A mid-point review of the state of the marginals:
With two weeks to go, almost every indicator shows that this will
be a very close election. It is tempting to defy the conventional
wisdom and predict a landslide one way or the other; every commentator
longs to be bold & unconventional. But the price of that fleeting
fame, as Gary Morgan discovered in 2001, is the risk of being
spectacularly wrong.

Your psephologist therefore is constrained to agree that this one
really is too close to call. Saturday’s Nielsen poll, which showed a 3%
swing to the government, is so far an isolated case, and has to be
regarded at least provisionally as a rogue sample (it does happen:
Newspoll had one back at the beginning of June).

When the margin is within a percentage point or two either way, an
election pretty much comes down to luck. Political parties don’t like
to admit this – they talk about “superior marginal seat campaigning” –
but in fact no-one knows what random influences might swing the last
handful of votes in a key seat.

To sound another clichéd theme, it brings up the real chance of a hung
parliament. No, it probably won’t happen. But commentators who go out
on a limb to dismiss the possibility, such as Peter Brent & Dennis
Shanahan, are being reckless. With three well-entrenched independents,
and the chance of more joining them, it is very possible that neither
side will command an absolute majority after 9 October.

So here goes, with a state-by-state breakdown of how things look heading into the final stage of the campaign.

New South Wales

New South Wales has been a disappointment to Labor in recent
years, with the glory days of 1993 now just a distant memory. The
polls, however, suggest that there is some movement in its favour this
time. There are targets in plenty: six Coalition seats with margins
below 3%. Of them, Parramatta & Richmond look the most likely to
fall. But beyond that might be difficult; Dobell & Paterson have
new sitting members, with the advantage of incumbency this time, while
Eden-Monaro & Page are the sort of semi-rural seats where
incumbents are notoriously hard to dislodge.

Partly offsetting those difficulties is the fact that Labor has
possible gains out of seats with nominally greater margins. Lindsay
(5.5%) is probably the best chance among them, although Robertson
(7.0%) should also not be written off. But at present you would have to
back the government to hold seats like that.

The only Labor seat at risk is Greenway (3.1%): they are worried about
it, but since the state as a whole seems to be moving in its favour,
they must be considered favourites.

Among the independents, Peter Andren in Calare & Tony Windsor in
New England are both safe. Green Michael Organ will probably lose
Cunningham to Labor, although it looks like being close. And Peter King
is in with a real chance in Wentworth (yes, I have changed my mind
about this), with Malcolm Turnbull perhaps just having his nose in

That amounts to a net gain of two seats for Labor, plus Cunningham from the Greens.


Commentators have so far tended to downplay the importance of Victoria
compared to New South Wales & Queensland. It is true that the polls
show very little movement either way, but there are plenty of marginal
seats in Victoria and in a close election they cannot be ignored.

The Coalition has 7 seats with margins under 6%. Even with no overall
swing in its favour, Labor looks set to take McMillan, notionally
Liberal (2.9%) only due to the redistribution. It has hopes of another
five: roughly in order of probability, McEwen, Corangamite, La Trobe, Deakin & Gippsland. Only Dunkley (5.2%) looks secure.

If a swing is on, some or all of those 5 will fall. At present, Labor
is odds against in even the closest of them. The sheer number of them,
however, makes it seem reasonable to think that Labor will gain one,
let’s say McEwen.

Further up the pendulum, there’s not much scope for surprises; Flinders
(7.4%) is the only one being talked about, but no-one really thinks
Labor will get that far. On the Labor side, there are five seats below
the 6% line, but the closest of them is Chisholm at 2.7%. Very few
Liberals rate it as a serious chance, although if the swing is towards
the government it could be close.

Prediction: Labor to gain 2.


It is being said on all sides that Labor must win seats in Queensland
to win government. Of course it is not true: it is easy to build a
scenario where Labor wins without any Queensland seats shifting. But
there is no doubt that big gains in the state would be a major plus for

So far, the polls have been equivocal, some suggesting that a swing is
on, others that Labor is failing to make headway. Queensland is also
the most difficult state for a Victorian to understand – or at least
this Victorian. So at this stage I am reluctant to predict anything

There are 7 Coalition seats at risk: 6 below the 4% mark, plus Dickson
(6%), which is less safe than it looks. Three of them look particularly
good prospects: Herbert (1.5%), the most marginal; Hinkler (2.2%), a
classic sugar seat; and Bowman (3.1%), a Labor-held seat that has
become notionally Liberal (and whose sitting member has fled to
neighbouring Bonner).

With no definite indications of a swing, however, it is best to be conservative, so I am tipping only one Labor gain, Herbert.

There are also 2 possible Liberal gains, Brisbane (1.0%) and the new
seat of Bonner (1.9%). I am tipping that both will hold, but not with
any great confidence.

Predicted gain to Labor: 1.

Western Australia

Most polls agree that Labor is not doing well in Western Australia.
Only one Liberal seat is seriously under threat – Canning (0.4%), held
by the execrable Don Randall – and it seems likely that it will hold.

Labor has three marginals that are causing it concern: Stirling,
Hasluck and Swan. The first two should hold – Hasluck because Labor has
a new sitting member, and Stirling because the Liberals have had their
own local problems. Swan, however, is looking like a probable Liberal

Prediction: Labor to lose 1.

South Australia

A popular first-term state government has made South Australia Labor’s
most promising state. There are 3 Liberal seats held by less than 4%
(Adelaide, Hindmarsh & Makin); the safest of them (Makin, 3.7%) is
well within the swing that the polls are indicating, and also has a
local member in particular difficulty, so I expect that all 3 will
fall. The Liberals, however, although they have pretty much written off
Hindmarsh, still rate themselves a chance in the other 2.

One other Liberal seat, Boothby (7.4%), has to be rated a chance, but I expect Labor will fall short.

Labor has 2 marginals, Kingston and Wakefield (the old Bonython), but
in view of its good position in the state it should not be seriously
troubled in them.

Prediction: 3 gains to Labor.

The rest

Tasmania has only one seat in play, the Labor marginal of Bass (2.1%).
It’s a bit of a toss-up, but the prime minister’s new-found enthusiasm
for forest protection probably isn’t helping him, so Labor must be
favourites to hold.

There are no marginal seats in the A.C.T., but the Northern Territory
seat of Solomon is the most marginal in Australia (Country Liberal,
0.1%). Labor is odds-on to take it this time.

Predicted gain to Labor: 1.

What does it all mean?

I am forecasting a net gain to Labor of 8 seats, plus Cunningham,
giving it 72, against 75 for the Coalition and 3 independents. Although
technically that is a hung parliament, in practice it means the
government would be returned: Labor could not get a majority, even with
all the independents, and the government would only have to persuade
one independent to become Speaker to give it a majority on the floor.
Tony Windsor would be the logical choice; he may not have made much of
an ambassador, but he would be a perfectly competent Speaker.

Even the best seat-by-seat prediction, however, is rarely spot on, and
two weeks out the chances of being exactly right are negligible. If a
couple more marginals fall over the line, or if Peter King gets up in
Wentworth, or if the swing is unexpectedly big in just one state, it’s
a whole different ball game.

What it really means is that this election is still up for grabs.
Campaigns usually don’t matter much, but when it is this close, the
last 2 weeks are going to decide it. If Latham continues to perform
well, he will emerge the victor.