Sunday, June 28, 2015

Singh Dumped To Fourth On Senate Ticket

Summary Of Where We Stand (As Of 30 August)

With more figures coming through from this ballot I thought I should put a summary of some of the claims that have been made and their status:

CLAIM 1: That a factional deal caused Lisa Singh to lose preselection for a winnable spot. 
STATUS: False.  (Whether a deal existed or could have caused Singh to lose is beside the point. Factions voting for their own preferred candidates, and Singh not having enough primary votes, explain the outcome without any such deal being taken into account).

CLAIM 2: That had only the rank and file voted, Singh would have been preselected third instead of fourth, meaning that Singh was effectively demoted by union delegate and state conference voters against the will of the rank and file.
STATUS: Apparently true.

CLAIM 3: The outcome was determined by the preferences of rank-and-file voters.
STATUS: Misleading, since it was only the presence of other votes in the ballot that caused rank-and-file preferences to be distributed before Lisa Singh could win the third place.  

Also, this article maintains that the choice of quota for the ballot was incorrect.

Today Labor in Tasmania announced the remainder of its federal preselections.  None of these came as any surprise, but the one attracting by far the most attention is the dumping of sitting Senator Lisa Singh to the stereotypically "unwinnable" fourth position on the party's ticket.

While there was grassroots party involvement in the preselection via a state vote of members, it did not shift a long-expected outcome: the other sitting Senators Anne Urquhart and Helen Polley retained the first two positions and the Manufacturing Workers Union state secretary John Short was placed in the third position occupied by Singh at the 2010 election.  While it was widely claimed Singh was the victim of a left-right factional deal in which in return for Short being supported for the third position, the Left agreed to encourage its members to support Polley, the real cause of Singh's demise seems to be the Left supporting Short ahead of Singh, and Singh basically just not having enough primary support to withstand a preference flow from Anne Urquhart to Short as a result of that.  (See the detailed comments by Adam Clarke below, although I am not currently as of 30 August convinced that the result would have been the same based only on the rank-and-file votes.  I have also made some comments about the voting system.)

The reported deal attracted many appalled and bewildered reactions on social media, especially after the outcome was announced, with some commenters drawing comparisons to the electorally disastrous elevation of Joe Bullock to the top of the WA Senate ticket.  It's been seen as a typical example of the faction-hack disease that has been blighting the ALP for several elections, in which candidate quality takes a back seat to internal candidate support.  Singh appears well regarded by the media and we can expect a fairly large amount of media interest in this especially following on from the negative publicity for the party's internal workings in The Killing Season.  A case like this provides plenty of ammunition for those who want to argue that Labor has learnt precisely nothing about the damage its internal power games do to its wider standing.

Lisa Singh's Record

Senator Singh is a former one-term Tasmanian state MHA and state Minister.  Singh polled 5760 votes in Denison at the 2006 state election, more than (for instance) Julie Collins who is now federal MHR for Franklin.  However, in 2010 Singh's vote declined as a result of the general decline in the Labor vote at that election and competition from newcomer Scott Bacon.  With Bacon's election, both Singh and fellow incumbent Graeme Sturges lost their seats as Labor dropped from three to two seats in Denison.

Singh was preselected for the Senate in third position on the ballot.  The 2010 election was the last at which Tasmania had a very high percentage of below-the-line Senate vote and Singh polled 9,132 below-the-line votes (6.7% of the ALP's total).  This was, in fact, the highest vote total for a number three Senate candidate in Tasmania since above-the-line voting began.  It was not, however, the highest percentage of a party's vote obtained by a number three candidate, as the Liberals' Guy Barnett polled 7.5% of all Liberal votes at the same election.  It did Barnett no good as he was defeated by Singh with the state returning a 4-2 left-right split (3 Labor 2 Liberal 1 Green) for the second election in a row.

Singh's appeal is especially to left-wing, young, environmentally-inclined Labor voters, and likely includes some voters who do not normally vote ALP.  At her exclusion in the 2010 state election cutup, an enormous 27.5% of Singh's vote leaked rather than flowing to Scott Bacon, with the Greens' Helen Burnet alone getting nearly a 10% leak.

