In case you missed it (another good reason to subscribe), I was in yesterday’s Crikey with a state-by-state summary of the prospects in the Australian half-Senate election, with particular attention to what might happen if preference exchange between Labor and the Greens should break down – as Phil Coorey suggested last week. Here’s my summary of that issue:
… while Labor may decide to preference against the Greens … it has no possible rationale in terms of Senate numbers for doing so. It would simply be indulging its dislike of the Greens at the risk of handing an Abbott government a Senate majority.
Conversely, the Coalition has nothing at stake in the Senate as between Labor and the Greens and is therefore free to indulge its ideological or other predilections. Because the Coalition starts out so far ahead on Senate numbers, there is a basic asymmetry: Greens v Coalition contests matter to Labor, but Greens v Labor contests don’t matter to the Coalition.
But the story also raises in passing the question of how to think about Senate numbers. Because the senators who are up for re-election this time were elected in 2007, it’s natural to use that election as the comparison and so to say, for example, that the Greens might pick up a seat in Victoria at the expense of the ALP – since Victoria in 2007 split three Labor and three Coalition, so there is no sitting Greens senator.
If this were, say, the United States, that would be a completely logical way to look at it. Senate elections in the US are staggered in much the same way as ours (a third every two years, rather than half every three), but because each state elects only one senator at a time they are high profile individuals, often capable of defying a party trend. So if you want to analyse next year’s Senate election in New Hampshire, you start by looking at the 2008 election, when the incumbent was elected, rather than the most recent Senate election held there, in 2010, when a different senator (from, as it happens, a different party) won.
But my argument is that in Australia this is the wrong way to look at things. Our Senate is not a chamber of well-known and powerful individuals; it is very much a party house, filled largely with unknown hacks whose name on the ballot paper means nothing to the average voter. The huge majority of voters don’t even vote for individual candidates at all – they just number a single party box above the line.
So I suggest there is no reason not to just take the most recent Senate vote (in this case 2010) as the baseline and calculate gains or losses relative to that, rather than to the current numbers in the Senate, which partly reflect the more remote 2007 election.
In the case of the Coalition’s seats, the difference doesn’t really matter since the total comes out the same – 34 seats – whichever way you do it (although it’s composed differently; in 2010 the Coalition dropped a seat in Victoria and picked one up in South Australia). But when you’re talking about Labor vs Greens there’s a big difference. In the current Senate they have 31 and nine seats respectively, but based on the 2010 figures alone those figures are 28 and twelve.
In other words, although the Greens have nine sitting senators, they only need to hold their existing level (not just in votes, but in relative positioning against the other parties, including preferences) to emerge with twelve seats after 1 July 2014. Since the Greens have evidently declined somewhat from their peak of 2010, that number will probably come down – especially if Labor preferences are directed elsewhere.
Whether a Greens loss, however, represents an actual defeated senator, such as South Australia’s Sarah Hanson-Young, or merely the loss of a notional seat, as in New South Wales or Queensland, is in psephological terms a distinction without a difference. The individuals concerned, of course, might be forgiven for not seeing it that way.
Disclosure: I have done some consulting earlier this year on Senate preferences for both the Sex Party and the WikiLeaks Party, and would happily do the same for others if they paid.