When Julia Gillard announced August 21 as the election date rather than August 28, one of the consequences was that the electoral roll closed a week earlier than it ordinarily could have – providing very little time for younger voters to get themselves on the roll. As we know, younger voters – especially the 18-24’s – have historically voted for the ALP in two party preferred terms at a much higher rate than the broader population, so today we’ll have a squiz at the size of the sacrifice in national two party preferred vote share Labor could expect to experience as a result of smaller numbers of their most favourable age cohort being enrolled than would have been the case if the election was called for a week later.
In April of this year, the AEC estimated that there were around 430,000 Australians aged between 18 and 24 that were not enrolled to vote. At the end of June, the latest period we have enrolment data for, there were 1,450,267 18-24’s on the roll – meaning that around 77% of that cohort was enrolled. There was also 12,451,573 people aged 25 and over on the roll, giving us a total electoral roll of 13,961,671 people.
We also know that 18-24s in the polling have been supporting Labor at a rate of between 54% and 62% for the last couple of years. At the 2007 election they actually voted for Labor somewhere around the 64% mark.
To estimate the consequences of 18-24 enrolment size on the ALP national two party preferred vote share, if we assume that 18-24’s vote at 58% Labor (the mid point of the polling) and hold all other age cohorts constant at 50% ALP vote – we can look at how the ALP two party preferred vote increases as those 430,000 odd young voters are added to the roll in 1% increments.
It doesn’t really matter what we hold the voting level of people 25 and older at, as that only changes the level from which the vote estimates change, not the actual rate of change in TPP that the 18-24s create – which is what we’re actually interested in here. We’ll also use 95% as the maximum enrolment level we could expect to see among the 18-24’s, as Peter Brent (of Mumble fame) and Simon Jackman suggested that 95% was about the level of enrolment we saw for the 25 years and older cohort in the 2007 election.
So, under these assumptions (58% TPP for 18-24’ and 50% for everyone else), the results come in like this.
With the current 77% enrolment level of 18-24’s, we’d get an ALP two party preferred of 50.83%. If that enrolment jumped from 77% up to, say, 85%, we’d expect the ALP two party preferred to increase by only 0.08% to come in at 50.91%.
So the boost that the ALP would be expected to get as the number of young people on the electoral roll increases, isn’t particularly large.
If we run the numbers again, but this time use three different assumptions on the size of the 18-24 ALP two party preferred vote – a low of 52%, our 58% assumption above and a higher 65% level – and look at just the size of the increase in the national ALP two party preferred these three voting levels would have across the 18-24 enrolment spectrum, this is what we get (click to expand):
As we can see, the higher the ALP vote among the 18-24s, the larger the impact they ultimate have on the national ALP TPP level. With what we might call a “full enrolment” level of 95% among the young, at 52% TPP they would boost the national ALP two party vote share by a meagre 0.04%, at 58% TPP they’d boost the national vote share by 0.17% and at 65% TPP they’d boost the national Labor vote share by 0.32%.
The decision by Gillard to have the election on August 21 rather than August 28 doesn’t look like it will have a particularly large TPP vote cost in terms of smaller numbers of 18-24 voters being enrolled. In 2004 when the rolls were open for a week after the election announcement, 156,000 voters enrolled over that period – with a large proportion being 18-24s.
From that earlier Brent/Jackman link – in 2007 when 3 working days were available, only 77,000 new voters came on the roll over the period, again, where a large proportion was 18-24s
With only 1 working day this year being available for people to get on the roll, it would be surprising to see a massive level of late voter enrolment. Looking through the possible numbers:
If just under 54,000 young people got on the roll between July 1st and Monday July 19th (when the rolls closed), 18-24 enrolment would hit the 80% level – a big ask BTW. At this level the ALP would expect to get around a 0.03% boost to their TPP, maybe up to 0.05% if historical voting patterns were repeated.
If the rolls were left open for another week and a well funded AEC advertising campaign was run, we may have seen 90K to 100K of additional 18-24’s jump on the roll from the end of June figures – which would have boosted the ALP TPP by around 0.05%, maybe up to 0.09%.
In the broader scheme of things, that’s probably worth less to the ALP in vote outcomes than a mild gaffe by a Minister in the last few days of the campaign.
However – not all electorates are equal, as some electorates have much larger numbers of 18-24 year olds living in them than others. Unfortunately, the ABS hasn’t updated their census data yet for the 2010 election boundaries, preventing us from using that as an easily accessible source, and while the AEC has enrolment numbers with age breakdowns by the new seat boundaries – it doesn’t help us as we’re trying to find the numbers that weren’t enrolled at the end of June rather than the numbers that were. One thing we can use as a proxy is some data from the Australian Parliamentary Library, where those fine folks have ranked seats (on the new boundaries) by the proportion of 15-24 year olds (as they were at the 2006 census) living in the electorate.
If we assume this is a fair proxy for the current proportion of 18-24 year olds by electorate, we can look at the seats which have both a relatively high youth ranking (at least 15% of the electorate being 15-24) and are quite marginal – giving us a peak at which seats, if any, may be more likely to be decided at the election as a result of closing the rolls early.
We’ve left the seat of Melbourne off the list here even though it had the highest proportion of 15-24 year olds in Australia, coming in at 21.5% – but as this seat will likely be a contest between the Greens and the Labor Party, and where both parties have high levels of support among the young – it gets awfully complicated.
If the election results in these seats are within a thousand votes or so, this is where the choice to go for an August 21 election rather than an August 28 may become decisive in terms of which major party ultimately wins them – with the seats of Herbert, Swan and Ryan being particularly susceptible.