The Victorian election is living up to its billing as the latest battlefield in new paradigm politics, with the Liberals finding themselves shunted from the front pages by a stoush between Labor and the Greens. At issue are the professional activities of the Greens candidate for Melbourne, Brian Walters, SC, who has been targeted over his legal work for accused war criminal Konrad Kalejs and a company associated with coal mining. After a furious response from the legal fraternity and the liberal end of the Melbourne media (The Age playing a tellingly distinct role in the controversy from the Sunday Herald Sun), most have concluded Labor’s attack has badly misfired, with Andrew Crook of Crikey going so far as to argue it has doomed Melbourne MP Bronwyn Pike to certain defeat. The correctness of this view depends largely on the resolution of the campaign’s other Greens-centric controversy: the split in the Liberal Party over whether to continue placing the Greens ahead of Labor on how-to-vote cards.

The behaviour of major party preferences has been little studied, as in the normal course of events they are not distributed. Of much greater interest has been minor party and independent preferences and their bearing on major party outcomes. The only substantial interruption to this picture in recent times came with the emergence of Pauline Hanson’s One Nation, who Labor and eventually the Coalition parties both saw fit to put last. Hanson herself topped the primary vote when she contested the new seat of Blair at the 1998 federal election, but was thwarted when about three-quarters of Labor preferences went to Liberal candidate Cameron Thompson. When state Labor advised voters to simplify matters with a “just vote one” strategy in 2001, made possible by Queensland’s optional preferential voting system, the rate of exhausted Labor votes shot from a third to three-quarters. These episodes confirmed what scrutineers had long known about major party voters’ observance of how-to-vote cards.

Even more helpfully, a ballot paper study conducted by the Victorian Electoral Commission after the 2006 election (thanks to Peter Brent of Mumble for alerting me to this) encompassed all four of the electorates under consideration, and found the rate of obedience among Liberal voters ranging from 30 per cent in Richmond to 45 per cent in Brunswick. With those Liberal voters who didn’t follow the card favouring the Greens over Labor about 60-40, the total rate of preferences to the Greens was consistently around 75 per cent, or slightly below the 80 per cent recorded in Melbourne and Batman at the federal election. As a rough guide, it can be inferred that a change in the Liberals’ how-to-vote policy would cut their preference flow to the Greens from the high 70s to around 40 per cent.

The likely impact of this is best considered in light of recent voting patterns. The table below shows how the state electorates voted both at the 2006 state election and the recent federal election (results from the latter being derived from booth results with slight adjustments made to account for declaration votes). While the latter figures have the advantage of being more current, they are unavoidably contaminated by specifically federal factors.

 
GRN 2PP
  ALP GRN LIB LIB PREF NO PREF
2006 STATE
Melbourne 45% 27% 22% 48% 40%
Richmond 46% 25% 20% 46% 39%
Brunswick 48% 30% 17% 45% 40%
Northcote 53% 27% 15% 42% 37%
2010 FEDERAL
Melbourne 36% 37% 22% 57% 49%
Richmond 39% 37% 20% 55% 48%
Brunswick 46% 31% 19% 48% 41%
Northcote 46% 33% 17% 49% 42%

On the basis of the 2006 state election, the order of dominoes would look to be Melbourne, Richmond, Brunswick and Northcote, with some distance separating the last two. The federal election results tell a slightly different story, with the Greens in a substantially stronger position in Melbourne and Richmond than in Brunswick and Northcote – remembering that the former two constitute most of federal Melbourne, where Labor suffered the loss of Lindsay Tanner’s personal vote. By the same token, it should be remembered that Labor is losing incumbent Carlo Carli in Brunswick, where the contest could be further complicated if former federal independent Phil Cleary runs as an independent. The last two columns in the table project the Greens’ two-party vote in scenarios where they do and don’t receive Liberal preferences, and herein lies the rub. With Liberal preferences, they look to have Melbourne and Richmond in the bag, as well as being highly competitive in Northcote and Brunswick. Without them, they could yet emerge entirely empty-handed.

Personally, I would be very surprised if a party in a system as adversarial as our own saw fit to grant such a huge free kick to their real enemy. But at the very least, it will be interesting to see if the Liberals can do better this time in Greens preference negotiations which have traditionally been entirely fruitless for them.

UPDATE: Sam Bauers in comments makes the good point that the VEC study shows how Liberal voters behave when the how-to-vote card reflects their expectations: a change in policy might increase the rate of rebellion.