The NSW special minister of state, John Graham, yesterday announced that the government will ask the state Joint Standing Committee on Electoral Matters (JSCEM) to hold an inquiry into rates of eligible voters participating in NSW state elections, with a goal of achieving 100% participation – every eligible NSW resident casting a formal vote at state elections.
Graham then linked that goal to concerns around electoral disinformation, attempting to sow a lack of confidence in our electoral institutions.
Making it easy for voters to cast a vote is important. But NSW is currently a national leader on policies to address these issues, which have led to a very high rate of enrolment and relatively low rates of informal voting. I struggle to think of other policies that would make a significant difference.
And if we’re going to talk about things that affect voter confidence or representation, there are other parts of the electoral system in dire need of reform that have a vastly bigger impact than the technical question of whether every eligible voter has a vote counted.
When we talk about the numbers, we need to clarify a few things.
Firstly, I have been unable to find any estimates from the NSW Electoral Commission of the size of the voting-eligible population (VEP) at the time of NSW state elections. Thankfully the AEC has been publishing estimates of the VEP for each state on a quarterly basis since 2013, and slightly less often from 2010 to 2012. So I have used the March figures from each state election year for my analysis.
Secondly, Graham’s office’s numbers seem off. I thought it was strange that they only reported a raw figure of almost 800,000 people who didn’t have a formal vote cast (either by not being enrolled, not turning out, or by casting an informal vote), rather than reporting percentages, and making no attempt to give historical context of how those numbers compare to those in the past.
By my estimate, the number of voters missing from the roll is roughly in the right area, but Graham has underestimated the number of voters on the roll not casting a vote by about 100,000. About 660,000 voters didn’t turn out, not 553,000. And he overestimated informal votes by about 4,000. Overall it adds up to about 900,000, or 16.2% of the VEP, who didn’t cast a formal vote.
But how does that compare to the past?
This chart shows six different metrics which combine formal votes, turnout (formal and informal), enrolment, and VEP
The share of enrolled voters who are casting a vote, or casting a formal vote, has been declining. The formal rate as a share of votes cast has been more steady, with the informal rate slowly increasing, but then dropping in 2023.
But those first two metrics are misleading because they miss a big piece of the puzzle – NSW has been making remarkable progress in finding and enrolling eligible voters who once weren’t on the roll.
NSW passed legislation in 2009 for what is usually called direct enrolment. This is a process whereby an electoral commission can use government data to identify potential voters and add them to the roll without the voter specifically signing a form.
The federal government followed NSW later on. Generally the rolls are consistent between federal and NSW, although differing laws on direct enrolment meant some voters may only appear on one roll.
The AEC publishes vastly more data on participation rates than NSW, publishing the proportion of the estimated VEP in each state who are on the roll on a quarterly basis. This chart shows that measure for the federal roll in NSW since 2010.
At the start of the direct enrolment process in the early 2010s, enrolment rates bounced around between 91 and 92%. By 2020, they had cracked 98%, and in 2023 they reached new heights of about 99%. I estimated that 98.4% of the VEP was enrolled for the NSW state election in 2023.
This is honestly getting close to completion on this measure. We will never have everyone, but almost everyone is now on the roll.
So when you look at the first chart, you can see that the formal vote and the total turnout both look much more impressive when measured against VEP, not enrolment. Unfortunately I only have this data back to 2011, but we saw a noticeable jump between 2011 and 2015. 2023 was slightly worse than 2015 or 2019, but still better than 2011.
The other excellent measure we have in NSW is the right of eligible residents to enrol to vote on the spot and then immediately cast a vote on election day. This policy is far from universal across Australia, but NSW is using it.
As for the informal rate, NSW is the only Australian state to not force voters to number every box on their lower house ballot. This has led to significantly lower rates of informal voting.
AEC surveys in the past have suggested that about half of informal votes appear to be accidental informal votes, and the other half are deliberate (ie. blank ballots). The former category is mostly eliminated in NSW, and thus we see a NSW informal rate just over half the federal rate. Perhaps it could be better. I’m sure we could do more for education and voter literacy, but I think NSW is doing quite well.
While I think compulsory voting is a fine measure, and it is important to make it easy for people to vote, I don’t think 100% participation is a reasonable goal. I think it is fair enough that voters have the option of not exercising a choice by casting an informal vote and it shouldn’t be a goal of government to eliminate those cases.
But if we are concerned with voter confidence and representation, there are much bigger problems.
I know I have talked a lot before about proportional representation, so let’s pick a different measure of the weak representation of the NSW lower house electoral system. This chart shows the proportion of formal votes which were cast for the winner in each seat.
This number first dipped below 50% in 1999, but then climbed again as One Nation declined. The 2023 election saw the lowest level in at least a quarter of a century, and I suspect much longer. More than half of all voters voted for someone else.
The figures are even worse for the House of Representatives, with just 42.3% of voters casting a primary vote for a winner in 2022.
In comparison, Tasmanian elections saw 61.2% cast a vote for a winning candidate in 2024, and over 90% cast a vote for a party that won at least one seat in their electorate.
A PR system would mean that vastly more voters would have a local representative they voted for.
And I’m sure people will talk about two-party-preferred figures and the value of preferences. It’s good that we use preferences, but they’re not the same. It’s also true that once you factor in exhausted votes, a major party has only won a majority of the 2PP in a NSW election twice since optional preferential voting was introduced in the 1980s (in 1988 and 2011).
We could also do more modest things that would slightly improve representation, such as changing ballot instructions and advertising to more strongly advocate for the marking of multiple preferences, even while we allow voters to ultimately just vote 1.
NSW elections could be a lot better, but I think Graham is looking in the wrong places, trying to fix things that are working really well and ignoring the real problems.