During the 2006-10 state parliament, Singh notably did not support Labor's Pulp Mill Assessment Act (an attempted fast-tracking of the now moribund Bell Bay pulp mill proposal by the now insolvent timber company Gunns), unusually claiming it to be a "moral issue".  After Singh unsuccessfully requested a conscience vote on the matter, she was permitted to abstain from the vote.  While contributing to the cult appeal she currently enjoys among Tasmanian politics-watchers, this has probably not done wonders for Singh's reputation as a team player or her support among industry groups influential on pro-forestry swinging voters.

It is not easy to measure the profile of state Senators but Google News returns about almost as many hits related to Singh as fellow Tasmanian ALP Senators Anne Urquhart and Helen Polley combined.  Singh is mentioned over three times more frequently than Senator Carol Brown and about five times more often than extremely obscure Senator Catryna Bilyk, whose very existence may well be news to some of my readers.

Jobs for the boys?

One aspect of the debate I don't take too seriously is the suggestion that Singh has been pushed aside for Short because of her gender.  To the extent that gender might have been a factor, Tasmania currently has five female Labor Senators and no male, and previously had a 6-0 female-male split.  The idea that Short is being picked as a token male is hardly the whole explanation when Singh is not the only female Senator.  Of more interest in my view is the loss of ethnic diversity (Singh is Australia's first MP of Indo-Fijian ancestry), and the fact that it was Singh and not Helen Polley who was demoted.

Polley Trail Goes Silent

I freely and even proudly disclaim that I'm not usually a fan of Senator Polley's politics.  Socially reactionary union-linked types are not a group I have much politically in common with, and I think it's a shame they continue to wield so much political influence within what is supposed to be a vaguely modern party.  This problem was most acute when Julia Gillard's political debts to Shoppies boss Joe de Bruyn stymied what could have been a mandated Labor vote on same-sex marriage in the previous parliament, as a result of which Australia has avoided any chance to get there before even such notoriously religious polities as Ireland and the USA.

My own biases aside, what is interesting in the context of Singh's dumping is that at a certain time in the current Senate term, Senator Polley appeared "embattled" by more than just her antiquated views, and seemed to be on the way out.  In mid-2014 Polley came under scrutiny after spending tens of thousands of dollars in expenses on charter flights between Hobart and Launceston.   This came about six months after even more serious news that Polley's office had been investigated over bullying and harassment claims with a number of negative findings (as detailed in that link) and with three staff members taking successful action.  Senator Polley also came under scrutiny over the grand Tasmanian political tradition of employing family members in political offices.

At the time of these controversies I saw a few comments along the lines of: not to worry, she probably won't get preselected again.  Surprisingly these concerns about Polley's suitability to continue what is already a two-term career just vanished and didn't even resurface once the anti-Singh deal was struck.  I have not seen any mention of any of them in recent media coverage of the preselections, most likely because Polley's retention of the #2 position was so much a done deal that there was no point in opponents even trying to bring these matters up again.

What Is A Good Senator?

Defenders of Polley might say that despite these rough edges and reservations about some of her views, we should look at the work she does as a Senator, both for her constituents and in combating the government.  And likewise, they might say that Singh's media profile and appeal to a certain voter type is not necessarily a sign that Singh is an effective Senator for the party.

Unfortunately it is almost impossible to say who is a successful Senator by any objective measure.  At least with MHRs it is possible to look at their performance as a candidate from election to election compared to the fate of their party nationally and locally, and from this to get some idea of which candidates have strong personal votes.  But for the Senate there is no such yardstick and most voters just vote for a party without seeming to care much who its state Senators are.   This means there is really little obstacle to internal deals determining Senate positions.

What Will The Impact Be?

The impact of demoting Senator Singh on Labor's chances depends not only on the reaction but also on what electoral system is in place for the next Senate election, and we don't as yet know what that will be.

