We have learned (or had reinforced) several things about the ACT electoral system in the past couple of weeks.
First, there is an unacceptable level of chance in the system. This time it favoured older men against younger women in the Liberal Party and it favoured candidates on the right of the Labor Party over the left. But it was just luck that it happened that way.
Second, the system gives voters an opportunity to promote women in a way not readily available in other systems, and that many voters, presumably women took advantage of it.
Third, it is a complicated system, both in the voting and the counting.
Let’s deal with these separately.
A frequent criticism of the Hare-Clark system is that it is complicated and too hard to understand. I think a lot of that is unfamiliarity. This is only the second Hare-Clark election. The system does not get the same level of criticism in Tasmania where it is used for much the same reason as in the ACT. A small place needs a form of proportional representation to prevent wild swings of seats from one party to another. Opposition Leader Wayne Berry, who has been a critic of some elements of Hare-Clark, acknowledges this. Also, pure proportional system across the whole territory or state would lead to some very fringe groups getting in with a quota of 1/17th of the vote, or 5.9 per cent.
Complexity should not disqualify a voting system. One of my first cars, a Morris Minor, had a simple engine. Even I could pull it apart and (usually) get it back together. I now look under the bonnet of my Subaru and cannot hazard a guess at what most of the cables and boxes do. But I know which is the superior engine and which does the better job.
More importantly, independent mechanics who do understand the intricacies of the Subaru engine will universally tell you it is superior to the Morris Minor engine in design for the job.
Few people understand the count for the Senate voting system, yet it is universally regarded as fair. Indeed, before World War II under much simpler systems there were bizarre outcomes in the Senate, sometimes with one party taking all the seats in one state.
Complexity should not preclude fairness. Not many people understand how the AFL or rugby league finals series work. And even fewer understand the weighting of raw probability of getting into the grand final against position on the ladder at the end of the season. Yet both those systems are exquisitely fair. The higher up the ladder a team is, the better chance it has of getting into the final. The complex design of the finals series achieves both fairness and purpose (providing entertainment).
Incidentally, the finals series have something else in common with the Hare-Clark system. The outcome of a tussle between two teams, say Team A and Team B, has a direct bearing on the fate of a third team, say Team C.
In the AFL Essendon (5th) might stay in the finals series if Collingwood (2nd) beats Carlton (6th), but be out if Carlton wins. Just as Simon Corbell awaits the outcome of the tussle between Marion Reilly and Steve Garth.
I think, though, that we have seen an unacceptable level of chance in the Robson rotation element of the Hare-Clark system that should be fixed.
Berry sees the solution has having a party box. If you tick the party box you get the party’s candidate order. I think that in the ACT context, this gives too much power to too few people.
It may be fine at the national level for the Senate where there are a very large number of party members and a fairly rigorous system of pre-selection. In a small place like the ACT where the party has only 1000 to 1500 members it is too easy to hijack the pre-selection and concentrate too much power in too few hands.
It is more democratic to put it in the hands of voters. This election quite a number of voters exercised that option, particularly in the absence of how-to-vote cards at the polling booth and in the absence of any formal preference of candidates by the parties themselves. They relied on voters to chose the order of the parties’ candidates.
Unfortunately, the absence of the parties’ preferences meant that where there were no strong sitting members, some voters opted for the easy way. They went into the polling booth, choose the party they wanted and marked 1 to 5 or 1 to 7 down that column as it was presented to them. I call this a party linear vote rather than a donkey vote (because the voter has deliberately chosen a party and is not a donkey who would vote top left to bottom right through the whole ballot paper).
Under the Robson rotation system each candidate gets an equal number (give or take a few) of ballot papers with his or her name at the top of the column. Before the ballot papers are printed each candidate is assigned a number, according to how many candidates there are in the column, typically, 1 to 5 or 1 to 7, but sometimes fewer. Then the candidates names are printed in the order shown in the tables.
This can have a perverse effect if a significant number of people do a party linear vote. And on February 21 quite a lot of people did. Go to the five-column table. Just say Candidate 2 gets most votes, followed by 4, 3, 1 and 5. So 5 will be excluded first and then 1. But whenever their names appear at the top of the column you can see that 3 appears ahead of 4. So 3 will get a big flow of preferences. Even though 3 gets fewer votes than 4 from the discerning people who have bothered to deliberately order their candidates, 3 wins the seat because the preference order which in turn depends on the random ballot of assigning numbers to candidates. That is not desirable.
There will always be an advantage to one candidate or another where there are an uneven number of candidates. No-one knows how the advantage falls until after the primary vote comes in.
This election the party linear advantage probably determined four seats. On the Liberal side, Vicki Dunne lost to Harold Hird and Louise Littlewood to Trevor Kaine. On the Labor side Andrew Whitecross and probably Simon Corbell lost to John Hargreaves and probably Steve Garth.
The perverse system could be fixed by increasing the number of sample ballots so two randomisations are printed for each top candidate.
Sure, this time it hurt women, but that need not be the case. Indeed, this election the party linear effect was considerably lessened because many voters deliberately favoured women in giving not only their first but also subsequent preferences. Marion Reilly got more than double the preferences than otherwise expected after Tania McMurtry was excluded.
I have got several very large spreadsheets from the ACT Electoral Commission that detail the count (they are published in a booklet after the election). They show many examples of women getting more preferences from other women than you would expect. And scrutineers say there were many ballot papers with numbers only against women candidates across party lines. It can only be explained as a women-voting-for-women effect. Such a strategy can only be effective in a system like Hare-Clark where you can vote all over the ballot paper and where the easy way out of a party box is not available.
I’m sure there is a PhD thesis in it. For now, I have run out of space.