Wednesday, August 21, 2019

Expected Scott Bacon Recount

Resigning MP: Scott Bacon (ALP, Clark)
Recount from 2018 state election for remainder of 2018-22 term 
Contest between Madeleine Ogilvie and Tim Cox
Ogilvie likely, but not certain, to win [UPDATE: Ogilvie has narrowly won.]
Ogilvie may sit as independent and share effective balance of power with Sue Hickey, or may rejoin Labor. [UPDATE: Ogilvie has said she will sit as an independent.]

Recount updates will now be added at the top

Previous Party-Hopping Cases:

As noted below Ogilvie's (under unique circumstances for Tasmania) is the first case of a Lower House MP deserting their party mid-term and sitting with a different party status in 38 years.  However prior to that, this was a more common event.  Here is a not necessarily perfect list since World War II:

* Carrol Bramich (1956) Labor to Liberal (policy tensions and internal issues).  Re-elected as a Liberal.
* Reg Turnbull (1959) Labor to IND (kicked out after refusing to resign as Minister). Re-elected with massive support, later Senator.
* Bill Hodgman (1960) Liberal to IND. Defeated.
* Tim Jackson (1960) Liberal to IND (leadership change fallout). Defeated.
* Charley Aylett (1963) Labor to IND (quit after being disendorsed). Defeated.
* Kevin Lyons (1966) Liberal to IND (preselection issues). Later formed Centre Party and was re-elected.
* Nigel Abbott (1972) Liberal to IND (policy dispute). Defeated.
* Doug Lowe (1981) Labor to IND (leadership change fallout). Re-elected.
* Mary Willey (1981) Labor to IND (leadership change fallout).  Defeated.
* Madeleine Ogilvie (on recount 2019) Labor to IND (multiple factors)

All of the Bramich, Turnbull and Lowe/Willey cases precipitated state elections.

There is also the case of Gabriel Haros (Liberal) who lost preselection for the 1986 election and ran as an Independent, and probably other similar cases.

It is interesting to note the weak performance of some of these independents at elections.  In the 1964 election Bill Hodgman (Will's grandfather) managed only 475 votes and Charley Aylett only 102.  This didn't stop Bill Hodgman going on to become a two-term MLC for Queenborough (1971-83).

Thursday 12 Sep:

The main takeaway from Ogilvie's Fontcast interview is that Ogilvie has said that she thinks the government should be able to see out its term and that no confidence motions are generally stunts (unless someone does something pretty serious).  It's not a guarantee of confidence to the Liberal Party but it's close, and if that's the way it is then it removes Sue Hickey from the stability equation, since no-confidence motions will now generally be lost 11-13 on the floor depriving Hickey of her casting vote.  This also applies to any matter on which the Liberals and Ogilvie agree.

Wednesday 11 Sep:

Apparently the Mercury will be revealing Ogilvie's decision at midnight. (Done: see here)

UPDATE 9:06: Ogilvie will sit as an independent for the remainder of the term according to the ABC's Emily Baker.  If she does so even for a day she will be the first independent in the House since Bruce Goodluck (elected as an independent for the 1996-8 term, at the end of which he returned to retirement).  She will hence be the first independent as such in the 25-seat parliament that has existed since 1998.  As far as I know she will be the first MP to be elected on a recount but not immediately sit with the party that she stood for.  I think she will also be the first political defector in the House since Doug Lowe and Mary Willey quit the Labor Party following Lowe's removal as Premier in 1981.  (However there have been changes of party allegiance by MPs in the Upper House since, including Terry Martin's expulsion from the Parliamentary Labor Party in 2007.)

People are bound to ask whether she should be allowed to do this, as has been the case with a number of recent insta-rats in the Senate (some of whom were considerably more provoked).  However if a rule is passed that MPs cannot leave their parties, some will stay in their parties while voting however they like, so such a rule may not make any difference.  If the party can force MPs to leave parliament by expelling them (which would be the natural result of such behaviour) then parties could use the same power to exercise absolute control over their MPs, including expelling them over minor disagreements.  So I'm not sure what the solution is.

Tuesday 10 Sep: 

Today's the day.  The recount will be on today and there is a fair chance of it finishing or at least the winner being reasonably clear by 6-7 pm, but it's also possible the TEC will have to come back tomorrow.  Candidates for the recount won't be finalised til after midday.  At this stage I am not involved in the recount but will relay any information that I get.

2:45 The TEC has posted the initial distribution from the recount.  Ogilvie has 42.1% and Cox has 37.4%.  Sherlock has 15.1% and the rest 5.4% between them.  Cox would need a rather steep 61.4% of the remaining votes to go to him to win if none exhausted, which means he must have done badly (or at best about evenly, which isn't enough) on the Bacon-Haddad votes as these have been distributed.  Even more difficult for him now.

