Crikey psephologist Charles Richardson has been studying all the polls and preference deals for the Senate and has produced some firm predictions on how the state’s house will look after October 9. See the completed series below, including his final national Senate predictions:

Victorian Senate

The Victorian Senate race is broadly similar to New South Wales. Labor and the Coalition each have a chance at 3 seats (in 2001 they both got there), and the Greens and now Family First are their main challengers for the last two places.

There are three key differences from NSW: the total left vote in Victoria has been a good deal higher in recent years; there’s no Fred Nile in Victoria pulling in a strong fundamentalist Christian vote; and because the ALP’s third candidate, Jacinta Collins, is herself on the hard right, she is getting preferences from the religious groups, including both the Catholic DLP and the Protestant Family First.

However, Family First have the same advantage as in NSW, that they will get preferences from a raft of small parties: the DLP, One Nation, Liberals For Forests, Meg Lees’s group, and the Christian Democrats. Even if their primary vote is only 2 or 3 per cent, they are very likely to be in the last four, along with the Greens, the third Coalition candidate, Judith Troeth, and the ALP’s Jacinta Collins.

None of these will have a quota on their own, so the result will depend on who gets eliminated. If Troeth goes out, her votes will elect Family First’s Steve Fielding, and the Greens and ALP will fight out the last vacancy (with the Greens as favorite). Vice versa if Collins goes out: Family First again gets up, and the Greens have to beat the Liberals for the last place.

(So much for the great ALP-Greens preference deal: it only took Labor a day to rat and give key preferences to the religious fundamentalists instead of their supposed allies.)

If Family First comes fourth, their preferences will scatter (because they come from many different primary sources), but the most likely result would be election of both Labor and Greens, with Senator Troeth missing out. And if the Greens were to be eliminated, their votes would flow to Labor, and Labor’s surplus would then elect either Troeth or Family First.

All those scenarios are possible, but they point to the most probable winners being 2 Labor, 1 Liberal, 1 National Party, 1 Green, and 1 Family First (Assemblies of God).

NSW Senate

New South Wales is often the most interesting Senate race. It has the most votes and the most candidates, so it takes the most work to analyse. And it’s the only state, since the expansion of the Senate in the 1980s, that has ever thrown up anything other than a 3/3 left/right split in a half-Senate election.

NSW has done so twice: in 1990, when right-wing groups directed preferences away from Chris Puplick and elected a Labor senator in his place, and in 1998 when the Coalition directed preferences away from One Nation and elected Democrat Aden Ridgeway instead of David Oldfield.

This year, preferences are even more confusing than usual due to the presence of the Families First (Assemblies of God) ticket, which has had remarkable success in sewing up preference deals. Assemblies of God will receive preferences from Hemp, No GST, Lower Excise, Australians Against Further Immigration, the New Country Party and Liberals for Forests.

These are all tiny parties, but between them in 2001 they had a fifth of a Senate quota. That is a big chunk of votes. If those votes are enough to put Assemblies of God ahead of its minor-party rivals, then further preference deals will come into play: from Fred Nile (Call to Australia), One Nation and the Australian Democrats. In that event, it is very likely that these parties between them would have enough to elect a senator ahead of the third Coalition candidate, Senator John Tierney.

If Assemblies of God are eliminated too early to benefit, then Fred Nile could conceivably be elected instead, since he will also get a third of any preferences from the ALP. But the Democrats cannot perform the same trick, since Nile and One Nation both flow to the Coalition ahead of them.

None of this would disturb the 3/3 left/right split; it is likely that Labor and the Greens will still have 3 quotas between them, electing 2 ALP and 1 Green. But if Labor is doing badly overall, Tierney and Joan Wood, the Assemblies of God candidate, might both sneak in (one of them getting the surplus votes of the other) at the expense of the Greens.

So my call is 2 Coalition, 2 Labor, 1 Green and 1 Assemblies of God, but the third Coalition candidate and Fred Nile both in with a chance.

Political editor Christian Kerr adds:

Family First are organised. Organised and cashed up. They are wheeling out a series of television ads that will start running from the AFL Grand Final at a media event in Adelaide tomorrow and promising to reveal “unique preferencing arrangements for the House of Representatives”. The plot thickens.


Queensland is traditionally the Coalition’s strongest state. In 2001 it elected 2 Liberals and 1 National (the parties run separate tickets), and it is probable that they will again have three quotas, or very close to it, between them. It is just possible that the Liberals could win three seats, but to do so they would have to get five times the Nationals’ vote: the fact that this is conceivable is a measure of what bad shape the National Party is in, but it still seems pretty unlikely (in 2001 it was not quite four to one).

The ALP, of course, will win 2 seats, and the sixth is being contested by a bewildering array: the third Labor candidate, the Greens, Democrats, Family First, One Nation’s Len Harris, his former leader Pauline Hanson (now an independent), and independent child abuse witch-hunter Hetty Johnston.

