The 2017 votes - making sense of the numbers

Monday October 09, 2017 Dr Ganesh Nana

Yes, it's probably not core business and/or economic research. But following the true BERL ethos of Making Sense of the Numbers we comment as below.


A simple comparison of votes between the 2014 and 2017 elections as per the chart below can be spun by either National or Labour operatives as positive.


berl graph 1010172


National gained more than 20,500 votes, which after 9 years in government and gunning for a rare 4th term is pretty impressive.


Labour gained more than 351,000 votes, which for a leader of only 7 weeks standing is also pretty impressive.


However, there are buts for both cases.


In Labour's case, the impressiveness pales when considering the starting point - beginning from a miserable 604,000 (just on 25%) votes, any increase was going to look impressive. In addition, some of this came through the reduction in the Green's vote - so really can't be counted as an unambiguous addition to their camp.


Similarly, for the National's case the outcome is more nuanced, as there are reductions in ACT and Conservative votes that need to be factored into the equation.


The 'new' votes


But, firstly, we need to note and allow for the increase in the total number of votes (of about 186,000) between 2014 and 2017.


So, how to proceed? True to our upbringing as economists, we make some assumptions and then see how the story pans out.


For starters, we assume that the 186,000 extra - or 'new' - votes were split between the parties according to the shares for the 2017 Special votes then we get the following table. The justification for this assumption is that many 'new' votes would have been from voters who registered late. Hence these are likely to have voted more in line with 'Specials' than the overall vote.


berl graph 101019


As the table shows, this suggests that the comparable vote gain for Labour is a more muted, but arguably still impressive 272,000. But, we should also subtract the 110,000 comparable loss in Green Party vote.


In contrast, the comparable vote movement for National between 2014 and 2017 is a loss of more than 47,000 votes. This performance is exacerbated by the loss of more than 94,000 Conservative Party and ACT Party votes.


The 'true gains'


Overall, therefore


  • 'true' gain for Labour-Green camp = 162,000
  • 'true' loss for National-ACT-Conservative camp = 141,000


If, we assumed that the 186,000 additional 'new' votes were split according to the overall vote share (instead of by the share of Specials), then the above numbers become larger.


  • 'true' gain for Labour-Green camp = 176,000
  • 'true' loss for National-ACT-Conservative camp = 157,000


Either way, irrespective of assumption, this election resulted a significant loss in votes (compared to 2014) for the National-ACT-Conservative camp; and a significant gain for the Labour-Green camp. Progressing further gets, admittedly, murkier.


Reverting to our original assumption, we note the loss (totalling about 77,000) in NZ First, Māori, Internet-Mana Party, United Future and 'other' party votes. Where did these go? And then there is the new TOP vote to consider. Where did these come from?


More assumptions


No doubt, students of political science will be able for comment from a more informed position. But, here are some primitive assumptions to progress the narrative.


  • the reduction in Māori party votes are gained by Labour
  • the reduction in United Future party votes are gained by National
  • the reduction in New Zealand First votes are gained, in proportionate shares, by National and Labour
  • the reduction in Internet Mana party votes are gained by TOP
  • the remainder of the TOP party vote comes proportionately from National and Labour
  • half of the reduction in 'other' party votes are gained by TOP, with the remainder gained in proportionate shares by National and Labour.


The end result


The full list of changes by Party between 2014 and 2017 under these assumptions is provided here. The consequential narrative is that


  • National's gain of 20,500 votes was made up of
    • gaining 67,600 from 'new' voters
    • gaining 94,300 from ACT and Conservative
    • gaining 3,600 from United Future
    • gaining 17,500 from NZ First
    • gaining 1,700 from 'other' parties
    • losing 12,400 to TOP
    • losing 152,000 to Labour


  • Labour's gain of 352,000 votes was made up of
    • gaining 79,300 from 'new' voters
    • gaining 110,600 from Greens
    • gaining 14,500 from NZ First
    • gaining 1,700 from 'other' parties
    • losing 10,300 to TOP
    • gaining 152,000 from National


Changes to the assumptions, of course, have the potential to change the narrative.


But, making sense of the numbers suggests - after allowing for 'new' voters and controlling for shifts within each of the 'left' and 'right' camps - there was a significant shift in the number of voters from National to Labour.