How have countbacks performed in NSW?

I wrote a blog post last weekend about how the NSW government has been sneaking in a change to how council vacancies are filled into a bill that is billed as primarily ending the privatisation of council election administration.

Since then I’ve had the chance to do a bit more analysis on the topic and I wanted to provide more data to inform the debate. This was too late to be included in my inquiry submission so hopefully it can be found useful by MPs anyway.

Since countbacks were first introduced in 2022, we now have 30 vacancies that could have been filled using countbacks across two council terms, and it gives us quite a lot of data. I have also looked at examples of where councils made a decision about whether to use by-elections or countbacks when they knew that someone was due to resign, examples of how absurd it is to give this to councils to decide in that moment.

I’ll also look into what proportion of councils would move to a party-replacement system, and how many would switch to by-elections, under Minister Hoenig’s new system. Hoenig’s numbers are technically accurate, but actually quite misleading.

Under current NSW law, each local council gets to decide at the first meeting after a council election as to whether they will use countbacks or by-elections. If the council opts for countbacks, they can only be used in the first 18 months of the term. For the remainder of the term, by-elections are restored as the method of filling vacancies, but councils can ask to leave seats vacant if the vacancy falls in the last 18 months of the term (and those requests are always approved). So in practice in a standard 4-year term, vacancies the first 18 months would by filled by countbacks, the middle 12 months would be by by-elections, and the last 18 months would leave the seat vacant. It leaves plenty of opportunity for the timing of vacancies to be manipulated.

The 2021-24 term was a shortened one, with just 33 months between elections due to the COVID-19 pandemic.

Putting aside five by-elections caused by insufficient candidates nominating at the general elections, and results being voided in three councils due to the iVote failure, 18 vacancies were triggered during the 2021-24 term. 16 of these went to countbacks, and 2 went to by-elections.

Exactly half of these eighteen vacancies were from councillors elected in a group – relevant because the new Hoenig system would treat these councillors differently than others. Grouped councillors would be replaced by someone in their group, while ungrouped councillors would be replaced via by-election.

Of the eight grouped councillors replaced via countback, every single one was replaced by someone from the same group. For the one replaced via by-election, the Liberal Party gained a seat off Labor in Ryde.

For the eight other countbacks, the outgoing councillor was not part of a group, and neither was their replacement. In most of these cases, all councillors were ungrouped. These are the exact circumstances where Hoenig’s new system would revert to by-elections, but the countbacks worked. There was just one by-election for an ungrouped councillor’s seat, in the D Ward of Lachlan council.

It has been less than ten months since the 2024 local government election, yet there has already been 16 resignations triggered by councillors retiring or dying. Just one was a by-election.

In four of those 15 by-elections, an ungrouped candidates was replaced by another. In ten of the other eleven, the replacement came from the group of the departing councillor. The only exception was Inverell, where the departing councillor’s group had no more eligible councillors and he was replaced by an ungrouped candidate.

The only exception was the Mid-Western council, which made a controversial decision to go to a by-election in the full knowledge that Labor’s Sharelle Fellows was about to resign due to poor health. Labor ended up winning the by-election, but it was seen as a transparent attempt to alter the balance of the council.

I have also found three other examples where councillors are likely to have known that a specific councillor may be leaving soon after the council makes that decision at the first meeting, and may lead them to make the decision one way or the other – if you’re in a council majority, you’d be tempted to use a countback if the departing councillor is in the majority and a by-election if they are not. It is a clear conflict of interest and leads to biased decision-making.

In Upper Lachlan at the first meeting in 2022, John Stafford announced his resignation. I don’t know how that influenced the decision on countbacks.

In Shoalhaven in 2024, Shoalhaven Independents member Mitchell Pakes resigned the day after the meeting where the countback decision was made. This was criticised by former councillor Paul Ell, a Liberal member who likely would have been Mayor if the party had managed to submit his nomination form on time. While the councillors making the decision may have had information that influenced their decision, Ell was not neutral either – he would have had a good chance of coming back to council if there was a by-election.

Finally, in Glen Innes Severn in 2024, Rob Banham lost the mayoral election 4-3, and resigned soon after. Presumably Banham was aware of the chance he might retire, and possibly some of his allies on the council.

