A Message to IRV Fans - Look Beyond the Hype

Instant Runoff Voting (IRV) is touted for its ability to take away the spoiler effect that small parties can have on an election, thereby also taking away the fear of voting for them, and giving a realistic idea of how much support they have.  But it doesn't always live up this promise, and there are vastly better methods that do.

It is ironic that many people believe IRV allows voters "more choices".  IRV has produced two-party domination in all four countries where it has been substantially used; namely Australia, Ireland, Malta, and FijiAustralianpolitics.com says IRV "promotes a two-party system to the detriment of minor parties and independents."  Since Ireland's presidential post began, in 1938, the Fianna Fáil Party has won all but once, when the Labour Party's Mary Robinson won in a phenomenal fluke; so IRV has produced a virtual party monopoly there.  Fiji instituted an IRV-based system for their House of Representatives in 1999.  On 2005 December 21, Fijian Prime Minister Laisenia Qarase voiced concerns over their system, and called for consultations on a possible return to plurality.  And approximately two dozen U.S. cities (the largest being New York City, in 1936) have implemented STV (IRV is STV applied to a single-winner election), only to have almost entirely reverted to plurality.

Of the 27 countries we know of that use "top-two runoff", on the other hand, 21-23 have broken free of two-party duopoly.  The same is true even of non-IRV branches of government within IRV countries.  The Australian Senate, for example, uses STV, and had 9 of its 76 seats occupied by alternative parties in the 2004 elections.  Yet Australia's House of Representatives, which uses IRV, is a two-party body.  Maybe that's why Australia's minor parties want to replace it, and why the Libertarian Reform Caucus calls it a "bullet in the foot".  So if third parties would like to have a chance of succeeding in single-winner elections, they had better steer clear of IRV, because it is absolute suicide for them.

It is also important to consider the kind of nightmarish paradoxes that can occur with IRV.  For instance, consider the following set of ballots:

% of Voters How They Voted
34% Nader > Gore > McCain
15% Gore > Nader > McCain
17% Gore > McCain > Nader
34% McCain > Gore > Nader
Don't think this example is "realistic"? Think again.

With IRV, Gore is eliminated first, giving 32% of the ballots over to their second choices.  McCain then wins with 51% of the vote.  But wait!  66% of the voters prefer Gore over McCain!  Imagine the outrage from the Gore and Nader voters if they were to discover that McCain was elected, even though 66% of the voters preferred Gore!  IRV chooses a "wrong winner" because it ignores the 15% of Gore voters who "become" Nader voters after the first round, as well as the second choices of the Nader voters, who prefer Gore over Nader.  So much for the myths that IRV prevents wasted votes and "IRV makes your vote count."

Now consider what happens if Nader drops out of the race.  Most of the Nader supporters would vote for Gore as their first choice, and Gore would win with a 66% majority.  With Nader in the race, McCain wins.  Nader is a spoiler.  So much for the myth that IRV eliminates spoilers.

In this example, Nader takes first-choice votes away from Gore, thus "splitting" the liberal votes, and causing Gore to be eliminated in favor of a conservative.  If you're more conservative than liberal, simply swap McCain and Nader, and the example works works the same.  If you don't believe that a third party candidate like Nader (who would be considered far from center, even in non-partisan electoral systems) could beat a "mainstream" candidate like Gore, then you are admitting that IRV leads to two-party domination (or "center squeeze" in non-partisan systems).  In that case, substitute someone more "realistic" like Dean for Nader, and this example becomes totally plausible.  So much for the myth that IRV eliminates vote splitting.

Does IRV eliminate the incentive to vote strategically?  Sorry, that's another myth.  In the example, McCain wins, which is the worst outcome from most of the Nader voters' viewpoints.  But if a few of those Nader voters strategically vote Gore first, Gore wins, which is a better outcome for them.  Thus, strategic voting sometimes pays with IRV, just as it sometimes pays with plurality.  Note that strategic voting causes the first-choice vote results to be distorted; in this example, strategic voting reduces the number of first-choice votes for Nader and increases the number for Gore.  So much for the myth that IRV accurately measures the support for third-party candidates.

