MOST people are expecting a new election for the Senate in Western Australia following the debacle of the lost 1370 ballots in an election that was decided by fewer than a score of votes. But do not be so sure.
It is quite possible that the $13 million cost can be avoided and that a new result emerge that the people should be reasonable confident about.
How so?
The High Court of Australia, sitting as the Court of Disputed Returns, resumed its sitting in the case this week. It is hearing a critical preliminary point in deciding what do about the challenged election.
The 1370 ballot papers were lost by the Australian Electoral Commission sometime after the first count and before the second count.
Also in dispute are 946 other ballot papers. They are in dispute over whether they satisfy the Australian Electoral Act’s requirements for formality because of the way voters marked them. Various candidates assert that some declared formal should have been declared informal and vice versa.
They can be dealt with in a fairly straight-forward way. If enough move from a successful to unsuccessful candidate, the court can reverse the result without the need for an election.
The missing votes are different.
At first glance it should be obvious. The election was decided by a handful of votes (on both the first count and the recount). The handful of votes were fewer than lost number, so the result of the election was tainted and there should be a new election because some voters were disenfranchised.
This is what happened in the Queensland state election in the seat of Mundingburra in 1995. The seat was won very narrowly by Labor, but some votes of soldiers serving overseas were not included. There were more of them than the winning margin, so the Court of Disputed returns ordered a new election. It was won by an independent who supported the Coalition, thus ending the Goss Government which had a majority of just one.
But the WA Senate case is different from Mundingburra. In Mundingburra, the soldiers’ votes were never opened and never counted. No-one had any idea how they voted.
In WA, however, these 1370 votes were counted in the first count and recorded electronically, but because the ballot papers were lost, they were not included in the recount. However, from the electronic counting record it is possible to determine precisely how those people voted.
Overall, at the first count Labor’s Louise Pratt and Zhenya Wang from the Palmer United Party won the final two spots – the fifth and sith seat.
At the recount asked for by the Greens, the Greens sitting senator, Scott Ludlam, and the Australian Sports Party’s Wayne Dropulich were elected in their stead.
In that recount, 14 votes moved from the Fishers Party, 12 of them to the Christians. That changed the preference flow and caused Pratt and Wang to be excluded before Ludlam and Dropulich.
The first four Senate positions went three Liberal and one Labor. But those four are caught up and are parties in the court because if a fresh election is called for all Senate seats, they will have to fight another election.
In court this week the Australian Electoral Commission argued for a new election, using the disenfranchisement argument.
But is that really necessary?
During the count, officials separate the above-the-line ballots from the below-the-line ballots. The Aboves are entered into a computer. It is just one number for each ballot – a fairly simple task. It transpired there were just 14 mis-entires in the more than 1.1 million votes. But they were critical.
The Belows, on the other hand, are each entered by two operators and mismatches are repaired immediately – virtually no chance of error.
Once entered electronically, some computer software churns through the preference flows and spits out the result.
In the recount only the above-the-line votes were recounted by re-entering them in the computer.
In the recounting process, the commission worked out where the missing votes came from and by comparing the booth-by-booth figures which were recorded electronically in the first count, worked out the actual votes on the missing ballot papers: Liberal 886; Labor 164; Greens 112; Animal Justice 21; and then 11 parties each with fewer than 15 votes. All were above the line. No below-the-line votes went missing.
In the first hearing before Christmas, Justice Kenneth Hayne put the pertinent questions. Can the court take evidence of how those missing votes were cast? He said that that was a matter of law.
Then two things flow from that. If those votes would have resulted in a different result at the recount (almost certainly), does the court as a matter of law have to declare the recount bad and order a new election? Or can it instead add the missing votes to the recounted votes and get a whole new result?
That last path seems to me to be practically and fairly the best course. But whether it is one that the law permits is another matter. One of the difficulties with electoral law is that in the quest to be cut and dried and utterly certain, there is no room for good sense and discretion.
Applying the missing votes (and the preferences that flow from them) to the recounted above-the line vote and the already counted below-the-line vote would correct both errors and give an accurate result as if no ballot papers went missing.
The acid question is how reliable is the evidence of how the missing votes were actually cast.
Some of the parties suggest there is a very high degree of accuracy. This is because just 14 votes were found to be wrong in 1.1 million – so how many are likely to be wrong in 1370? None or one, a statistician would say. But a statistician is not a lawyer.
Aside from the $13 million cost, there is the question of to the fairness to the top four candidates who were clearly elected. Why should they have to run again?
A partial election is simply not possible because the major parties would take the seats which would not reflect total electoral will.
It is unfortunate for the Australian Electoral Commission that the missing votes happened in such a tight election. But that’s life.
Justice Hayne asked (again pertinently) whether the telling of how the 1370 votes were cast and what effect they would have had on a recount was a matter for hazy expert evidence or whether it could be shown fairly conclusively from the Bar table as a logical deduction.
If the latter, of course, logic would demand that they be applied to the recount and the new result be the result of the election without a new vote.
If that happened, the 886 Liberal votes would be decisive. Of the four affected parties, the Liberal preference ticket put the Sports Party first, then Palmer, then Labor and then the Greens.
It could mean that the Palmer candidate Zang would get the seat instead of the Greens’ Ludlam and the Sport Party would get the remaining seat. And it would be by a few hundred votes, enough to have reasonable confidence that that was how the votes fell.
You could fairly easily work out how the vote would fall if you had a spare few hours.
Of course, the whole exercise calls out for full-scale reform of the Senate voting system. The Palmer and Sports Party got so few first preferences, that no-one can seriously argue they deserve a place in the Senate. But that’s another story.
CRISPIN HULL
This article first appeared in The Canberra Times on 1 February 2014.