What primary vote is too low for majority government?


I have seen a lot of commentary in the last week about the low polling position of both major parties, and whether that might make it impossible for either side to form a majority government.

While it is true that chances of a hung parliament are higher now, and that recent polls suggest we could be on track for a record low primary vote for the combined major parties, the two are not directly linked, and it is quite possible a major party could form a majority government off a very low primary vote.

It’s not a new thing to hear claims that “Labor can’t win majority government without a primary vote with a 4 in front of it” or similar arguments.

Indeed the two Queensland major parties polled less than 70% of the primary vote at the 2017 state election, but Labor still managed a slim majority off just 35.4% of the primary vote.

These arguments can be theoretically valid if they assume a particular mix of the non-major party vote, although I don’t think either major party needs a 40%+ primary vote to win a majority now. Assuming no major change to minor parties since 2019, it’s probably true that a 33% primary vote wouldn’t be enough for Labor to win, after they polled similarly in 2019. But when the minor party and independent vote is growing and changing, the criteria for majority government also change.

I wrote last week about the long-term trend of a declining major party vote. At a certain point this may translate into massive losses of seats, but that hasn’t happened yet. It has resulted in fewer and fewer major party senators, but for this post I’ll be focusing on the House, where government is formed.

Up until this point, the relevant factor is the relative positions of the major parties, and how well each does on preferences. The position of the Greens has given Labor an advantage, although there are other sources of preferences as well. So while we have had one hung parliament at a federal election, it hasn’t happened regularly. While the polls are predicting a lower major party vote in 2022, it’s not that much lower than what we’ve seen over the last decade.

So why doesn’t a declining major party vote lead to neither party winning a majority? It’s about the single-member electoral system.

We have a majoritarian voting system, which tends to produce lopsided majorities for the biggest party. If we had a proportional voting system, neither Labor or the Coalition would have won a majority in their own right for decades.

But we don’t have a single national election, we have 151 separate contests. Each seat can only elect one person.

If both major parties have a vote over one third in an electorate, it is impossible for anyone else to win, since they can’t get ahead of them on the second-last count when there are three candidates in play. So if the vote was relatively even, you could see little or no crossbenchers elected despite a major party vote dropping down to about 67%.

The single-member system rewards parties who concentrate their vote, winning a majority in a smaller number of areas rather than winning small minorities spread around. This is why the Nationals hold so many more seats than the Greens.

It’s not just the Greens – most of the higher-polling minor parties in Australia tend to spread their vote around relatively evenly. One Nation polled just over 3% with just under 60 House candidates, while United Australia polled 3.4% running candidates everywhere. The existence of the Senate encourages this sort of behaviour, since you can win a Senate seat with votes from all over a state, and that’s much easier than winning a high enough concentration to win a House seat.

You can see the disconnect between total primary vote and winning seats by looking at the results for the Greens in the large east coast states over the last decade. The Greens now hold three seats each in New South Wales and Victoria, and two in Queensland, when they held none prior to the 2011 NSW state election. There has not been a major increase in the Greens vote in that time, but what has happened is that it has been concentrated.

The advantage for the bigger parties is made worse when you consider that “the minor party vote” is not a monolith – these voters are quite different to each other. One Nation and United Australia and Greens voters and those voting for independents. These voters won’t necessarily direct preferences to each other (indeed, the recent history of Senate elections teaches us that voters won’t do it even when their parties might prefer that they do) – if the Greens are trying to defeat a Liberal MP, many of those other minor voters will likely prefer the Liberal.

So you can rack up quite a large minor party vote for candidates who have no chance of getting elected. The large field of nominations in 2022 will likely depress the major party vote further without electing new crossbenchers. One Nation is now running 2.5x as many candidates as in 2019. They could well win 6% or 7% without winning any House seats.

The Liberal Democrats are running one hundred House candidates, up from ten in 2019. They may well chip 1% off the Liberal Party where they are running, but it won’t help them win a single House seat. I doubt it will even have much impact on the contest between the major parties.

