely are taught in school, and few Americans follow the
details of other nations' elections. Some might notice that winners
of athletic awards are often elected by journalists who rank the
contenders, but even these elections are misleading. They are
weighted systems, where a first choice might mean five points, a
second-choice four points and so on -- systems that in public
elections would lead to most voters only ranking one candidate due
to not wanting support for a lesser-preferred candidate to count
against their top choice.
However, several developments are drawing attention to instant
runoffs. Major candidates are getting knocked out more frequently by
spoilers. Support for independent and third party candidates is
growing. Money is playing an increasing role in politics. The
downsides of costly, time-consuming runoffs are becoming better
understood; in the past two years, voters in two American
jurisdictions -- Santa Clara County (Calif.) and Vancouver (Wash.)
-- have passed stand-alone charter amendments to permit replacement
of two-round runoffs with instant runoffs.
Addressing Concerns
In spite of the instant runoff's obvious benefits, concerns are
natural. Although recognizing that no voting system is perfect, most
of these concerns are easily answered.
Instant runoffs are not too confusing: In November 1999, San
Francisco's Board of Supervisors came close to putting the instant
runoff on its March 2000 ballot. The proposal ultimately was sent
back to committee for further study because of some supervisors'
belief that the system was �too confusing.� (The postponement also
was closely tied to the fact that at the same time the legislation
was under consideration, the sponsor, Board president Tom Ammiano,
was waging a write-in campaign for mayor against incumbent Willie
Brown and winning a spot in a runoff election.)
The system in fact is simple for voters. They can choose to vote
as they do now or rank some number of candidates in order of choice:
1, 2, 3. Voters in Australia and the Republic of Ireland, where
turnout is far higher than in the United States, have used the
system for decades. If the system isn�t too confusing for them, why
should it be too confusing for Americans?
Indeed Americans have used rank-order systems in many parts of
the country with little trouble. Recently, high school students in
Vermont were surveyed about how they liked the system after
participating in mock instant runoff elections. More than nine out
of ten students said that instant runoffs were not too difficult.
Only 1% said it might make them less likely to vote after they turn
18, while 46% said it would make them more inclined to vote. At the
largest mock election, at St. Michael's College in 1998, 197
students voted in a mock instant runoff election without casting a
single invalid ballot.
In the recent debate in San Francisco, some argued that voters
are capable of casting equally-weighted votes on several candidates
in an at-large election for six seats, but would be unable to
rank-order more than one or two candidates. In fact, people make
decisions through ranking choices all the time: in choosing what
meal to order at a restaurant, what video to watch, what television
program to watch and so on. In elections, most people in elections
choose a candidate to support after an internal process of weighing
and ordering candidates. They mentally divide candidates into �worth
considering� and �not worth considering,� then further divide the
candidates into serious contenders for their vote and long-shots,
and finally settle on one, their top pick.
Consider asking a small child about her favorite ice cream.
Chocolate, she might way. And what if there is no chocolate, you
ask? Then she'll have strawberry. And if there's no strawberry,
she'll settle with vanilla. The child just ranked three candidates:
chocolate, strawberry, vanilla. That�s all there is to the instant
runoff.
Finally, note that in plurality and runoff elections, some voters
have to make complicated strategic decisions about whether to vote
for their favorite candidate or for a candidate with a better chance
of winning. In an instant runoff, voters are free from such
calculations, which in turn makes voting decisions easier and more
enjoyable.
Instant runoffs do not give extra votes to supporters of losing
candidates: In the instant runoff, most voters typically only
have their vote count for their first choice candidate because their
candidate won't be eliminated before the field narrows to two. Given
that other voters will have their second-choice or subsequent
choices count, some allege that supporters of eliminated candidates
get more votes than supporters of more popular candidates.
But of course every voter only has one vote count in any given
round. Just as all voters get one vote in both rounds of a
traditional runoff, everyone's ballot counts as one vote in each
round of an instant runoff.
Some then rephrase the argument: only supporters of losing
candidates get a chance to switch their vote to another choice. But
surely these voters would rather their vote remain with their first
choice than have to count for a lesser-preferred candidate. Just
like in a traditional runoff, only supporters of losing candidates
must switch their votes, while supporters of more popular candidates
can continue to support their top choice.
Elimination of candidates in the count is not arbitrary: Some
express concern that the process of eliminating candidates is
arbitrary and can change the results. But eliminating candidates in
order of their strength of support is a sensible standard consistent
with our political traditions.
Of course most of the candidates who are eliminated during the
course of an instant runoff are too weak to have any chance of
winning the election. However, it is possible that a third-place
finisher in a three-candidate race could be the second choice of
most supporters of the other two candidates and thus could have been
able to defeat either of the other two candidates in a hypothetical
one-on-one race.
But such a candidate of course also would lose in a plurality
election or a two-round runoff, and indeed such results occur in our
system -- some believe that Ross Perot could have won the 1992
presidential race if matched against either George Bush or Bill
Clinton. The only reason we would even know of such a possible
result in an instant runoff election is because the system allows
voters to provide more information about their preferences than
possible in plurality voting or runoffs.