Evidence suggests that countbacks have been enormously popular with councils. In the current term, all but 28 councils have opted for countbacks, and one quarter of those 28 councils are those where the council was entirely uncontested, so a countback is not possible. There are 100 councils which are using countbacks in the current term.

It’s not surprising why. They are much cheaper than countbacks, both for the council and for political participants. They maintain the political balance on the council. They are also democratic, because the political equality between voters is maintained. A by-election doesn’t do that – some voters remain represented by their original choice while getting another bite at the apple.

I want to respond to a couple of the specific claims Minister Hoenig makes in his speech.

The minister said:

The use of countbacks to fill vacancies can result in councillors being elected who received little to no support at a general election. The current system also discourages councillors from resigning from office when their life circumstances change. They may be replaced by a political opponent, resulting in a change in the political balance on council. This has the effect of keeping councillors around longer than they may wish. Maintaining stability within councils is imperative to good governance and ensuring that councils make decisions that are in the best interests of the community they are elected to represent.

There’s quite a lot that is wrong here. Plenty of councillors have resigned due to countbacks, and it is almost always the case, where political affiliation is clear, that councillors are replaced by like-minded representatives. This provides tremendous political stability compared to by-elections. Hoenig is implying a level of instability that just doesn’t exist with countbacks.

I want to push back on the idea that we should be encouraging maximum convenience for councillors to resign when they don’t want to be on council anymore. Taking on a role in public office can be a burden, but it is something people have freely chosen to take on. Wanting to resign but not doing so because of your likely replacement suggests you don’t really need to resign, and maybe you just need to finish your term. Making it maximally easy to replace a councillor with just about anyone leaves you with the Brisbane City Council situation, where no councillors retire at an election and a large proportion of councillors first took office without anyone voting for them. Countbacks do often make it relatively easy for a councillor to resign, and be replaced by a political ally.

Those elected at countbacks have done so because they have a certain amount of political support, even if it’s not first preferences. It is false to imply that people are elected with no popular support.

Most of the time, countbacks would produce the same result as the same-group replacement method Hoenig is proposing. This procedure would be universal where a councillor is in a group, and someone else in that group is in a position to take the seat. So in one sense it would be an improvement because it eliminates the silly process of a council deciding which procedure to use each term. But it wouldn’t always produce the same result, and when it doesn’t, the countback makes more sense.

We have an example from just a few days ago, in the Mortdale ward of Georges River. Labor councillor Ash Ambihaipahar resigned from her council seat after being elected to the federal seat of Barton.

Mortdale was one of the wards where the Liberal Party failed to nominate in 2024. Labor won two seats, while the Georges River Residents & Ratepayers Party (GRRRP) won one. Labor’s second candidate defeated the second GRRRP candidate by just 363 votes, or about 2.5% of the total formal vote. But when the countback was conducted earlier this week, the third Labor candidate defeated the second GRRRP candidate by just seven votes.

While the Labor above-the-line votes all flowed from the ineligible Ambihaipahar down to the third Labor candidate, she also had 778 below-the-line votes. If we compare the original count to the countback at the point where there was just one Labor candidate and one GRRRP candidate left standing, Labor’s vote is down by 199 votes, and GRRRP are up by 150 votes.

The minister would have you believe that if GRRRP had closed that seven-vote gap and won the countback, that it wouldn’t have reflected a deviation from democracy. But I don’t think so – it reflects that there were people who chose to vote for that individual Labor councillor but not for her third-placed colleague.

There’s one more quote from the minister that I would like to respond to:

The changes will allow for vacancies to be filled without a by-election at any time during a council term for candidates who ran on tickets. The result is that, for the 85 per cent of residents in New South Wales who live in council areas where councillors are predominantly elected from tickets, by­elections will become an unusual occurrence.

This is a misleading statistic. In urban areas and large regional councils, grouped candidates are more common. but many countbacks are currently being conducted in small rural councils, and they work well. Those areas would be condemned to by-elections. And because there are a lot more councillors representing those areas, the numbers of such vacancies is likely to be a lot more than those councils’ share of the state population. This table shows that wards where all successful candidates were from a group did make up 85% of the electoral roll in 2024, but just 50% of councillors.

The minister’s speech and the explanatory note do nothing to explain what is wrong with how countbacks work now, and why they can’t continue to be used for small rural councils. I hope to have the chance to make these points to the Committee.

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