Advance Summary
1. Traditionally, on contested votes the Hobart City Council is loosely divided between "pro-business" councillors and councillors who stress environmental issues and/or the interests of impacted residents.
2. This term of Council has been highly factional and these patterns have been stronger overall than in previous terms, especially since early 2024.
3. While councillors vote independently on particular motions, all councillors across this term bar one can be classified based on how they most often vote as members of either the dominant "green/left" cluster of councillors or the minority "blue/right" grouping.
4. The only councillor who does not belong to a cluster across the whole term, Louise Bloomfield, initially tended somewhat to vote with the "green/left" side but later in the term has voted far more with the "blue/right" side, and would be a clear member of the blue/right cluster if just the later 2.5 years of data were used.
6. A possible ordering of councillors from "greenest" to "bluest" for the whole term is: Kitsos, Reynolds, Burnet (not currently on council), Sherlock, Dutta, Lohberger, Posselt, Harvey, Bloomfield, Kelly, Behrakis (not currently on council but a candidate), Coats, Elliot, Zucco
7. Each of the "green" and "blue" clusters includes members who are more or less diehard but in this term there was no relationship between this and whether that councillor is a known member of a party or not.
-------------------------------------------------------------------------------------------------------------
Mercifully to all, the noisiest and most contentious term in the recent history of the Hobart City Council will very soon be over. The election held by post from October 5-27 will be the second since the introduction of compulsory voting but the first since the reduction in councillor numbers from twelve to nine passed parliament with less resistance than might have been expected. There has still been nothing done about the ludicrously low bars for running for local government, so the councillor numbers cull won't do much to stop there again being far too many candidates. But that said the toxicity of this term might, with a conspicuous shortage of new female contenders at this stage. Darcy Murphy's tracker already lists 28 declared candidates, though that is down five on the same point in the previous cycle.
Council politics isn't party politics, and even when candidates are elected together under a party banner they are not expected to necessarily vote with each other all the time, and don't. But it can be factional, and this tendency has waxed and waned in HCC voting over the years. Amid what I here find to be a fully justified belief that this term of council has been very tribal, some candidates have been distancing themselves with labels such as "crossbencher", "moderate centrist" (is there some other kind?) and the good old "independent" (comes in "true" and vanilla flavours). But the numbers tell a rather different story.
As I noted in the 2018-22 edition, it's often hard to judge new candidates, who have no voting records to classify them alongside the incumbents. For this reason we should be very wary of new candidates boasting that they'll be above all this; it is untestable and it's probably not true. But for the incumbents there is a vast store of data. I've been classifying HCC voting patterns since the 2005-7 term (see also on this site 2011-2014 and 2014-2018. While councillors always display some degree of independence from each other there has long been a tendency for most councillors in any given term to fall into one of two camps.
These I've called the "blues" (the right flank of council, a generally "pro-business" cluster which often includes Liberal and ex-Liberal party members) and the "greens" (the left flank, a generally "pro-environment" cluster which often includes past and present Greens.) Some councillors who align strongly with one of these clusters don't actually have any party history, and some councillors in any given term may not align strongly with either. Councillors with Labor connections might be found in either cluster, and over the course of a career some councillors move in and out of clusters or in very rare cases change sides. (OK, that's happened precisely once, and he's now done it with the river too.)
I used to write that the blue/green divide most often arises on development issues but in this term it's become clearer that it arises in different directions on different kinds of developments. That is that there are development proposals that the "blues" like and the "greens" don't, and vice versa, depending on whose politics they are attractive to. This is not to say that most of the items the council votes on have anything directly to do with developments.
Historically the council tended to have more "blues" than "greens", although the "blues" struggled to win the leadership positions until the 2011-2014 and 2014-2018 terms. In the last two elections the "green" side has won Lord Mayor and Deputy and has been increasingly competitive at councillor level, to the point that in this term it dominated the council.
Methods
The great majority of council motions are unanimous. Often a proposed development doesn't actually bother anyone but council approval is required; these motions aren't politically interesting but it is worth noting there's a lot that councillors of different viewpoints do work together on productively.