While the number of independents is not higher than it was in 2019, my opinion based on checking the websites of so many is there are a lot more who are credible enough to poll in the high single digits without being electable. That may well show up in a big uptick in the independent vote (which was 3.4% in 2019). But the handful of candidates who can actually win will be a small share of that vote.

To illustrate this point, the following table shows the share of the total national formal vote for elected and unelected candidates for each party.

PartyUnelected candidates (%)Elected candidates (%)
United Australia3.43
One Nation3.08
Katter’s Australian0.230.26
Centre Alliance0.060.27

A slight majority of Coalition and Labor votes were cast for a successful candidate, but just 1.4% of the formal vote was cast for one of the six successful crossbenchers. Another five independents could get elected and the vote for them wouldn’t register in the national polls, even though it would have a dramatic impact on the parliament.

There are 897 candidates running who are not in the Coalition or Labor – the odds of a hung parliament are increasing because of a narrow subset of these candidates, and their votes, while consequential, won’t be a major factor in the total non-major party national vote.

A minor party or independent candidate can only get elected if they make it into the top two after preferences. That usually means making it into the top two on primary votes, although we’ve had a handful of cases of candidates winning from third (Andrew Wilkie in 2010, the Greens in Prahran).

The Australian Electoral Commission has a category they call “non-classic” which covers races where the two-candidate-preferred count is not Labor vs Coalition. I charted this in my chapter in the Morrison’s Miracle book about the 2019 election.

The number of non-classic races peaked at 17 in 2016, and dropped slightly to 15 in 2019.

As the number of these races increases, the election becomes less about a Labor vs Coalition contest and more about a complex mixture of different contests. We are seeing this more in states like New South Wales, and as non-classic races become more common the two-party-preferred vote and the standard pendulum will become less useful.

I expect the number of independents will increase the number of non-classic races in 2022, and could well see some independents win.

As the number of crossbenchers increases, the odds of a hung parliament do go up, but they still depend on a relatively close election result. If there are 3 crossbenchers, then a major party needs to win 51.3% of the remaining seats for a majority. If there are ten crossbenchers, they need to win 53.9% of the remaining seats.

There is some relationship between the general increase in the primary vote for the minor parties and independents and the chances of some of them getting elected, but it is not that close. It’s possible we could see a surge of support for independents and yet not see any of them gain a seat. Coming close doesn’t help produce a hung parliament.

I think we will see a very high vote for minor parties and independents in 2022, and there’s quite a good chance it will be a record high. I also expect we will see a surge in independent support in places where they can win, and some of them may well win.

If Labor wins the election comfortably, say with the 53% two-party-preferred predicted by this week’s Newspoll, I expect they will win a majority in the House, even if the Coalition loses a number of seats to independents. They will likely do that on a lower primary vote than any previous Labor government.

If things are closer, though, there is a real chance of a hung parliament, particularly if some new independents win their seats. While that may come alongside a record high minor party primary vote, there will still be a lot of voters left unrepresented.

Liked it? Take a second to support the Tally Room on Patreon!


  1. The furphy about the minimum necessary primaries has been around for decades – I remember David Oldfield (remember HIM?) jeering at the ALP’s low primary at the time of the 1990 election. To give Uhlmann his due, he did eventually get around to discussing preferences. At the moment, based on previous preference flows, the polling seems to have steadied (or at least paused) on a mean of 53.2% and a Monte Carlo simulation suggests an ALP seat number of 78 to 81. The polling for the Teal seats is very dodgy and it is not looking too crash-hot for most of them. Thus, as Kevin B has pointed out, a Hung Parliament, though flavour of the moment, is quite unlikely. For those interested in what might happen should Parliament be left hanging, a paper by Anne Twomey “The GG’s role in the formation of Government in a Hung Parliament” [Sydney Uni Law School Research Paper 10/85] is of interest. I was speaking to someone who knows Hurley well and they said he was well aware of the issues and the precedents discussed in Twomey’s paper

