There will not
be a double dissolution: there was never going to be. For my
detailed reasoning on that see my article "Do the sums: double
dissolution threat is hollow" published in The Australian on July 1,
2002.
These triggers were put there by John Howard purely for purposes
of propaganda. Since the gun was always pointed at his own head, the
Prime Minister was never going to be stupid enough to pull the
triggers.
I confidently predict that there will be a result, and that
result will lie somewhere in a range between a landslide to Howard
and a landslide to Latham. Those people who snigger when I say that
need to understand the true function of the Mackerras pendulum,
which makes very clear that a system of single-member electorates is
landslide-prone.
I have no idea how the Australian people will divide their votes
and nor does anyone else. However, this ignorance of mine does not
lead me to predict a close result. In truth, a close result is
merely one of the possibilities. It is just as likely to be a
landslide one way or the other.
Consider the Howard landslide first. In November 2001 the
Australian people divided their two-party preferred votes 51-49 in
favour of the Coalition. Suppose that in November 2004 the division
is 54-46, a swing to the Coalition of 3 per cent. My pendulum tells
me that the result would be a 42-seat Coalition majority which
compares with the present notional majority of 20 seats.
Sceptics will tell me this piece of arithmetic assumes a uniform
swing. Not so. It merely assumes that the deviations from uniformity
will cancel out. We know from experience they do cancel out.
For example, suppose Labor performs relatively badly in Victoria
and Queensland but relatively well in NSW and South Australia. It
could be that Labor retains its seats of Banks and Kingston, wins
its notional seat of Wakefield but loses its apparently safer seats
of Ballarat, Bendigo and Lilley in addition to its more marginal
seats of Brisbane, Bonner, Rankin and Chisholm. The numbers would
then be exactly the same as if the swing were perfectly uniform.
By the way, I do see Victoria and Queensland as being Labor's
difficult states, regardless of the size and direction of swing. I
agree with the conventional view that Labor has taken a big gamble
on Latham.
Suppose, however, that this year sees him connecting very well
with ordinary Australians and that his themes resonate well. Suppose
also that Howard is widely seen to be a dishonest, old, shop-worn
liar who has outstayed his welcome. And suppose that the situation
with housing affordability is so bad there is serious discontent
even with our strong economy.
In that event our nation could easily divide 53-47 in Labor's
favour, a swing to Labor of 4 per cent. My pendulum tells me Latham
would be our next prime minister and sitting on a 24-seat majority,
the best result for Labor since Bob Hawke defeated Howard by 24
seats in July 1987.
So much for the Mackerras pendulum as a tool of overall election
understanding. Now for some technical details about the pendulum and
the accompanying tables.
The four seats held by "others" (Calare, Cunningham, Kennedy and
New England) are shown two seats each on either side. That is
sensible and tells us Howard enjoys a notional majority of 20 seats.
The Prime Minister keeps telling his supporters that his
Government is "eight seats from oblivion". That is a propaganda
statistic designed to make his election victory (if he wins) look
better than it really would be.
I do not know where he gets that number. For my part I would say
"11 seats from oblivion". Presumably he has ignored the fact of
major boundary changes in Victoria, Queensland and South Australia.
Presumably also he assumes that the members for Calare, Cunningham,
Kennedy and New England are all equally passionate Howard haters,
all of whom would bring him down if they get the chance. A look at
Table 1 suggests how he gets his statistic.
For me it is far more sensible to say that a Labor government
takes office when Labor makes a net gain of 11 seats from the
Coalition. That is why the 11th weakest Coalition seat is my
"point-of-the-pendulum" seat.
It is Eden-Monaro (NSW) for which Labor needs a swing of 1.7 per
cent. Coincidentally, it is the case that Eden-Monaro has been won
by the nationally winning party at every general election since
1972.
Now notice something else. Only one of those 11 seats
(Parramatta) is in Latham's western Sydney territory. Furthermore,
we need to go all the way up the pendulum to Lindsay on 5.5 per cent
to find another case.
For the record, I should explain how I have placed the seats of
Calare, Cunningham, Kennedy and New England.
The Calare placement is determined by the shares of the
two-candidate preferred vote between the Independent winner, Peter
Andren, and the losing National Party candidate.
Cunningham was won by Labor in November 2001 but lost to the
Greens at a by-election in October 2002. The statistic shown is that
needed by the Liberal Party to win Cunningham from Labor, based on
the 2001 general election vote.
Both Kennedy and New England are normally strong National Party
seats. Their placement is determined by the swings Labor requires to
win the seats from their notional National Party winners.
I seem to be using the word "notional" a lot. It is normally used
when there have been boundary changes.
In Victoria, McMillan is held by Labor but becomes notionally
Liberal as a consequence of boundary changes. In Queensland, Bowman
is likewise. However the new seat of Bonner is notionally Labor so
Queensland Labor stays at seven seats.
South Australia loses a seat, Bonython, which is abolished.
However, much of Bonython moves into Wakefield, which therefore
changes from actually Liberal to notionally Labor. I confidently
predict that Wakefield will be won by the sitting Labor member for
Bonython, Martyn Evans.
In terms of seat names coming and going, only Labor is affected.
Labor loses Bonython (SA) and Burke in Victoria but the two new
seats are both notionally Labor, Bonner (marginal, Queensland) and
Gorton (safe, Victoria).
Consequently the different rows for Victoria, Queensland and
South Australia in Tables 1 and 2 are explained by the above.
However, on these tables I should make a further explanation about
NSW and the Northern Territory.
In the case of NSW the rows for the two tables are the same. I
did think of the idea of showing NSW Labor as 19 in Table 2 because
of the Cunningham loss at by-election. However, since the heading is
"Seats Won, 2001, Notional to New Boundaries" I have shown the NSW
rows to be the same in the two tables.
In fact, the NSW distribution now is 21 Liberal, 19 Labor, seven
National, two Independents and one Green in Cunningham.
For an entirely different reason, the rows for the Northern
Territory in the two tables are the same. The actual situation is
that the population growth of the Northern Territory during this
parliamentary term has been slower than that for Australia as a
whole. Consequently, the statistician has told us that the Northern
Territory loses its second seat, reducing the total size of the
House of Representatives to 149 members.
At the 2001 election, Warren Snowdon won Lingiari (NT) for Labor
with a majority of 4631 votes while David Tollner won Solomon (NT)
for the Country Liberal Party by a mere 88 votes. So the Liberal
Party would lose a seat.
However, it has become clear that the Liberal Party has found a
contrivance whereby the second NT seat can be saved. The contrivance
in question has not yet been implemented by legislation but I have
no doubt that will be done well before we head to the polls.
Consequently I show the Northern Territory seats as being unchanged
because that is what I expect.
Associate Professor Malcolm Mackerras teaches in the school of
politics at the Australian Defence Force Academy in Canberra.
PRINT-FRIENDLY VERSION