My interest here is in the politically informative contested motions, especially those with more than one vote on both sides. I went through the Council minutes that have been posted for this term and recorded voting on every non-unanimous motion with a "voting record" showing. (This can arise through divisions or through a small number of councillors asking that their dissent be recorded.) Where a vote repeats the same pattern as an earlier vote on the same agenda item I ignore the duplicate unless it has a different winner (this mainly happens where there are ties with the voting lineup flipped - which is relevant to my assessment of which Councillors are most often on the winning and losing sides.) If an agenda item occurs across multiple meetings (eg a resubmission of a rejected proposal) I treat the two items as distinct as frequently the conditions of the item will have changed.
In previous articles I have only used the main full Council meetings. In this term however the structure of council meetings changed. Routine development applications of the form "16 BEASLEY BOULEVARD: CHANGE OF USE TO NOODLE BAR AND DUST REMOVAL" now appear in the Planning committee meetings at which, in theory but often not in practice, all councillors attend and vote. My sample of a total of 512 motions includes 87 of these and another 11 from the Hobart Workshop Committee which also has pretty high attendance, sometimes. Of the 512 votes used, 405 were "multiply contested", meaning at least two voters on each side. I use these as my main measure of voting alignment - the lone dissents can be idiosyncratic and can be more about whether a particular councillor can be bothered having their dissent recorded, though there are some patterns within them.
Behind planning applications, a rather high number of contested motions this term were about council policies for the city, and internal governance also featured prominently. CBD issues (such as Collins Street bike paths and scramble crossings), responses to state government on local government related discussions and short stay accommodation were also common items. The 2022 election was strongly influenced by the University of Tasmania city move but it has been a fairly minor player in voting patterns in this term. For all the noise heard from time to time about council involvement in overseas issues (eg Gaza) these accounted for under 1% of all contested votes in this term.
For each pair of councillors I find all multiply contested votes where they both voted and find how often they agreed and disagreed with each other. This results in an agreement matrix which I then sort into loose groups, calculate agreement ratios involving councillors from each group and then put in order. This Council term saw two mid-term changes, first with Will Coats replacing Simon Behrakis and secondly with Gemma Kitsos replacing Helen Burnet. This means Behrakis was never on this Council together with Coats or Kitsos and Burnet was never on it with Kitsos. I've assigned some estimates of how often they would have voted together based on surrounding percentages for the purposes of the largely useless principal components graph below (78%, 26% and 80% respectively)
My sample for this council goes to the end of August. The manual data entry is not guaranteed to be error free but any errors should be immaterial to the overall patterns. I don't have time to do articles like this for all councils but am again intending a smaller scale article for Clarence subject to time and will consider doing them for other councils if somebody else constructs an appropriate spreadsheet of votes from those councils' minutes (contact me to discuss if interested but I will be offline a lot between Sep 17-30).
Participation in contested votes
The following is my record of how often each councillor voted on contested motions (including lone dissents) that are recorded in my sample (which as discussed above omits most duplicates). General council meetings are now monthly. The other committees I include meet more often but will not always generate any contested votes at a meeting. Councillors are often absent from contested votes because they are on leave, either travelling away from Hobart (often on official council business including "professional development") or with other commitments. Louise Elliot was also suspended for some meetings in the term. Councillors must leave the room to avoid voting on motions they have conflicts of interest on. Conflict of interest is, mostly, quite broadly defined and can even include just having commented strongly on an issue before it came before Council. However, absence (for all or part of a meeting) is by far the larger cause of not voting for most councillors.
These figures are not the same as attendance figures (simple ratios of how many meetings a councillor was present at), which tend to be somewhat higher. Raw attendance figures can be somewhat misleading because some councillors are more likely than others to attend only part of a meeting for whatever reason, and some are more likely to abstain through conflict of interest. On the other hand, a councillor with a high attendance rate might just happen to miss the meeting at which there were many contested divisions. I give what I get, unofficially, as the rate of recorded full Council meeting attendance up til the end of August 2026 in the final column.
Despite criticism of some councillors' attendance rates - in particular Coats was at one stage called a "FIFO councillor" because he missed a number of planning meetings while working in Melbourne - the average "Voted On" score in this council is actually identical to the previous term at 90%.