4:10 A few minor candidates have been excluded and Cox now needs 62.1% assuming zero exhaust.  Really we are waiting for the final exclusion (Sherlock) to decide it all.

4:40 Down to the last exclusion of Sherlock and Cox now needs 64.8% assuming zero exhaust.  Extremely unlikely.

5:02 Report from Alex Johnston (WIN) that Ogilvie has won by about 200 - if confirmed, that's closer than expected but a win is a win is a win.

5:25 Ogilvie has won by 201 votes (51.0-49.0) after a strong, but not quite strong enough, 56% flow from Sherlock to Cox.  This is the closest Assembly recount for 24 years and the third closest I can find (though records of pre-1950s recounts are hard to come by online) - a Labor recount in Franklin 1995 was decided by 54 votes.  A Liberal recount in Bass 2000 was nearly as close as this one (218 votes).

6:00 Ogilvie has posted a victory video on Facebook.  It doesn't say whether she will rejoin Labor but does comment about "the stability of our important Parliament" and "good governance" - whatever that means.


(Graph above - if the rate of decline since 2017 continues - not that I expect it to - then by the 2026 election there will be no blokes left in the Tasmanian Parliament!)

Tasmanian politics is abuzz with news that popular Clark Labor MP Scott Bacon has said he will resign from parliament for family reasons very soon, triggering a recount for his seat from the 2018 state election.  This article explains how the expected countback (confusingly officially called a "recount") works, pending confirmation that former MP Madeleine Ogilvie will contest it.  It is already known that former radio host Tim Cox will contest the recount.  A Cox win is no problem for Labor, but an Ogilvie win could be a big headache for the Opposition.

A Hare-Clark recount is based solely on the votes that the resigning member had at the point of them being elected.  How close someone did or did not get to being elected in the original count, or how many primary votes they got in the previous election, is not directly relevant.  There are cases where being one of the last candidates excluded is a disadvantage; this isn't one of them.  This recount will be won by one of the three defeated Labor candidates from the last election - but will the winner sit as a Labor MP? 

This is one of the cases where there is the most clear information to look at, because Scott Bacon was elected with over a quota on the first count.  He had a surplus of 932 votes in the original count, but in the recount all his papers will be thrown again at a total value of a quota.  After adjusting for a small loss due to fractions in cutting his original vote down to the 932 votes, the breakdown of his second preferences is approximately as follows:

33.1% Madeleine Ogilvie (ALP)
28.4% Tim Cox (ALP)
17.3% Ella Haddad (ALP) - elected, ineligible for recount
10.5% Zelinda Sherlock (ALP)
3.9% Cassy O'Connor (Green) - elected, ineligible for recount
1.7% Elise Archer (Lib) - elected, ineligible for recount
1.5% Sue Hickey (Lib) - elected, ineligible for recount
3.6% other candidates (all non-ALP)

In the recount, all Bacon's votes are distributed to the candidate who the voter has ranked the highest among the candidates contesting the recount, so Ogilvie will get at least 33.1% if she contests (for example).  If no-one gets more than half of them initially, the lowest vote-getter is excluded and their votes distributed again, and so on, like a single-seat election.

The votes that were 1 Bacon and 2 for an eligible candidate who contests the recount become starting primaries for the recount.  Where an ineligible candidate, or someone who chooses not to contest the recount, was at 2, the vote goes on to whoever was at 3, and so on.  As Ogilvie and Cox each have nearly a third of the total with nearly a quarter to be distributed again, they will clearly be the final two if they both contest; Zelinda Sherlock has a theoretical chance but in practice will not overtake either.  (She would need her share in a three-way split of votes to be 47 points higher than Cox's to even get into second, let alone win.)

Assuming it is Ogilvie versus Cox, 38.5% of the total vote for the recount consists of votes that were 1 Bacon and 2 for somebody who is either ineligible or cannot realistically win.  The question then is simply whether Cox can get a big enough edge on these 38.5% to overhaul Ogilvie's advantage on Bacon's known second preferences.

If no votes exhausted from the recount, Cox would need just over 56% of those remaining votes to flow his way to beat Ogilvie.  A small number of votes will exhaust, so the share he will need of those that don't will be slightly higher.

It will come down mainly to how those voters who voted 1 Bacon 2 Haddad and 1 Bacon 2 Sherlock saw the relative merits of Ogilvie and Cox.  The voters who voted 1 Bacon slightly preferred Ogilvie. Voters whose votes were with Sherlock when she was excluded (these are only an indicator of possible 1 Bacon 2 Sherlock votes) slightly preferred Cox. The preferences of the voters who voted 1 Haddad (again only an indicator of how 1 Bacon 2 Haddad votes might flow) between Cox and Ogilvie are unknown to me.

