by Malcolm Mackerras

As I have indicated in past articles for this website, when it comes to commenting on election results I prefer to KNOW the result before commenting. This taste of mine applies most recently to the Griffith by-election held on Saturday February 8.
 
At the September general election last year the result after preference distribution was 45,805 votes (53 per cent) for Labor’s Kevin Rudd and 40,604 votes (47 per cent) for Bill Glasson, the candidate standing for the Liberal National Party. That represented a swing of 5.4 per cent against Rudd.
 
The final by-election result after preference distribution is 40,049 votes (51.7 per cent) for Labor’s Terri Butler and 37,340 votes (48.3 per cent) for Glasson. In other words there was a further swing of 1.3 per cent against Labor.
 
In my Switzer article, published on Friday January 24 my prediction was that the vote would divide between Butler and Glasson 53-47, the same as at the general election. Essentially my prediction was close to the mark and, in that article, I gave reasons to justify my arithmetic.
 
But the question which now arises is: who really won? The answer is that Labor won but it was important for Labor that they could be seen clearly to have won on the night. Had it been close enough to be in genuine doubt on the night it would have been possible for Liberal spin doctors to make great play on the fact that there was a further swing in their direction.
 
I am sure readers know the line of the spin doctors: it is asserted that the average by-election swing over the past fifty years has been four per cent against the government of the day. Therefore, it is also asserted, Labor needed at least 57 per cent of the two-party preferred vote to be able to claim to have had a respectable result.
 
Since Labor’s final percentage was five points lower than what it should (on this argument) have been, Griffith has generally been taken to be a bad result for Labor which, when combined with bad national polls, spells trouble for Bill Shorten.
 
I cannot say I agree with this reasoning. In my opinion by-elections over the past four years have not shown swings against the government of the day. In every case the by-election has been caused by a resignation and in almost every case the swing has been against the party holding the seat.
 
Since Griffith has been the sole federal by-election my proposition can only be illustrated by looking at state by-elections. When that is done the Griffith result simply becomes standard for recent by-elections.
 
Going round the states, there has been only one by-election in Queensland. When Anna Bligh lost office she immediately resigned her seat of South Brisbane. Labor retained the seat but, in a result remarkably similar to Griffith, suffered a further swing against it. Incidentally state South Brisbane lies wholly within federal Griffith.
 
There will be a state by-election in Redcliffe next Saturday. Labor needs a swing of 10.2 per cent to re-gain the seat which was Labor’s until the Bligh disaster of March 2012. I believe Labor will win Redcliffe in a large swing which will be both against the government of the day and also against the party of the resigning member.
 
In New South Wales there have been six by-elections in the period I am considering. In four of these there was a swing against the party/independent member of the resigning member, in Penrith in June 2010 under Labor (when the swing was 26 per cent) and, under the Coalition, in Clarence, Northern Tablelands and Miranda. In this last case the swing was 28 per cent, a record.
 
The two cases when the resigning member felt vindicated were the resignations of Kristina Keneally in Heffron (who, unlike Bligh and Rudd, remained in her seat for two years) and clover Moore in Sydney whose resignation was, effectively, forced on her by the O’Farrell Government.
 
In Victoria the two most recent by-elections saw swings against Labor when members resigned. In Melbourne in July 2012 the swing to the Greens (in a seat where the contest is always Labor versus the Greens) was 4.7 per cent. In Lyndhurst in April 2013 the swing to Family First (whose candidate became the de facto Liberal candidate, the Liberal Party having decided not to contest the seat) was 6.9 per cent.
 
In South Australia in the term just ended there were two by-elections, caused by the resignations of former premier Mike Rann in Ramsay and former deputy premier Kevin Foley in Port Adelaide. Both by-elections were held on the same day in February 2012. Since the Liberal Party did not contest either seat independents became de facto Liberals and the swing was against Labor.
 
There has been no by-election during the current term in Western Australia and, of course, Tasmania does not have by-elections under the Hare-Clark system.
 
Consequently my proposition is that the normal by-election these days sees a swing against the party of the resigning member. This is partly due to the loss of his/her personal vote and is partly a protest by ordinary people against the member who promises to serve a full term and then breaks that promise.
 
Looked at in this way I think a case can be made that Tony Abbott and Bill Glasson bungled Griffith by passing up a golden opportunity to humiliate Shorten. As I pointed out in my January 24 article it is largely chance that creates the fact that we need to go back to 1920 to find a case of a conservative federal government taking a seat from a Labor opposition at a by-election
 
Since then there have been three cases of it coming close to it happening,  the first Griffith by-election in 1939 (which Labor retained by just eight votes), the first Bendigo by-election in 1960 and the second Bendigo by-election in 1969.
 
Liberal spin doctors should be mourning a missed opportunity rather than celebrate a result which, in the very short term, looks good for them. In any event Labor now has, in Terri Butler, a personable young new member and she will very quickly establish a personal vote of her own.
 
(Malcolm Mackerras is a visiting fellow at the Public Policy Institute in the Australian Catholic University’s Canberra campus. malcolm.mackerras@acu.edu.au)