Of Campaigns and Crosswords: Pluralities as Poor Political Reflections
October 4th, 2007
Paul
Paul Fidalgo was Communications Director at FairVote.
A FairVote Innovative Analysis
_ _ _
Facts in the Spotlight
Gubernatorial general elections won without a majority between 1990 and 2006: 44
Currently serving governors elected at one time without a majority: 16
_ _ _
Elections are like Scrabble. Winning requires strategy, a command of language and a sense of your opponent’s weaknesses, there are very often more than two players, and it doesn’t hurt to pull a seven-letter bingo from out of nowhere for a 50-point bonus (wait, I think that last one is only in Scrabble).
But elections and Scrabble are also similar in what it takes to be declared the winner. In Scrabble, even when there are four players, there’s no expectation that the winner should have a majority of all points, just more than anyone else. The same, very often, goes for elections. Officials can be elected with less than 50% of the vote, so long as they have more votes than their opponents. And hey, if it’s good enough for Scrabble, it’s good enough for government, right?
Minority rule and accountability - Of course, there are few leadership positions doled out to the winners of board games - in elections the stakes are very real. When an election can be won with a mere plurality (less than 50% of the vote) it stands to reason that more than half, a majority, voted against the winner and may not be happy with the results. In board games, we can accept "first past the post." In a democracy, the goal should be majority rule. Otherwise, the principle of accountability is lost: if more than 50% don’t want an incumbent returned to office, that majority shouldn’t lose just because it happened to divide its vote between two or more candidates.
Non-majority winners occur more often than you might think: Between 1990 and 2006, 44 gubernatorial elections have been won with mere pluralities. These pluralities can be relatively high, sometimes scraping the 49 percent mark. But often, and especially in the case of strong third party candidacies, governors take office with vote percentages in the 30s, such as the 2006 reelections of incumbents John Baldacci of Maine (with 38%) and Rick Perry of Texas (with 39%), whose lower pluralities were in part the result of two strong independent candidacies in both races.
Party nominations for governor are particularly vulnerable to plurality wins as multiple candidates are the norm. In Maine’s close fought three-way race for the 2006 Republican gubernatorial nomination, for example, the primary was won with barely 39% by conservative State Senator Chandler Woodcock, who many believed was a weaker general election candidate than would have otherwise been nominated had the vote not been split three ways.
A means to consensus - Candidates taking office with less than a majority begin their terms with far less political capital with which to pursue an agenda than if the majority had supported them. Very often, plurality winners may well have been the consensus choice of the majority, but the current system offers no way for voters to express an inclination for candidates other than their first choice. If states used instant runoff voting (IRV), however, voters would be able to rank candidates in order of preference. If their first choice (say, a strong independent candidate) did not have enough first-choice votes to compete, their second choice candidate (someone closer to their political viewpoint than the other remaining candidates) would automatically receive their vote. Using IRV, voters can rest assured that whomever does take office has the support of the majority.
Primary pluralities - Plurality winners do not pose a danger to gubernatorial legitimacy alone. Presidential primaries are a constant reminder that successful candidates are not always the consensus choice. Take the 1996 New Hampshire Republican presidential primary, where Pat Buchanan won the contest with a mere 26%, while an absolute majority almost certainly would have preferred Bob Dole or Lamar Alexander. If trends continue, the 2008 GOP Iowa caucuses and New Hampshire primary will be won with a small plurality as well (with recent polls showing frontrunners Mitt Romney and Rudy Giuliani in the mid-20s in Iowa). Even the Democrats’ current battle over Iowa seems locked in a three-way tie between the front-runners, all in the low 20s. Can a win with such a small number ever truly reflect the will of the majority? And can we count on results truly reflecting voters’ preference if they have decided their first choice can’t win and don’t want to "waste" their vote?
The power of preference - These problems are eliminated, however, when voters can rank their choices, expressing preferences for other candidates if their first choice doesn’t make the cut. That kind of power can mean a lot to all sorts of voters, from those who back the Democrat who scores a close third place in Iowa, to the supporter of candidates, as Chris Dodd has put it, "competing with the margin of error."
One more difference between success in board games and elections: in Scrabble, it takes a seven-letter word to score a bingo. For elections, we need only three: I-R-V.
* * *
To learn more about instant runoff voting, you might enjoy the recent Christian Science Monitor commentary by our chairman John Anderson, the 1980 independent presidential candidate.
You can also see a new video prepared for cities using instant runoff voting in North Carolina for the first time this fall.
For more general information on IRV visit fairvote.org/irv, and read more about plurality elections throughout US history right here.
Previous editions of Innovative Analysis can be found here.
Other posts by Paul
- Nomination by Attrition - October 25th, 2007
- Cary Voters Sing the Praises of IRV - October 25th, 2007
- FairVote Announces Upgrade Democracy Video Contest Finalists! - October 23rd, 2007
- The Senate Loses its Cool: The Undemocratic Ways States Fill Senate Vacancies - October 12th, 2007
- Of Campaigns and Crosswords: Pluralities as Poor Political Reflections - October 4th, 2007
- The Primaries' Premature Nomination Problem - September 28th, 2007
- Barry Bonds, Blast-offs and Ballots: What the Fate of a Baseball Teaches Us about Voting - September 20th, 2007
- The President of Everyone? Crucial Issues Suffer when Most Voters Don't Matter - September 13th, 2007
- Upgrade Democracy Video Contest News Update - September 11th, 2007
- The First Shall Be Last: The Dangerous Decline in Primary Turnout - September 7th, 2007
October 4th, 2007 at 5:59 pm
>Gubernatorial general elections won without a majority
>between 1990 and 2006: 44
I’m interested in this data.
