Legislative Council Voting Patterns 2018-22

 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 involving the current (and one recently retired) MLC in the last four years.

2. Although there is a degree of independence in all Legislative Council voting (outside of party blocks), the Council continues to have a fairly clearly defined "left" side consisting of independents Rob Valentine, Meg Webb and Mike Gaffney

3. However, Labor (which currently has three MLCs pending a by-election) and independent Ruth Forrest no longer tend strongly to vote with the left grouping and in the last four years are on average best considered as only slightly left of centre.

4. During his brief period as a no-longer-caucusing Labor MLC, Bastian Seidel voted similarly to the left grouping.

5. The four Liberal MLCs and independents Rosemary Armitage and Tania Rattray form the "right" side of the cluster, though Armitage and Rattray only vote with the Liberals moderately often.  Data so far is consistent with Dean Harriss being part of this group but with nowhere near enough evidence to be confident of that yet.

6. A possible left-to-right sort of the council is Valentine, Webb, Gaffney, (Seidel pre-retirement), Forrest, the three Labor MLCs in no particular order, Armitage, Rattray, and the four Liberal MLCs in no particular order.  Harriss is not placed in this list yet pending further data.

7. The increased tendency of Labor to vote with the Liberal government in recent years has meant that the Government has suffered relatively few (and generally minor) defeats on the floor of the Council in the last year and a half.

------------------------------------------------------------

Last week saw what seemed to be an unusual Legislative Council voting pattern.  The Government was having trouble getting its third attempt at laws against obstructive protests over the line and accepted a bunch of amendments designed to narrow the Bill's scope and reduce some of its penalties.  Finally on the third reading the Bill passed 6-5 (plus 1-1 paired) with independents Ruth Forrest, Dean Harriss and Tania Rattray voting with the Government while Rob Valentine, Rosemary Armitage, Meg Webb and Mike Gaffney voted with Labor.  Given Armitage's no vote, and on the (perhaps no sure thing) assumption that Labor would have voted the same way if it actually mattered, the Bill's passage depended on all of these three things being the case: (i) Forrest voting yes, (ii) Labor being down one because of the resignation of Jo Siejka (iii) Labor's Craig Farrell not voting because he has the casting vote, and the casting vote can only be caused to break a tie and not to make one.  (That's just wrong in a house of fifteen members, by the way - all MLCs should have a deliberative vote only - but I digress.)

With the Legislative Council about to have another Pembroke by-election, and with a desire to do a practice run prior to my much larger analysis of Hobart Council voting patterns, I decided this was a good time to revisit my analysis of Legislative Council voting patterns, which I last looked at in early 2021.  Since the 2021 article, the Liberals have gained Windermere on the retirement of Ivan Dean, while Labor has lost Huon after Bastian Seidel resigned and independent Dean Harriss beat Labor in the by-election.  Now Labor has had an incumbent resign, so for the moment there are four Liberals, three Labor and seven independents.  Of the seven, four have in the past been assessed as on the left, with two usually landing on the right and one expected to do so, so the chamber appears finely balanced.

Methods

Mostly this analysis follows the same methods as previous years but I have made a few changes because of the presence of some very long debates that threatened to swamp the dataset,  Firstly for each new piece of legislation I now include a maximum of ten votes, giving priority to the second and third reading votes (if any) and then filling up the remainder of the quota of ten from the beginning, including only those votes that do not have an identical pattern to any earlier vote on the same bill.    (The best way to handle this would be to include every division and downweight amendments so that no one issue swamped the results, but that's more work than I easily have time for at the moment, and wouldn't lead to greatly differing results.)  Also I am no longer treating new identical votes on the same issue on different days as distinct.

As usual I'm including only the last four years of voting on a rolling basis.  I'm continuing to treat Labor MLCs as one voting entity and Liberals as another, with the exception of the period in which Bastian Seidel left the Labor caucus and started voting independently.  Although Seidel has now left the parliament I thought it would be interesting to assess how he had voted in that period.  There were no more conscience votes with splits within a party in the last year and a bit (though there is one vote where Jo Palmer is erroneously recorded as voting on both sides) so I preserved the same treatment of those still in the dataset.

The current dataset includes 98 divisions on which there was a contested vote with two or more MLCs on each side, of which 36 were not in last year's dataset.  The new votes include the maximum sample of ten apiece for poker machine and anti-protesting laws, two sources of marathon debates with many amendments.  The remaining divisions were, as usual, an eclectic mix of mostly minor issues.

