Once I'm on holiday, I hope to post a fuller explanation of the reasoning underpinning my assessment of Paula's Peril - why I think, if an election petition went against her, that she'd be out of Parliament, but in the interim - and I know it might not sound like fun - but please go and read Judge Adams's recount judgment.
No jokes. It's 14 pages of democracy.
An (unlikely) scenario:
Paula Bennett requests a judicial recount of the votes in the Waitakere electorate. And after the judge (or the officer appointed by the judge) conducts the recount, the numbers have changed, and Ms Bennett has won by a few votes. The Electoral Commission publishes the final result, declares Paul Bennett the MP for Waitakere, and Raymond Huo a Labour List MP.
Carmel Sepuloni, unsatisfied with the conduct of the recount in the District Court, or concerned that people's votes were improperly counted (or were impermissibly not counted), launches an election petition. Some time in the new year, the High Court hearing the election petition, able to enquire into the qualification of electors in a way a simple recount cannot, for a variety of reasons, finds that Ms Sepuloni has won: not because of a corrupt practice or anything like that, but because on a close analysis, she just had more valid votes.
What happens?
Well, the Court certifies that Carmel Sepuloni won. Carmel Sepuloni is sworn in as a member of Parliament, and Paula Bennett ceases to be the member for Waitakere. And ceases to be an MP.
Not Raymond Huo, though, he sticks around. Same with Cam Calder, Paula Bennett wouldn't take the list spot he fills. National's parliamentary strength has fallen by one from the official result, and Labour's is up, meaning there is no longer a parliamentary majority for the partial sale of state-owned assets, and National can't pass legislation with only the support of Peter Dunne and John Banks.
Why is this? Well, a policy decision was taken that finality was more important than proportionality, and the possibility that an election petition (or by-election) could change multiple seats (e.g. by removing a party from Parliament because it no longer passed the one seat threshold) months after an election was thought to be the greater evil. Of course, if it's certainty you want, this doesn't really provide it, as Paula's Peril shows. But the hilarity.
There probably isn't a good answer to this conundrum. But whether the balance has been struck correctly is something I think I'll be asking the MMP review to address. I can accept that proportionality may be lost through a by-election, but an election petition related to a general election is a different beast, and the distinction may be enough that different rules should apply. And if that means that - on very rare occasions - we have to wait a month or more before we're absolutely certain of the result, that may be a price worth paying.
The special votes are in, and my rudimentary analysis was pretty good. I was right overall (National down one, Greens up one) and on the small scale as well: getting the Green, New Zealand First, Maori Party, and United Future votes correct to two decimal places; and also correctly predicting that Mana would pass ACT.
We won't get the full details for a couple of months (split vote analysis which show various oddities, like the 1039 voters at the last election who party voted National in Helensville but wouldn't vote for John Key personally, or the 262 Epsom voters who wanted ACT, but not Rodney Hide) but there are some things to note.
1. If either John Banks or Peter Dunne had lost their electorates, Labour would have picked up a list seat, and National couldn't have governed without the Maori Party (same if the overhang had been one higher).
2. The Special Votes wouldn't have altered have altered the no-threshold counter-factual. National couldn't have governed without either the Maori Party or the Conservative Party.
3. The Maori Electorates were massive MMP supporters (Te Tai Tonga at 78.9% for keep, the only one below 80% support); Mangere, Manukau East, Rongotai, Wellington Central and Manurewa the strongest MMP supporters among general seats, each with 70%+ for keep. Clutha-Southland were the biggest supporter of change, at 55.4%. 14 of the 70 electorates voted for change, all National-Party held.
4. The massive number of informal votes recorded in Part A of the Referendum in the Botany electorate after the count of the advanced votes has been reduced to being in-line with everyone else (leaving me to wonder if one of the people counting the advance votes there on election day made the mistake the Electoral Commission warned others about, and thought you had to vote in both parts).
5. First Past the Post topped the valid vote in Part B in 68 electorates, with Epsom voters favouring Supplementary Member, and Wellington Central voters favouring Single Transferable Vote (and SM).
6. Labour came third in the party vote in Wellington Central.
7. Parliament has three new "backdoor MPs" - incumbent MPs who were re-elected on the list after losing their electorates - Chris Auchinvole and Paula Bennett from National, and Clayton Cosgrove from Labour. Auchinvole was beaten by Damien O'Connor, who was a "backdoor MP" in the last Parliament.
Finally, it seems likely there will be a couple of judicial recounts, with Paula Bennett behind in Waitakere by 11 and Brendan Burns behind in Christchurch Central by 45. Despite their name, judicial recounts don't actually have to be conducted by judges, but can be conducted by an officer of the Court with assistants without the judge even present.
