Tuesday, October 31, 2017

Section 44: Could Parry Peril Unelect McKim?

Warning:  Section 44 has now reached Wonk Factor 5/5.  

The Section 44 citizenship crisis affecting the Federal Parliament has reached a new level of electoral law murkiness today with the news that a Tasmanian Senator, Senate President Stephen Parry, is investigating whether he may be a UK dual citizen by descent.  Parry's father was born in the UK but moved to Australia as a child.  If it turns out Parry is ineligible, he is expected to resign from the Senate immediately, and following a referral the High Court would presumably follow the precedents set in Re Canavan regarding Barnaby Joyce and Fiona Nash and give him the flick.

That might all be the straightforward end of proceedings.  The possibility of a special count for Parry's seat (a recount of all votes as if the departing Senator had died before the poll) creates new problems in dealing with the replacements for ineligible Senators.   The reasons for this concern two unusual features of the Tasmanian Senate count: the very high rate of below-the-line voting and the extremely close result for the final seat.  Incumbent Nick McKim (Green) held off Kate McCulloch (One Nation) by just 141 votes.  McKim should have been re-elected easily but the Greens vote was damaged by the successful campaign to save Lisa Singh, who had been dumped to a normally unwinnable position by her party.  

In most states, a major party Senator being found ineligible just results in their seat being taken by the next major party Senator on the list, with only very minor impacts on the rest of the count.  But in Tasmania's case, the impacts on the count are major, because of Richard Colbeck's high below the line vote.  In the original count, after the election of the first three Liberals, neither David Bushby nor Richard Colbeck had a quota and both remained in the count for a long time.  Colbeck was finally excluded in 14th place with 0.648 quotas to distribute.  In the special count, Grahame Bowland's simulations show that Colbeck (as Liberal 4) is now elected directly on the surplus of Bushby (now Liberal 3) leaving only one Liberal, John Tucker, in the count.  Colbeck received a lot of below-the-line preferences, and his votes were also rather leaky, and Tucker is eventually excluded also in 14th but with only 0.551 quotas.  There are many other changes created through the count and the upshot of them all is that this time, McCulloch beats McKim by 227 votes.  Note that Grahame's simulations for the Bob Day recount - based on programming of data files including all actual votes - ended up with exactly the same numbers as the AEC's.

So What Does That Mean?

So here we have the question: what happens when a Senate special count for one candidate's casual vacancy returns not just one Senator who was not originally elected but two, and fails to return one who was originally elected?  

Firstly, there's the question of whether the court could examine McKim's position without a specific reference to do so. In Re Wood, the court noted [22] that in the absence of a further reference concerning other Senators, they could only rule on a replacement method for the Senator who had been disqualified (in that case Robert Wood).  However, a special count showing that Kate McCulloch was the "winner" of the final seat once Parry was removed (as if he had died before counting) from the original election, could give rise to a further referral to determine whether McKim was rightly elected.  Also, in the single judge hearing to settle the recount process for Heather Hill's seat (long AEC account here), other parties were allowed to be represented in court in case "any questions arose about disturbing the election or non-election of other candidates" (as it turned out that wasn't an issue there.)  So there do seem to be ways in which the matter might be heard, or perhaps even must be heard, if Parry is ineligible.

[EDIT: Michael Maley in comments picks up Graeme Orr pointing to sections 360 and 379 of the Act which seem to grant the Court of Disputed Returns wide powers.]

If the final seat matter was heard, the argument could be run that McKim's election by a very narrow margin was a product of the presence in the count of a candidate who was ineligible to stand and that McCulloch had been unfairly denied victory by the presence of this ineligible candidate.  Even if one might run the hypothetical that had the Liberals known Parry was ineligible they would have acted differently, McCulloch could then claim that in that case they may have preselected Colbeck higher causing her to win anyway.  Courts generally do not explore such what-ifs.