Under the Local Government Act, if a councillor remains in the room and does not vote with either side of a motion then this "abstention" is classed as a vote against. Such "abstentions" have been extremely rare on Hobart Council in the past despite a culture of them happening often on Clarence council. However this term saw a record total of twelve - five by Mike Dutta, two each by Bill Harvey, Elliot and Louise Bloomfield and one by Ben Lohberger. Those by Dutta were all around the investigation into the long-running home snacks affair involving Zelinda Sherlock (his daughter); these "abstentions" were rightly controversial (and I believe have since ceased) as in some cases they were converted to a vote against taking action. One of the Harvey "abstentions" was also contentious as it caused a motion to double rates on whole-home short-stay accommodation to fail, which led to friction between Harvey and Ryan Posselt especially and unsuccessful calls for the Greens (who at state level support reining in short-stay) to disendorse Harvey.
Lone Dissents and Losing Votes
Some other things I look at include who casts the most lone dissents and who is most often on the losing side of motions. The following table is ordered by who is least often on the losing side of multiply contested motions (those which have at least two votes on either side). It also includes each councillor's rate of lone-dissenting (out of all motions they voted on) and how often each councillor is on the losing side of all motions. I have changed the ordering to least often for a reason.
In this term Marti Zucco was by far the commonest lone dissenter with 36 lone dissents; Posselt had 14 and Coats ten. At the other end of the scale Lohberger and Sherlock had one each and I didn't find any for Kitsos. Kitsos was the least often on the losing side of a motion, and what is unusual about that is that in past councils often the councillor who is least often on the losing side is a key swing vote in the centre of the council. But in this council, the sorting of least to most often on the losing side very closely matches the left to right sort I give below, and the lone dissent rate also matches to a degree. (Posselt is a bit of an exception on both fronts, and that is not surprising given that he is an ALP member whereas the rest of the left is mostly Greens or councillors with past Green connections).
Despite the larger sample there were only 27 tied votes in this term, down from 44 in the previous. The change to nine councillors will get rid of tied votes in cases where everyone is voting, but they can still happen when an odd number of councillors are not voting.
Agreement Matrix
The following matrix shows how often (as a percentage) each councillor voted with each other councillor, subject to them both being present and to there being at least two councillors on each side of the vote. For instance, the matrix shows 2022 mayoral rivals John Kelly and Anna Reynolds voted together in such contested motions 32% of the time. With apologies to colour-challenged readers, I've used dark green and dark blue to highlight agreement scores above 75% (a commonly used cutoff for finding clusters in this kind of work) and lighter green and blue to highlight scores that are close to that level, in the range 70-74%.
A strong disclaimer again here: "green" for the alignment scores means small-g and not the party; I've even taken out the capitals.
Party affiliations are shown in brackets and by colour where known, though only the Greens formally endorsed candidates on any scale. If any councillors are no longer in the party shown, please let me know and I will update or remove that party listing. Elliot - who was briefly a Liberal - joined the Tasmanian Nationals in September 2025 but with the party recently polling 0.1% in a state poll there is speculation about whether this membership still exists, together with speculation that she could join One Nation and seek state or federal office for said party. Nothing has been confirmed though there appears at least to be some level of social media co-ordination from time to time. If any of the above party memberships were incorrect as of 12 Sep 2026 then please let me know and the table will be corrected.
The table shows the usual clusters. In the top left is a cluster of eight councillors with agreement percentages with each other generally over 70%, in most cases over 75% and in one case (Kitsos-Reynolds) rising to 90%. In the bottom right is a cluster of four councillors with agreement percentages with each other all between 76% and 85%. The only councillor not belonging to either cluster for the term as a whole (and not close to belonging to one either, but see below) is Bloomfield. I find all the rest to very clearly belong strongly to one cluster or the other across the term, though there are still differences in the strength of this (eg Kelly is much closer to the centre than Zucco). I've classed them all as "strongly" to one side to emphasise the difference with previous councils.
On the right hand side the "score" column gives a ratio between how often the councillor agrees with the two clusters. (For the purposes of calculating this, agreement percentages with Burnet and Behrakis - except with each other - have been downweighted by 70% given they were off council for most of the term.) If the councillor agrees more with the green/left cluster it is coloured green while if they agree more with the blue/right cluster it is coloured blue ... noting here that "right" and "left" are relative to this council.
