Thursday, March 31, 2016

Legislative Council Voting Patterns 2012-6

Advance Summary

1. This article presents a revised analysis of voting patterns in the Legislative Council (the upper house of Tasmanian Parliament) based on contested divisions in the last four years.

2. Although there is a degree of independence in all Legislative Council voting, the Council continues to have a clearly defined "left wing" consisting of Craig Farrell (Labor), and independents Mike Gaffney, Ruth Forrest, Kerry Finch and Rob Valentine.

3. Excepting Adriana Taylor and Rosemary Armitage, the remaining MLCs have leaned significantly to the right to varying degrees over the past four years.  However, Tania Rattray displayed a centrist voting pattern in the last year specifically.

4. Voting in the Legislative Council in 2015-6 was much less polarised than in previous years, probably because of a low proportion of high-profile partisan political issues.

5. The revised analysis again shows most of the conservative MLCs to be between Liberal and Labor positions, albeit closer to the Liberals.

6. The Liberal Government lost very few divisions in the 2015-6 Council, however this will change should Labor win the two seats being elected this year.


We're just over a month away from elections for two of the fifteen seats in the Tasmanian Legislative Council.  Independent incumbents Tania Rattray (Apsley) and Adriana Taylor (Elwick) are being challenged by Labor in both seats, and there may be other challengers too.  Attacks on candidates' voting records - although apparently ineffective - are an increasingly common aspect of Legislative Council campaigns and I hope this piece assists in explaining where the Upper House really lines up.

I'll be putting out a preview article for these two seats very soon, but for time reasons I've decided to go with what seems to have become an annual update of Legislative Council voting patterns first.  For previous articles on this, see last year's edition, which links to an earlier piece.

Every time I do this review, I include just the last four years of data.  I'm interested only in those votes where there is actually division on the floor of the Council, and not the many unanimous votes. Since the last review there have been 29 divisions with at least two votes on each side, counting identical lineups on the same issue at the same session as a single division.  Issues covered in these votes have included gun control, discrimination in religious school enrolments, politicians' pay, a code of conduct for local government, industrial hemp, land use planning and even Lonely Planet travel advice.  Generally, however, there have not been many high-stakes political-culture-war type issues in the last 12 months, and this has provided more chances for Legislative Councillors to behave independently.

I've again aimed to produce a couple of different descriptions of the observed voting patterns.  One of these is a two-dimensional graph and the other is a left-right sort.  Again this mainly follows the methods spelt out in the ultrawonky PDF attachment to an old Tasmanian Times Hobart Council article.  Again I had to assume an agreement value for Jim Wilkinson with each of Leonie Hiscutt and Robert Armstrong (based on averages of MLCs who tended to most agree with the latter two).  Actually I thought about leaving Wilkinson out since, as President, he hasn't voted except to break ties in three years, but I decided to retain him in the analysis for now.

There is, again, a minor judgement call in the left-to-right ordering, discussed later.

In Two Dimensions

The following is a two-dimensional view of the voting patterns of the fifteen current MLCs over the last four years.   For those unfamiliar with graphs of these sorts, a principal components analysis aims to represent patterns in 2D with as little distortion as possible.  Both the angle of different lines to each other and the distance of different data points from the centre are relevant here.  The angles between different candidates indicate whether or not they display different kinds of voting patterns and the distance indicates how strongly each pattern is realised.  Even if two Legislative Councillors appear opposite each other, if one is close to the centre they will still agree often.  If two Legislative Councillors are at a similar angle and a similar distance from the centre then it is likely their political views are rather similar. The two axes chosen by the analysis do not necessarily mean anything in particular and are not predetermined by me, but it's obvious in this case that the x-axis corresponds pretty closely to "left-right".

(Arm = Rosemary Armitage, Ast = Robert Armstrong, Dean = Ivan Dean, Farr = Craig Farrell (ALP), Fin = Kerry Finch, Forr = Ruth Forrest, Gaff = Mike Gaffney, Good = Vanessa Goodwin (LIB), Hal = Greg Hall, Hisc = Leonie Hiscutt (LIB), Mul = Tony Mulder (Ind Lib), Ratt = Tania Rattray, Tay = Adriana Taylor, Val = Rob Valentine, Wilk = Jim Wilkinson).