If it is the current system then there will probably not be a great impact.  While Lisa Singh polled a large personal vote at the 2010 election, the 2013 election saw a great reduction in the rate of below the line voting in Tasmania, on account of the record number of parties running for the election.  With this continuing, several of those who cast a personal vote for Singh last time would still just vote 1 Labor this time.  Even on the very plausible assumption that some portion of them defect to the Greens or left-micros in disgust, I don't think there's much likelihood of a net impact above a 1% fall in the ALP vote.

There is also no point in the current system in placing too much hope in a below-the-line rebellion to save Singh's position.  Labor voters are especially obedient of how-to-vote cards and especially averse to voting BTL.  In the 2014 WA by-election, despite the enormous controversy over Labor's placement of Joe Bullock above Louise Pratt, Labor's BTL voting rate was still lower than the average across all parties, and barely a third of the BTL rate for the Greens (who had no such candidate issues).  Pratt herself received only 2% of the state Labor vote.

It may be that several left-wing parties including the Greens and various micros all decide to preference Singh above the other Labor candidates, but if so this won't make much difference.  The Greens are unlikely to have a large (if any) surplus, and if they do then Singh will likely be excluded before their second candidate.  We don't like to say things are impossible round here, but the chance of Singh winning from fourth in the current system seems vanishingly low.

Under the proposed JSCEM model for Senate reform, things could be a deal more exciting in Tasmania.  Because of the state's strong familiarity with multi-candidate voting, it's possible that optional Senate preferencing would lead to a high BTL voting rate in Tasmania specifically.  This would create the opportunity for spurned major party candidates to try to save themselves through BTL campaigns, but even so it wouldn't be easy.  Voters who currently vote below the line in Tasmania are not only doing so because they prefer to order candidates within parties differently.  Many are doing so because they prefer to preference parties in a different order to their own party's lodged group tickets.  This type of voter would instead use the above-the-line preferencing option.

If there is lasting anger about what has happened in this case then I think the most likely impact is a small loss of ALP votes to a range of left alternatives, including the Greens.  I don't think we're likely to quite see a repeat of the WA Senate disaster as a result; not only did Joe Bullock have a lot more "form" than Senator Polley, but also that was a by-election.  There is also a long time between the announcement of this decision and the next election.

As for Labor's chances of winning a third Tasmanian seat at all, under either system that will not be easy.  They did this in 2007 (when Labor did very well nationally) and 2010 (when the Liberals ran a poor Tasmanian campaign) but those are the only times a 4-2 left-right split has happened in the state. It's not something to be confident about this far out but there seem to be good prospects for some level of rise in the Green vote, making the Greens very difficult to beat.  Three seats for Labor isn't impossible, but it is difficult enough that Singh would have been facing a difficult battle even had she just not been promoted.

I may have some more comments on this issue after seeing media reaction and whether there are any serious moves to have the decision overturned.

PS: A couple of posters on Twitter have drawn my attention to Polley being the sole survivor of what was otherwise a left-wash of the Tasmanian ALP preselections, with all the House of Reps candidates being from the left of the party.

NB: Some parts of this article have been revised following the comment from Adam Clarke below.

Another Update: James Brady at The Examiner has provided some figures! Brady gives rank-and-file votes of Urquhart 221 Polley 123 Singh 110 Short 74 (the votes for the remaining candidates are not stated) and partial figures for other vote sources.

In comments below I point out a deficiency of Labor's implementation of Hare-Clark, which is that they treat the ballot as an election for four positions, when in fact the fourth position is the "unelectable" spot and should not be treated as a win.  On this basis had what I think is the correct quota been used (25% for three winnable positions, rather than 20% for four) Polley would certainly have fallen short of quota on the first ballot of the rank and file members.  Depending on the split of Urquhart's votes between Short and Polley it is possible Short would have then overtaken Polley and been elected second, if the election had used the correct quota and only the rank and file votes.  So at this stage it is unclear to me that Polley's retention in second place is really the will of the rank and file.