Overall Cox's task is pretty difficult here and it won't help him that virtually every 1 Bacon vote that wasn't 2 for him was 2 for a female candidate (all else being equal, voters who put males 1 and 2 would be more likely to also put a male 3).  However, it's close enough and Ogilvie was controversial enough that unless Labor have detailed scrutineering data on the contest, I can't be sure that Ogilvie would win.  She is however, in my view, the favourite.

If Ogilvie Doesn't Run

If Ogilvie doesn't run, then Sherlock would need to be ahead of Cox on at least 64.6% of the votes that were 1 Bacon and 2 for neither Cox nor Sherlock. In this case, gender factors would assist Sherlock but given her low profile compared to Cox it would be very surprising if she beat him.  Sherlock has said she will contest.

Ogilvie - Labor Or Independent?

Madeleine Ogilvie's sole term in parliament from 2014-2018 was now and then marked by severe friction with elements of the left and pro-LGBTI forces within the party, chiefly over Ogilvie's conservative positions on social issues of Catholic concern.  Following her loss to fellow Labor candidate Ella Haddad, Ogilvie left the Labor Party.  She ran unsuccessfully as an independent for the Legislative Council seat of Nelson.  (She finished fourth by a narrow margin, and may have won had she managed to overtake eventual winner Meg Webb). 

If Ogilvie contests the recount, wins, and chooses to sit as an independent, then there is nothing Labor can do about it.  Section 232 of the Electoral Act allows a party to order a single-seat by-election if it has run out of candidates willing to contest a recount, but the test is "none of the candidates who were included in the same registered party group as the vacating Member are available to contest the vacancy".  Ogilvie was included in the same registered party group, so Labor could not activate this clause by getting Cox and Sherlock to sit this one out and then claiming they had run out of candidates.

Ogilvie as an independent would further unsettle the already messy dynamics of the House of Assembly, in which the government in theory has a majority but frequently loses votes on the floor because of its rebel Speaker Sue Hickey.  However Ogilvie as an independent would be a bonus for the government, which would have a second route to pass social issues legislation opposed by Hickey.  (Note that this would be no help on issues on which Ogilvie sided with the left - indeed she was one of the early movers on Labor's now abandoned anti-pokies policy.)

Ogilvie as a Labor MP would also be a bonus for the government.  On some social issues conscience votes Ogilvie might still vote with the government, and the fact that Labor had accepted her back into the party after she resigned from it would provide the government with ammunition.  Ogilvie's Nelson run would look like a cynical case of fake independence and Ogilvie would have a lot of explaining to do concerning how she had run for election on a platform of helping the Hodgman government pass its bills (oh yes she actually sorta said that) but was now opposed to them.

Perhaps Ogilvie's presence and some conspicuous public fence-mending with her would help Labor with its current attempts to pivot back to the centre but even that would be embarrassing and the kinds of within-party conflicts involving her as an MP have run deeper than just ideology.  All up, Labor's best scenario here is that Ogilvie would prefer to spend the next 30 months studying space law and posting Youtube music videos on Facebook.

And whoever wins, being rid of such a star performer as Bacon from Labor's Clark lineup is a bonus for the government at the next election, making it harder for Labor to win three Clark seats.

News and other snippets will be added.

Update Thursday 2:40: Re contesting the recount, Ogilvie has said " I’m going to just take little time to consider, and discuss with my family, so I’ll get back to you hopefully in a few days."

Added: Timeline 

There is a two-week period for nominations to be received, which means that the earliest possible date for the recount is September 6 (update: it is actually on on September 10).  This gives the government an extra bonus because Labor will be down one for the first three days (at least) of the September sitting.  There is no pairing convention for casual vacancies (see my article on Rene Hidding's departure re this) and so it appears that the Government will enjoy a 12-11 floor majority in the first week of September and be able to pass legislation without relying on the casting vote of Speaker Sue Hickey.  It will be interesting to see what use, if any, the government makes of this.  (Probably anything all of Hickey , Labor and the Greens would vote against would not get through the Legislative Council anyway.)

Added: Changes of Numbers after Recounts

This recount is unusual in having the potential to change the party numbers on the floor of the Parliament.  As best I can determine recounts started from the 1922 election (replacing by-elections) and there have been 85 of them.  Nearly all have been like-for-like replacements at party level.

The best known example of someone outside the vacating party winning a recount was Bob Brown (IND) taking Norm Sanders' (DEM) place in 1983, despite other Democrats contesting the recount.  There is one other example - in 1961 Reg Turnbull (independent who had been Labor Treasurer in the previous parliament) quit to run for the Senate.  Over 40% of Turnbull's surplus had gone to Labor candidates and only 20% to his independent running mate, so a Labor candidate won the recount.

