Splitting the progressive vote
Whenever the topic of voting for the Greens comes up, the question arises of whether this involves “splitting the progressive vote” or “wasting your vote”. In large measure, this suggests results from confusing the Australian system with the “first past the post” (plurality of first preference) system that prevails in the US and Britain. In all Australian voting systems it is possible to vote Green with a second preference for Labor. Obviously, if you don’t do this (by preferencing the Lib/Nats or by exhausting your ballot where this is an option) the result will be to reduce the chances of a Labor candidate being elected. But are there any other cases where this is a risk?
For single-member constituencies, the most common case is where the Greens are sure to run third. In this case, their preferences are distributed, and a Green vote with a Labor second preference is functionally equivalent to a Labor first preference, although the views they express are different. In the rare cases where Labor is sure to run third, the opposite is true. The only interesting case is where the order in which Greens and Labor are going to finish is unsure. In this case, the outcome that gives the best combined Labor-Green chance is that the candidate with tighter preference flows should finish third. That way, there is less leakage of preferences to the conservatives. In practice, I don’t think this is a big deal, but to the extent it is at all relevant, it usually suggests that progressives should vote Green. But there is a problem with this, in that, if people really acted on this analysis, there would be a reward to the party whose voters were more likely to split their preferences.
Broadly speaking, the same analysis applies in the Senate. Most of the time, the total number of Green and Labor Senators elected will be the same regardless of how preferences are arranged between them. Given the complexities of the system, especially with voting both above and below the line, anything can happen, but since these complexities are almost impossible to predict in advance, they can’t really influence a decision on how to vote.