[EM] Rob Richie's criticism of Approval Voting
jan.kok.5y at gmail.com
Tue Jun 13 13:18:58 PDT 2006
I encourage people who want to know the details of what is going on
with Denver's election reform to subscribe to the irv-l list via
irvdenver.org . The traffic is about 2 or 3 posts per day.
The message attached below is Rob's reply to a nice post by Mahendra
Prasad, included below. (Thanks, Mahendra.) I will respond to Rob's
One bit of help I could use: Is anyone familiar with the article Rob
mentions in this paragraph?
> Jack Nagel, a UPenn professor, has an important new article out this year called "Burr's Dilemma" that goes into how this flaw played out in the 1796 and 1800 presidential election, where the presidential electors (who at that time had two equally weighted votes), made strategic mistakes with major consequences in both elections. Nagel used to say approval voting was better, but now says IRV is better.
Did they really use Approval Voting? It sounds more like cumulative
voting, or a restricted form of AV where the number of votes was
limited to 2 max. Could someone post a summary of the relevant
Please direct replies to this (my) post to the EM list.
---------- Forwarded message ----------
From: Rob Richie <rr at fairvote.org>
Date: Jun 11, 2006 10:30 AM
Subject: Re: [irv-l] I agree with Jan Kok's Concerns
To: irv-l at irvdenver.org
I'm hoping this list will be a practical list about focusing on
whether the coming year is the right year for IRV in Denver and if so,
focusing on how to achieve that goal. If people want to have questions
about the theory of voting methods (something I can talk about all
day, but prefer not to when my day has to be filled with things to
help achieve reform), I would hope that a separate list could be
I will say that there are reasons why approval voting has been
dropped by at least one major professional society, why it hasn't been
adopted in any US jurisdictions and why it doesn't have any major
political advocates while IRV has won support over the years from
people like John McCain, Howard Dean, Barack Obama, John Anderson and
Approval voting has two glaring flaws that are far greater than any
alleged flaw in IRV:
* Voting for ("approving") for a lesser choice counts against your
first choice. Voters have great incentives to "bullet vote" vote for
one only) and campaigns certainly have great incentives to tell core
supporters to bullet vote. The end result is that approval voting will
look like a plurality voting election for most sophisticated voters
and results may depend on the lack of strategic know-of of the
Jack Nagel, a UPenn professor, has an important new article
out this year called "Burr's Dilemma" that goes into how this flaw
played out in the 1796 and 1800 presidential election, where the
presidential electors (who at that time had two equally weighted
votes), made strategic mistakes with major consequences in both
elections. Nagel used to say approval voting was better, but now says
IRV is better.
* A candidate can be the first choice of 51% of voters -- someone who
would win on the first round in Denver's current system -- and lose if
some of those voters err by "approving" of a second choice. You can be
sure that those voters in the 51% group would have wished they hadn't
"approved" of someone else if this is clear.
Even when working on a t heoretical level (something I don't think
will happen in any meaningful election where there are high-stakes)
approval voting puts a premium on acceptability. IRV balances
acceptability with having core support. Its "flaws" are all based on
not necessarily electing an acceptable candidate who is unable to
attract as much first-choice support as other candidates. I think we
want leaders, not just followers- -- and leaders need at least some
degree of first-choice support in my book.
At 12:49 PM 6/11/2006, you wrote:
Voting reform is a real issue and not an academic exercise.
Therefore, whatever reforms are made, we want to make sure they are
good ones. Whatever changes occur may stick with us for a while. IRV
would be a real improvement over plurality voting with runoff.
However, it does not guarantee the election of the candidate with the
greatest support and that concerns me.
First, I want to thank Rob Richie and Kathleen McKenzie for the
efforts to improve Denver city council elections. However, I hope
that we will consider other voting systems such as proportional
representation or approval voting due to problems with IRV detailed in
the PS. I know that Rob already supports proportional representation
and has provided arguments, so I will not go into detail there.
Another voting system I hope we will consider is approval voting.
In approval voting, each voter is asked to mark each candidate on
their ballot that they are willing to form a consensus around.
Whichever candidate is marked by the greatest number of voters is
elected. This ensures that the candidate that most voters are willing
to form a consensus around ( i.e. the candidate that has the greatest
support) wins the election.
There are many benefits to this voting system compared to IRV.
First, it is much easier to understand. All voters have to do is mark
every candidate they like, and whichever candidate is liked by the
greatest number of voters wins. Second, it is much cheaper and easier
to implement than IRV. In order to practically implement IRV for
thousands of voters, it requires computers to perform all the
tabulations in quick time. If computers breakdown, it requires much
checking to verify final results. With approval voting, even if the
computers breakdown, it is much easier to count the ballots and decide
the winner. Also, there is no need for significant ballot change
because ballots would look similar to existing ballots that voters are
familiar with. This will save us taxpayers MONEY. Third, approval
voting will guarantee the election of the candidate that is supported
by the greatest number of voters. Approval voting is used by various
groups such as:
The Mathematical Association of America (MAA) with over 32,000 members
The Institute of Management Science (TIMS) with over 7,000 members
The American Statistical Association (ASA) with over 15,000 members
The Institute of Electrical and Electronics Engineers (IEEE) with over
UN Security Council to elect the UN Secretary General
Mathematical and science organizations use approval voting because
they understand the technical problems surrounding the various voting
systems and know which works best for them. Though IRV is superior to
plurality voting, I am worried that it is complicated, to costly to
replace ballots, and can elect candidates who are not supported by the
most voters. I want to urge all of us to push for proportional
representation or approval voting, because the reform we make maybe
stuck with us for a while.
PS: A Simple of Example of How IRV can elect an alternative that is
not supported by the most voters.
Suppose the candidates are Adlai, Bob, Casey, and Dwight. There are
five voters: Helen, Ian, Janet, Ken, and Leah. Their preferences are
Helen: A, D, C, B
Ian: A, D, C, B
Janet: B, D, A, C
Ken: B, D, A, C
Leah: C, D, A, B
If we were using IRV, Dwight would be dropped in the first round
because it has the least first place votes. In the second round,
Casey would be dropped. In the third round, Adlai would beat Bob 3 to
2. But notice that if it had been a one-on-one competition between
Dwight and Adlai, then Dwight would have won over Adlai 3 to 2,
because Janet, Ken, and Leah all prefer Dwight over Adlai. (Further
notice that Dwight would have beat every other candidate in one on one
contests). Nevertheless, IRV eliminated Dwight in the first round!
irv-l mailing list
irv-l at irvdenver.org
F a i r V o t e
The Center for Voting and Democracy
6930 Carroll Avenue, Suite 610
Takoma Park, MD 20912
rr at fairvote.org
irv-l mailing list
irv-l at irvdenver.org
More information about the Election-Methods