The counter-arguments could then be that the result (the unseating of an eligible Senator who had been declared elected and done nothing wrong) was perverse and potentially conducive to unnecessary instability concerning the makeup of the Parliament, and that therefore the use of a special count to fill a vacancy that only exists because of the conducting of a special count was not appropriate.  But then it might be countered that McKim was no different to any MP who had been wrongly elected as a result of, say, a counting error, and that the fact that it wasn't his fault makes no difference.  And so on.

In Re Wood, the High Court made some negative remarks about a case Hickey v. Tuxworth (1987) 47 NTR 39 in which an election had been ruled void because a failed candidate had been ineligible, and had that candidate not stood, it was arguable that the results could have been different.  This argument nearly raised its head again in a seat at the 2015 Queensland election, only for the issue to be avoided when the ineligible candidate's votes could not have changed the outcome.  It is worth quoting the Court's remarks in Re Wood in full:

The decision of Nader J. in Hickey v. Tuxworth (1987) 47 NTR 39 was cited in aid of the submission. That was a case where a challenge to the validity of the election by preferential voting of a qualified candidate in a single member seat was declared void. Nader J. held the election void because, an unqualified candidate's name being on the ballot paper, it was impossible to be satisfied (having regard to the closeness of the voting and the order in which preferences were distributed) that the result would not have been different if the unqualified candidate had not taken part. With respect, the conclusion in Hickey v. Tuxworth cannot be accepted, at least for the purposes of the Act. If the unqualified candidate's nomination in that case was formally correct and his name was properly on the ballot paper, it is difficult to see how the election miscarried. The problem of want of qualification arises under the Act if an unqualified candidate is elected, but an election is not avoided if an unqualified candidate stands. If it were otherwise, the nomination of unqualified candidates would play havoc with the electoral process, for the ministerial officer who accepts nominations has no general power to refuse a nomination in due form: see s.172 of the Act. In any event, in the present case, it is unreal to suggest that the presence of Senator Wood's name on the ballot paper has falsified the declared choice of the people of the State for any of the first eleven candidates.

The differences in this case are firstly that it is very real to suggest the presence of Senator Parry's name on the ballot paper has at least raised a question over the declared choice, and secondly that this is a less common case in which the unqualified status of a candidate can be shown to lead to a different outcome based on the results as cast.  However, since we already know that seven people who were elected at the 2016 election were ineligible to stand, it stands to reason that there would be many unelected candidates who are also ineligible.  

If a declaration of ineligibility regarding Parry could lead to a special count in which McKim's seat was at stake, then the same might be argued regarding failed candidates.  This could in theory, for a really close outcome, lead to a series of recounts with a Senator constantly being elected and unelected.  For instance, suppose Parry was unelected and replaced by Colbeck, then a few months later Colbeck was found to have been ineligible too.  In that case, McKim would be unelected and replaced by McCulloch ... who would then be unelected and replaced by McKim. That's only considering two of the candidates who actually were elected!  (That said, in this case it is very unlikely that the loss of any unelected candidate who isn't Colbeck would make a difference.)

Even if the court decided that McKim's place was unimpeachable in this manner, an argument might also be run that although it is "obvious" that Colbeck should replace Parry, nonetheless the special count does return two new Senators in place of one, both of whom "should" be declared elected, and provides no objective test for who should be the replacement.  By analogy with the WA 2013 Senate voiding, an argument might then be tried that the whole Tasmanian Senate election is void beyond repair and should be rerun.  

An Actual Unelection

Earlier this year we saw the only case I am aware of of this innocent-victim scenario actually happening.  (See Malcolm Baalman's detailed writeup.) The luckless victim was Michael Caiafa, then a Melbourne City Councillor.  Unfortunately, Melbourne City Council continues to use the undemocratic Group Ticket Voting system.  Cr Brooke Wandin was found to have been ineligible because she did not live in the municipality.  The rerun of the election with Wandin removed unsurprisingly elected her ticketmate.  However, because Stephen Mayne's group had submitted a preference ticket that flowed to Wandin but then did not flow directly to her ticketmate, the countback cutup for other councillors was substantially different and the rerun no longer elected Caiafa.  Instead it elected a member of Team Doyle, giving Team Doyle a majority on the Council.  Thus, Caiafa was unelected as a result of not one but two oversights by others over which he had no control.  