What is notable here is how large these ratios are, much higher than I am used to finding. It is not that the values at the ends of the chart are extreme; indeed they're not. Rather there are no cases of moderate members of either cluster who might agree with some members of the other cluster well over 40% of the time and hence get a ratio of something like 1.5; indeed there is no cross-cluster agreement score between any pair higher than 36%. This is a sign that this Hobart council is the most factionalised I have seen, beating the tail end of the previous council in that regard.
Is this just because the moderate or unusual councillors from both sides in the previous term lost their seats and the previous councillors are still voting the same way with respect to each other but have just been joined by new less moderate councillors? No. Looking at the agreement percentages involving six councillors still on council who served in the previous term (Harvey, Dutta, Reynolds, Sherlock, Coats and Zucco) the average agreement between the four on the green side is up from 74.3 to 79.8. Their average agreement with Coats and Zucco is down from 29 to 22.9. (That said Coats/Zucco is down from 85 to 81).
Looking at Bloomfield's voting patterns one notices the unusually high agreement percentage with Burnet and the low agreement percentage with Behrakis, when based on the rest of Bloomfield's voting one would expect the opposite. And there is a reason for this, which is that Bloomfield's voting behaviour from early 2024 on has been strikingly different from 2022-3. If I just take the period when Burnet and Behrakis were both still on Council, Bloomfield voted with the average "green/left" councillor 65.6% of the time and the average "blue/right" councillor 42.5% of the time, for a ratio of 1.54 to the green side, so I would classify her as green/left leaning for this period. However if I take the period after both Burnet and Behrakis had been replaced, Bloomfield voted with the average green/left councillor just 36% of the time and the average blue/right councillor 73% of the time, which would make her a clear member of the blue/right cluster in this time with a ratio of 2.21 the other way. (Just shy of Kelly across the term, but in this part of the term Kelly himself was more polarised and would score a ratio over 3).
On that basis looking at the Council as it finishes the term, there's an argument that everyone is now on a side most of the time and there aren't any serious crossbenchers, though some vote across the aisle much more than others. Defections on the left side on individual votes often prompt a lot of criticism especially if they cause that side to lose the vote - I mentioned Harvey on short-stay accommodation above, another was Lohberger in May 2024 on the soon-after modified and approved St Vincent de Paul Argyle St proposal. But these are the exceptions rather than the rules.
In two dimensions, for what it's worth
Something I've always done with these voting pattern charts for Council is a two-dimensional principal components analysis graph to see if it picks up anything different on trying to squash the voting patterns into two variables instead of one. Last time this was interesting, but what I generally find is that the first variable that the graph picks up is the left/right split and often the second variable doesn't explain anything much and is just driven by the voting pattern of a councillor who isn't part of a cluster. And that has happened again this year with the second axis being basically how Bloomfieldy a councillor was or wasn't given what else we know about them, and that in turn is determined mostly by when they were on council. The left-right axis picks up 92% of variation in voting patterns so the 2D plot is probably only useful for amusement but anyway here it is.
(The 6 councillor cluster on the left is, from top to bottom, Lohberger, Dutta, Kitsos, Sherlock, Reynolds, Posselt)
The road ahead
The Council is being culled from twelve to nine but with Reynolds and Kitsos retiring and Coats not visibly recontesting at this stage it looks like nine incumbents will seriously recontest for nine seats (six of them running for Lord Mayor alongside at least four off-council candidates and three running for Deputy alongside Behrakis and at least three other off-council candidates).
If everyone who appears to be running again retains then the new numbers would be a 5-4 or 5-3-1 (depending on how one classifies Bloomfield) majority for the left, but I think it is unlikely all will retain, and despite Hobart's leftness as a city I am not certain the green/left side will retain five after a pretty turbulent term. Aside from any chance of a shift in numbers back towards the blue side, there are significant challenges from within on both sides (on the blue side obviously Behrakis given his profile and past HCC results, but there are others.) On the Lord Mayor front if we take the 2022 votes as a starting point then Kelly seems the candidate to beat (though Harvey could poll a high primary vote as the Greens candidate in a Reynolds-free field); has anyone else picked up enough of a following in the 2022-6 term to do so?
I will as usual be covering the elections with a candidate and election guide in the near future and coverage of the counting.
Disclaimer
Any comment about this article in any form that reaches either me or anyone related to me will be considered on the public record, especially if stated otherwise. This is to prevent candidates from attempting to covertly influence my commentary.
