This post was written with my colleague Devin McCarthy.
For decades, the Oscars have picked nominees for all major categories using the fair representation form of ranked choice voting (RCV), as reported at FairVote's OscarVotes123 blog. Since 2009, when the Academy of Motion Pictures increased the number of Best Picture nominees to ten (it can now be between five and ten), it has also used RCV - sometimes called instant runoff voting or preferential voting - to select the Best Picture.
Analyzing Oscar voting is tricky, because the Academy doesn't release actual vote totals. As a result, any investigation into how an Oscar election played out can only be based on the announced winners and anecdotal evidence of what voters were thinking.
This year, Entertainment Weekly made an attempt to quantify Oscar votes for Best Picture and the other major awards. Their methodology is ambiguous ("based on previous awards won and conversations with insiders and Academy members"), but it provides some hard numbers for examining the 2013 Oscar race. Assuming that real Oscar votes will be comparable to EW's predictions, we can infer how the voting system may affect the outcome of the election.
EW's projected first choice percentages for Best Picture reveal one obvious fact: it's a good thing the Academy uses ranked choice voting. No movie is projected to receive more than 20% of first choices, and the difference between the frontrunners (Gravity, 12 Years a Slave, and American Hustle) and the pack is a mere few percentage points -- well within the margin of error of this exercise.
Here are EW's projected first choice percentages (these numbers are only available in the print version):
12 Years a Slave:18%
The Wolf of Wall Street:13%
If Best Picture voting were conducted using a plurality system, as are most American elections, Gravity would be projected to win a narrow victory. But that result would not demonstrate that Gravity is Academy voters' favorite movie. In theory, 81% of the Academy could hate Gravity, yet it could still win. Indeed, any of the top four movies could slip into the lead even if not broadly supported.
In fact, these numbers may look familiar to those who recall the notable 2002 French presidential race. In the first round of that election, the results were:
Le Pen (far-right):16.9%
Jospin (Prime Minister):16.2%
Rest of Field:40.2%
France has a runoff system rather than a plurality system, so all the candidates except Chirac and Le Pen were eliminated. Le Pen, who was only 3% behind in first choices, ultimately lost the runoff by a whopping margin of 82.2% to 17.8%.
The French runoff shows that a majority threshold in the first round matters for a fair outcome. It also illustrates the problems of a rigid two-round system where only two candidates advance, as the much stronger candidate (Jospin) failed to make the runoff and lost his opportunity to win a runoff.
Back to this year's Best Picture election. Conventional wisdom says that Gravity and 12 Years a Slave have the most overall support, with American Hustle maintaining an outside shot at the award. The Wolf of Wall Street may be close in first choices, but few analysts consider it to have much overall support. Yet under a plurality rule, it could plausibly win an undeserved victory -- just as Le Pen could have won in a plurality system in 2002.
Ranked choice voting is similar to a runoff system, but even better. When votes are counted, movies are eliminated one by one starting with the nominee with the fewest first choice rankings, and the eliminated movies' votes then go to the next choice on each ballot.
In our hypothetical Oscar election, suppose that after the bottom six movies are eliminated, the field is narrowed to three. As is likely to occur, no movie has even 40% support at this stage. The count after six rounds of elimination might be:
12 Years a Slave:36%
American Hustle would then be eliminated, producing a final round runoff of:
12 Years a Slave:52%
We'll never know how the voting really broke down. But we know with certainty that whatever movie wins, it will have majority support among Academy voters who ranked either it or its top competitor somewhere on their ballots.
Interestingly, the Academy has chosen not to use RCV for any awards other than Best Picture. Our hunch is that while RCV ensures a more representative result, it does not always make for exciting television. A plurality voting system increases the probability of apparent "upsets" by allowing relatively polarizing nominees to win, even if most voters didn't think they were that good.
The non-Best Picture categories only have five nominees instead of nine, so the projected first choice share of the winners tends to be slightly larger. But notably, even races with clear frontrunners do not have a majority of first choice support. According to the Entertainment Weekly survey, for example, Jared Leto will win Best Supporting Actor (with a projected 38% of the vote, ahead of his closest competitor's 20%) and Cate Blanchett will win Best Actress (projected 35% of the vote, 10% more than second-place Amy Adams). EW does not project any other category to be won with more than 30% of the vote, including the tight race between Gravity's Alfonso Cuarón and 12 Years a Slave's Steve McQueen for Best Director.
If there is a rare Best Picture/Best Director split this year, it might be because voters think more highly of Cuarón's directorial work on Gravity than the movie itself, but it could just as easily be due to the difference in voting systems between the two awards. If Gravity has more first choice support but 12 Years a Slave is preferred when compared one-on-one by most of the electorate, 12 Years would likely win Best Picture while Gravity would deliver the Best Director Oscar to Cuarón.
Melena Ryzik of the New York Times unfortunately dismissed RCV this week as "complicated" and "unglamorous." Meanwhile, Entertainment Weekly picked Gravity to win Best Picture purely based on its projected advantage in first choices, displaying a lack of understanding of the fact that the winner will really be determined by the final one-on-one runoff. Ranking choices and conducting a series of runoffs is, in fact, easy - and produces a much fairer outcome.
For more informed coverage of the voting system, we urge readers to peruse OscarVotes123 and the writings of The Wrap's Steve Pond. Finally, keep in mind that a growing number of American jurisdictions are adopting ranked choice voting to ensure our elected officials are fairly elected. Those jurisdictions, as well as the Best Picture voting process, should serve as a model for other elections to emulate - from the entertainment industry to American's highest political offices.