With a Three-Way Tie in VA Primary, Consensus is Off the Radar

by Paul Fidalgo // Published June 5, 2009
Following what should have been a well-attended gubernatorial primary in New Jersey this week (but most certainly was not), new polling shows that Creigh Deeds is defying expectations in Virginia's Democratic gubernatorial primary contest. Essentially, there is a three-way tie among Deeds, former DNC chair Terry McAuliffe and last year's presumed frontrunner Brian Moran.

If recent history is any guide, turnout will likely range between really low and really-really low. In the Democrats' 2006 hotly-contested primary for US Senate won by Jim Webb, turnout was less than 3.65% of registered voters -- and it was only 2.6% in 2001's four-way lieutenant governor primary. So it's anyone's guess who will end up the nominee on Tuesday, but what is almost beyond doubt is that whoever wins will not have passed a majority threshold -- and may not be the candidate most Virginia Democrats preferred.

Public Policy Polling (PPP) asked questions that suggest who that candidate might be. Last month, just as Deeds' numbers were beginning to trend upward (watch the purple line at Pollster.com), PPP speculated about voters' preferences using the candidates' favorability numbers among their opponents' supporters. PPP blogged:

-67% of voters with a favorable opinion of McAuliffe are planning to vote for him.

-50% of voters with a favorable opinion of Moran are planning to vote for him.

-Only 44% of voters with a favorable opinion of Deeds are planning to vote for him.

What's going on here is that Moran and McAuliffe's supporters hate the other candidate, but tend to be fine with Deeds because he's stayed above the fray.

[ . . . ]

It appears for most McAuliffe and Moran voters Deeds would be their second choice...but in a state without runoffs that's not going to do him much good. It's just another example of how this being a three candidate field may be what propels McAuliffe to the nomination with well less than a majority of the vote.

The point (for us) isn't whether or not Deeds should wind up the winner or who hates whom. The important point is that with a plurality system like this, in which the eventual nominee could potentially be elected with a little over one-third of the votes, there's no way to determine who might truly be the consensus choice of Virginia Democrats. PPP is guessing that Deeds would be that choice if preferences could be measured, but it is only a guess, though an educated one. Whatever happens, something close to two-thirds of the party faithful will have voted against the winner.

This wouldn't be the case if Virginia Democrats were using instant runoff voting (IRV). With IRV, primary voters could indicate a second choice on Tuesday. Whoever wound up in third place at the close of polls would be eliminated, and his supporters' ballots would be divvied up between the remaining two candidates based on who was listed second, producing a winner who had majority support among the voters with a preference between those two candidates.

If IRV were being used in Tuesday's contest, the winner would almost certainly be someone whom more primary voters could live with, particularly in this case where, according to PPP, Moran and McAuliffe supporters tend to "hate the other candidate." And also important: no one would have to come back to the polls for a separate, expensive runoff election (which isn't an option anyway, as PPP notes). As our study of federal primary runoffs shows, the odds are overwhelmingly high that turnout would plunge -- of the 116 regularly scheduled federal election runoffs from 1994 to 2008, 113 (all but three) saw turnout declines. The mean turnout decline for the period was more than 35%

Some Virginia Democrats already have embraced IRV. Last month, the Democratic Party of Charlottesville held a "firehouse primary," an open caucus to nominate its candidates for various local offices, and for the first time they used IRV to reach consensus choices, in an election with turnout almost quadrupled from previous primaries. Local Democrats had the chance to express their preferences in the case that their first choice didn't make the cut, as in the contest for sheriff in which a majority was not reached on the first tally, and IRV allowed the caucus-goers to reach consensus without having to return to the polls. Many private organizations beyond political parties see the benefit of using IRV to elect their leadership (check out a list of highlights here), and it's even recommended by Robert's Rules of Order for elections where repeated, in-person voting isn't possible.

Ranking one's preferences is not exactly an alien concept to Democratic voters outside of Virginia, either. The party's single biggest name, President Barack Obama, was a prime sponsor in 2002 of legislation to use IRV for Illinois primaries. The presidential caucuses in his neighboring state of Iowa have a key similarity to IRV, although more protracted (and decidedly public). In the caucuses, a candidate needs to achieve a certain threshold of support within a room to be "viable." If it turns out that a candidate can't reach that threshold, his or her supporters disperse throughout the room to join another candidate's group -- a physicalization of ranking one's preferences. If my first-choice candidate isn't going to make the cut, I get to lend my support to my second choice, and candidates with broad, consensus appeal tend to pick up support. There's no "fifty-percent-plus-one" threshold in the Iowa caucuses, so winners can still be declared with small pluralities (and they often are), but after the battle royale that was the 2008 Iowa contest, we at least have all become intimately familiar with the concept of ranking preferences.

This year, Virginia Democrats will have to do without that opportunity. As of now, there are too many factors to consider (eventual turnout, mobilization efforts, last minute ad blitzes, etc.) to make predictions of the outcome hardly more reliable than sport -- even FiveThirtyEight.com's Nate Silver agrees. But we can predict that without a ranked voting system like IRV in place, the true consensus choice of Virginia's Democratic voters will likely remain a mystery long after the polls have closed.