I have not found any precedent for a member being elected on a recount for the seat of a ticketmate from the original election and then not taking their seat as a member of that party.  However, there is an unsourced Wikipedia comment suggesting that Brian Crawford had left Labor when he initially won a recount for a Labor seat in 1962.  Crawford's win of the recount was successfully challenged in the Court of Disputed Returns on residency grounds (which fortunately for Tim Cox no longer exist), preventing him from taking the seat in any case.  More recently we had Brenton Best saying he would sit as an independent if he won Bryan Green's 2017 recount, but he didn't win.

Sunday - Monday: Ogilvie in

Ogilvie has confirmed she is contesting but has not said whether she will seek to rejoin Labor, and has suggested she will make that decision only if elected.

However, Ogilvie has also made some comments, especially a claim that she would take a "more intellectual approach" to parliamentary tactics, that will do nothing to endear her to those within Labor who do not want her back.  (Update: on Monday there is more of this stuff with Ogilvie claiming that being a "social democrat" means "that I’m allowed to be intelligent and think things through.")

Matthew Denholm has reported that the recount is on September 10, meaning that the government will have at least four days with a floor majority (possibly five as the recount may not finish until noon the following day).  He also reports Ogilvie as seeking confidence that Labor would share her priorities (including traffic congestion) and that there be an end to 'hostility' (Denholm's word, not Ogilvie's) from the Left.


  1. Hi Kevin, do you have an explanation or view on the design/policy of this recount system? Why do the votes of electors who got a preferred candidate elected at the last election (Haddad, O'Connor, Archer, Hickey) get a second go, while the votes of electors left in the count with no candidate elected get left out? Are their different systems in different jurisdictions? And are some systems better than others for a proportional outcome? Thanks

    1. In this case the only votes that are included are those that were 1 for Scott Bacon, because he got a quota. All the 1 Bacon votes are included in the recount, whether they ended up with an elected candidate, with the last unelected candidate (Ogilvie), or with exhaust.

      The main theory behind the system is that if the departing MP's votes are used to choose the departing MP's replacement then that should provide the closest thing to like-for-like replacement of the MP who has left the parliament.

      One alternative is to recount the whole election with the departing candidate's votes re-allocated to others. This is what is done with Senate disqualifications. For retirements it has a few problems though: 1. in close cases it may elect a replacement from a different party, which might discourage an MP from quitting in a case where they really should quit. 2. in some cases it could "unelect" one of the other MPs who was originally elected and 3. it is a lot more work and takes longer.

      The current system is better than recounting the whole election for those reasons, but it is not perfect. In cases where the retiring candidate is elected partway through the preference throw (not the case with Bacon) there can be an advantage for candidates excluded early in the original count over those who were excluded late.

      State PR houses that do not use Hare-Clark use the same casual vacancy method as the Senate (in effect appointment by the party); ACT which uses Hare-Clark uses the same vacancy method as Tasmania.

    2. Abroad the main methods for filling Hare-Clark/STV vacancies are:

      * A by-election of the entire constituency
      * Co-option of the outgoing member's party's nominee
      * Co-option of a name from a list of designated successor replacements supplied by the elected member as a candidate

      Off the top of my head by-elections are used for the Republic of Ireland's Dáil (lower house) and the university seats in the Seanad (upper house). The Seanad Vocation Panel members are normally elected by an electorate of members of the incoming Dáil, the outgoing Seanad and members of local government. By-elections see the electorate restricted to the current Dáil and Seanad members. Co-option is used for local government and, I think, the European Union Parliament (each EU member state determines its own voting system for that).

      Scottish local government uses by-elections.

      Northern Ireland uses co-option for the Assembly, the European Parliament and local government with the designated successor method for independents and, if none are available, a by-election for local government.

      (There's been one by-election in nine years, for a local government seat where the sitting Independent died and all on the designated list declined.)

      I think Malta uses co-option for its parliament.

      All have benefits and drawbacks. By-elections can distort the proportionality of a constituency's representation, with small parties losing seats mid-term. Northern Ireland's political structures build the community divide into the system so midterm variations can significantly distort things (and there are those who argue that casual elections could be politically damaging at key moments, a hangover from the 1970s attempt at power sharing). Co-options can preserve the party balance but bring all the risks of high member turnover, party patronage, lack of voter accountability and can also encourage parties strategically arranging resignations and co-options to bring forward new candidates so they can fight a first election as an incumbent.

  2. Hi Kevin, in the list of defections, what about Gabriel Haros (1986)?

    1. Haros lost preselection and ran as an indie at the 1986 election. Did he actually sit as an indie in the parliament though? I haven't yet found evidence that he did, although he may have done.