South Australia’s tiny wards

South Australian voters are currently going to the polls to elect their local councillors for the next four years. Voting takes place by mail, with polls closing on November 10. Incidentally this means South Australian council elections should continue to closely mirror Tasmania’s council elections, which take place this month on a four-year cycle.

I was asked if I would do a podcast which I considered but it’s going to be too much of a lift to be up on the issues and teams running in even the largest councils. I remain convinced that political divisions, the potential for a current majority to lose power, exist in many non-partisan councils, but it’s much harder for me to analyse those stories in the absence of party labels.

Instead I decided to examine the electoral systems of South Australia’s councils, and how they have changed over the last decade and a half.

In theory, South Australia uses “proportional representation”, but as I have explained in the context of New South Wales and the Northern Territory, the level of proportionality depends a great deal on the magnitude of the district – ie how many representatives are elected from a single area. If you use STV methods on a single district, it basically turns into preferential voting, but that’s not proportional. Magnitude-2 elections aren’t much better, and there is a huge difference in proportionality between magnitude-3 and magnitude-9.

There are 67 councils in South Australia. I have collected results data back to 2006, which is now in the data repository, and in that time there have always been 67 councils, apparently with no boundary changes (although a couple of councils have changed names).

There are seven local government jurisdictions in Australia (the ACT does not have local councils). Some form of proportional representation is used in five of these jurisdictions. Queensland uses single-member wards with preferential voting in most urban councils, with block voting in most rural councils. Western Australia uses first past the post voting for single-member wards and block voting (a similar plurality system) for multi-member wards.

At least four jurisdictions use a process of independent “redistribution review” coordinated by the electoral commission to make periodic decisions (usually after every second council election) about whether the ward structure for the council is appropriate and to redraw ward boundaries as needed: Queensland, Victoria, South Australia and the Northern Territory. By contrast, while there are rules about how much population can vary from the average in New South Wales council wards, and referendums are required to change the number of councillors or magnitude of wards, the actual boundaries are left up to the council, and they are usually the key mover in any ward structure changes.

Amongst the 5 PR jurisdictions, there are different rules about what sorts of ward structures are permissible.

In Tasmania, there have been no wards at all since 1989. Since 2014 the entire council has been elected at once, which doubled the magnitude pretty much overnight so now all areas elect between 7 and 12 representatives.

In New South Wales, there is a requirement that all wards in the same council have the same magnitude. I don’t believe there is a prohibition on single-member wards, but there have been no such wards since Botany Bay council was merged with its neighbour in 2016, and Botany Bay had only used single-member wards for one council in 2012. I found no other single member wards used as far back as 2004.

In the Northern Territory, Victoria and South Australia there are mixtures of single-member and multi-member wards within the same council.

In this 2020 blog post, I documented the last two decades of change in Victorian ward structures, which saw a shift towards higher magnitudes before the current government imposed a preference for single-member wards prior to the 2020 council election. For the rest of this post, I’m exploring similar trends in South Australia since 2006.

First, to keep things simple, this chart shows the average number of councillors per ward across the state. There has been a gradual increase in magnitude, with the biggest increases in 2006 and 2022. Over 16 years, the average has increased from 3.27 to 3.68.

Why has this change taken place? For the next two charts, I have decided to use data which is weighted according to the share of enrolment, not a raw count of wards, which by its nature will tend to overcount rural councils but also overcount small-magnitude wards, since a council with 5 wards of 2 will count as 5 wards whereas a council with 10 councillors at large will count as one ward. ECSA only publishes enrolment data by council, so I have estimated enrolment by ward by dividing up the council's enrolment based on the number of councillors per ward.

This first chart shows the different types of ward structure: a mixture of single-member and multi-member wards, multi-member wards of the same magnitude or different magnitudes, or undivided councils. There are no councils consisting solely of single-member wards (as in Victoria). If this was New South Wales, you would only have multi-member same and undivided.

Single-multi mixtures are not very popular, and have declined substantially. The number of undivided councils have increased, and the number of councils with consistent ward magnitudes has also gone up. This has caused a big decline in variable-magnitude councils.

This next chart shows the number of voters who vote in a particular magnitude ward:

It's worth noting for this one I have excluded the at-large councillors from the City of Adelaide. This council has an unusual structure where there are councillors representing wards, and voters cast another vote for councillors representing the whole council. The same system was also used for the City of Campbelltown in 2006.

About half the state voted in M2 wards in 2006, but this number climbed up to 60% as of 2018, and has now dropped slightly.

M4 wards have dropped significantly in popularity since 2014, while M3 wards have become more popular after dropping in 2010. At the higher end of the spectrum, undivided councils with ten members have lost ground to M9 councils.

It's worth noting that a majority of SA councils have a directly elected mayor - this number has increased from 70% in 2006 to 79% in 2022 - and that position is not counted in these calculations.

While the enrolment-weighted data is more informative, there are some interesting differences when you look at the raw counts. A majority of SA councils are undivided, even though they only cover about 20% of voters. And there was a decline in the number of consistent-magnitude multi-member ward councils from 2006 to 2010, and then increasing since then. I think this inconsistency is because there are quite different trends in urban councils (which make up a majority of the population but a smaller proportion of councillors, wards and councils) and rural councils.

In 2010, a number of rural councils switched from wards to undivided structures, and that trend has continued more slowly since 2010. While this has happened, only two urban councils have undivided structures. There has been a slight shift from different-magnitude wards to same-magnitude wards in urban SA, but mostly those councils have the same structures as they did in 2006.

And if you look at the proportion of wards of a particular magnitude, the number of M2 wards in urban SA has increased from 72% in 2006 to 83% in 2022, while the same statistic in rural SA has dropped from 36% in 2006 to 17% in 2022.

And when you go back to the original chart, but split the data between urban and rural, you see very clearly two different trends.

Average magnitude in urban SA has declined slightly from 2.45 to 2.32, while in rural SA the magnitude has increased from 3.85 to 5.

This is due to urban councils shifting more to M2 wards, while rural councils have become much more likely to go without wards.

So what to make of this approach? I'm not a fan of magnitude-2 wards. They are barely proportional at all, and they unnaturally divide councils up into tiny areas. Particularly in urban areas, I don't see why councillors can't represent the whole area, or at least a larger ward. South Australia outperforms Victoria, which has now shifted decisively back towards single-member wards, but it's low magnitudes in urban SA compares unfavourably to New South Wales (where the average magnitude has varied between 4.6 and 4.84 over the 2004-2020 period) and Tasmania (where the average magnitude is 9.07).

While proportionality is reasonably strong in rural SA (where two thirds of councils have no wards), the same cannot be said of the large urban councils which contain most of the populatiion.

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