Tuesday, September 2, 2014

Hobart City Council Voting Patterns 2011-2014

Advance Summary

1. Traditionally, the Hobart City Council is loosely divided between "pro-development" aldermen and aldermen who stress environmental issues and/or the interests of impacted residents. 

2. The current term of Council was expected to be dominated by the "pro-development" grouping which I refer to as the "blues".

3. The first half of the current Council term supported this expectation, with seven aldermen displaying a blue voting pattern and an eighth displaying a tendency to support them.

4. An ordering of aldermen from "greenest" to "bluest" up til the end of 2012 was: Cocker, Burnet, Harvey, Ruzicka, Foley, Freeman, Thomas, Sexton, Hickey, Briscoe, Zucco, Christie.

5. In the 2013-4 half of this term of Council voting behaviour changed, with both the Green and the blue voting clusters becoming much less cohesive, so that it is not even accurate to classify some aldermen as still in the blue cluster.

6. An especially notable shift in this period was that both Jeff Briscoe and Ron Christie moved away from the "blues" and became much more Green-friendly than before, while John Freeman became more hardline.

7. An ordering of aldermen from "greenest" to "bluest" since the start of 2013 is: Cocker, Burnet, Harvey, Ruzicka, Foley, Christie, Briscoe, Sexton, Hickey, Thomas, Freeman, Zucco.

8. Some of these changes are explained by changes in the issues mix, but by no means all.  Positioning for the upcoming Mayoral contest may explain some of the others.

(This article is long and some bits are technical.  However the really scary stuff has been shuttled off to a PDF link buried in the dark recesses of Tasmanian Times.)


===========================================
Changes Coming

The Hobart City Council elections will be a major focus for this website over the next couple of months.  There may also be coverage of other council elections, but there is only one of me, so I wouldn't count on there being much of it.  As a result of various legislative changes recently, local government elections in Tasmania generally will be quite unlike those we have seen in recent years.  The most important changes include:

* All councillors are elected at once, instead of the old system in which half the aldermen would face the music every two years.  This means that the percentage quota for election is quite low, but also that fields will be very competitive because of the numbers of incumbents fighting for re-election.
* Mayors and deputy mayors are now elected for four years instead of two.
* There is no longer a requirement that a candidate for mayor or deputy mayor must have already served on a Tasmanian council.
* For election watchers, some of the councillor counts could be very different, though this concerns only the count process and not the results.  Some councils including Hobart could be counted using computerised data entry rather than manually.  This could  mean real-time preference distribution estimates when primary figures are updated, as in ACT elections.

That's where we're going in the next few months, but those voting in Hobart council may also want to know where we have been.  One of the biggest problems in local politics in Tasmania is that with a relatively limited level of party-endorsed competition, it's not that easy for voters to know what policy positions their councillors have stood for.  I've tried to make information available about Hobart City Council voting patterns to try to remedy this issue, over the years.  Past findings are available on Tasmanian Times:


The top one is the most useful as it includes recaps from the previous terms (it also has its own attached methods paper!)  The bottom two have been chewed a bit by the TT archiving process, so what was the top paragraph for each is either missing or at the end.

The Council Political Landscape

Hobart City Council aldermen are mostly independents, and even those who are endorsed by parties do not caucus on decisions and aren't bound to vote together.  There aren't any formalised factions, voting blocs or caucuses, just clusters of aldermen who tend to see things the same way a lot of the time.  So what patterns can be observed on the floor of the Council are patterns that only tend to crop up (with most aldermen sticking to them usually, but not all the time) on contested major issues.  For as long as I've been observing it (26 years now) aldermen have been mostly loosely split between two main such clusters of aldermen at any given time.  The generally larger cluster through this time has generally been what I call the "blue" cluster.

Councillors assigned to the "blue" cluster typically portray themselves as pro-development and pro-business, or at least vote that way.  Historically they are slightly more likely than not to have had some links to the Liberal Party at some stage of their career, but these links are not always lasting.  In the 1990s the main opposition to the "blue" cluster came from a string of residents-group type electoral tickets often also supported by grassroots environmental groups.  Councillors who got their start running on such tickets became extremely successful at leadership level, but never controlled the Council, and over time endorsed Tasmanian Greens have come to hold more of the non-"blue" seats.  As the Greens' positions have been more radical than those of their precursors, aldermen who used to reside on the "left" of council have latterly tended to vote with the Greens and the blues more or less evenly.  

