[EM] Science Magazine, CVD IRV letter

[Bart Ingles]

DEMOREP1 at aol.com wrote:
> 
> D- The CVD folks sent the below to Science Magazine which apparently had a
> story about voting in May 2001.  [...]
>
> I have NOT seen the original Science magazine material.

You also evidently haven't seen Brams and response, appearing on the
next page.

I don't have the original op-ed that prompted the exchange, but you can
probably find it at most libraries.  Following are the uncut versions of
the later exchange.

----------------------------------------------------

  The Science of Elections - A Rejoinder.

  Steven Brams and Dudley Herschbach are right about the
  defects in the plurality voting system used in most American
  elections (Editorial "The Science of Elections", vol.292, 25
  May 2001, page 1449). But, on both theoretical and practical
  grounds they are wrong to tout approval voting over instant
  runoff voting (IRV).

  Used for decades in Australia and Ireland and considered in 13
  state legislatures this year, IRV lets voters rank candidates in
  preference order. A voter's best strategy is to sincerely rank the
  candidates. If no candidate gets a majority of first preferences,
  candidates at the bottom are sequentially dropped. Each ballot
  cast for those eliminated candidates is added to the totals of the
  next choice indicated on that ballot until a candidate achieves a
  majority. IRV duplicates a series of traditional runoffs, but
  without the need for additional elections that cost taxpayers and
  candidates more money and often lead to falloffs in voter
  participation.

  In contrast, approval voting is a binary system, where the voter
  can only indicate "yes" or "no" for each candidate. The
  problem is that voters rarely have binary views about a range
  of candidates. Let's assume a voter sees Z as most favored, Y
  as less favored, and X as unacceptable. By voting for
  "acceptable" candidates Y and Z the voter could cause Z to
  lose. But by voting only for Z, the voter makes it easier for
  unacceptable X to win. The voter will be torn between voting
  defensively against X or strategically for Z.

  Approval voting has another important real world flaw.
  Political behavior has much to do with what is rewarded by the
  election system, and approval voting would exacerbate one of
  the worst aspects of American campaigns: avoidance of
  substantive policy debate.

  Because a candidate could lose despite being the first choice of
  an absolute majority of the electorate, smart candidates would
  avoid controversial issues that alienate any significant number
  of voters. Smiling more and using policy-empty themes like "I
  care" will not clarify the important choices leaders must make.
  Those rewarded by approval voting should be characterized as
  "inoffensive" more than "centrist."

  IRV strikes a better balance. It rewards candidates who stand
  out on policy enough to gain first-choice support, yet
  encourages coalition-building and fewer personal attacks as
  candidates seek to be the second-choice of other candidates'
  supporters.

  These arguments help explain why IRV is used and proposed
  far more often than approval voting, including in upcoming
  ballot measures in Alaska and San Francisco that would
  implement IRV for their major elections (see
  <www.fairvote.org/irv> for details). The American Political
  Science Association even uses IRV to elect its president. It is
  the right system for America's high stakes elections with a
  single winner.

  (Rob Richie is the executive director of the Center for Voting
  and Democracy. Terrill Bouricius is the Center's New England
  regional director. Philip Macklin is a professor emeritus of
  physics at Miami University in Oxford, Ohio.)

  Rob Richie            
  The Center for Voting and Democracy
  6930 Carroll Avenue, Suite 901
  Takoma Park, MD 20912
  (301) 270-4616
  rr at fairvote.or

  Terrill Bouricius
  56 Booth Street
  Burlington, VT 05401
  (802) 864-8382
  terryb at fairvote.org

  Philip Macklin
  211 Oakhill Drive
  Oxford, OH 45056
  (513) 523-5644
  macklipa at muohio.edu

-----------------------------------------------------------
RESPONSE

IRV is a special case of a voting system proposed by Thomas Hare
of England (and others) 150 years ago.  It sounds attractive but,
when compared with approval voting (AV), has some decidedly
unappealing features, including:

* its propensity to lose majority candidates, especially centrists, who
may do poorly when challenged from both the left and right.  Even when
there are only three candidates, it is not uncommon that the centrist
comes in third, which means that he or she loses under IRV.  By
contrast, AV tends to help such candidates, because they draw approval
from their opponents' supporters on both the left and right, who want to
avoid at all costs helping the candidate on the opposite side of the
political spectrum.