A "Lynn" who returned my message to OaklandIRV.org, a pro-IRV group created to support Measure O in Oakland's 2006 election, told me over the phone that IRV won't help third parties to win in the first place.  As she puts it, the point is to take away the fear of voting for them, so that we can get a realistic gauge of voter opinion, and so that we can stop the spoiler effect.  I encouraged her to make sure that alternative party voters and candidates know the reality—that IRV produces two-party domination.  But she seemed to think this was obvious.  When I explained to her that IRV produces spoiler effects of its own, and does not necessarily measure voters' real preferences accurately, she simply dismissed such possibilities as unlikely.  Yet in the real world, where IRV consistently produces two-party rule, these effects are apparently quite real.

How about the claim that IRV ensures that the winner is chosen by a majority of the voters?  Unfortunately, that's both false and misleading.  In the example, if most voters vote for their first choice only, no candidate gets a majority of the votes.  Even if most voters indicate a first, second and third choice, it is possible that no candidate gets a majority of the votes, if there are many candidates.  (In fact, this happened in 6 of the 12 San Francisco Board of Supervisors elections in 2004 and 2006.  The winner was marked as the voter's 1st, 2nd or 3rd choice on less than half of the original eligible ballots.)  The claim is misleading because there are multiple ways to manipulate the ballots to form "majorities."  In the example, IRV finds that McCain is supported by 51% of the voters.  But the truth is that 66% of the voters prefer Gore over Nader, and 66% prefer Gore over McCain.  Gore is supported by a majority against both other candidates.  Why shouldn't Gore be declared the winner?

Consider also the expense associated with implementing IRV.  Many IRV advocates claim that IRV would save money.  But consider this assessment by Princeton math doctorate, Warren D. Smith:

Election expense will certainly increase by using IRV rather than voting systems which can use present-day plurality-type voting machines not connected together via a computer network. It may be that the cost "decrease" they had in mind was versus plurality with a second runoff election. It is true that a single IRV election is cheaper than two elections (original plus runoff), if all other things are equal, which is the point of the word "instant." However, because most places that require runoff elections only need them rarely, the expense ratio on average is not anywhere near 2-to-1, and hence the expense of switching to IRV would usually exceed the savings for a long time (and considering the need to continually replace machines, perhaps forever).

Of course, we can compare the individual properties of voting systems ad infinitum, but that's a bit like comparing the engines, tires, and aerodynamics of two race cars.  The ultimate metric is simply, when the rubber meets the road, which one gets the best time?  The analogous test for a voting method is called "voter satisfaction index" (or "social utility efficiency", by economists).  In lay language, it is simply the percentage of the maximum potential voter satisfaction achieved by a voting method.  A theoretical process that could read the voters' minds, and choose the candidate who would be the most pleasing to the most voters would have a VSI of 100%, by definition.  Rigorous experimentation has shown that Range Voting (RangeVoting.org) produces about twice as great a voter satisfaction index as IRV or plurality, even when voters are extremely strategic instead of honest.  This means that Range Voting gives us as much improvement over plurality and IRV as either of those methods gives over non-democratic random selection of the winner.  This means that Range Voting effectively doubles the happiness brought about by the invention of democracy!  This also means that using Range Voting would produce a greater improvement to democracy than the total eradication of fraud, in any country where the effect of fraud does not produce worse results than random selection would produce.  Incidentally, in highly strategic electorates, IRV tends to produce as poor a VSI as plurality voting.

With Range Voting, each voter simply assigns a score (say from 0-10) to each candidate, and the candidate with the highest average score wins.  It's simple and intuitive, and suffers far less harm from the use of strategic ("insincere") voting than other known methods, like plurality and IRV.  It also has the enormous benefit of giving third party supporters a chance to always express their sincere first choice preferences (or put another way, with Range Voting, a vote for Nader is NEVER a vote for Bush, as it easily can be with plurality or IRV).  The Libertarian Reform Caucus acknowledges these benefits, and as of 2006 began to support range voting, and use it for internal processes as well.  And unlike IRV, Range Voting can be implemented on all standard U.S. voting machines!

It's time for voters to get educated about the various alternative voting methods that exist.  Read more deeply into the facts and myths surrounding election reform.  Not every idea associated with reform is a good one, and IRV happens to be particularly problematic.  There are those who say, "But IRV has so much more momentum than anything else."  Well, global warming has more momentum than global cooling.  Does that mean we should support global warming?

Voters who care about choosing the candidate who will bring about the greatest overall satisfaction for society should push for the adoption of Range Voting.  Meaningful, quality democracy requires that we do.

The folks at the Center for Range Voting

Please contact co-author Clay Shentrup at 206.203.6619 with any comments or questions.

Return to main page