The 2009 election saw the Greens win their first leadership position with Helen Burnet elected Deputy Lord Mayor.  However at the 2011 election the "blues" recaptured both leadership positions with Damon Thomas narrowly defeating Burnet for the mayoralty and Ron Christie becoming his deputy after shading Marti Zucco at the critical exclusion point by seven votes.  Another sign that the post-2011 council was going to be a very "pro-development" one came when Rob Valentine resigned to run for the Legislative Council and his seat was taken on countback by John Freeman, who Valentine had originally defeated to win the top job.  

Looking at the lineup of the Council in early 2012 with Freeman back on Council after losing his seat, it would seem that the rest of the term was set to be one of the most straightforwardly "pro-business" terms in the council's history, with the blue cluster holding the top two jobs and seven of the twelve seats.

Yet that is not the way it has turned out ...

Some Introductory Stuff

Before I get on to the agreement matrices and PCA graphs and so on, just a few words about how I do all this.  In measuring the voting patterns on council, I'm interested in who agrees with who on the biggest share of contested motions.  I use motions that had at least two aldermen on the voting for each side for the serious stats, because unanimous motions don't tell us anything (and there are lots of those), while some aldermen may be more inclined to measure lone dissents than others, for reasons more of political style and alignment.  

In this term (up to a self-imposed cutoff of the mid-August meeting) there were 289 distinct non-unanimous votes (I count each vote on the same matter as distinct if it happens at a different meeting to another or if it has a different voting lineup).  Of these 89 were lone dissents.  As a proportion of all votes at which they were there, the aldermen most likely to lone-dissent on any given non-unanimous vote were Burnet (10%) and Ruzicka (7.3%).  They were followed by Freeman (4.4%), Cocker (3.6%), Foley (2.6%), Valentine (2.4% of a small sample), Briscoe (1.9%), Christie (1.5%), Harvey (1.5%) and Zucco (1.2%).  Three aldermen - Lord Mayor Damon Thomas, Peter Sexton (again), and Sue Hickey, cast no lone dissents in this term.  Burnet, and Ruzicka were both more likely to lone-dissent in this term than the previous one (in which they were also the most frequent lone dissenters) and Cocker, Harvey and Thomas less so.  It seems that Lord Mayors are less likely to lone dissent than other aldermen while in that role.  

Some aldermen notably lone-dissent on particular matters.  For instance Foley's lone dissents were mostly on compulsory voting in local government, while Freeman's often included motions on social or broader political issues (gay rights, the death of Nelson Mandela, Section 18C etc).

Of the 289 contested votes, no alderman attended all of them.  Aldermen often miss votes because they are required to declare conflicts of interest and abstain, or because they are representing the Council overseas.  They may also be missing for health reasons or because of clashing commitments of whatever kind.   In this term Cocker attended 280 votes, Ruzicka 274, Foley 272, Harvey 268, Briscoe 267, Thomas 262, Christie 260, Hickey 259, Burnet 250, Zucco 242, Sexton 236, Freeman 225 (of a possible 246) and Valentine 42 (of a possible 43).  

With the 89 lone dissents out of the way, I move on to the 200 more seriously contested motions.  61 of these had two aldermen on the losing side, 65 had three, 37 four (including two 4-4 ties), 32 five (two 5-5 ties) and there were five distinct votes that resulted in a 6-6 tie.  The aldermen most often on the losing side were as usual Cocker 56% Burnet 55 Harvey 45 and Ruzicka 38.  The rest were Freeman 28 Foley 27 Christie 23 Zucco 22 Briscoe and Hickey 21 Sexton 14 and Thomas just 11.  For aldermen other than the Greens, this figure tended to go up as the term progressed (for reasons that will be apparent very soon.)

200 motions is a massive data set.  While I was processing it (which took me many hours) I stopped half-way through to see how the patterns were going.  When I started again I was surprised how different the voting was in the last 20 months.  So different, in fact, that I've decided to split the results into two.

The Council in 2011 and 2012: 

In doing the stats for this section, I decided on the following treatment for votes including Valentine.  As he was only there for a short time in this period, I decided to include data for votes he took part in (even if only he and one other alderman dissented) but to not include him in the PCA graph or agreement matrix.  The reason for this is that one alderman for whom there is much less data could affect the results too much for others.  

The following is the agreement matrix for the period from the 2011 election to the end of 2012 for the twelve current councillors.  Each cell shows the percentage of contested votes for which both councillors were present, and on which the two councillors agreed.  For instance, Ruzicka and Sexton agreed on 51% of such votes.