The entry into the House of Representatives of "Unelected List MPs" is a common complaint about MMP.
And while the response: they were elected by the party vote, and voters were told who was on the party lists before they decided their party votes is a pretty good one, it's not enough for a lot of people, so mechanisms by which voters can have a greater influence over the exact membership of the House are often proposed.
A frequent suggestion is that, instead of having a list, party vote seats (I guess, you can't call them list seats!) in the House would be filled by the best runners-up from electorate races: by the candidates for that party who scored the greatest percentage of votes in electorate races, but without winning those electorates. It is argued that these people have established some evidence that they have personal support to be in Parliament (or at least more personal support than those who were candidates for their party who didn't do as well). There is something in this argument, although I'm not a convert to it. At least yet.
Given the (hopefully) upcoming review of MMP, and the likelihood of this being proposed in submissions, I thought I'd take a look at this through a counter-factual, using, as an example, the newly elected Green caucus. Except it's quite different. If, instead of a list, empty parliamentary places were filled by best losers, the Green Caucus would have been:
Russel Norman
Metiria Turei
Catherine Delahunty
Dora Roimata Langsbury
Kennedy Graham
James Shaw
Jack Taotokai McDonald
Holly Walker
Steffan Browning
Eugenie Meryl Sage
Sue Coutts
Joseph Burston
Robert Moore
and the possibility of Aaryn Barlow
And gone would be:
Kevin Hague
Gareth Hughes
David Clendon
Jan Logie
Denise Roche
Julie Anne Genter
and the possibility of Mojo Mathers
If I get some time, I may look at the other parties (they're slightly harder to do, because the other parties included list-only candidates), but I suspect there would have been a number of unexpected MPs, perhaps including some welcome returns. As you can see from the Green Party above, there might be a very different parliament: one that might have included our fastest ascension from the Youth Parliament!
Of course, this somewhat random group shows one of the major problems with using a best runners-up system (and with open list systems): the number of MPs a party gets was determined by its party vote, but without a closed list, no-one can know who those MPs will be. Given that I'm inclined toward some form of open list, this is something I'm going to have to think about (although at least an open list has the advantage that the overall list is determined by the voters across the country, rather than in particular electorates, where perhaps the reason someone did well, was that they were one of only three or four nominees).
Having spent last evening at an election results watching party, where I got to explain to bunch of politically-savvy people, who figured they'd have known if anyone, exactly why we won't get to know the final result on the referendum until December 10, I thought I'd go through it here.
The Electoral Act 1993 governs the counting of votes in a general election, and it requires that on polling day, all the advance votes are counted, and all the votes cast on polling day are counted and a provisional result is announced. They don't count special votes, and they don't do the checking to ensure that no-one voted twice, but they give everyone a pretty good idea about the result.
The Electoral Referendum Act is different: it required local returning officers to count the advance votes (which they can start counting from 2pm), but forbade them counting normal votes on election day; instead of individual polling places conducting a preliminary count, they must instead send the voting papers to the returning officer before they are counted at all. The Act requires that each returning officer count the referendum votes "As soon as practicable after the parcels of referendum voting papers are received by a returning officer."
I'm not entirely sure what that means, because despite advice from the Electoral Commission that they won't be releasing any results until the final official count, I don't see that it prohibits them from releasing a preliminary count of the ordinary votes in the days after polling day. A final result - waiting for overseas votes to reach New Zealand and the like - will take as long as the parliamentary elections, but it seems pretty practicable to me to count the ordinary referendum votes in the next few days.
The count of referendum votes is a lot easier than the count of general election ballots: the usual protections of traceable ballot papers weren't imposed, so even if someone double-voted, all their referendum votes will count, because there's no way of finding the ballots they cast. Even the protection of scrutineers was dispensed with, and with the requirement to get the votes to the local returning officer pretty quickly, there's seems to be no reason after the "as soon as practicable" count is completed, why a fuller preliminary count couldn't be released. The law may not require it, but it doesn't ban it, and the principle of open government, and transparent elections, seems like an excellent starting point.
A final note about special votes: it has been pointed out in comments in other places, but bears repeating, that the numbers of special votes listed for each electorate are the number of special votes cast in that electorate. These may be special votes from people who aren't on the printed electoral roll (because they're on the unpublished one, or they enrolled after the rolls were closed for printing), but they also include votes cast within that electorate in respect of another electorate. In an electorate like Auckland Central, which has a lot of special votes, many will have been cast from people in neighbouring electorates who found it more convenient to vote while in the city on Saturday.