The recent Tasmanian election was very interesting and complex to follow, right down to the final preference distributions, but it also did a good job of translating votes into seats.
The House of Assembly was expanded in size from 25 to 35, and with that became a system more likely to produce a highly proportional result.
Now that the results have been finalised, it’s clear that this was a very proportional translation of votes into seats.
If we just start by looking at the 2024 results by party, they line up quite neatly.
The translation of votes into seats was almost perfect for Labor, the Greens and the independents. A proportional seat share for the Liberal Party would have been one seat less than the actual result, which is a pretty small winners bonus.
The Jacqui Lambie Network was slightly over-represented, but a majority of that deviation would have gone away if they had stood in Clark. Based on a comparison of how the JLN polled in the 2022 Senate election and the 2024 state election, I would expect they would have polled 4-5% in Clark and not won a seat, but that would have brought their total vote close to their seat share.
There were a few small parties that polled around 2% and didn’t win seats, but that is typical in most proportional systems.
So this information made me wonder: how did this result compare to past elections? There’s a reason to think that an increase in the district magnitude and an increase in the total number of seats would both lead to a closer translation of votes into seats.
The standard metric for measuring disproportionality is the Gallagher index. It compares the vote share to seat share for each party. If there is no disproportionality, the score is zero.
Tasmanian elections generally have much lower Gallagher scores than you get for more majoritarian elections for the federal parliament or the big mainland states. The worst score in the last four decades was almost 10 in 1998, but that is in the typical range of disproportionality for the House of Representatives, and much less than the score of 16.5 for the 2022 federal election.
A system with less proportional features can still fluke a very proportional result on occasion, but it won’t be able to do it consistently.
It does look like there was an uptick in disproportionality in 1998, when the Assembly was cut from 35 to 25. The 2002 and 2010 elections were fairly proportional, but the 2014, 2018 and 2021 elections were all amongst the least proportional.
We will need to wait and see if the 35-seat system will continue to produce such proportional outcomes, but I suspect it won’t bounce back to the disproportionality we saw last decade.
There are often arguments around whether Single Transferable Vote counts as a proportional system, but there’s no doubt in looking at these results, either statewide or in a single electorate, that it can be proportional. If there is limitations in its proportionality, it is because it is usually used in very low magnitudes. All PR systems will be less proportional when magnitudes are low. And the small size of Tasmania’s parliament means that random variations from the statewide trend are less likely to be cancelled out.
At some point in the near future I’ll be trying out some calculations to see how Tasmania’s votes might have been translated into seats under some other electoral systems.