Financial Review  
July 10 2004
Home  |  Register  |  Subscribe  |  Help  |  Feedback     Legend:   Premium = premium  |  Logout  

C O M M E N T   A N D   O P I N I O N 

Lies & Statistics
Jul 10 2004
Feedback Peter Brent

How many votes does Mark Latham need to win the next election? The Morgan poll on Friday had Labor on 51.5 per cent two-party preferred (against the coalition's 48.5 per cent). Would that be enough?

The Mackerras pendulum puts the ALP's required vote at 50.7 per cent. But before the 1998 election the pendulum predicted Kim Beazley would win with only 50.3 per cent; he got a touch more than 51 per cent and was still seven seats short.

There have been 22 federal elections since 1949. Imagine sorting them, from largest to smallest, by the two-party preferred vote of the winning side.

Atop the list, with 56.9 per cent, is Harold Holt's 1966 victory. John Howard in 1998 sits at the bottom with 48.9 per cent. Four others are below 50 per cent.

But the vote per se doesn't tell us how close the contest was.

For instance, when the ALP got 49per cent in 2001, that, in uniform terms, was 1.7 per cent short of victory.

And in 1998 Beazley would have needed 52 per cent to win.

We could instead sort our list by the two-party preferred vote that would have produced a different election result.

We then get a more interesting story. At one extreme, the coalition made best use of its votes in Robert Menzies' 1955 landslide when it needed only 46.5 per cent. (It got 54.2per cent.)

At the other end was Bob Hawke's 1987 victory, when Labor received 50.8 per cent but anything over 47.4 per cent would have sufficed.

At only two other elections since 1949 would the ALP have won with 50 per cent of the vote or less, and Hawke was PM at both. Coincidence? No - the benefit of incumbency.

Governments win with less than half the vote when they do well in the marginal seats, and vital to this is the types of people who tend to live there.

Marginal electorates fall into two broad categories. One is the relatively affluent outer suburban mortgage belt. People here are difficult to shift, especially early in a government's term or if the economy is healthy.

The other is a less homogeneous clutch of electorates, but they're outside the cities, the median incomes are lower and the four-wheel drives actually see dirt.

Both groups have stuck with Howard for three elections.

Now back to Latham. History alone tells us that Labor will probably need significantly more than 50per cent to win, a situation exacerbated in the mortgage belt by the economy; punishing the man who presided over a tripling of your house value doesn't come easily.

But the other side of the marginal coin is more volatile, and contains a potentially cranky demographic similar to that which surprised everyone by tossing out Victorian premier Jeff Kennett in 1999.

There are 12 government-held regional seats with margins of less than 3per cent but only seven urban ones.

The regions are Latham's best chance. But if he makes no ground there, he could lose with a very high vote indeed.

Peter Brent is editor of

[Go to related tables]

Email to a friend Email to a friend
Email to a friend Printer friendly version

stock quote
Stock code:
  Code look-up
Select details:
news search Keywords: 
Advanced search

Subscriber log-in

f2 Network Privacy Policy | Conditions of Use | Member Agreement | Copyright