1. What is the denominator? How many won the general election with a majority?
2. How many elections (/ what percent of elections) for governor used a runoff election? In your 44 did you count those who did not achieve a majority in the general election but then did win a majority in a separate (eg non-instant) runoff?
>One more difference between success in board games
>and elections: in Scrabble, it takes a seven-letter
>word to score a bingo. For elections, we need only
>three: I-R-V.
As someone who prefers several other election methods to IRV, the game that I associate with IRV is Sorry.
October 4th, 2007 at 8:54 pm
What’s sorry is the amount of time that range/approval zealots spend trashing IRV instead of agitating for PR. District magnitude is the key.
October 4th, 2007 at 11:49 pm
I’m generally in favor of PR, although I have not gone on record as supporting any specific PR system and have not fully researched the systems. I tend to favor ideas that allow voters to select who their representative is rather than vote for their representative.
I was not previously familiar with the term ‘district magnitude’. It apparently refers to the number of representatives to be elected in each electoral district. I agree that it is key. One thing that is clear though is that there are many cases where a single winner election is the sensible election and so district magnitude=1. So single winner election methods do matter and should matter even to strong supporters of PR.
Some US-centric examples:
president
senators
state governors
congressmen for single rep states
etc…
Even in cases where a district magnitude of > 1 makes sense it is not clear that PR is a win. My judgment is that it has to be about 5 or 6 for PR to be a helpful reform. This is an uneducated out-of-a-hat number that I have no real argument for. I’d be interested in reading published arguments for larger or smaller district magnitudes.
When the magnitude is too low, there remains no way for a representative of broadly popular but locally weak views to become elected.
Also, when the magnitude is too low, some election methods that produce good results as PR methods for higher district magnitudes instead produce poor results. *ahem* STV / IRV.
October 5th, 2007 at 3:33 pm
PR is helpful, just not as helpful, with district magnitudes as low as 3. The political dynamics are different at lower district magnitudes, but there’s still value in it.
Good documentation is provided by the experience of the Illinois legislature from 1870 to 1980 (dates approximate), which was elected by cumulative voting from 3 member districts. This arrangement elected almost 0 members of parties other than the Republicans and Democrats. What it did do was elect some Republicans from the Chicago area and some Democrats from rural areas downstate. Almost every district was split 2-1 or 1-2. That meant that every supporter of a major party got representation by someone s/he had voted for. Granted, small party voters didn’t get anything.
And this was accomplished using a voting method, cumulative voting, that is really semi-proportional rather than proportional.
Visit http://www.fair.vote.org/?page=512 for links to a 2001 study that reviews this history (and recommends that Illinois go back to cumulative voting).
October 5th, 2007 at 9:44 pm
>[AAV] Even in cases where a district magnitude
>of > 1 makes sense it is not clear that PR is a win
I phrased this poorly. My intention was to compare PR systems for a small district magnitudes with a bunch of single member districts. Eg. 5 districts with DM=3 vs 15 single member districts.
Bob: Thank you for the Illinois reference. I was not aware of it. I sincerely hope that no-one is pushing for cumulative voting for multi-member districts.
>[BR]What it did do was elect some Republicans from the
>Chicago area and some Democrats from rural areas
>downstate. Almost every district was split 2-1
>or 1-2. That meant that every supporter of a major
>party got representation by someone s/he had voted
>for.
This is a positive, but not really a large one. I believe that most voters care about having their views represented in legislature and do not care very much about having their own representative share their views. In other words the situation of having ‘non-representative representatives‘ is not nearly as harmful if the viewpoints of all the representatives are close to the viewpoints of all the voters.
October 8th, 2007 at 5:21 pm
AllAboutVoting: … if the viewpoints of all the representatives are close to the viewpoints of all the voters.
But (and it’s a huge but) in legislatures elected from single member districts, this can’t be true in general. It can be approximately true in the special case where (a) partisan allegiance corresponds closely to geography, and (b) there are only two significant parties. Deviations from (a) lead to the legislature being more lopsided than the electorate (in the extreme case where all districts have the same political makeup, one party can win all the seats). Deviations from (b) complicate this picture, leading to bizarre elections in which one major party wins a majority of the seats with fewer votes than one of the other parties. Example: British Columbia election of 1997, where the Liberal Party got 42% of the vote, but the New Democratic Party got 52% of the seats on 39% of the vote.
October 9th, 2007 at 1:04 pm
Bob: I fully agree that single winner districts can lead to very distorted representation.
My point was that the quality of representation can be measured in two ways:
1. how well the population as a whole is represented by the the legislature as a whole
and
2. how well each individual is represented by their specific representative.
I think that #1 is a dominant concern but that #2 matters to a degree as well. #2 is important in cases where access to one’s representative is valued.
Single winner districts and multi winner districts that use the ‘cumulative’ election method flunk on both #1 and #2.
Many PR methods pass #1 but dodge #2 to a degree in that an individual voter cannot identify which representative is theirs and may not have special access to a specific representative.
I tend to lean towards PR methods where property #2 is met as well since the voter truly feels that they have a representative who is accountable to them and to whom the voter has access.
October 12th, 2007 at 5:57 pm
AllAboutVoting: 2. how well each individual is represented by their specific representative … is important in cases where access to one’s representative is valued.
Meaningful access to one’s representatives is based on political compatibility to a far greater extent than it is on geographical proximity. How well I am represented by my specific representative really depends on whether I voted for that person (or, in a ranked ballot system, gave her a high ranking). Opponents of PR argue otherwise, but I think they’re wrong.
I think AllAboutVoting’s first and second criteria are mostly incompatible. Quality of representation and meaningful access to one’s representatives both depend on proportionality. For the ombudsperson functions of legislators (untangling red tape), having a representative who is “one’s own” in the very narrow sense provided by single-member districts might be somewhat meaningful. But that doesn’t provide representation on policy questions.