It’s no surprise that the informal rate increased at the recent Tasmanian state election, but I wanted to explore how high the rate went, and what might explain it.
Tasmania has a rule that voters must number as many boxes as there are seats to be filled. That was seven up until 1996, five from 1998 until 2021, and then seven again this year.
In part to match this rule, the three biggest parties each run seven candidates in each electorate, even though they would never have a chance of winning all seven. It means that a voter can number 1 to 7 just within their group and those preferences will stay with the group until that group has no candidates remaining in the count, and that they can be sure that their vote is formal.
(Running surplus candidates is also helpful for filling vacancies later in a parliamentary term, but it doesn’t explain why the Greens would run seven candidates rather than, say, three.)
But for smaller parties that don’t run the full magnitude of candidates, they need to ensure that voters preference outside the group. If someone were to just preference the three JLN candidates in a single electorate, their vote would have been informal. Thus the increased magnitude is a higher barrier for those minor party or independent voters. Someone who voted for Craig Garland needed to find four other candidates to preference in 2021, but six in 2024.
So, how much did things change in 2024?
For most of this analysis I’ve gone back to 1989, but I wanted to go back a bit further here to get more examples of elections with magnitudes of 7.
The 2024 election has produced the highest informal rate in at least fifty years. Indeed in the entire history of Hare-Clark, the only election producing a higher informal rate was 1946, when the informal rate spiked to 10.1%. I’m not sure why – if anyone knows what happened that year, please post in the comments. Apart from 1946, the informal rate has often cracked 5% but never 6%, until the 6.3% informal rate in 2024.
The magnitude is definitely a major factor in the informal rate. The four other highest informal rates in the last fifty years were all in the magnitude-7 era from 1982 until 1996. The informal rate never quite reached that level again in the magnitude-5 era, although it did climb through the last decade.
The informal rate was not consistent across the state, as the next chart shows.
The two southern electorates have traditionally had lower rates of informal voting, dating back to at least 2002, although the five electorates all had very similar rates in 1996 and 1998.
All five electorates went up in 2024, but the rates were much higher in the three northern electorates. The rates in Franklin and Clark returned to a similar level to the 1996 election, but the other electorates reached new highs not seen in the last thirty years.
So next I wanted to understand what other factors might play into the informal rate. We know that in single-member electorates under compulsory preferential voting the total size of the ballot paper is a factor. But it’s more complicated under Hare-Clark.
Should we look at the total number of candidates? That figure is highly influenced by the increase in magnitude. Labor, Liberal and the Greens ran six extra candidates between them in each electorate without necessarily increasing the difficulty in voting for anyone. I also looked at the total number of columns, or the number of columns running less than the magnitude of tickets.
But one thing I realised is that 2024 stood out on every metric. I would find somewhat strong correlations (up to about 0.5), but when I looked at it on a chart, it was clear that 2024 stood alone. When I separated the data into magnitude 5 and 7, any correlations disappeared. They particularly disappeared when I excluded 2024 entirely.
So this has led me to conclude that there is something particular about 2024 – it wasn’t just the return to magnitude-7 elections, although that did play a role. It may be that informal rate was higher because voters weren’t used to the higher requirement and will fall at the next election, or that the increased field of viable small parties and independents made it easier to screw up and number less than seven boxes.
Tasmania could partly deal with this problem by looking to the ACT. The ACT electoral system is very similar to Tasmania, although they still elect five members per electorate. The ballot papers do still advise voters to number five boxes. But in practice a ballot paper with a single clear 1 will still count.
Such an approach may result in a few more votes which end up being counted as formal and then end up exhausting more quickly, but I think it would still be an improvement. As long as parties and the Commission continue to advocate for voters to number at least seven preferences, it should allow more votes to count without lowering existing preference flows.
Sometimes the simple answers are the most likely ones and in this case I would chalk up the higher informal vote in 2024 to political disengagement. A sizeable fraction of the population (indeed anywhere not just in Tas) have no interest in politics and totally tune out all political information including, in this case, the fact there were more seats up for grabs hence the minimum number of boxes needing to be ticked had risen. For many elections they have turned up & ticked 5 boxes and turning up this year, having paid no attention to the news about additional seats, ticked the same 5 boxes, whether randomly, along some party line or as the proverbial donkey. It may be hard to believe for readers of this blog, but there are plenty of ‘voters’ who don’t even know who is the PM or what party is on office (and yes, I know some of them).
As for the 1946 anomaly, given the date, given that WW2 had ended the previous year and there was an an influx of refugees and returned soldiers, one would expect all sorts of anomalies in 1946 & the following few years.