I've included both Bastian Seidel (non-caucus period) and Dean Harriss in the assessment, but the data for them is limited to 16 and 12 divisions respectively, so I've only given their voting patterns a 50% weighting in assessing the alignment of other MLCs.  Harriss's 12 divisions included only three separate matters so I think there is not enough data to form any reliable judgement about his voting behaviour yet, but I show his agreement percentages to this point anyway.

Agreement matrix and left-right sort

This chart shows how often the Legislative Councillors agree with each other on contested votes.  For instance, the chart shows that Rattray and Forrest currently agree 53% of the time.  As usual I've highlighted agreement scores over three-quarters in certain colours, and scores close to that mark in paler colours ... and here's the big change - I have had to change the colours!  The left cluster is now shown in green (not red) while the right cluster is still shown in blue.  

I've done this because while the Labor Party used to belong at least loosely to the left cluster, after this year's voting it no longer does so.  In the last four years Labor's rate of agreement with the Liberal Party in contested divisions (46%) is not much different to its rate of agreement with the lefties Valentine, Webb, Gaffney and (post-caucus) Seidel.  The nearest thing to clustering (and it's not that close) involving Labor is a 66% agreement score with Ruth Forrest, but Forrest is also clearly no longer a member of the left cluster and over the last four years now comes out as just a little left of centre. These assessments can move around depending on the issues mix (eg Rattray and Armitage have both moved around between the right and centre over the years), and Forrest has often been different to the other left-independents on law and order issues, but it's not just that - eg the recent anti-protesting bill is only 10 votes out of 98.

I think the changes this year have shown that Valentine, Gaffney and Webb are really more like teals or Greens than they are like Tasmanian Labor.  Labor's four-year rolling agreement share with the Liberals has done this:

...and in the 36 divisions I added, the major parties voted together 20 times, including all ten divisions that I included on poker machines.  Some of these saw the remarkable spectacle of both major parties on one side and every single independent (left or right) plus Seidel on the other.  When I first noticed a resurgence in "Laborial" voting it appeared to be largely pandemic-driven but there seems to be more to it than that.  Pokies is, of course, an especially "Laborial" issue, but others have appeared in the mix in the past (eg forestry) without this effect becoming so apparent.

The upshot of Labor's move to the centre is that the government has a few different paths to winning votes - it can get support from Labor, it can get support from Forrest or one of the clearly left independents, or there can be votes where Labor opposes it but nobody else does.  The government did lose six of the 36 divisions in this year's sample but none of those was a big deal.  (Labor was on the losing side in ten.)  On the other hand, the government folded on a number of potential divisions that it would have lost in the anti-protest legislation debate, in order to expedite passing what of the Bill it could.

Here is this year's chart:

Webb and Valentine are the government's most frequent opponents, voting with the Liberals on just 12% and 13% of contested votes respectively in the last two years.  There is very little difference between their voting patterns - the ratio method I use has again given Valentine the nod for the apparently somewhat coveted title of leftest of them all, though that is purely on the basis of a higher agreement percentage with Gaffney, so arguments could be made there either way.  

The post-caucus Seidel voted rather similarly to Gaffney (clearly left on the issues voted on, but not as much as Valentine and Webb) and some left-wing observers were impressed with his performance and wished he would stay there and keep voting like that.  As mentioned above, Forrest and Labor are just very slightly left of centre in this assessment (Forrest a little bit more so than Labor.)

Aside from Rattray's high agreement score with Harriss off an insufficient sample of votes, Rattray and Armitage vote most often with each other.  They also vote reasonably often with the Liberals these days and well less than half of the time with Valentine, Webb and Gaffney, so they are clearly on the right (though not that strongly so), despite Armitage's third reading vote on anti-protest laws.  Harriss is so far also voting on the right but as mentioned above the sample is woefully insufficient so more data is needed to classify him.  He voted differently to the government on a couple of anti-protest amendments in my sample already.  And while MLCs to the right of the Liberals have at times existed in the past, at the moment there clearly aren't any.  

I am used to talking about the numbers of "left" vs "right" MLCs but this year's results suggest it is better to talk about "non-right" vs "right" or "left and centre" vs "right".  Anyway if Harriss is, as expected, at least somewhat right of centre, then that means the right will regain a majority if it wins Pembroke (in practice through the Liberals' Gregory Brown, since the Shooters do not have a chance.)  That wouldn't be a massive difference though - in the last year and a half there wasn't even one single vote in my sample that had all the "right" MLCs together on one side and everybody else on the other.  Once upon a time, such votes were pretty common on major legislation, but things have changed!