The Victorian Civil and Administrative Tribunal decision may be downloaded here, and cites Re Wood and the more recent Culleton cases as examples of precedent, but with very little discussion of possible objections to "unelecting" candidates compared to what would come out in the High Court.  Caiafa did not appeal the decision to the Supreme Court.  The finding - which is under different laws and at a much lower level than the High Court - so presumably nothing by way of precedent - did however state that both the newly "elected" candidates needed to be considered, given that both would have been elected had the ineligible person been excluded from the start.  

In my view it is quite unclear what the High Court would do about the "unelection" of a Senator on a special count and whether they would follow VCAT or not.  The "vibe of the thing" would be that unelecting people is just silly, but some people also thought that of the technical disqualifications just seen.  Weighty arguments can be made in either direction. 

How This Could Be Avoided

The real problem here is that "special counts" (recounts of the whole election) are a silly system for filling vacancies.

In Tasmanian Hare-Clark recounts, this "unelection" situation cannot arise.  Each departing member is replaced by another by a recount of only the votes they had when elected.  This can create issues of its own (like the "recount bug" I have discussed here often in which a candidate is disadvantaged for polling too well in the first election) but in a Senate context these would be irrelevant and a disqualified Senator would generally just be replaced by the next on the party list.  I have frequently argued against using whole-election recounts for Hare-Clark precisely because of this "unelection" problem.  

In theory, the Tasmanian system could be adapted to the Senate, but there's a problem with it for that purpose.  That is that when a federal Senator is elected, they simply have a number of votes to their name and do not retain any particular ballot papers at part or full value.  The value of the ballot papers is recalculated (and wrongly distorted by the unweighted Inclusive Gregory system) as they are passed on to the next candidate. A rejig would be required so that votes were defined as remaining (at a certain value) with the elected candidate, and this value could then be used for recounts.  This would be an excellent opportunity to get rid of Inclusive Gregory distortion too.

[Note added 7/11: An argument against the above solution is that it might in theory result in a party deliberately and knowingly running an ineligible but popular candidate, with the intent that when that candidate is disqualified, their own votes will elect a less popular candidate who otherwise would not have won.]

What A Mess!

That these issues are even arising with Parry, even should he escape with his seat intact, shows that the Liberal Party's vetting was not sufficient to completely avoid problems.  Thus, it is not just a problem with One Nation, the Greens and Nationals (all minor parties.)  Thus far the Labor Party has escaped having any of its MPs referred, but it's hard to tell if that is really because its processes are bulletproof, or if perhaps any Labor MPs who might be in trouble are lying low and not disclosing it.  The case of Justine Keay (ALP, Braddon), who correctly applied to have her UK citizenship revoked but appears to have needlessly been too slow doing so to have it revoked by the date of nominations, at least suggests that Labor is not perfect either.

How many more? 

I may have further thoughts later.  There were two articles, one presentation and one report I meant to write today - this wasn't one of them!

Not Referring Parry

It's possible Parry could resign and not be referred, with the state party nominating Colbeck, the argument possibly being that this just avoids a High Court outcome (Colbeck's election) that is inevitable anyway.  This might be argued simply as a way to save money.  Labor and the Greens would not want to see McKim bumped so would have every reason to be agreeable to this.  However, One Nation would thereby be denied any effective chance to have their claim for a seat considered by the court, and might therefore argue that it was a partisan non-referral for political reasons.  I am confident proper process will be followed and he will be referred.

