WA’s upper house counting in the dark

We have a lot of information about how the count is going for the Legislative Assembly in Western Australia – there’s a few close races, and places where we’re waiting for full distributions of preferences, but the information is reasonably clear.

Unfortunately there’s a lot less useful data for the Legislative Council, and there’s bunch of ways in which the data reporting is falling short of what we’ve got in other states using similar electoral systems.

For this I’m going to explain the data that is missing, why that makes it harder to track the count, and I’ve also put together a map showing the vote share for the six highest-polling parties (at least for the vote categories where we can identify the seat).

When New South Wales counts Legislative Council ballots, they do two successive rounds of counting. The first round totals up the number of above the line votes for each group, as well as the total number of votes that are either informal or below the line. This is pretty similar to the first count being conducted for the WA Legislative Council. So far, so good.

Unfortunately, when you look at the vote breakdowns for that upper house count, there’s something odd going on. Firstly, the WAEC is not publishing breakdowns below the statewide totals on their public website. I have managed to gather the information from the media feed. But when you look at this data, there is no mention of electorate.

For results at each booth, it is possible to identify the electorate that the booth is contained within. It’s even possible in the case of dual booths to see two different sets of data, and if you sort them by the number of votes cast you can work out which seat each half of the booth is part of.

But for special votes – mobile polling, postal votes, absent votes and any other declaration votes – there is no seat breakdowns at all. It’s just one big total for each category cast anywhere in the state.

Unfortunately those vote categories are going to be reported at an inconsistent rate. It would be really useful to know which seats have finished counting, and which ones still have more to come, which would help us understand how the total vote shares are likely to change as the remaining votes are added. But we can’t do that.

I asked the WAEC media team about this and was told:

The publishing of Legislative Council votes has changed due to a legislative amendment. Votes are now counted as Whole of State, not by Region, so separating votes by region is no longer possible. 

While it is true that the regions are no longer around, the districts (or electorates) do exist. All ballot papers are cast in a particular electorate. It shouldn’t be impossible to disaggregate voting figures by seat.

It’s also important for transparency, because it allows us to analyse patterns and find mistakes. It’s much harder to tell if one batch of votes has been transposed between the wrong groups if it is lumped in a much larger pile of votes.

Then we get to the second count, where ballot papers will need to be data entered. At the end of the process we will have below-the-line figures for each candidate.

When New South Wales does this second count, they publish daily updates of total figures for the second count, including below the line votes, and with reports on which batches of votes have been entered. This then allows you to compare like-with-like between the first and second count to identify the impact of below-the-line votes on the count. Other jurisdictions do other things – the ACT even conducts a daily interim distribution of preferences with all of the digitised results produced so far!

We will not get any of that for WA. The WAEC has said:

Until the result is declared only results by above-the-line parties and groups will be available on the website.

So there really isn’t too much to do for the Legislative Council. Special votes will continue to be added to the upper house count, but until they have all been added we won’t know what bias there is between different parts of the state. And we won’t know anything about the role of below-the-line votes until this count is over.

So all of that is unfortunate, and I wanted to explain the things that will be hindering election analysts.

I’m working on another blog post for tomorrow explaining all of the issues with the conduct of this election that demonstrate the need for better processes to review WA state elections, so I’ll come back to this topic then.

In the meantime, I’m going to publish the one bit of interesting analysis looking at upper house votes by district. I won’t try and calculate a swing, because these numbers include no postal, absent or other declaration votes. But they have some interesting patterns, most of which are not unexpected.

You can toggle between maps for the Labor, Liberal, Nationals, Greens, One Nation and Legalise Cannabis parties.

Of course Labor’s vote is highest across the metro area, while the Greens have particular peaks in some inner city seats and the south-west. The Nationals of course did best in regional areas, although did surprisingly well in Bateman and South Perth. The Liberal vote peaked in metro seats close to the Indian Ocean.

One Nation did better in regional areas, but unlike the Nationals they also did well in outer suburban seats. And Legalise Cannabis generally did better in the outer suburbs and some regional seats, doing particularly well in Kalgoorlie.

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