Len Harris can be ruled out immediately. Family First’s prospects also look hopeless: Hetty Johnston and the Democrats are exchanging preferences, which will lock them out, and Johnston is also likely to eat into their support base. Johnston will also get preferences from Hanson and One Nation, so if she has enough primary votes to stay in contention it is possible that she will get over the line.

If Johnston is eliminated, her preferences will flow (after Democrats & Family First) to Pauline Hanson, which could conceivably be enough to elect her, but only if she has done unexpectedly well on primaries. As Malcolm Mackerras has pointed out, both Hanson and Johnston will be handicapped by having no party identification on the ballot paper.

The Democrats will not get preferences from Hanson, but could benefit indirectly: if they are still in when she is eliminated, her preferences will go to the Coalition, and their surplus would then elect Democrat John Cherry over the Greens’ Drew Hutton.

None of that will happen, however, if Labor and the Greens can garner three quotas between them. Even if they don’t, the Greens (who look like being well ahead of the third Labor candidate) should still win a straight contest against Pauline Hanson. It is even possible that Drew Hutton and Hetty Johntson could both get up, at the expense of the Coalition’s third seat.

In other words, Queensland is a real lottery. My best guess is 2 Liberal, 2 ALP, 1 National and 1 Green, but the Democrats and Hetty Johnston are both in with a chance, and a rough shot for the famous red-headed one.

WA Senate

Western Australia was another good state for the Coalition in 2001, and the polls suggest that Labor has not made up much ground. There will probably again be three Liberal senators elected; the National Party runs a separate ticket, but its only function is to funnel a small number of preferences to the Liberals.

Labor will elect two senators, but a third seems beyond them; the Greens are the only ones giving them preferences, and they are confident that their Rachel Siewert will be well ahead of the third Labor candidate, Emiliano Barzotto.

Also in the running will be Families First (Assemblies of God) and the Australian Democrats, who are once again swapping preferences. Most of the smaller parties are also favouring one or both of them.

Remember the start of the campaign, when Democrats & Greens proudly announced they were preferencing each other across the board? Yet now the Democrats are parties to a preference strategy clearly designed to lock out the Greens. Democrat Senator Brian Greig must be starting to worry about how far to the right his party is drifting: the Democrats are even getting One Nation preferences ahead of the Coalition.

Antony Green says that “a victory by Greig requires him to pass the Greens but still be behind the third Labor candidate.” He is right to say that is unlikely, but there is another possibility: Democrats, One Nation and Family First, together with some preferences from micro-parties and surplus from a major party, might put together a quota between them.

In that case, whichever is ahead out of Democrats and Family First (probably the former) would be elected at the expense of either the Greens or the third Liberal. While this is a real possibility, it seems more likely that the Coalition and Labor/Greens will each have three quotas, so electing 3 Liberals, 2 ALP & 1 Green.

SA Senate

South Australia is on all accounts the strongest state for the fundamentalist Assemblies of God party, Family First. But because their real strength comes from preference deals, not primary votes, that does not necessarily make it their best prospect for winning a seat. Two things work against them in South Australia: firstly, the Coalition will probably have something close to three quotas, and so very little surplus to distribute (it was 3.19 in 2001, and their support seems to be dropping); secondly, it is also the Democrats’ strongest state, raising the possibility that the Democrats, with preferences from their renegade senator, Meg Lees, could get ahead of Family First and knock them out.

Ironically, the backlash against the Democrats from their preference deal with the fundamentalists may be the very thing that drags down their vote and allows their preferences to put Family First within reach of a seat!

As in Western Australia, it seems likely that the ALP and the Greens, with a few preferences from minor parties (Greens from the Trots, Labor from, of all places, One Nation), will have 3 quotas between them. If the Greens are ahead, they should get up on Labor preferences. But if the Greens go out, their preferences will go first to the Democrats, which would elect them, and the surplus (coming partly from Lees and Family First) would elect the third Liberal instead of the ALP.

So while the 3/3 left/right split will probably be preserved, the third ‘left’ spot could be any of Greens, Labor and Democrats, while the ‘right’ could be either a third Liberal or Family First. There is even a slight chance, if the Liberals are stranded just below 3 quotas, that the Democrats and Greens could both get up: something that has never happened before in any state.

Still, the single most likely outcome is the rather boring one of 3 Liberals, 2 ALP & 1 Green. It just might be a bit of a roller-coaster ride to get there.

Tasmanian Senate

Superficially, Tasmania is very straightforward. The Liberals and ALP will each have somewhere around two and a half quotas, and the Greens will have about one or a bit more. It’s the one state where five seats are clear: 2 from each of the 2 majors, and the Greens’ Christine Milne.