Update (8 Nov)

Help for McKim could come from the possibility that Senator Jacqui Lambie could be a dual citizen (father born in Scotland).  If Lambie is also excluded, her #2, Devonport mayor Steven Martin, is no longer elected with quota on the first count, as a result of the massive leakage from Lambie's huge below the line vote.  This then creates its own knock-on effects as a result of which McKim romps it in by an enormous 615 votes!

This raises questions of timing concerning whether all possible issues arising from the continuing Section 44 investigations will be referred to the High Court and heard at once, or whether they might be heard and resolved at different times.


  1. It seems that the Liberals might be just getting Parry to resign and be replaced by Colbeck as a casual vacancy - possibly the simplest way to go, even if not strictly the most lawful.

    1. (This raises the question though, why would the Liberals baulk at a possible chance to get rid of McKim?)

    2. Of course that wouldn't change Parry being ineligible in the first place, meaning that if he was referred all the questions about McKim would be opened. But if the Liberals took the position not to refer it, which Labor and the Greens would have good reason to concur with, One Nation would be left fuming on the sidelines. They could perhaps try a common informer petition but this could be complicated by the replacement Senator not being constitutionally ineligible (and the provision is a mess anyway.)

  2. Are you saying Colbeck was also ineligible or do you mean Parry? If Parry resigns and no one refers him to HC could he get renounce his offending UK citizenship and then be nominated by the Libs as his own replacement?

    1. I meant Parry so I deleted the comment and reposted it, ta. In theory that could be the case, but the Government would have to explain why Parry had not been referred and hence allowed to take advantage of that loophole when all the others were referred and disqualified. I suspect the other parties would gang up against such a solution and prevent it. If he is ineligible he should go like all the others.

      There might also be a common informer petition against an originally ineligible Senator who was reappointed to their own casual vacancy in this way.

    2. I suspect that the common informer provision does not extend to Senators reappointed to possitions, so long as they are not disqualified to be a parliamentarian or otherwise declared incapable of being a parliamentarian (possible example: not compliant with the party provisions in section 15). Although I am not sure about this, not being a lawyer.

    3. In any case the common informer provision is a quagmire; there are a lot of doubts about how it will go with Gillespie.

  3. Another excellent comment by Michael Maley:

    Thanks for your interesting post on the Parry problem. Here are a few thoughts/comments.

    The logic that the Courts have been trying ever since the Wood case to bring to bear on this issue is getting ever more twisted. In Wood, the Court observed among other things that “A Senate election is not completed when an unqualified candidate is returned as elected.” And it said that the appropriate way to complete the election was to count the votes as if the disqualified candidate had been deceased. But if that is the basic principle, the notion of a candidate being unelected comes with it, and needs to be accepted. If the notion isn’t accepted, the basic principle has to be watered down – but to what?

    Of course, un-electing a candidate may encounter any number of preliminary legal obstacles, including that the 40 days for challenging that candidate’s election is over, and that a question of qualification doesn’t arise. And in Wood, the Court was at pains to dismiss the possibility that any seat other than that of Wood himself could be in doubt. Graeme Orr has nevertheless suggested on Twitter that it may be open to the Court to declare a candidate unelected when dealing with a reference from the Senate concerning another candidate, citing section 360 of the Commonwealth Electoral Act 1918, as picked up by section 379.

    If it be supposed that Senator McKim for whatever reason cannot be unelected, there’s then the question of what to do when a court-ordered recount to replace Senator Parry in fact “elects” two new candidates. Antony Green has been quoted as saying that “The court will almost certainly choose Richard Colbeck, as he was the earlier elected candidate, and also because the election of a replacement from the same party is in line with the casual vacancy provisions of section 15 of the constitution.”

    That’s all very well, but I find neither of those considerations very convincing.

