We now have two Governments in Australia which got power with a minority of votes and which owe their existence to the donkey vote. The Carr Government in NSW and the Goss Government in Queensland both scraped home with one-seat majorities.
The counting in Queensland ended on Tuesday with the last seat won by Labor by 16 votes. In March, the NSW Labor Government won its majority with the last few seats being determined by a few hundred votes. Let’s go first to the minority-vote issue; then to the donkey vote.
In NSW Labor got a majority with only 48.8 per cent of the vote. In Queensland it got a majority with 47.7 per cent. (Estimates vary a little, but all are at least 1 per cent and 2 per cent under 50 per cent respectively.)
It is, of course, the result of Labor’s marginal-seat strategy _ concentrating campaigning where it matters and not wasting campaigning and votes in seats which are held by wide margins and are not likely to change hands, no matter what.
In 1990 it resulted in Labor getting government federally while getting a tad fewer votes than the Coalition after preferences. It also resulted in Labor spreading itself more thinly over successive elections as it ignored more comfortable seats, thereby converting them to marginals. The effect is illustrated by the increasing number of Labor marginal seats (held by under 5 per cent) from 11 in 1983 to 27 in 1990. (A large increase even allowing for an increase in the size of Parliament.)
Given modern communications and marketing techniques and increasingly sophisticated computer programming, marginal-seat tactics are likely to continue to cause distortions. Some changes are needed so that the national or state-wide sentiment is more accurately reflected, perhaps with an ACT-style Hare-Clark system or some form of top-up of nationally contested seats. At present, little or none of the Green, Democrat or Independent vote gets reflected in seats.
The Queensland and NSW position is no where near as bad as it has been in some states in the past _ South Australia under Playford, the Western Australian Upper House, and Queensland under Joh.
The latest results come not from some in-built unfairness that gives votes unequal weight or where boundaries are drawn irregularly to favour one party. Rather it comes from the ability of one major party to extract the most from what is on paper a fair system. None the less fairness in a system is not dependent solely its abstract formal structure. It is always also dependent on how it operates practically on the ground.
Though Queensland now is not as bad under as under Joh and not the result of deliberate rorting of the formal system, the words of Tony Fitzgerald (who inquired into government in Queensland) are still important:
“A fundamental tenet of the established system of parliamentary democracy is that public opinion is given regular, free, fair elections following open debate.
“A government in our political system which achieves office by means other than free and fair elections lacks legitimate political authority over that system. This must affect the ability of Parliament to play its proper role . . . .”
Now to the donkey vote. The donkey-vote problem is perhaps more pernicious than the marginal-seat effect because it is a function of the system and can be easily corrected without affecting fundamentals.
The donkey-vote is where some apathetic voters just number in order down the ticket. Its exact size is unknown but can be estimated when two utterly unknown independents run with one at the top of the ticket, and also by scrutineers who see the phenomenon when there is no how-to-vote-card explanation for it. Estimates vary up to about 1 per cent. Even at half that, it would be 175 votes in a NSW seat and about 100 votes Queensland. It means that a donkey vote could affect seats won by up to 350 and 200 votes respectively.
In NSW if just 131 voters had voted the other way in Gladesville, Carr’s majority government would be either in minority or in Opposition. In that seat, the ALP candidate drew a slot above the Liberal Party on the ballot paper and got the donkey vote. If the donkey vote had been just half of one per cent (very likely) then the result of the election would have been different. There were no counter-balancing seats where the donkey vote would have provided a Liberal seat that otherwise would have gone to Labor. That is the luck of the draw; but elections should not depend on it.
In Queensland, if just nine voters had voted the other way in Mundingburra the Liberals instead of Labor would have won the seat. Labor drew top spot on the ballot paper in that electorate with the Liberals at No 2. It is almost conclusive that the donkey vote decided that seat and hence the whole election. If Liberal Frank Tanti’s name had been drawn out of the hat before the name of Labor’s Ken Davies when ballot positions were being decided, Rob Borbidge would now be forming a minority government. Elections should not be decided that way.
The donkey vote most likely decided the seat of Whitsunday as well. In that seat Labor drew a higher spot on the ballot that the National Party. If just 29 voters had voted the other way, the Nationals would have won. There were almost certainly 29 donkey votes. That would have given Borbidge a majority government. Similarly, Mulgrave, requiring 96 changed votes, was also most probably was decided in Labor’s favour by the donkey vote.
Off-setting this is the Liberals’ win in Greenslopes which would have required just 33 voters to go the other way to have been won by Labor. The Liberals were higher on the ticket and would have got the donkey vote.
There are no other seats where you could argue that the donkey vote would realistically have changed the result. In summary, if you concede a donkey vote of just one sixth of one per cent, you have to acknowledge that Wayne Goss is Premier because of it.
Labor supporters might be saying thank God for the donkey vote; but elections should not be decided that way.
It could easily be fixed with the ACT’s Robson-rotation system, even in single member electorates. Several different batches of ballot papers would be printed for each electorate so that each candidate got a equal number of top spots. If that makes how-to-vote cards more difficult to work, too bad.