(Note added August 9: see also Wonk Central: How Should Parties Conduct Member Ballots For Senate Tickets, for a full technical discussion of voting systems for this kind of election.)

Update Aug 26: Some new figures have emerged.  In today's print Mercury, Nick Clark notes the total rank-and-file vote as 542, meaning that other candidates polled just 14 rank and file primaries.  After the addition of 211 state conference delegate votes, totals were Urquhart 288 (+67), Polley 242 (+119), Short 158 (+84), Singh 153 (+43).  The changes exceed the 211, which might be a result of weighting or might mean there were other votes added that haven't been noted.

There is then a comment "He [John Dowling] said Senator Urquhart's excess rank and file votes then distributed with a transfer value under the Hare-Clark system giving a final tally for Short of 426 to Singh's 177".  This doesn't make sense by itself (at least not in Hare-Clark as we know it) since Urquhart would have distributed four more votes than she had, so more detail (probably concerning weightings) is needed to make sense of what actually happened.

Really, there's enough of this stuff in the public domain already, why not just release the whole scrutiny sheet?

Update Aug 29: A further dribble of information (mixed with confusion and error) about the count process with Mark Kenny reporting that 200 union and state delegate votes carried a weighting to make them equal in value to the 542 rank-and-file votes, meaning that the total should be 1084. Singh's tally is given at 154 not 153 meaning that 842 votes are accounted for; the rest would presumably be for the minor candidates or informal.

Kenny's report then states that Short being four votes ahead of Singh meant that Short won the final position, but this is clearly false.  The strong flow from Urquhart's surplus (including her rank-and-file votes) to Short was the decisive factor rather than Short's four-vote lead but we have still not seen exactly how this happened.

I should note that based on the (incorrect) quota used for the ballot, had the election been purely of the rank-and-file votes then Urquhart, Polley and Singh would have all made quota on primaries (Singh by a whisker) and been elected 1,2,3.

I have also received a claim about the use of how-to-vote tickets in the ballot, which I hope to look into soon.  Two other aspects of the issue that remain underexplored are the extent to which rank-and-file membership might overlap with union and state delegate voting, and the extent to which the rank-and-file might itself be factionalised.


  1. I was at the ballot count on Friday night. It is worth noting that Anne Urquhart and Helen Polley were both immediately declared elected 1 and 2 after attaining quota solely on their primary votes.

    A distribution of Urquhart's surplus then allowed John Short to attain a quota and he was declared elected 3.

    Lisa Singh had still not attained a quota, however as it was mathematically impossible for the two other candidates to surpass her she was declared elected 4.

    There are a few important points to be noted from this.

    Firstly the suggestion that Helen Polley was re-elected as part of a Left-Right deal is factually incorrect. Helen received enough number 1 votes in her own right to attain a quota. Necessarily this meant that no votes distributed from Anne Urquhart flowed to Senator Polley.

    The only distribution to take place in the count was that of Urquhart's surplus. There were no other exclusions or distributions.

    The result would have been exactly the same had only the rank-and-file votes been counted, rather than the current system of an electoral college that grants rank-and-file members 50% of the vote and conference delegates the remaining 50%. (It is also worth noting that even under the current electoral college system the Tasmanian branch preselection process is by far the most open to rank-and-file members of any Labor state branch. The effective union component of the vote is less than 25%).

    Given that the vote was conducted entirely by secret postal ballot, using the Hare-Clark proportional system and the fact the outcome would have been exactly the same based solely on the rank-and-file vote (with a one member one vote weighting) it is unclear just what further reform could be possible that would changed the result. It's what the members voted for.

    No factional deal could have unelected Senator Polley. She received a quota from the membership.

    Of course people can be unhappy at the outcome. They can say the membership got it wrong. But to argue for a reversal of the result would be to ignore the votes of the membership in a robust democratic process. It would be to argue for exactly the kind of centralised process that enables the factional deals people are arguing against to occur.

    It is a centralised, closed shop process that produced the WA senate ticket with Joe Bullock at the top. That was a factional deal. Had the membership been given a vote in WA I am confident Louise Pratt would have won the first spot.

