In the wake of the (still to be finalised) recent election results there has been quite a bit of commentary about making changes to the senate. I’m in fairly broad agreement with the opinions given by Antony Green here, but do have a few comments of my own to add.

So what is the problem? The main motivation for changes seems to be that candidates with very small primary votes are looking like winning seats. I would suggest that this on its own is not necessarily a problem. For example consider an election for 6 senators (so a quota of about 14%) where 30% vote Labor, 30% Liberal, 16% Green and the remaining 24% of the vote is spread roughly evenly between 12 candidates. Clearly 2 Labor, 2 Liberal and 1 Green would be immediately elected but then each has a surplus of 2% to go towards their next candidate and there are another 12 each with about 2% so you have at that point 15 different candidates each with about 2%. Who should the last candidate be? There is no obvious winner without distributing preferences, but if the 24% who voted for small parties predominately prefer that anyone except the big three get elected, then a perfectly democratic outcome would be for one of the candidates with a very small primary of about 2% to win. Now suppose a similar setup, but with one of the 12, candidate A, quite popular with 8% and the remaining 16% shared between 11 candidates, so after electing 5 senators we have 3 bigger parties on 2%, 1 on 10% and 11 with 16% between them (let’s say none bigger than 2%). Then it appears that candidate A should win, but suppose that all of the other small parties have similar interests but don’t like candidate A, and the big parties don’t like them either. Then of the remaining 28% or so (there is a rounding error that stops it from adding up precisely here) of votes, 8% like candidate A but 20% would prefer for any of the other candidates to win, so it is perfectly reasonable for a candidate with about 2% to beat the candidate with 8%. So a candidate getting up to win with a small vote is not necessarily bad. To some extent this is what has happened in the current election, many people have voted for small parties and do not want senators from larger parties, quite often the biggest of these votes are for the likes of the Sex Party or Family First, but most people voting for one of these do not like the other, however they, and various other small party voters, would prefer the likes of the Motor Enthusiasts or the Sports party to the larger ones. It is not completely unreasonable that such candidates are the preferred choice of this large constituency disillusioned with the more established parties.

The real problem is that the preferences as allocated might not reflect people’s true preferences due to Above The Line (ATL) voting. In principle every voter should give a full choice of preferences for all candidates, in practice this tended to result in a level of informal voting deemed to be unacceptable so the alternative was to vote for one party above the line, with your full vote then coinciding to a preference ticket provided by the party to the AEC. Some people are rather misguided in raging against secret deals and complaining that their votes have been traded around by the parties. This is not how it works, the parties must provide their preference tickets in advance, they are published before the election and are available to view at polling booths, they are not secret. You still have full control over your vote if you choose to vote below the line. I don’t think any individual has reasonable grounds to complain about what happened to their own vote, however if lots of people aren’t aware of these things and some preferences are allocated in unusual ways this is something that one can reasonably be concerned about when taking an overall view of the outcome. For example there is the possibility of parties being formed with a name to appeal to certain voters which then channels their votes to someone they wouldn’t normally preference.

Tied in with this is the problem that ballots are becoming very large and difficult to deal with, with a disincentive to vote below the line due to the time taken and higher possibility of failing to number it correctly and rendering it informal.

So how to deal with this issue? Some people think that preferences themselves are inherently the problem and are unfair, they take a simplistic view that if only each voter had just one vote then it would be fair. This is very, very wrong. First past the post is one of the worst voting systems and would lead to less democratic outcomes.

Another suggestion is to have a minimum quota of first preferences for someone to be elected, but as I pointed out above, people being elected on small votes isn’t really the problem, so I don’t see this as the solution.

The real problem is above the line voting.

The simplest way around this is to allow for preferences above the line. A voter would have the choice of either numbering every party above the line or every candidate below the line. This would make the count more complicated but I think would be worthwhile.

Something else often suggested is allowing optional preferential voting. I tend to prefer full preferential voting, but in an election for multiple candidates such as in the senate, then optional preferential with a reasonable minimum number of preferences required would be acceptable. My problem with optional preferential is that it allows for many votes to exhaust, so at later stages of the count candidates can be elected from a much smaller subset of voters. This is balanced by the problem that in the senate there are so many candidates that voters know very little about, so it is not necessarily meaningful for them to rank them in order of preference.

I do think that lots of people want optional preferential for the wrong reasons, for example someone might resent that they have to number the Liberal candidates because they are a Labor or Greens voter, but leaving a number blank is not a vote against someone, it just means that once all the ones you do number are elected or eliminated you’ve given up your right to have a say. With optional preferences I would still allocate a preference to the Liberals even though I don’t like them because there are others I like less, for example, if it ended up coming down to a choice between Liberal and One Nation then I’d certainly still want to have my say and would happily prefer the Liberal. In such a scenario those who exhausted their preferences might be quite happy to say they didn’t vote for either, but they could effectively be helping One Nation to win the seat. By allocating as many preferences as possible you don’t have to make guesses as to who might realistically have a chance at each stage of the count, you get to express preferences regardless of what happens. Something like this is happening in the current Tasmanian count, after the Greens candidate is elected there is a surplus to be distributed and the remaining candidates are Liberal, Palmer United, and Liberal Democrats. Most Greens voters probably don’t like any of them, but at this stage one of them has to be elected, not because of the distribution of Greens surplus, but because the vast majority of other voters who are still being counted like one of them. In such a situation, I would rather choose between three options I don’t particularly like then say nothing at all.

So basically, I think that a combination of above the line preferencing, and optional preferential but with a high enough minimum number to stop too many votes exhausting would be a reasonable reform.

A further reform to help keep ballot size reasonable and to stop people trying to exploit the system with ‘front’ parties would be some tightening of the rules for candidature. It would seem reasonable to require parties to be registered for longer before an election and to stop the same people registering numerous parties. I think the required number of party members should be counted for a single party only as well (i.e. parties can still allow dual membership but their submitted list of members to qualify for party status on the ballot cannot include members counted for another party). I would be wary of setting the bar too high though (particularly with regard to monetary cost) so as not to make it too difficult for genuine new parties to contest elections.