  2. A rough rule of thumb.. liberals need 40% to win at a global figure.. also in an individual seat if the liberal vote falls below 45% they are in serious danger of losing. If Labor outpolls the liberals in primary votes they will win. OPINIONS PLEASE

  3. On a general note, after analyzing some 300 contests between INDs vs Incumbent of either colour, I would reckon:
    IND has a 50% chance of winning at about 42% of primaries
    INCUMBENT has a 50% chance of losing at about 38% of primaries
    These are quite rubbery and there are some exceptions to the rule … most notably recently, when Phelps won in Wentworth in 2018 with only 28% primaries

  4. It’s illuminating to note that most of the independent psephologists such as Antony Green, Malcolm Mackerras and Jill Shepherd believe that the mostly likely result is a Labor win with an outright majority. The party that’s really talking up the prospect and dangers of a hung parliament is the Liberal party. It’s running a scare campaign that if Labor won it would be beholden to the Greens and other Independents. We have had hung parliaments before that have worked successfully. The thing is that the the Liberal party if re-elected would be beholden to the National party. We are seeing now as it’s being played out by Matt Canavan and others what this would mean on matters such as climate policy, with a reluctance on the part of some Nationals to commit to a net zero target.

  5. It’s illuminating to note that most of the independent psephologists such as Antony Green, Malcolm Mackerras and Jill Shepherd believe that the mostly likely result is a Labor win with an outright majority. The party that’s really talking up the prospect and dangers of a hung parliament is the Liberal party. It’s running a scare campaign that if Labor won it would be beholden to the Greens and other Independents. We have had hung parliaments before that have worked successfully. The thing is that the the Liberal party if re-elected would be beholden to the National Party. We are seeing now in this election, as it’s being played out by Matt Canavan and others, what this would mean on matters such as climate policy, with a reluctance on the part of some Nationals to commit to a net zero target.

  6. Teals will only win at the expense of Liberals, and Teals winning in an election that Labor is on track to win won’t lead to a hung parliament. Boothby is a possible exception with Jo Dyer and Labor both targeting it

    The real question is Greens. They could win up to 6 seats off Labor, though we will be able to rule Canberra, Cooper and Wills in or out as soon as Liberal HTVs are revealed (I suspect the latter). So if Labor are on track to win 76-78 seats, then maybe a risk of hung parliament from Griffith, Richmond and Macnamara.

    Interestingly I wouldn’t rule out Mayo, Indi and Warringah becoming notionally Labor 2PP. Labor has no hope of beating incumbents in exclusion order, so it’s trivia. But it could be interesting for hung parliament discussions, although the incumbents can easily manipulate result with HTVs.

  7. 1. With the potential exceptions of Fowler and Boothby, the serious independent candidacies tend to be in otherwise safe Coalition seats and so they will not reduce the ability of the ALP to win but will reduce the ALP primary in the otherwise safe Coalition seats.

    2. Major party candidates only need to get just over 33.33% each on 3CP and where there are 4 or more candidates they will get preferences from eliminate candidates, so it is entirely possible for major parties to stay in the 2CP with primaries a bit under 33.33%.

    3. A major example of wining majorities from low primary votes was the Bolte Liberal Government in Victoria. Its primary vote was always bellow 40% and twice bellow 37% but it never had less than half the seats in the Legislative Assembly and on all but 1 occasion had more than half (including in 1970 when the Country Party directed preferences to the ALP and gained them 3 MLAs from western Victoria and Gippsland Province in the Legislative Council).

  8. Given the lopsided flow of preferences from Greens to Labor, the combined Greens/ALP vote is more important than the ALP primary alone in indicating how successful or otherwise they’ll be. On the other hand, the minor right-wing parties don’t tend to preference the Coalition in anywhere near the same numbers (ONP -> LNP preference flow is usually around 60% or lower.) Thus the question about whether or not X party can win with Y% primary vote is far more pertinent to the Coalition than to the ALP.


Please enter your comment!
Please enter your name here