    Ultimately the outcome of the Tasmanian Senate ballot was determined by rank-and-file party members.

    1. Thanks very much for this detailed comment; Adam. As is my usual habit I have edited the original text to get rid of those comments that were inaccurate.

      I do wonder though whether the use of a PR system to preselect in this way - if I've understood the description correctly - is correct. Four candidates are being selected to run on the ticket meaning that the quota is just over 20%, so if (say) 21% support Polley then she is guaranteed to get quota. But which position she would get as a result of getting quota is very heavily dependent on the distribution of primary votes between the remaining candidates.

      In a completely imaginary example in which Polley had 21% primary support and everybody else put her last, it would be possible for Polley to finish anywhere from first or second (=automatic win) to fourth (=unelectable) depending on the breakup of primary votes between the other five candidates. The system, if I've got it right (or even if I've not quite got it right) seems exceptionally sensitive to whether or not a majority vote tactically. In most proportional-voting elections, order of election carries relatively little importance, with the most important aspect being that someone who has sufficient support is elected at all. In this case, however, order of "election" is everything.

      I also think that "electing" four places when the fourth candidate receives an unwinnable position is less than ideal. The top three places are by far the most critical, perhaps even the top two. Fourth place could just be filled by the first "unelected" candidate.

      Very likely these issues make no difference and Polley would have been elected second no matter what. But this just doesn't seem an ideal usage of Hare-Clark to me, though finding a better solution for a rank-and-file ballot as opposed to a central process, while stopping the dominant faction filling all places, may be extremely difficult.

    2. Hi Kevin - your understanding of the process is essentially correct. It's also the same process we use locally to elect the party President and three Vice Presidents and nationally to elect the President and two Vice Presidents. Either the highest vote getter or the first to attain quota is elected President.

      As you note it is difficult to consider an alternative process that would not simply result in candidates from the same group winning all positions. It would certainly be interesting to explore alternative electoral method suggestions however.

      Actually in addition to the above we also have Affirmative Action rules that apply to all of our proportional ballots. At least 50% (rounded down) of those elected must be women, should sufficient nominations be received. This also allows for some strategic voting. The election proceeds normally but if AA is not satisfied then enough successful male candidates are unelected in reverse order to then elect female candidates to those positions.

  2. Thanks for the extra detail. I'd certainly increase the quota so that the election is for the realistically winnable positions (who fills out the ticket below third generally being no big deal to anyone). Beyond that it is very difficult, perhaps impossible, to find a method that is proportional but that does not make the order of election depend on whether or not certain voters use strategic voting.

  3. The point that is being missed here is that having pre-selections decided by ballots of paid-up members is no better than having them decided by conference delegates or elected committees or a group of union secretaries meeting at the pub. In fact is some ways it's worse. Most of the ALP's paid-up membership is either 60 year old retired teachers or ethnic stacks. The only real solution to the ALP's problems is to empower Labor VOTERS, not Labor members, through open primary elections.

  4. Labor did attempt to set up closed primaries for a few of its more minor preselections by starting a system where you could register as a supporter for free and vote without being a member. There was a dismal lack of public interest in the concept.

    Open primaries are an interesting idea (and one I thought about too while writing this article) but I can see one potential problem. In the US if a candidate wins an open primary by doing well off the other party's support base then that probably means they will compete well with whoever the other party puts up. But in Australia (and Tasmania especially) I suspect an open Labor primary would be stacked by Greens supporters who would drive the party's choices way to the left, without any real benefit to Labor since they get 80-odd % of Green preferences no matter who they put up.

  5. Adam, re "The result would have been exactly the same had only the rank-and-file votes been counted,", is that actually correct? The widely reported rank-and-file figures have Singh with 110 out of 542, meaning that if only the rank-and-file votes were counted then Urquhart, Polley and Singh (just) would all have made quota on primaries and been elected 1,2,3. Urquhart's surplus flowing massively to Short would have made no difference as Singh would have already been elected third.