    There’s really no reason why a candidate elected late in the count is less deserving of a seat than a candidate elected earlier. Both candidates will have received a quota of votes, and indeed, it’s quite possible that a later elected candidate, when compared with an earlier elected one, may (i) have a larger surplus; and (ii) have in his or her quota a greater proportion of first preferences. Trying to order elected candidates as more or less deserving really takes one outside the basic logic of STV.

    And in relation to section 15 of the Constitution, the problem is that it’s designed to maintain the proportionality of representation at a validly conducted previous election, not as a rule for who should have been elected in the first place.

    Finally, while in this case those two ad hoc principles might point to the same outcome, it’s entirely possible that in future they would point to two different outcomes.

    My guess is that the High Court would be aghast at finding itself confronted with such a dilemma: the last thing it would want would be to be seen to be choosing a senator itself, rather than giving effect to the votes of electors in accordance with the law. The problem is that the law is silent on what ought to be done in that specific situation.

    For that reason alone, I wouldn’t be surprised if the Court were to see the implementation of the result of the recount, with Senator McKim unelected, as the least messy and most defensible in principle of the various possible steps open to it.

  4. I would think kicking out a Senator who has not been disqaulified and replacing them with their polar opposite would look very political and messy. I think a new election is the better outcome if an inelligable candidate being removed removes an elligable candidate, although it does raise other issues.

    If the Tasmanian Senate Election was rerun, it would raise the question as to whether or not the new election created the terms of the senators elected, throwing Tasmanian Senate elections out of kilter with mainland Senate elections until the next DD. Section 13 would likely need looking at by the High Court.

    If the Senate terms were thrown out of kilter, until the next DD, it might give Tasmanian premiers a taste of deciding half-Senate election dates (the more likely the more separate half-Senate elections occur), which may continue even after term realignment. Xenophon becoming SA Premier (probably not likely at the 2018 election but he might become Opposition leader and then win a subsiquent election) would increase the chances of SA independently choosing half-Senate election dates (especially if Tasmania was choosing its own half-Senate election dates) as he would have fewer pressures in favour of conformity.

  5. On Lambie “uncomplicating” the situation: I don’t know what the simulations say, but I think it’s at least mathematically possible that [Parry DQ’d > Parry recounted and replaced (While Lambie is eligible) > Lambie DQ’d and replaced] produces a different result to [Parry and Lambie simultaneously DQ’d and replaced in a single recount]. In which case, it’s possible that if the HC takes an anti-unelection approach, someone might miss out on a seat that they would’ve had if Parry had been referred earlier. I have no idea how that might be handled.

    1. The end point is the same in either case since both post-Lambie special counts would be minus both Parry and Lambie. (This is different to the Tasmanian system in which there is no provision for recounting multiple places at one time and the order of recounts could make a difference.) However it is certainly possible for someone to be temporarily one of the "winners" if the countbacks are done separately as opposed to together.

    2. My understanding/assumption is that Tasmania`s coutback is a casual vacancy provision and that an ineligability issues could result in a full recount (Although lacking a state level equivalent section 44i there are far fewer chances of this).

    3. At the 1979 state election three Labor MPs in Denison were found to have been ineligibly elected by reason of breaching spending restrictions. Rather than holding a full recount, a by-election for all seven places in Denison was held as a result of the passing of special temporary legislation to that effect.

    4. I forgot about the particularls of thatthat. Although my understanding is that the by-election/special election was legislated to avoid a partial election, where only the ineligable candidates seats were up for election, rather than a recount.

    5. I believe that is correct. I am not sure if a full recount was specifically canvassed as an option.

    6. I was before the High Court came down against anything other than special elections or special recounts in the case of issues with a proportional election.

  6. Lambie's replacement gets bumped down to 9th elected, instead of her 4th. is there any prospect of the term times getting reassigned? what about if 2 senators are replaced simultaneously who have different terms lengths, who get which term? though i know that isnt the case here as both Parr and Lambie have (had) 6 year terms.

    Also i see there are questions of numbers 2 and 3 on JLN ticket possibly having issues with being an office of profit of the crown. if that is the case, is there any risk of McKim (or others) coming under question again?