With apologies for vision-impaired readers, I've used colours to highlight some cases of agreement percentages of 70 and above.  Agreement percentages of 75+ are commonly used to define clusters in this sort of work, and I count 70-74 as close enough for borderline cases.  The councillors are sorted from top to bottom using methods as described in the 2009-11 methods PDF (I warn casual readers that that attachment is Extremely Boring, and rated at Wonk Factor 5) and a score for how blue or green they were is given, where 1 equals total neutrality and the higher the score the more blue or green.  (It's deliberately set so the bluest alderman and the greenest alderman are as polarised as each other.)  

The period from 2011-12 was the most polarised for several years, being 39% more polarised at the ends than the 2009-11 term.  

Here's the matrix:


Different ways of calculating the ordering of aldermen will give very slightly different results.  For instance, the ordering of Zucco and Christie is debatable.

The clustering is pretty obvious here.  The three Greens form their usual cluster and then most of the other pairs of aldermen agreed with each other 75+% of the time.  The main exceptions are Ruzicka (between 43% and 60% for everyone) and Foley (who had 78% agreement with Thomas and 73% with Freeman, and was generally more likely to agree with the blues than the greens, but doesn't reach the cluster threshhold for most of them.)  So seven blues, one blue-leaning, one intermediate, three greens. Piece of cake.  This is the 2D graph:

A PCA graph like this one shows which aldermen vote similarly to other aldermen by clustering those with similar voting patterns.  The direction of the arrow shows the kind of pattern the alderman displays (as compared to the others) and the length of the arrow shows how strongly that pattern is displayed.   So in this period of Council, Christie, Zucco and Briscoe were not only taking a more hardline position than Thomas, Hickey and Sexton, but they also tended to respond to a slightly different mix of issues.   The axes have no preset meaning but are found by the maths involved; the horizontal axis is likely to be close in meaning to "conservation vs development" while the meaning of the vertical axis is rather challenging to unravel.  

There's no big surprises in all this considering the previous term, in which the order of the continuing councillors was almost identical (Burnet, Cocker, Harvey, Ruzicka, Sexton, Thomas, Briscoe, Christie, Zucco).  Hickey was always likely to display a blue voting pattern given Liberal connections and business involvements, and while no-one knew quite where Foley would fit in it was expected his position might be something like that of the late Darlene Haigh (and it is.)

(For the record, the 24 votes for Valentine in this sample would have placed him in the blue group, with 88% agreement with Thomas, 86% Sexton, 83% Briscoe, 80% Zucco, 79% Hickey, 75% Christie, 73% Foley, 71% Ruzicka, 50% Harvey, 32% each Burnet and Cocker.  That's a bit surprising, but it is a small sample, and Valentine had been gradually moving more towards the blues and away from the Greens over his Council career.)

The Council in 2013 and 2014: 

After doing the above work for 2011-2 and getting the above results as a half-time score I was a bit surprised to see what happened in the second half of the data set.  As I was processing it I found that votes I would have previously considered weird turned up a lot of the time.  The Greens sometimes barely seemed to be voting like a party anymore, and certain hitherto "blue" aldermen were quite often joining them.  Aldermen rarely seen on the losing side of a motion were more often seen voted down. Just about everyone was becoming less predictable.  Here's the agreeement matrix:


In comparison to the 2011-2 matrix, the clustering at both ends has weakened greatly.  There's still a small cluster of Green results in the top left, but Harvey only votes with Cocker 68% of the time although they are both endorsed by the same party.  At the bottom of the table Sexton, Thomas and Zucco all cluster together, and Freeman can be added to that cluster based on very strong agreement with Zucco, a near miss with Thomas and a 68% with Sexton.  However Foley, Christie, Briscoe and Hickey each do not cluster strongly with any other aldermen.  They variously cluster weakly with some (but not all) of each other, and in some cases with some (but not all) of the blues. 

This phase of Council was much less polarised than the previous (and indeed it was slightly less polarised at the extremes than the 2009-11 term, previously the least polarised term I've examined.) But that is not all that is going on here, since the order of the seven councillors classed as blues for 2011-12 has changed.  Christie (the most hardline blue in the first half of the term) and Briscoe (the third most) are now far more agreeable with the Greens and are no longer really in the blue cluster.  On the other hand, Freeman has gone from being the most moderate of the seven in his first year and a half back to being one of the least moderate (and is the only alderman who the decrease in polarisation seems not to have affected.)  Some of the changes are massive.  Christie is on average 22 points more likely to agree with a given Green, and 24 points less likely to agree with Zucco.  Briscoe is 28 points less likely to agree with Zucco and 22 points more likely to agree with Burnet.  Sexton is 20 points more likely to agree with Cocker, and 19 points less likely to agree with Hickey.  And so on.

