The current LNP state government in Queensland was elected on a promise of restoring the Optional Preferential Voting (OPV) system used in the state up until the 2015 election.
Under this system, voters are not required to number every box on the ballot. In contrast, the current Compulsory Preferential Voting (CPV) system requires voters to effectively number every box.
CPV does lead to higher rates of informal voting, but ensures that all formal votes remain in the count until the end. OPV does give voters the option to ‘exhaust’ their preferences, but it can also be used by cynical parties to encourage ‘just vote 1’ strategies as a way to reduce preference flows to their opponents.
When CPV was introduced in 2016, it was in an environment where Labor benefited most from preferences, and thus would benefit from a system that would encourage stronger preference flows.
Yet that hadn’t always been the case, and indeed it may no longer be the case.
The current Queensland state government has been slow to move on a change to the electoral system – this blog post will show some evidence for a conservative government to hesitate before changing systems in a way that could hurt them in some parts of Queensland.
While the Labor government of 2016 was a beneficiary of Greens preferences, that hadn’t always been the case.
Preferential voting was first introduced in 1919 at a federal level to address the issue of farmers candidates running against the governing Nationalists – these groups would eventually coalesce respectively into the National Party and the Liberal Party of today. Through much of the 20th century, these parties were able to run against each other while not worrying so much about splitting the vote. Indeed the Country Party sometimes chose to run multiple candidates for a single seat, using the preferential system to let their voters make the final call.
The Whitlam government in the 1970s was an advocate of OPV, and OPV was introduced in New South Wales in 1981 under a Labor government.
At the 1998 Queensland state election, Labor faced off against three different conservative forces. These three parties polled almost 54% between them, while Labor polled just 38.9%. Yet Labor almost won a majority, and returned to government with a minority government, facing off against an opposition split between Liberal, National and One Nation.
In 2001, the Beattie government campaigned with a ‘just vote 1’ slogan, but it wasn’t directed at their own voters: instead it was aimed at convincing conservative voters to not bother to use the full power of their ballot paper.
Within the decade, One Nation had largely disappeared and the Greens had emerged as the third party. I witnessed the newly-merged Liberal National Party using ‘just vote 1’ messaging in inner-city electorates in Brisbane in 2009.
I would argue that we are seeing a shift in direction again, with preferences becoming more useful for the LNP in some areas than others. Arguably CPV now benefits the conservatives in some areas, and progressives in others.
I have analysed data dating back to 1998 for this blog post. To start, this chart shows the number of seats at each election which were won by a candidate who wasn’t in first place on primary votes.
No more recent election has produced as many come-from-behind wins as 1998. One Nation won six seats, the Nationals won six, and a conservative independent won one. Labor came first in most of these seats, but One Nation did lose three seats to the Nationals. Generally conservative parties favoured each other over Labor, but Labor preferences saved some Nationals seats from One Nation.
The Labor just vote one strategy totally eliminated this issue in 2001. Labor did poll 49%, but there was also a notable uptick in exhaustion.
Come from behind wins were relatively rare over the next decade, slightly favouring Labor.
2015 (ironically the last election before Labor switched to CPV) saw Labor come from behind in nine seats.
Since CPV was restored in 2017, the pattern has been more balanced.
As a simplistic measure of who benefits more from preferences, I subtracted the Labor and LNP primary vote from the estimated 2PP for the whole state to produce a figure of how much each party benefited from preferences. This can only be done back to 2006 – the 1998-2004 elections did not feature 2PP estimates.
Labor has usually benefited more from preferences than the LNP, but often by slim margins. The collapse of Labor support in 2012 led to the LNP doing better. Labor widened their preference lead in 2015 and 2020, but it was quite narrow in 2017 and 2024. It’s worth noting that 2017 was the year when One Nation first re-emerged in Queensland state politics, polling 13.7%. This chart doesn’t cover the 1998-2004 period, but it would have been undoubtedly more favourable to the LNP.
It’s also worth noting the total volume of preferences has continued to expand over this period, peaking in 2017. Although the number would have been much higher in 1998, when One Nation outpolled both the Liberal and National parties (separately).
For the rest of this analysis, I have focused specifically on seats where Labor and the LNP, or its predecessor parties (who I will refer to as the LNP unless it’s necessary to distinguish them), made the final preference count, and where at least three candidates stood. Firstly, I looked at what proportion of preferences in these seats went to Labor, the LNP or exhausted.
I should also note that there are quite a few seats missing at the 1998 and 2001 elections. These seats were mostly those where One Nation made the final preference count, and thus tend to be more conservative than the statewide average. I suspect if a statewide 2PP had been conducted in those years, the LNP’s lead on preference flows would have been even greater.
The LNP did better out of preferences in 1998 and 2001, but since 2004 there has been an advantage to Labor. It’s worth noting that the proportions between the parties remained roughly the same when the system changed from OPV to CPV, but with those exhausts now splitting in similar shares between the two parties.
The 2004, 2012 and 2024 elections were ones where the LNP were able to get close to Labor’s share of preferences. The 2024 gap is particularly small considering that there were no exhausts.
The rate of ballot exhaustion shot up dramatically in 2001 when Labor first ran the ‘just vote 1’ campaign. It dropped slightly but then surged again in 2012. I suspect 2012 can be explained by the Labor government being particularly unpopular. Voters for a party like the Greens were less likely to bother to preference when Labor was so clearly set to lose the election. A similar trend can be seen at the 2011 NSW state election.
Ironically the 2015 election, the last before the switch to CPV, saw the lowest exhaustion rate since the turn of the century. At the time there was a concerted effort to encourage left-of-centre voters to mark their preferences, and Labor gained much stronger preference flows than they had in prior elections. I often point to this election when people try to claim that OPV is functionally equivalent to first-past-the-post (FPTP). While OPV doesn’t require people to mark preferences, the option is still there, and political campaigning can influence what people do. It’s not a law of nature that OPV leads to low rates of preferencing.
But I don’t just want to look at this issue in aggregate across the whole state. Rather I want to look at how it might be playing out differently in different seats. This next chart shows how many seats at each election gave more preferences to Labor than LNP, or vice versa.
Back in 1998 and 2001, there were a lot more seats where the LNP had the advantage.
2015 stands out as the year when Labor had an advantage in every seat, but otherwise there has always been some seats where preferences favour the LNP.
The LNP had a preference advantage in more seats than at any other election since the early 2000s at the 2024 election. Labor had the edge in 49 seats, and the LNP in 34.
This is the point I am getting to. As the vote for right-wing minor parties like One Nation and KAP has been climbing, it has shifted the preference landscape in some Queensland seats, but not all of them. In the south-east of Queensland, Greens preferences still play an important role in boosting Labor. But in the regional parts of the state, the opposite is true.
If One Nation’s support were to continue to climb, they would take votes predominantly from the LNP. As long as One Nation stays in third place, the LNP will be needing those votes back as preferences.
I think Queensland politics is now too fragmented to give a single answer as to “who would benefit” from a switch to OPV. It is likely Labor would benefit in the biggest cities, and the LNP would benefit elsewhere. But that exact benefit would depend on the specific circumstances of each election.
In this post I have focused on the political advantage of a change in electoral system, but I don’t think principles should be lost. A party that tries to make this switch because they consider it in their self-interest will eventually find that self-interest reversed.
Instead, I suggest we should consider what is the best electoral system based on principles of ensuring the freedom of the voter, encouraging the use of preferences, and minimising informal voting. I believe the best system is neither CPV or OPV, but something lying in between. That will be the topic of my next post.
The LNP were using “just vote 1” in Brisbane in 2024 (for council, which still has OPV)