Every election system must reflect the priorities of its
designers. Plurality elections put a high premium on strong core
support. An alternative system might emphasize the importance of
having the broadest possible support even if the support were so
shallow that the candidate would be unable to ever win in a
plurality election.
The instant runoff is a compromise between these positions. As
with plurality voting and runoff elections, winners must have enough
core support to avoid early elimination. At the same time, in
contrast to plurality voting, winners must have the capacity to
reach out to supporters of other candidates to forge a real
majority. We believe that such a balance preserves important values
of our current system while creating conditions for cheaper, better
and fairer elections.
Modern voting equipment can accommodate ranked-ballots: Many
jurisdictions currently use voting equipment like lever machines and
punch-cards that is incompatible with instant runoffs. Administering
an instant runoff would require these jurisdictions to acquire new
voting equipment or use a time-consuming hand count. Current
legislation to enact instant runoffs in Vermont is founded on the
idea of handling the count of ballots beyond first choices in the
same way that a re-count would be handled, but some jurisdictions
would be unwilling to consider any counting of ballots by hand.
Fortunately, more modern voting equipment, such as optical
scanners already in wide use and computer touch screen equipment can
handle instant runoffs at no additional cost beyond a one-time
upgrade of software. Vote-by-mail elections, already the law in
Oregon and many localities, and of course internet voting make
ballot-counting much easier. Once jurisdictions modernize their
voting equipment for reasons of efficiency, speed and security, they
can adopt instant runoffs at no additional cost.
In the two localities where voters recently approved instant
runoff charter amendments, the amendments allow, but do not require,
the use of the system. In both places, existing voting equipment was
incompatible with ranked ballots, but now the jurisdiction is
empowered to use instant runoffs once the jurisdiction acquires
new equipment. For jurisdictions interested in exploring use of
instant runoffs, this is a wise approach if current equipment cannot
accommodate instant runoffs.
The instant runoff gives voters enough time to evaluate
candidates: Allowing voters additional time to scrutinize the
leading candidates is one of the strongest arguments for traditional
runoff elections. If voters don�t learn enough about all the
candidates in the first round, they can benefit from more time to
study the top two candidates.
Of course we don't have this extra time in most of our major
elections because runoffs are not used in presidential elections and
rarely to elect governors and members of Congress and state
legislators. And extra time doesn't necessarily change minds. The
winner of a runoff often has more to do with who bothers to vote a
second time rather than any change of support.
Runoffs of course bring their own problems. In places with a
November general election followed by a runoff, turnout generally
drops. The very dispersal of the decision -- with voters aware that
the first-round may not produce a winner -- can lead to decreased
attention by the media and electorate on the candidates. And the
extra time of campaigning may be devoted more to negative attacks
than informing the electorate about one's own positions.
In an instant runoff, on the other hand, voters have incentives
to learn about more of the candidates because of their chance to
rank them. All of the candidates have incentives to put their ideas
forward, and the media has reason to cover all the candidates.
Instant runoffs thus give voters the benefits of runoff elections --
additional scrutiny of the candidates -- without the costs.
Instant runoff can win at the polls: Some charter commissions
and elected officials might like the idea of the instant runoff, but
fear that it would lose at the polls. In fact, rank-order systems
have a strong history with voters, both here and abroad. Most
recently, voters in Vancouver (Wash.) in 1999 and Santa Clara County
(Calif.) in 1998 approved charter amendments to allow instant
runoffs; the leading newspapers and political forces in these
localities endorsed the idea.
Cambridge (Mass.) uses the proportional representation variant of
the instant runoff. The system survived several repeal attempts --
organized by those opposed to the fairer representation of diversity
provided by proportional representation -- and now is widely
accepted. Similarly, Ireland's proportional representation variant
of instant runoffs has survived two national referendums seeking its
repeal. There currently is no movement to change rank-order
balloting for presidential elections in Ireland or Australia's
parliament; polls indicate the system is very popular in both
nations.
Future Prospects
As we approach the end of the twentieth century, it seems to us
highly unlikely that Americans will maintain antiquated rules like
plurality voting very long in the coming century. Our society
evolves so quickly in nearly every area but our political
institutions. That can be a strength, but not when old rules like
plurality voting and two-round runoffs fit poorly with the modern
world.
Instant runoff voting has burst onto the political landscape in
this decade. In 1997, Texas became the first state to consider a
statute on instant runoffs. By 1998, legislation to enact instant
runoffs for statewide and federal offices had been introduced in New
Mexico and Vermont. In 1998, a charter commission in Santa Clara
County (Calif.) placed an amendment on the November 1998 ballot that
explicitly allowed instant runoffs to replace runoffs in future
county elections. The measure won after gaining the endorsement of
the San Jose Mercury News, Chamber of Commerce and other key
political players.