Here's the 2D graph:


Unhelpfully for anyone wanting to compare the two, the maths of the PCA calculation has produced a mirror image of sorts here, with Ruzicka in the opposite place to before.  That change is meaningless and Ruzicka's voting pattern is actually pretty much the same as in 2011-2.  

What we can see here though is that last time Briscoe and Christie were on the opposite side of the blues to Foley, and now they are rather close to him.  Meanwhile Freeman has gone from the Foley side to near Zucco.  All pretty much as would be expected from the agreement matrix.  We also see the three Greens behaving much less similarly to each other (again, as expected).  One thing the graph says is that Sexton is much more moderate than the other blues (something also readable from the matrix through his higher agreement with the Greens).  Another thing is that not only has Briscoe moved away from the other blues in terms of voting pattern, but his new direction is quite a strongly developed one.  

So what has happened here?

At this stage I am not entirely sure, but it is interesting!

Partially the changes I refer to can be attributed to the issues mix.  For instance from the Greens' perspective, an issue on which they are united (gambling) came up repeatedly during the first half of the term, while an issue on which Alderman Harvey generally votes differently to the other two (AFL football) was a subject of several motions in the second half.  There were also actions taken during this term to control the Greens' fondness for moving Notices of Motion about matters that most aldermen consider to be duplicated by other spheres of government, but I can't see too many motions in the first half of the term that would have fallen foul of the rules that were brought in.

However, the issues mix does not explain why there were such dramatic shifts in ordering within what used to be the blue group.  Looking at, for instance, the increasing tendency (from a very low base) of Christie and Briscoe to vote with the Greens, there are a few signal issues (such as the cable car and admission costs to the Taste of Tasmania) but overall it just seems like these aldermen have started voting with the Greens more often across a wide range of mostly minor issues.  

Councillors can probably explain some of these changes better than I can, by reference to personality conflicts within the Council, but another possible driver of some changes is Mayor election product differentiation.  No-one thinking of running for Lord Mayor against the incumbent or supporting an opponent -  wants to be seen as just a yes-sayer to anyone they are running against.  And quite a few aldermen have been cited as thinking of doing so (with Hickey a confirmed challenger from the blue side and Christie and Briscoe also considering, though earlier it seemed more likely Briscoe would support Hickey).  A chicken-and-egg argument can develop as to whether an opponent runs against the Mayor because they have moved away from his position, or moves away from his position because they've decided to run against him.

An opponent may feel they need a critique and an ability to say they have the voting record to show they would do things differently.  Tensions created by a brewing electoral contest can also drive prospective candidates to disagree with each other.  It can even be helpful for would-be challengers to vote in a way that creates random havoc so they can argue the incumbent isn't in charge and isn't controlling the Council. It seems little accident that the aldermen continuing to maintain a very high agreement percentage with the Lord Mayor on the floor are those who don't seem to have an eye on his job.  Without being sure that this is the cause of the strange inversion of the blue voting layer, I'd say the evidence is quite consistent with political positioning being a part of the puzzle. 

The greening of Christie is especially striking because in the 2005-7 and 2007-9 terms he was rated the bluest alderman, and in 2009-11 second to Zucco, before again rating as the bluest (arguably) for 2011-2. At one stage, cases of Christie agreeing with the Greens on anything were usually caused by him opposing something they also opposed, but for the opposite reason (eg they did not think it went far enough while he thought it was a bad idea at all.) In Briscoe's case, he has always had maverick moments even in his bluest and most Liberal days, and this is not really new territory for him (indeed he was first elected to Council in 1994 as an anti-cable-car leftie on one of the above-mentioned residents-group type tickets).  

Comments welcome ... but!

Comments and suggestions for things I should look for in the data are very welcome, provided they are within relevant electoral provisions, and also defamation law.  Comments may be emailed for inclusion if necessary.  I should also add my pet disclaimer when publishing an article about HCC that all responses to it in any form (including verbal) are on the public record, especially if stated otherwise.  So anyone thinking of sending or saying anything "confidential" in reply to this article, just forget it - whatever you say in this case may be reported.

I look forward to providing more coverage of the Council elections. Some time in the next couple of weeks I intend to kick off a general preview article and list of declared candidates.

No comments:

Post a Comment