Last week we looked at the striking growth in pre-poll voting in Australia over the last couple of decades, and the issues that has raised about how and when votes are counted. Today I want to look at what we find when we do count them.
The electoral commission classifies all votes as either “ordinary” or “declaration” votes. Declaration votes are those that require the voter to fill out their name and details on an envelope and seal their ballot paper(s) inside, so that their eligibility can be confirmed before their votes are opened and counted. Ordinary votes don’t: the voter’s name is just checked off on a roll and their votes go straight into a ballot box.
Declaration votes are then broken down into four categories: absent, postal, provisional and declaration pre-poll. The commission reports election results by these categories – for example, here are the primary vote totals from the last House of Representatives election.
As you can see, there’s a pattern to them. Postal voters vote more strongly Coalition than any other category. Absent voters vote less strongly Coalition than ordinary voters, although in primary votes the difference mostly goes to the Greens rather than Labor. Provisional voters (those whose names can’t be found on the roll) favor Labor, although there are so few of those that I won’t bother mentioning them again. And declaration pre-poll voters don’t differ much from ordinary voters.
The pattern holds in every state and across most divisions. Looking at two-party-preferred numbers (I can’t find a table that totals these, but you can easily do it yourself from this one), we get the following breakdown:
|Vote type||Coalition 2PP|
Pre-poll votes, remember, started out as a sort of subset of postal votes: votes for people who weren’t going to be able to get to a polling place on election day, so applied to get their vote beforehand. The difference was just that they did it in person instead of by mail. And postal votes have always favored the non-Labor parties; people who are rich enough to go on holidays and well-educated enough to plan ahead are disproportionately likely to support the Coalition.
But pre-poll voting has now been normalised, as seen from the huge number of people doing it. In principle they are still supposed to have some reason for not voting on polling day, but the commission has made it clear that it knows most of them don’t and it doesn’t care. So it might seem that these figures confirm that: pre-poll voters are a representative cross-section of the population and vote much like everyone else.
Well, not so fast. While the above classification makes perfect sense from the electoral commission’s point of view, it doesn’t necessarily reflect anything important about voting behavior. When you go into a polling place, you’re unlikely to know whether you’re going to have to fill out your name on an envelope or not.
The distinction between declaration pre-polls and the rest of the pre-polls – which are treated as just a species of ordinary vote – is administratively meaningful but otherwise unimportant. As a voter, you just know that you’re showing up at a pre-poll centre to vote: as far as your decision-making goes, there’s no difference at all.
So where does the difference come from? Once upon a time, all pre-polls were declaration votes. But since 2010, they’ve been treated as ordinary votes – that is, put straight into a ballot box – if they’re cast at a pre-poll centre within the relevant electorate, or at one that’s sufficiently close to it, or sufficiently important, to maintain a roll for that electorate.
The only time you’ll cast a declaration pre-poll vote is if you’re at a pre-poll centre that’s well away from your own electorate. In other words, it’s the equivalent of an absent vote, but cast before polling day. But as far as voter behavior goes, the same can be said of many “ordinary” pre-poll votes – if you go to a pre-poll centre some distance from where you live, you don’t know in advance whether they maintain a roll for your electorate or not.
In an effort to try to capture how they appear from the voter’s point of view, I did my own sorting of polling places into four categories. The basic idea is to distinguish between, on the one hand, polling places that are in the electorate, or so close to it that voters might easily find them the most convenient to where they live, and, on the other hand, polling places that are sufficiently far away that any voter who turns their mind to it must realise that they’re not on home turf.*
The former constitute the first two categories, divided by whether the votes are cast on the day – “fully ordinary”, let’s call them – or as pre-polls. The latter forms my third category, absent and out-of-electorate pre-polls. The fourth category is everything else; mainly postals, but a few from polling on the day in non-conventional polling places (hospitals, mobile booths, etc.).
Voting results from the last election according to the new breakdown are as follows:
|Vote type||Coalition 2PP|
|Fully ordinary – in electorate & on the day||49.6%|
|Absent & out-of-electorate pre-poll||48.8%|
|Postal, hospitals, etc.||57.3%|
So once we’re careful to compare like with like, the Coalition’s advantage among pre-poll voters magically reappears! It’s not as great as with postal voters, but it’s still substantial. And adding the other sorts of out-of-electorate votes, including pre-polls, to the absent vote wipes out most of Labor’s advantage there.
These results are very robust. The in-electorate pre-polls, for example, are ahead of the average Coalition vote in 133 of the 151 electorates. The fully ordinary vote is behind the average in all but ten electorates.
Two things appear to be happening at once. Voting early is associated with a greater tendency to vote Coalition; voting in person outside one’s own electorate is associated with a greater tendency to vote Labor (at least in two-party-preferred terms). Declaration pre-polls show both happening at once, so they roughly cancel out.
So does this just mean that Coalition voters are more likely to pre-poll in the first place, or is there something about pre-polling that makes people more likely to vote for the Coalition? Most observers assume the first is the case, and they’re probably right. But it’s a difficult thing to prove.
Moreover, if the second should happen to be true – if the pre-poll experience itself somehow swings voters to the right – that would at least provide an additional explanation for the Coalition’s unexpected victory a year ago.
* I used a ratio of 5:1 as a rule of thumb to gauge this.