In Australian politics, the two-party-preferred vote (TPP or 2PP) is the result of an opinion poll or a projection of an election result in which preferences are distributed between the two major parties, being the Labor Party and the Liberal-National Coalition, e.g. "Coalition 50%, Labor 50%". The preference distribution is usually based on the results of the most recent election, with votes for other candidates being distributed between the two major parties.
As such, the TPP is a rough indicator of voting intent that focuses on determining the likely majority in the lower house. It is compared to previous values to predict the swing and hence the likelihood of a change in government between the major parliamentary parties.
The TPP assumes a two-party system of government, i.e. that after distribution of votes from less successful candidates, the two remaining candidates will be from each of the two major parties. It provides no indication of the number of representatives of other parties or independent views on the cross-bench, and as the proportion of votes for other candidates increases the TPP becomes less useful. It cannot predict a hung parliament as it does not quantify the alternatives to the two main parties.
The TPP is often confused with the two-candidate-preferred vote (TCP). The TCP is the electoral penultimate result for an electoral division where preferences have been distributed using instant-runoff voting. The winner of the contest is the candidate with over 50% of the TCP vote.
Unlike the TCP, the TPP is informative only and has no direct effect on the election outcome. It is an indicator used for analysing results above seat-level, such as a national or statewide TPP. For individual seats the TCP is the preferred indicator, because when the final two candidates are from the major parties, the TCP will have the same value as the TPP, and when at least one of the final candidates is not from a major party the TPP is misleading, not informative.
The full allocation of preferences under instant-runoff voting is used in the lower houses of the Federal, Queensland, Victorian, Western Australian, South Australian, and Northern Territory parliaments, as well as the upper house of Tasmania. The New South Wales lower house uses optional preferential rather than full preferential voting – with some votes giving limited or no preferences, TPP/TCP is not as meaningful. TPP/TCP does not occur in the Tasmanian lower house or the Australian Capital Territory due to a different system altogether, the Hare–Clark single transferable vote system. Aside from Tasmania, TPP/TCP is not used in any other upper houses in Australia, with most using the single transferable vote method of proportional representation.
History Australia originally used first-past-the-post voting, as is currently used in the House of Commons of the United Kingdom. Full-preference instant-runoff voting was introduced at the federal level after the 1918 Swan by-election, and has been in use ever since. In that by-election, candidates from the Australian Labor Party, the Nationalist Party government (predecessor to the United Australia Party and Liberal Party of Australia), and the emerging National Party of Australia (the then-"Country Party") all received around a third of the vote. However, as Labor had a plurality of three percent, it won the seat. The new system allowed the two non-Labor parties to compete against one another in many seats without risking losing the seat altogether.
It is increasingly uncommon for seats to be contested by more than one Coalition candidate. For example, in the 2010 federal election, only three seats were contested by more than one Coalition candidate. With the popularity of parties such as the Greens and One Nation, preference flows are very significant for all parties in Australia.
Not distributing preferences was historically common in seats where a candidate received over 50 percent of the primary vote. Federal seat and national TPP results have only been produced as far back as 1937, though it was not uncommon in subsequent decades for major parties at federal elections to not field a candidate in a few "safe" seats, but since 1972, all seats at federal elections have been contested by the major parties. Full preference distributions have occurred in all seats since 1983.
Until recently, South Australian state elections had boundaries strategically redrawn before each election with a fairness aim based on the prior election TPP vote, the only state to do so. The culmination of the historical state lower house seat malapportionment known as the Playmander eventually saw it legislated after 1989 that the Electoral Commission of South Australia would redraw boundaries after each election with the objective of the party that receives over 50 percent of the TPP vote at each forthcoming election forms government. Nationally in 1983/84, minor gerrymandering by incumbent federal governments was legislated against with the formation of an independent statutory authority, the Australian Electoral Commission.
Procedure Under the full-preference instant-runoff voting system, in each seat, the candidate with the lowest vote is eliminated and their preferences are distributed; this process repeats until only two candidates remain. Whilst every seat has a TCP result, seats where the major parties have come first and second are commonly referred to as having a TPP result. In a TCP contest between Labor and the NSW/Vic Nationals and without a Liberal candidate, this is also considered a TPP, with the Nationals in these states considered a de facto major party within the Liberal/National Coalition. In seats where the major parties do not come first and second, differing TPP and TCP results are returned. When only one of two major parties contest a seat, such as at some by-elections, only a TCP result is produced. Swings in Australian parliaments are more commonly associated with the TPP vote. At the 2013 federal election, 11 of 150 seats returned differing TPP and TCP figures ("non-classic seats"), indicating a considerable two-party system.
