Inside Story

Is the enemy of my friend always my enemy?

Do all Labor voters prefer the Greens to the Liberals? Do National Party voters opt for the Liberals if their own party isn’t running? What evidence we have suggests the answer isn’t straightforward, writes Paul Rodan

Paul Rodan 7 March 2013 1088 words

Greens senator Lee Rhiannon and her lower house colleague Adam Bandt. Greens MPs

Updated 9 August 2013


WITH the relationship between the Labor Party and the Greens more fraught then ever, it’s a good time to consider how Labor’s support base regards the party’s bedfellow on the “left” of Australian politics. Opportunities to quantify these attitudes are rare, but when Labor decides not to contest a by-election in a safe conservative seat, leaving the Greens as the main flag-carrier for the progressive side of politics, we get a glimpse.

In a detailed analysis of one such contest – the Higgins by-election in 2009 – Rob Hoffman and Brian Costar conclude that “in the absence of an ALP candidate, certain types of Labor voters will not vote Green regardless of the often-asserted ideological compatibility of the two parties.” At one level, this finding has very few electoral consequences. Unlike in Higgins, anti-Greens Labor voters will usually have a Labor candidate on the ballot, and will be able to vote accordingly. And, in Senate contests, all but the most energetic anti-Greens Labor voter will take refuge in above-the-line voting, where Labor can usually be relied on to rank the Greens above the coalition.

But at another level, problems arise if these voters see a too-close relationship between Labor and the Greens as a reason to reassess their core commitment to the ALP. If they drift to a conservative party as their first preference, then Labor is in trouble. However, in this context, one needs to look closely at the evidence.

On the other side of politics, the attitude of Liberal voters to the Nationals (and vice versa) receives little or no attention. Because the two parties are long-term coalition partners, there’s an assumption that supporters of each will automatically preference the other. But the available evidence paints a more complex picture.

Historically, the two conservative parties have had a sometimes difficult relationship, varying from state to state. This longer-term instability is easy to forget in an era that has seen the parties merge in Queensland and coexist harmoniously – more or less – elsewhere, especially at the federal level. Significantly, three-cornered contests, where the two conservative parties oppose each other, are now much more the exception than the rule, and usually only occur where no sitting Liberal or Nationals member is contesting the relevant seat.

The 2010 federal election saw a small number of such contests, with the results affording an opportunity to examine preference flows between the coalition partners (assisted by Australian Electoral Commission data, which is superior to that provided by state authorities for their elections). In seven seats, both the Liberals and the Nationals fielded candidates, and in six of these the eventual contest was between a conservative and Labor. In the other seat, O’Connor (in Western Australia), the final contest was between the sitting Liberal, Wilson Tuckey, and the Nationals’ Tony Crook, with the latter prevailing after securing 84 per cent of Labor preferences.

In the remaining six seats, a clear pattern was evident. In the four seats where Nationals preferences were distributed (Durack, Forrest and Pearce – all in Western Australia – and Throsby in New South Wales), the highest share managed by the Liberals was 80 per cent of the Nationals vote in Throsby, followed by 77 per cent in Durack and Pearce, and 73 per cent in Forrest. Averaged, this level of support is less than what the Greens deliver to Labor when they preference that party. It is perhaps understandable that the weak flows occurred in Western Australia, a state where there has been long-term conflict between the Liberal and National (formerly Country) parties, and where coalition relationships rarely qualify for the adjective “amicable.”

In two NSW seats, Richmond and Riverina, Liberal preferences were distributed in seats where the effective contest was Labor versus the Nationals. In each case, 89 per cent of the Liberal vote went to the Nationals, a figure consistent with what the Democratic Labor Party delivered to the coalition between 1955 and 1974. Put more bluntly, more than one in five Nationals voters prefer Labor to the Liberals. On this evidence, the Liberals are considerably more effective at disciplining preferences than are the Nationals.

Apart from long-term historical motivations, it is feasible that such Nationals voters have been disappointed by their party’s marginalisation within the coalition. The dominance of powerful leaders like John McEwen and Doug Anthony is long past, and when did the party last win a policy battle with the Liberals? Indeed, when did it last have distinctive policies over which to wage a battle? Independent candidates have successfully exploited such themes – at state and federal levels – in defeating Nationals candidates in the past decade or so.

A larger number of three-cornered contests would allow for more confident conclusions. But even this limited evidence provides a cautionary warning against assuming complete ideological affinity among the supporters of notional allies.


HOW is this interplay of preferences likely to unfold for the Coalition on 7 September? Although nominations don’t close until 15 August, we have a reasonable idea of the field in those seats where there will be three-cornered contests between the Liberals, the Nationals and Labor.

All three parties are contesting O’Connor. If Labor preferences the Nats again, we would expect them to win. But with the longstanding and controversial Liberal former MP, Wilson Tuckey, out of the picture, the Liberals may do better. The local popularity of the candidates could well play a key role.

In Mallee, the Labor vote was under 20 per cent last time (when no Liberal ran). Assuming the Liberal candidate this time can do better than that, Labor preferences could be decisive. A cautionary note: as Antony Green points out, in 1993 Labor voters ignored the how-to-vote card sufficiently to see the Nats beat the Libs. (Liberal voters in the seat of Melbourne might be equally fickle this time around!)

In Durack, another retirement means that the Nationals and the Liberals are both in the field, but we would expect the Liberals to lead easily on first preferences, Labor to scrape into second place, and Nationals preferences to see Liberal home.

A Liberal and a National are in the field in Throsby, but the seat is safe Labor, so the fact that the Nationals’ candidate is Angry Anderson offers the only element of analytical interest…

Finally, Forrest, Richmond and Riverina, scenes of three-cornered contests in the recent past, will be contested by only one of the Coalition parties (Lib, Nat and Nat respectively). •