* its nonmonotonicity, which means that raising a candidate in one's
ranking can cause him or her to lose.  This can occur because of the
way in which candidates are sequentially dropped and their votes
transferred to those who remain in the race.  This perverse property
of IRV was discovered only about 30 years ago.  It is antithetical to
the very notion of democracy, in which expressing a stronger preference
for a candidate should help rather than hurt that person.  By contrast,
expressing approval for additional candidates under AV can never hurt
them and may help them.

* its complexity, which even mathematicians have not fully
understood, as witnessed by misstatements they have made about
the Hare system.  It is noteworthy that the American Mathematical
Society, after long debate, abandoned the Hare system for AV.  In
fact, none of the eight professional societies that have adopted AV
over the last 15 years has reconsidered its decision and chosen a
different voting system.

It is true that AV is a binary system, but not with respect to where
voters draw the line between acceptable and unacceptable
candidates.  Thus, if there are five candidates, a voter might
reasonably approve of one, two, three, or four out of the five
candidates.  The voter is sovereign in deciding who is worthy of
approval, whereas IRV forces voters to make a strict ranking, which
may be asking too much for those who do not know a great deal
about the candidates but do know who is basically acceptable and
who is not. 

It is also true that AV may not always elect the first choice of a
majority of voters.  But that result, surprisingly, is sometimes
desirable.  If, for example, 50 voters rank three candidates XYZ (in
that order) and 49 voters rank them YZX, AV will elect Y if the 50
XYZ voters approve of both X and Y, and the 49 YZX voters
approve of either Y or both Y and Z.  Is not Y the better social
choice, compared with the IRV winner, X, whom nearly half the
voters consider the worst choice? 

Our critics make two false claims.  A sincere ranking under IRV is
not always optimal-a voter can sometimes ensure the election of a
preferred candidate by not being sincere.  The American Political
Science Association (APSA) does not use IRV.  To the
embarrassment of one of us (Brams), a political scientist and a
member of the APSA, the APSA does not have competitive elections
for any of its offices.   

We think their charge that AV would force all candidates toward a
lowest-common-denominator position of blandness is erroneous.  In
a detailed study of the 1980 presidential election, which had a
significant third-party candidate (John Anderson), Peter Fishburn
and one of us (Brams) showed that Ronald Reagan would have won
under AV, based on both election and poll data (1).  We strongly
doubt that AV would have compromised Reagan's strong convictions
or his campaign behavior-or affected the outcome.  Indeed, trying to
be everything to everybody is likely to make a candidate not even
minimally acceptable to many voters and, therefore, not a smart
campaign strategy under AV.

Our critics point to the serious interest in IRV.  We would point to
the failure of the Hare system, after being adopted in several large
U.S. cities like New York about 50 years ago, to stand the test of
time.  The last city still to use the system in the United States is
Cambridge, MA.

Serious analysis of AV began only about 20 years ago.  Since then
AV has gained many adherents both inside and outside the scientific
community.  Both its compelling theoretical properties and its
simplicity commend it for practical use, which cannot be said for
IRV in those jurisdictions that do not already have electronic voting
equipment that would permit voters to rank candidates.  

Steven J. Brams1*. Dudley Herschbach2
1Department of Politics, New York University, New York, NY 
10003, USA.
2Department of Chemistry, Harvard University, Cambridge, MA 
02138, USA.
*To whom correspondence should be addressed.
 E-mail:  steven.brams at nyu.edu

References and Notes
1.  S. J. Brams, P. C. Fishburn, Approval Voting (Birkhäuser,
Boston, 1983).