Fact Check

This page serves to document some false and misleading talking points related to election methods. Until this new page is more replete with examples, you can read this page for numerous other examples.

Rob Richie criticizes Approval Voting on Twitter

On December 26, 2011, FairVote Executive Director Rob Richie wrote this post on Twitter:

@MikeSmithsonOGH MrsB wins "poster of year" at UK blog w/just 28% in approval voting election. Approval often devolves to plurality vote .

This was in reference to an Approval Voting election for "PB Newcomer Poster of the Year" on a blog called Political Betting. You can read the blog's official post about the results here. The numbers are as follows:

494 voters
MrsB 201 votes = 40.7% approval
Old_Labour 196 votes = 39.7% approval
Beverley_C 192 votes = 38.9% approval
Mick Pork 101 votes = 20.4% approval
Uniondivvie 39 votes = 7.9% approval
Average number of approvals per ballot = 1.48

What immediately jumps out is that, with 1.48 approvals per ballot, this election was not remotely close to "devolving to plurality". And this is a typical result which is consistent with numerous Approval Voting elections in the German Pirate Party for instance.

But what is not immediately obvious is where Richie's bogus 28% figure came from. MrsB was approved by 40.7% of the voters, not 28%. It eventually occurred to me that he had incorrectly divided MrsB's total by the total number of votes rather than by the total number of voters. The blog's author clarified this to Richie in a subsequent Twitter post:

@Rob_Richie With approval voting you divide the winner's votes by the total number of voters. In this case 40.7% not, as you assert, 28%.

Attacks on Bucklin voting

Bucklin voting is something like a hybrid between Approval Voting and Instant Runoff Voting.  Unlike IRV, candidates are never eliminated from consideration; if there is no winner based on first-choice votes, all second-choice votes are simultaneously added to the approval totals. It was used by millions of voters in over 50 U.S. cities in the early 1900's.

FairVote has repeatedly criticized Bucklin as a proxy for Score Voting and Approval Voting. Their argument is that Bucklin effectively degenerated to Plurality Voting, because of the prevalence of voters "bullet voting" only for their favorite candidate. They argue that this is an indicator of what would happen with Score Voting and Approval Voting in large contentious elections. Here are some examples.

Bucklin was found to be defective as it allowed a voter’s second-choice vote to help defeat a voter’s first-choice candidate. With Bucklin voting, most voters refrained from giving second choices, and the intent of discovering which candidate was favored by a majority of voters was thwarted.

Voters quickly recognize that voting for an alternate choice may help defeat their first-choice candidate. For example, if both your first choice and second choice advanced to the second round, your ballots would cancel each other out. For this reason, in high stakes elections in which voters have strong favorites, most voters opted to "bullet vote" and protect the interests of their favorite choice be withholding any alternate choices.

We delved into the history of Bucklin elections and found numerous cases in which a significant number of voters cast second choice votes. There is absolutely no evidence that Bucklin ever degenerated into Plurality Voting. Continuing from the above source:

A flaw unique to Bucklin, with its limit of two choices was that if a voter's second choice was not one of the top two initial count candidates, their second choice vote was wasted.

Simply false. See this counter-example.

 % of voters
their ranking 
49% X
26% Y > Z
25% Z

The first round has no majority winner — candidate X is in first place with 49%. So we add the voters' second choices, giving Z 51% after the 26% bloc has their second choice counted.

The previous quote continues:

As a result, the winner with the plurality of combined first and second choice votes could easily fall far short of a majority.

Of course it's possible that no candidate is the first or second choice of 50% of the voters. This has nothing to do with the voting system — it's just a potential state of voter preferences. The problem in this case is not with the Bucklin system, but with the two-ranking limit imposed by some cities.

FairVote members themselves have complained that the fiasco in the 2010 District 10 supervisor race could have been mitigated by allowing voters to rank more than three candidates. Example:

in D10, there were more exhausted ballots at the end of the count than received by the winner. (In fact, there were more exhausted ballots at the end than received by the remaining two candidates combined.) The best way to reduce the number of exhausted ballots, or at least the involuntary ones, is to get equipment that allows voters to easily rank more choices.
Steve Chessin, FairVote MemberPresident, Californians for Electoral Reform

Continuing in the same article:

In Alabama, for example, in the 16 primary election races that used Bucklin Voting between 1916 and 1930, on average only 13% of voters opted to indicate a second choice.. With 16 primary elections between 1915 and repeal in 1931, in no case did the addition of the second choice votes give the winner a majority (the purported goal of the system). And in only one case did the addition of the second choice votes change the outcome from the original first choice plurality candidate.

Those Alabama races did not use Bucklin. They used a substantially different system in which only the top-two first-round candidates got to have their subsequent votes added on. This is clear from the results.

Since no candidate won a majority, the second choice votes for Graves and McDowell were added up to determine a winner.
Alabama Governor Primary Balloting
Contender 1st choice 2nd choice total
Bibb Graves 61493 21978 83471
McDowell 59669 7943 67612
Carmichael 54072 20061
Patterson 47411 20893

Graves and McDowell were the only candidates whose supporters got to have their second choices counted. Whereas with Bucklin voting, Carmichael would have come in second with 74,133 votes.

We expect FairVote would argue that this was indeed Bucklin — just a variant which still warrants the same name. We observe that the system used in Alabama does not even meet the definitions of any of the variants described in the Bucklin article on Wikipedia.

The bottom line is that, regardless of what you call the Alabama system, its failings are certainly not an indictment of Score Voting or Approval Voting.