I believe that the upswing in informal voting in 1946 was to do with the introduction of columns on ballot papers and “some very confusing instructions” which voters had difficulty following.
The election was for 5, six-seat electorates and my understanding is that the candidates in each electorate were, for the first time in Tasmania, grouped into columns of Liberal (the Nationalist party at the previous election), Labor, Independent and Ungrouped (this column was only in the seats of Denison and Wilmot, now Clark and Lyons) and there were only between 1 and 4 candidates in the non-major party columns.
I don’t have a copy of the instructions for the voting on ballot papers with columns, but my understanding is that the instructions for voting between the columns was unclear and caused considerable confusion (Kevin Bonham did make a reference to this in his blog as well, but I haven’t been able to find the instructions that went with the change in layout of ballot papers).
I would make the point that requiring 7 votes for a formal vote means that anyone who onerously recalls the Federal Senate’s “fill in 6 preferences” will cast an informal vote.
Tasmania is crying out for a savings provision.
I really like the innovation of the South Australian state savings provision. They have CPV but ballots that partially follow a party ticket get saved at the point of exclusion and are presumed to follow the rest of the ticket.
I would run with that and make a more nuanced provision that takes it out of the hands of party tickets: all ballots that are excluded instead have their vote value reduced to zero and increase the vote value of the continuing ballots that have the same 1st preference as them.
Ie- it is presumed that cohort of voters who run out of relevant preferences have the same proportion of preferences between remaining candidates as the most similar cohort voter, the ones who had the same 1st preference.
This would reduce informal voting quite a bit and completely wipe out exhaustion with 1 rule.
That sounds pretty complicated – is letting a few votes exhaust in what would otherwise be CPV that big of a problem?
It’s not complicated for the voter. In fact it makes the election much less complex for the voters. Accidental informal votes would become nearly impossible. The fact that we strip voters of their 1st preference vote because of a mistake later on the ballot is the biggest flaw of CPV, that’s what savings provisions are about.
It’s in some respects more complicated for the electoral commission but in other respects a simplification.
It makes determining what a formal ballot is much easier. All it needs is a clear 1st preference. As above, for most ballots that’s all that’s required. No counting staff or scrutineer will quibble about whether a ballot looks like it has a 2 that looks like a 3 if it’s clear to everyone in the room that it’s never going to flow onward from it’s (now automatically formal) 1st preference anyway.
In single-winner elections the ballots involved in having their ballot value changed would be very small in number simply by the fact that the candidates who survive to be in the 2CP invariably have an overwhelming share of the primary ballots. None of those ballots will ever exhaust and therefore none of them will ever be increased or decreased in value.
In most multi-winner elections we already have concepts of surplus value that messes with ballot values as the count goes on, and in most multi-winner elections we do the DoP with data entry and a button press. If it’s a computerised count my savings provision would add to the total time basically not at all.
Whether exhaustion is good or bad and worth eliminating is a debated philosophical question. I personally believe exhaustion is a soft form of disenfranchisement. It’s the system allowing a voter to throw their vote away (usually without the voter fully comprehending the implications) and in multi-winner elections it messes with the equality of quotas required to win a seat (the higher the exhaustion rate the better it is for the party or candidate in the running for the “last seat” at the expense of the parties and candidates that won seat already at full quota). But this is a much more marginal concern than informality rate.
For the record, I’m fully in favour of savings provisions and agree with you the lack of them is an ugly flaw in CPV. I just think counting ‘saved’ ballots as they would be counted in OPV is the easiest to implement and least controversial approach. Definitely for single seat contests.
Multi-winner elections I can’t say my understanding of the pros/cons is up to scratch though.
*Pros/Cons of different saving provisions
I lobbied for the government to at least get the Electoral Commission to report on savings provisions (including suggesting that transitionally 1-5 be still accepted so that no vote that was formal in 2021 would be informal in 2024) but my submission was ignored – and in fact the Premier falsely stated that all submissions had been in favour of the return to 5×7 when I was only in favour of it if the informal voting problem was addressed. A previous submission I made in favour of adopting ACT style savings provisions was verballed. Unfortunately there appears to ba an ideological objection to savings provisions among some who are advising the government on these matters, motivated by spurious concerns about excessive exhaust if people are allowed to have their votes counted. Really hard for me to comprehend the sort of line of thinking that says that candidates getting elected with well less than a quota somehow so distorts the workings of Hare-Clark that we are going to exclude some voters’ intentions from the process altogether.