Graphs of the Day – a few Charts on Preference Flows at the 2019 Election

This is a post I’d been playing around with and published by accident.

It was an experiment in finding new ways to graph two-party preferred and preference flows.

Seeing I’ve published the post, I should add a few comments.

Composition of Two-Party Preferred Vote

The graph and table below was an attempt to break down two=party preferred vote by the origin of the vote as first preferences.

In the graph and table below, I’ve pulled apart preference flows in the three-cornered contests, and also split the first preference support for all third parties into their destination preference. So 8.55% of the Green’s first preference vote ends up part of Labor’s two-party vote, and 1.85% part of the Coalition’s.

For all the talk of preferences deciding elections, in the end who wins depends more on whether Labor or the Coalition have the higher primary vote.

In the last two decades, Labor’s first preference support has trended down largely because of the growth in Green support. The strong flows of Green preferences to Labor, and Labor to the Greens where counted, means that support for the two parties can almost be treated as a block, in the way we tend to ignore the increasingly rare three-party contests between the Liberal and National Parties.

If the Coalition trails Labor’s first preference support then Labor is best placed to win. The Coalition did trail on 39.5% in 1998 to Labor’s 40.1%, but the Coalition won that election with 49.0% of the two-party preferred vote through the vagaries of electing parliaments from single-member electorates.

Labor trailed 4% in 1990 and won narrowly, and trailed by 5.3% in 2010 and won narrowly. That Labor can win from behind as in 2010 comes down to the strong flows of Green preferences.

The 2019 election was interesting in that there was an increase in non-Green third party voters giving preferences to to the Coalition, or as shown in the graph below, of fall in the proportion giving preferences to Labor. That’s largely because of stronger flows of One National and UAP preferences to the Coalition than at previous elections. Whether that was a consequence of the 2019 election campaign or a longer term trend may be revealed by the 2022 election.

The above graph is taken from a previous post where I broke down preference flows from the 2019 election.

Detail on Composition of Two-Party Preferred Vote

to Coalition to Labor
Preference Source Votes Pct Votes Pct
Liberal/LNP first prefs 5,246,977 36.81 0 ..
National first prefs 575,704 4.04 0 ..
Labor 0 .. 4,752,160 33.34
Intra-Coalition preferences 70,130 0.49 14,064 0.10
Greens 263,380 1.85 1,219,093 8.55
One Nation 286.049 2.01 152,538 1.07
United Australia 318,413 2.23 170,404 1.20
Independents 194,826 1.37 285,010 2.00
Others 388,884 2.73 315,311 2.21
Total Two-Party Preferred 7,344,813 51.53 6,908,580 48.47

Note: All percentages are calculated as a percentage of total formal vote.

Preference flows at the 2019 Election

Below is a summary graph of the same data that I published in my previous post on preference flows at the 2019 election.

2019 Two-Party Preferred by Vote Type

The graph below shows the two-party preferred vote by vote type at the 2019 election. As the graph shows, Labor won a narrow majority in the two categories of votes cast on election day, Ordinary votes (50.4%) and Absent votes (53.6%). But the election was decided before election day, the Coalition winning clear majorities with Pre-Poll votes (53.8%) and Postals (57.6%), all adding up to an overall 51.5% for the Coalition once all the votes were counted.

3 thoughts on “Graphs of the Day – a few Charts on Preference Flows at the 2019 Election”

  1. Love the look behind the curtain! You should post drafts by accident more often 🙂
    If you’re still playing with visualisations for preference flows you should check out “Sankey Diagrams”. Initially used to represent energy flows they would be great for an “at a glance” view that you can also pick away at for detail.

    COMMENT: I have played with Sankey diagrams but I’m not convince they are that useful. You can show the formal distribution of preference with them, but analysing preference flow data is more useful than distributions. And I think stacked bar charts are more useful for preference flow data.

  2. I wonder how much of the advantage that ALP has in absent voting is because their poll workers have access to How to votes for all electorates. Libs often do not have HTV for other electorates.

Leave a Reply