We'll know the election night results very early on Saturday evening. But we likely won't know the election outcome until early October.

Now we’ve entered the last week of the election campaign, Saturday’s finishing post is in sight. Once the polls close at 7pm on that day, no further ballots may be cast.

Then the really fun part begins as we wait to find out what the public collectively has decided. As it is illegal to conduct “exit polls” of voters on election day, the media has no way of accurately forecasting what the result will be.

That fact leaves those covering the election, as well as those contesting it, just as much in the dark as are those of us watching and listening at home.

In 2017, we can expect the result to become clear both more quickly and later than we have been used to in the past. Let me explain what I mean by that apparent contradiction.

First of all, the ever-increasing number of votes cast in advance of election day itself means the election night result will become obvious sooner than we are used to.

Nearly half a million people have already cast an advance vote, with 45,822 making their decision yesterday - the first year advance voting booths have been available on a Sunday.

Election officials can begin counting “early votes” – advance votes cast by voters on the printed electoral roll for the electorate they voted in – at 9 am on election day. Their count of these early votes then becomes publicly available very shortly after the polls close.

If the Electoral Commission’s predictions are accurate, this means up to half of all votes may have been counted and made public at this point. And in the previous two elections, the preference distribution of these early votes quite closely matched the final election night result.

Furthermore, as we can expect fewer votes to be cast on election day itself, these can be counted and reported more quickly than usual after the polls close. It may therefore be an early night for those waiting to see a final election night vote tally.

However, this count is not the final election result. Not included in its total are “special votes” – that is, overseas votes and all ballots cast by voters whose name is not on the printed electoral roll for the electorate in which they are voting.

Such special votes are counted once the voter’s eligibility to vote is verified, so they only are included in the official result announced some two weeks after election day.

In 2014, there were some 300,000 valid special votes cast, making up one-in-eight of the total ballots. And the preference distribution of these special votes differed markedly from those counted on election day itself.

Once special votes were included, National lost a seat and the Greens gained one. That did not stop National from governing in 2014, but it might matter in 2017 if the election really is as close as the polls appear to forecast.

Furthermore, special votes could be even more of a factor this election. For the first time, the Electoral Commission is enabling unregistered voters to enrol at advance polling places and then immediately cast a special vote. Although this service only applies to advance voting; voters cannot enrol and vote on election day itself.  

It seems likely that this innovation will result in even more special votes than at previous elections. And again, if those casting special votes differ markedly in their preferences from those whose votes are counted on election day itself, it may result in the final election result differing even further from the election night vote count.

All of which may mean that even when the election is over this Saturday, it might not really be over. In fact, it may well be another two weeks before we can know for sure what has happened.

Comments (9)

by Ross on September 19, 2017

it may well be another two weeks before we can know for sure what has happened


You didn't mention that the main parties and some of the others are likely to be talking to one another to try to form a coalition. That could take longer than two weeks. But election night might give us a fair indication of a likely coalition.

by Alan Johnstone on September 19, 2017
Alan Johnstone

"the preference distribution of these special votes differed markedly from those counted on election day itself."

I presume from the Greens gaining an extra seat, the specials favoured them quite strongly.

Do you have the numbers for provisional vs final results ?


by Andrew Geddis on September 19, 2017
Andrew Geddis


Yes ... I meant "what has happened in terms of who is elected".


Yes ... this explains what happened.

Although special votes usually result in only minor changes to the parties’ final percentage share of the party vote, they can and do have substantial impacts on the final election outcome. In the 2014 election, although the National Party’s final share of the party vote dropped 1.02 percentage points to 47.04%, it resulted in the National Party’s initial allocation of 61 seats – an outright majority and provisionally the first under MMP – being reduced to a final entitlement of 60 seats. Consequently the National Party sought confidence and supply arrangements with its support partners in order to command a majority in Parliament. This one seat loss was a result of the increase in the Green Party’s share which rose 0.68 percentage points to 10.70%, allowing the Green Party’s initial allocation of 13 seats to increase to a final entitlement of 14 seats.  

by Katharine Moody on September 19, 2017
Katharine Moody

Are you saying, Andrew, that those uni students who both enrolled and then early voted on campus this year will be counted as Special Votes?  If so, interesting, as that would mean that these votes won't be counted in the election day/night vote count, right?

Has the EC given any indication whether these new enrolments have grown in number as a result of advance polling places being set up on uni campuses this election period?

by Andrew Geddis on September 19, 2017
Andrew Geddis


Yes - such student voters' votes won't show up on election night, but in the count of specials. Not sure of any statistics on the number of enrolments this time around.

by Katharine Moody on September 20, 2017
Katharine Moody

Andrew, the Electoral Commission released the updated stats on enrolments today;


Am I reading it right, that the 'Difference' column means growth in enrolments for that age demographic?  If so then we have an increase in enrolments of;

45% in the 18-24 yo age group
30% in the 25-29 yo age group
15% in the 30-34 yo age group

Whereas across all age groups, enrolments are only up 10%.

Is it fair to assume that the 'youth quake' might be a real thing?



by Katharine Moody on September 20, 2017
Katharine Moody

Nah, my bad - 'Difference' means estimated number not enrolled. Exact opposite of a 'youth quake'.  C'mon young ones - only two days to go!

Orange Man needs to get visible and make sure they know you can't do it on election day.

by Mina Crooks on September 21, 2017
Mina Crooks

Interesting that there was no advance polling place open on Sunday in Dunedin, when nationally almost 46,000 people voted.  


Post new comment

You must be logged in to post a comment.