In 1999, legislation to enact instant runoffs for statewide and
federal offices passed the New Mexico state senate, again was
considered in Vermont and was introduced and seriously debated in
Alaska. There are real chances for adoption of instant runoff in
those states in 2000 given the following circumstances:
Vermont: In 1997, Vermont passed a public financing law for
statewide offices that likely will result in more third party and
independent candidates winning votes. The state constitution
requires certain statewide offices, including that of the governor,
to be elected by a majority, and the instant runoff was proposed as
a means to accommodate increased candidate participation.
In 1998, the Vermont house established a citizens' commission to
advise the legislature on instant runoffs. The eleven members of
commission, appointed by the League of Women Voters and Common
Cause, included Democratic and Republican former members of the
Vermont House of Representatives, independents and unaffiliated
citizens. The commission issued a unanimous report in January
1999 that recommended adoption of instant runoffs for all statewide
elections beginning in 2000. Since that report, the list of formal
endorsers of instant runoffs for state elections include the Vermont
chapters of the League of Women Voters, Common Cause, the American
Association of University Women, the Grange and the Older Women's
League.
New Mexico: In 1997 and 1998, special elections were held in
two of New Mexico's three U.S. House seats. Green party candidates
won more than 15% in each race, helping "spoil" the races for
narrowly-defeated Democrats. A constitutional amendment to enact
instant runoffs for state and federal elections gained influential
supporters both in and outside the state in the course of passing
the state senate.
Nationally, House Minority Leader Richard Gephardt and
Congressman Patrick Kennedy were among those who spoke to New Mexico
legislators to advocate the legislation. Lydia Camarillo, the
executive director of Southwest Voter's educational arm, travelled
to Santa Fe to urge passage of the legislation. In-state backers
included former governors Toney Anaya (a Democrat) and David Cargo
(a Republican) and the Democrats' 1998 gubernatorial candidate
Martin Chavez.
Alaska: Just as Democrats were a driving force for instant
runoffs in New Mexico, Republicans have promoted the reform in
Alaska. The Republicans have large majorities in both houses of the
legislature, but have not elected a governor in more than two
decades -- in large part due to splits among conservative
candidates. Leading Republicans such as the state's representative
on the Republican National Committee convinced nearly the entire
party leadership of the merits of instant runoffs. A 1999 bill to
enact instant runoffs for all state elections passed one house
committee, but in a reversal of roles from New Mexico, the
Democratic governor was seen as likely to veto the legislation for
partisan considerations. As a result, a well-backed initiative
campaign was launched; the measure may well be on the November 2000
ballot and still has a chance in the legislature in 2000.
Instant runoffs are also on the table in other state
legislatures, including North Carolina, where an election laws
commission was created in 1999 in part to specifically research the
possibility of replacing runoff elections for state primary
elections with instant runoffs -- a change particularly pertinent in
light of interest in public financing of primary elections. At a
city level, a charter commission in Austin (Texas) in December 1999
recommended inclusion of instant runoffs in its charter package; the
city council will decide whether to place the package on the ballot
in 2000. Other local campaigns for instant runoffs initiated either
by citizens or elected bodies are being discussed in Kalamazoo
(Mich.), New York, Oakland (Calif.), Pasadena (Calif.), Prince
George's County (Maryland), San Francisco (Calif.) and San Leandro
(Calif.).
These developments indicate that when public officials believe
that instant runoffs solve a problem -- whether it be overcoming
spoilers, meeting a majority requirement, shortening an overly-long
campaign season or accommodating new campaign finance reform laws --
their support for reform can crystallize rapidly. Republicans are
promoting the instant runoff in Alaska and Utah, Democrats in New
Mexico, independent progressives in San Francisco and the Chamber of
Commerce in San Leandro (Calif.). And whatever the impetus for
consideration of the instant runoff, its ultimate strength is that
it is a non-partisan, good government reform that benefits everyone
-- the voter, the taxpayer and candidates -- over time.
Note also that a growing number of private organizations and
political parties use instant runoffs for internal elections. These
groups include the American Political Science Association, the
American Psychological Association and the Academy of Motion
Pictures for Oscar nominations (using the proportional variant of
the system). The Reform Party plans to use the instant runoff to
nominate its 2000 presidential candidate in a national, vote-by-mail
primary which may draw national attention to the system.
Conclusion
Plurality voting, runoff elections and instant runoff elections
generally elect the same candidate: the one with the greatest
support. But both plurality and runoff elections are susceptible to
breakdowns generating undemocratic results: plurality voting when
more than two reasonably strong candidates run and runoff elections
when turnout discrepancies are large.
Instant runoff voting prevents such breakdowns of majority rule,
but perhaps more importantly, it is about improving our politics. In
an era of shrinking participation in elections and government, it is
critically important to look at a simple reform that would take
money out of politics, promote positive public debate, save tax
dollars and encourage winners to reach out to more of their
constituents. The rapid rise of interest and support for the instant
runoff suggests that Americans are ready for a change.
|