    1. If Martin is ineligible Waterman is elected with no impact on McKim. If both are ineligible then the outcome hasn't been simulated yet but I'd strongly expect it to be that McKim and McCulloch are elected.

      In the Ludlam case the High Court held that Steele-John was elected in place of Ludlam. The Senate has not thus far attempted to rearrange their terms and there are serious questions about whether it can. So I am doubtful that we will see reassignments of term lengths based on changes to the six-seat recount outcomes.

      In a situation where two Senators are replaced at the same time, in most cases each will correspond to a party of a departing Senator and they will therefore be assigned to the place (and term) of the Senator from their own party.

    2. If both Martin and Waterman are ineligable, then things get very interesting because either JLN ticket votes are distributed to their next groups (they could really scatter, I presume) or the High Court potentially finds that a Senate group all being ineligable is the equivalent of an individual candidate being ineligable in the House of Reps and a new election is required.

    3. Details of where JLN ATLs go to next can be found on David Barry's site at https://pappubahry.com/pseph/aus_2016/tas_atl/ PHON is the biggest recipient. The Liberals get more than the Greens but not by enough to matter. I think a special count without any JLN candidates would see McCulloch elected 11 and McKim 12.

    4. Kevin,

      The Holly Hughes ruling today has something to say about when the election is deemed completed. Could this have an impact on when the first sitting of the senate following the election be deemed to be for the purpose of determining who gets long and short terms?

    5. That's an interesting possibility (ie that it might be deemed that the "first sitting" might be deemed to be extendable in a case where the electoral process continues indefinitely). I don't sense any political will to do anything about re-evaluating term lengths though, perhaps because there's so much else going on.

      A hypothetical in this light: if term lengths can be re-evaluated then it might be possible for a Senator originally elected to a six-year term to lose their seat during the second half of it as a result of a revised determination of the term lengths for that election. Which to me appears uncannily close to expulsion, given that the Senate can set term lengths however it likes.

  7. But if McKim stays, doesn’t logical consistency demand that that other temporary winner stay as well?

    And if the answer is that it doesn’t count because the two recounts were really close together, that temporary winner has been disadvantaged by Parry’s failure to get himself referred as soon as he became aware of his issue. In fact, taken to the extreme, it could create incentives for potentially ineligible Senators to hide their status (or rush their referral) to secure a more favourable recount depending on who else is getting disqualified and when.

    1. In the case of separate special counts for Parry and then some time later Lambie, there are two possibilities: either Senators can be unelected in special counts, or they can't. If they can, then it also follows that someone elected on a special count who wasn't previously a winner might be unelected on a subsequent special count, and that someone who was unelected on a special count might be elected on a subsequent special count. So if McCulloch could win a seat on one such count, she could lose it again on another one.

      The potential for "unelections" to create a disincentive for resigning in cases where a Senator should resign is one of the problems with the idea of the Court allowing Senators to be unelected. It is also one of the arguments for using the Tasmanian casual vacancy system instead of special counts. I did see some comment to the effect that a misconceived view of an impact on One Nation's chances of gaining party status may have affected the timing of Lambie's resignation.

    2. Resignations don`t cause recounts, only inelligability during the election process (Although in Hollie Hughes`s current NSW case that could potentially be ruled to be nomination day-present, causing another special count) and so resignations do not effect special counts.

    3. Resignations for reason of perceived ineligibility have thus far caused referrals to the Court of Disputed Returns which has then ordered a special count. The ineligible candidate could also be referred without resigning, but if they had not placed information in the public domain to justify referring them, then that might not occur. Hence the point about potentially hiding status and/or timing a resignation strategically.

  8. If the information justifying a successful reference later came to the Senate`s attention, the original election of the Senator would likely be thrown out and thus the csual vacancy be deemed never to have existed and the filling for the vacancy deemed void.