As we’ve discussed a few times, the Covid-19 crisis has forced a number of countries to reconsider the ways in which they hold elections. Postponements, expanded or universal postal voting, electronic voting and various sorts of procedural changes to attendance voting have all been in the mix.
One option, being implemented or considered in several states in the US, is the expansion of pre-poll voting. That doesn’t change the actual voting procedure, so it avoids the privacy concerns about expanded postal voting, but it spreads the load over a much longer period, making it easier to implement spatial distancing and other health measures.
Australia is already a world leader on this front. With no real public debate on the issue, we’ve already reached a point at which about a third of voters cast their votes at pre-poll centres before election day – a trend that shows no signs of slowing.
Yesterday I read a very interesting post by Antony Green on the subject. It’s a few weeks old now, but I somehow missed it at the time; if you did as well, it’s well worth going back and reading.
Green explains, with a wealth of supporting detail, just how we got to this point, and draws attention to some of the problems it has caused:
The growth of pre-poll voting has vexed candidates and parties as they struggle to staff pre-poll centres with people to hand out how-to-votes. It has also tested campaign strategies, the traditional crescendo of advertising and announcements at the end of the campaign being re-structured into a drip-feed of announcements.
But his main point is about counting. With such a large fraction of votes concentrated in large pre-poll centres, some taking more than 20,000 votes, counting them is a big task, especially if it’s all expected to be done on the same evening:
Polling places are staffed based on how many votes will be taken and need to be counted on election night. AEC [Australian Electoral Commission] research suggests polling places should take only around 2,000 votes for efficient operation.
… Counting 20,000 ballot papers is not just a matter of scaling up a 4,000 vote count. Large counts require more ballot boxes to be reconciled, and more experienced supervision to avoid the process getting lost in a blizzard of forms and ballot papers.
Green outlines two possible solutions to the problem: to cut back on the availability of pre-poll voting, or to start counting the pre-polls on election day before ordinary polling has finished.
I’m not as convinced as he is that the first option is a bad idea. The very short time between close of nominations and opening of voting strikes me as dangerous: voters should be given more time to find out who they are voting for and not encouraged to make low-information decisions. I think an extra week would be beneficial all round.
Green argues that political parties would use that time “to flood the electorate with postal vote applications,” which are less secure than pre-polls and subject to dodgy practices by the major parties. But I’m not so sure – if that were the case, you’d expect the rise of pre-polls to be accompanied by a drop in postal votes, which hasn’t happened. And party shenanigans should be addressed directly by legislation in any case.
That said, Green is right that the pre-polls cast in the first week are a small proportion of the total, so cutting back on the pre-poll period probably wouldn’t make much difference to the numbers involved. Cutting actual availability, by enforcing the rules about voters needing an excuse for not voting on election day, might do a better job, but the pre-poll option is now so popular that that is seen as politically impossible.
So Green endorses instead the option of starting the pre-poll count early, with appropriate security measures to ensure no results are divulged before 6pm. His discussion of the subject is very thorough. But while I have no problem in principle with it (I’ve managed university elections for many years with just such arrangements), I’m less than enthusiastic.
One reason is the limitation of finite resources. There are only so many trained staff, competent scrutineers and suitable premises to go around. If they are being employed to count pre-polls during the day, they can’t also be used to take ordinary votes at the same time. Green acknowledges the problem, but I don’t think he gives it sufficient weight.
In practice, less-qualified people will be employed, and more mistakes will be made. In most cases that just means that more work will have to be done later to correct things, but some mistakes are unrecoverable – for example, if ballot papers go missing as they did in the 2013 Western Australian Senate count.
There’s also a difference about what you think is being gained by finishing counting more quickly. It’s certainly true that counting very late into the night involves risks (tired workers make more mistakes), and having to start again the next morning offers another opportunity for things to go missing.
But if the only reason for pushing to get things done on the night is so that the public can learn the results before they go to bed, maybe it’s worth pausing and thinking about whether this is really a big priority.
Green and I are both professionally biased to think that it is. But I’ve felt for a long time that we rush things too much after an election: that our politicians are too eager to get back to work, and that our media are too impatient with any delays in finalising results or negotiating to form a new government. Perhaps we need to send the message that pundits should just chill for a day or two.
If (as will often be the case) packing up a count mid-evening and making a fresh start on the Sunday is going to be better for security and accuracy, I don’t think the AEC should be afraid to do it, despite the complaints of TV networks and political tragics. It’s worth noting that in Ireland, one of the few other countries with preferential voting, counting doesn’t even start until the following day.
Finally, I’m not convinced that Green’s two options are the only ones available for addressing the problem. Pre-poll centres don’t have to be so large; opening more of them, preferably in the same or similar locations to election-day polling places, would make counting easier as well as increasing the convenience for voters.
Political parties would be less keen on the idea, since they would have to find more workers to hand out how-to-vote cards. But that too is a matter of priorities; it seems to me that far too much of our electoral system is already designed for the benefit of the parties. If it means they have to do more recruiting and keep up their branch membership, that’s not such a bad thing.
That also leads to the question of the political effects of increased pre-poll voting – do pre-poll voters behave differently from those who cast their votes on the day? But this post is too long already, so we’ll get to that question in part II, later this week.