The sixth seat is the interesting one. The minor candidates are the Democrats’ Yulia Onsman, Assemblies of God’s Jacquie Petrusma, and Labor-turned-independent senator, Shayne Murphy. These three are mostly exchanging preferences, and it seems probable that between them they will outvote either or both of the third Liberal & ALP candidates.

The Democrats’ chances can be disregarded, since they are likely to be behind the other two & do not get Murphy’s preferences in any case. But either Murphy or Assemblies of God could put together a quota with preferences from the other small parties, plus whichever of the major parties is eliminated.

What makes Tasmania difficult is that its voters do not necessarily follow the tickets. In every other state, well over 90% of voters vote “above the line”, so their preferences are distributed automatically. In Tasmania last time it was only 80%, and much less for the minor parties – for the Greens, an extraordinarily low 57%.

That will hurt anyone relying on multiple preference flows from smaller parties. Even so, it will be a big ask for either of the major parties to get a third senator up, and an even bigger one for the Greens to get a second (something they occasionally fantasise about).

Murphy & Assemblies of God are better placed for the sixth seat, and if Murphy can get a reasonable primary vote – say something over 3% – he could be the one. He’s no Brian Harradine, but Tasmanians like independents, and the other choices are uninspiring.

Besides, your psephologist has been itching to say something controversial, so for Tasmania my forecast is 2 ALP, 2 Liberals, 1 Green, & Shayne Murphy.

The Territories

The Northern Territory is the easy one. As Antony Green says, “No need to look at the preference tickets.” The sitting senators from the Country Liberal Party and the ALP will each have more than a quota on primary votes, as they always do. The ACT, however, is much more interesting. Although the Liberals have always got a senator elected, it’s usually without much to spare. In 1998, Democrat Rick Farley came within about 3% of knocking them off: if he’d had One Nation preferences he would have won.

This year, the threat is from the Greens, whose candidate is local councillor Kerrie Tucker. She will get preferences not only from the ALP, but also from the Democrats and Meg Lees’s breakaway Australian Progressive Alliance. The only preferences that Liberal Gary Humphries will receive are from Fred Nile’s group. (Assemblies of God are not running in the territories.)

It could be close, and it is especially important this year because the Coalition has its best chance in a long time of winning effective control of the Senate. Defeat in the ACT would be a big setback for that – hence the importance that the Greens are placing on it.

To win, the Greens have to first get ahead of the second Labor candidate, and then keep the Liberals (plus Christians) below a third of the vote.

The first part should be fairly easy. In 2001, the Labor surplus was 0.26 of a quota, or a bit less than 9%. If the Greens, Democrats & Lees’s mob can’t do better than that between them, they might as well give up & go home. If anything, the Labor vote might go down this year, as Labor supporters vote tactically in an attempt to deny the Coalition control of the Senate.

So what it really comes down to is, can the Liberal & Christian vote be kept below 33.3%? In 2001 it was 36.1%, but there was also 2.2% for One Nation, who are not standing this year. So the swing that’s needed is probably 4 to 5%.

That’s not impossible by any means. My guess is that they’ll just fall short, and it’ll again be 1 ALP & 1 Liberal. But don’t be surprised if the Greens do sneak in.

National Senate predictions

Adding together all our state-by-state Senate predictions, we have the following Senate result (comparison with 2001 in brackets):

Liberal 15 (down 3)
Nationals 2 (no change)
Families First/Assemblies of God 2 (up 2)
ALP 14 (no change)
Greens 6 (up 4)
Democrats 0 (down 4)
Independent (Shayne Murphy) 1 (up 1).

The new Senate, including the long-term senators, would therefore have the following numbers from 1 July 2005:

Liberal 31
National 4
Families First 2
ALP 26
Greens 8
Democrats 4
Independent 1

That’s 38 left (Labor/Greens/Democrats), 37 right (Coalition/Families First), plus Murphy, who is somewhere in between but probably leans left. Neither side would have a majority, but the right in a sense would be better placed, since it would have less of a broad alliance to have to hold together: the ‘left’ total remember includes the likes of Democrat Andrew Murray as well as a fractious group of Greens.

As I have warned along the way, however, these numbers are highly speculative. I have tipped the Greens to win 6 seats, but at least 3 of them (Victoria, Queensland & South Australia) are touch-and-go – on the other hand, they could win a 7th in the ACT. Families First could win 3 seats or conceivably 4, but they could also miss out completely.

The Democrat vote is in free-fall, but they are still in the running in 2 states, Queensland & South Australia. The same uncertainty applies to the major parties. I have given the Coalition a 3rd seat in each of 3 states (Qld, WA & SA), but none of them is really safe. Yet if they do well, they could win not only those but also extra seats in NSW & Victoria.

Labor will not go below 14, but additional seats are possible in Victoria and South Australia. Even the traditional 3/3 left/right balance in each state is insecure: the preference deals between minor parties, often defying ideology, have thrown a number of wild cards into the game. It all goes to make this a very unpredictable election indeed.