The software I was using to do these graphs previously is so unreliable that I'm no longer using it.  I now do them using a free statistics package called Tanagra combined with an Excel biplot macro.

The following changes are especially noticeable compared to last year:

* The patterns of Liberals Goodwin and Hiscutt are now much more similar.  They have always voted together nearly all of the time (bar conscience votes) but over time most votes from when Goodwin was in the chamber but Hiscutt was not have fallen out of the sample.

* Both Dean and Mulder are closer to the two Liberals than in last year's 4-year sample.  However in Mulder's case this isn't because he is especially likely to vote with them, but rather because he votes with them much more than he votes with most others.  Hence the shorter arrow.

Overall the graph shows an obvious cluster on the left (the same cluster of Valentine, Finch, Forrest, Farrell and Gaffney that has appeared every time I have done this) while on the right there is a cluster involving the two Liberals and also Armstrong and Dean, with Hall and Mulder also close to this cluster.  Rattray leans more to the conservative cluster than the left cluster (but see more on this below) while Armitage and Taylor are more or less in the middle.

Left-right sort

The agreement matrix below shows some similar patterns to the PCA graph.  The matrix shows the percentage of contested divisions on which each pair of MLCs voted together.  For instance it shows that Finch and Mulder voted together 36% of the time.

The highest agreement percentages were 97% for the two Liberals, 90% for Forrest with Gaffney and 89% for Hall with Armstrong.  The lowest all involved Valentine, with Dean (20%), Hiscutt (22%) and Goodwin and Armstrong (25%).  These figures for Valentine would have been even lower had I included a lone dissent, but they are still not as low as some in previous years.

As usual I've highlighted agreement percentages of 75+% (a common cutoff for identifying clusters) and weakly highlighted those between 70 and 74.

This time I give two different alignment scores.  Score1 reflects how strongly the MLC tends to vote with the left of the Council (red) or the right of the Council (blue), rather than the other way around.  Taylor and Armitage are considered neither right nor left.  Score2 is based on ratios between the MLCs based on whether a pair of MLCs are more likely to vote with those to the right of them than those to the left (again this is explained in the ultrawonky HCC methods piece.)

It's immediately obvious that we still have the same left-wing cluster of five MLCs who all strongly tend to side with each other.  And, as usual the pattern gets a bit messier in the other corner, but we still have a clear cluster of Dean, Goodwin, Armstrong and Hiscutt.  Hall is also close to this cluster.  Mulder is much more likely to agree with this group than with the left group, but also less likely than most other MLCs to agree with anyone, so he still defies classification to some degree.  The remaining four centre to centre-right MLCs don't really belong to any cluster.

Score1 is directly comparable to last year's alignment scores, showing that the voting behaviour of several MLCs (especially Valentine, Rattray, Armstrong and Hiscutt) seems less polarised with the addition of this year's issues mix, which contained fewer big right-left issues than in previous years. The classic 9-5 voting pattern seen on many issues in the past few years was seen only once in the last year. Score1 and Score2 produce identical orderings except that Score1 places Goodwin to the left of Dean and Armstrong while Score2 places her to the right of them.  I have therefore placed her in the middle.

The only change in the ordering is that Farrell and Forrest have swapped places.  The only change in the alignment descriptions is that I have moved Tania Rattray from Right to Centre-Right.  This is based on quite a striking change - in the last 12 months Rattray has voted with the left-wing MLCs almost exactly as much as the right-wing ones.  The MLC she has voted with most often in that time by some margin has been Kerry Finch (72% agreement.)

The stakes in this year's election for two seats are actually quite high.  In the last year the government has had a good run, winning all bar four divisions.  Should Labor somehow manage to win both seats, the left of Council would swell from five to seven seats, giving it a blocking majority on those major issues on which it is united, assuming none of its members has to take the chair.  This would be a serious challenge for the Hodgman Government.  As my preview piece will argue, this isn't even a remotely likely outcome, but all the same, the possibility shows that Legislative Council elections are important.

No comments:

Post a Comment