The tallying of seat TPP results gives a statewide and/or national TPP vote. Non-classic seats have votes redistributed for informational purposes to the major parties so that every seat has a TPP result. Whilst the TCP is the determining factor in deciding which candidate wins a seat, the overall election TPP is statistical and indicative only, as swings in seats are not uniform, and a varying range of factors can influence marginal-seat wins with single-member electorates. Several federal elections since 1937 have seen a government elected with a minority of the TPP vote: 1940 (49.7%), 1954 (49.3%), 1961 (49.5%), 1969 (49.8%), 1990 (49.9%) and 1998 (49.0%).
As the TPP vote rather than the primary vote is a better indicator of who is in front with seats won and lost on a preferential basis, Australian opinion polls survey voter intention with a TPP always produced. However, these TPP figures tend to be calculated based on preference flows at the previous election rather than asked at the time of polling. The difference between the two is usually within the margin of error (usually +/– 3 percentage points). History has shown that prior-election preference flows are more reliable.
Three-candidate preferred (3CP) The three-candidate-preferred count (3CP) refers to the second-to-last stage of preference distribution, where three candidates remain. Historically this was not a figure of particular importance, as the primary vote for candidates outside of Labor and the Coalition was low, and thus the top two candidates usually had a lead over any other candidates in the count for House of Representatives seats. However, as the major party vote has fallen, it has increasingly been the case that the final two pairing will not be clear based off primary vote results, and three-candidate-preferred counts have been conducted to determine this.
The Australian Electoral Commission (AEC) first conducted 3CP counts at the 2022 federal election, in four seats. In Brisbane, the Greens candidate Stephen Bates had placed third on first preferences, but trailed Labor candidate Madonna Jarrett by only eleven votes. Both candidates would have defeated Liberal National incumbent Trevor Evans on the final count, but Bates overtook Jarrett after the distribution of preferences, receiving 30.1% to Jarrett's 28.4%, and polling 53.7% on the two-candidate-preferred against Evans after Jarrett's elimination. The Labor-held seats Macnamara and Richmond also saw gaps of less than 3.5% between the first and third-placed candidates. In both seats, a third-placed Coalition candidate overtook a second-placed Greens candidate after preferences, and Greens preferences easily elected the Labor candidate.
At the 2025 federal election, the AEC conducted 3CP counts in thirteen seats. In Ryan, the first-placed Liberal National candidate, Maggie Forrest, had clearly lost, but whether she would face Greens incumbent Elizabeth Watson-Brown or Labor candidate Rebecca Hack was unclear. The 3CP count showed Watson-Brown leading Hack by 0.78%, and thus Watson-Brown was re-elected. The divisions of Fisher, Forrest, and Grey were contests between the Coalition, Labor, and independent candidates, and it was unclear whether Labor would make the final count and be defeated by the Coalition, or whether the independent would make the final count and potentially win, as independents generally benefit more from Labor preferences than vice versa.
The increasing complexity of preference distributions can incentivise tactical voting. At the 2024 Northern Territory election in Fannie Bay, Laurie Zio of the Country Liberal Party defeated Labor incumbent Brent Potter, who placed third behind Greens candidate Suki Dorras-Walker on both primary votes and the 3CP count. Dorras-Walker was defeated by Zio after preferences were distributed, but the two-party-preferred count, between the Country Liberal and Labor parties, was won by Potter. This meant that, had 54 Greens voters instead first-preferenced Labor, Dorras-Walker would have been eliminated before the final count, and Potter would have defeated Zio. This phenomenon has been described as the Fannie Bay effect. It existed in reverse at the 2014 Victorian state election. In Prahran, the Greens candidate won a seat against the Liberal Party that the Labor candidate would have lost, judging by the two-party-preferred count. Additionally, a number of teal independents have performed better than Labor after preferences. However, as with the case of the Greens, this has not universally been true, as Labor was likely to have performed better against the Coalition in 2025 for the seat of Bradfield, where independent Nicolette Boele won with only 50.01% at the final count. Analysis After the count has taken place, it is possible to analyze the ultimate preference flows for votes cast for the parties that were ultimately excluded from the TPP calculation, in order to determine if the composite flow would have significantly affected the final result. Such an exercise is shown for the 2017 by-election in Bennelong:
+2017 Bennelong by-election – preference flow data Party Candidate First preferences % preference toVotes%LiberalLabor GreensJustin Alick5,6886.819.780.3 ConservativesJoram Richa3,6094.386.513.5 Christian DemocratsGui Dong Cao2,6263.172.427.6 ScienceJames Jansson1,0411.239.460.6 Sustainable AustraliaWesley Folitarik9951.248.951.1 Affordable HousingAnthony Ziebell7410.944.755.3 Liberty AllianceTony Robinson7190.979.021.0 ProgressivesChris Golding4250.542.157.9 People's PartyJames Platter1860.248.951.1 Non-Custodial ParentsAnthony Fels1320.256.143.9Totals16,16219.251.248.8
Preference flows in federal elections