Guest Post: By-election analysis

A guest post by Kiwi in America on the Christchurch East by-election:

and mainstream media commentators have said that the results in the Christchurch East by-election were bad for National and a negative portent for 2014. This analysis of the Christchurch East by-election compares the voter turnout and individual party's performance compared to the three most recent general electorate by-elections. By-elections (BE) are unusual for a variety of reasons:

  • There is no party vote, only a candidate vote and so candidate personalities assume far greater significance than in General Election (GE) campaigns because in a GE, only the Party vote matters. A poor candidate fielded in a GE can hide to some extent and ride the party's national coattails but is more exposed to negative media scrutiny in a BE.
  • Turnout is usually significantly lower in BEs thus changes in support for a particular party or an increase or decrease in turnout achieved by a party against the trend has a disproportionately greater effect
  • Because it is the only election going on, there is an inordinate amount of media focus. Thus particularly unusual local issues that normally might never see the media light of day in a GE assume greater s possibly skewering the result. This was certainly the case with Christchurch East where the issue of the slowness of insurers and the EQC to rebuild damaged homes after the earthquakes is a dominant but very local issue. Christchurch East is ground zero for residential damage in Christchurch having all of the residential red zone in its boundaries.
  • Similarly national issues of political significance can have a bearing on a BE result. A good example of this was National winning the normally reasonably Labour seat in the Timaru BE in 1985 because the BE was held right in the middle of the maximum pain inflicted by the implementation of Rogernomics.
  • Interest in BEs from the media, the major parties and thence the voters depend on how safe the seat is and how close the contest is likely to be. Of the 36 BEs since 1960, only one (Wellington Central in 1992) could be considered a marginal seat, the rest occurred in safe or relatively safe Labour or National seats.
  • BE results can be skewered by an emerging third party which lacks the resources to make an impact in a GE but can concentrate their efforts in a BE. There has been a long history of third party upsets (or near upsets) in BEs.
  • Whilst they are often a barometer of support for the of the day on the day of the BE, they are rarely a bell weather of the fate of the government of the day in the subsequent GE as many bad BE results for incumbent governments are reversed in the next GE.

 Acknowledging these trends and in comparing it with the previous three BEs (all held since National assumed office in 2008), Christchurch East's result was not exceptional.


At 39%, turnout in the Christchurch East BE is at the low end of the scale dropping from 72% in the 2011 GE. However turnout for the Botany BE in 2011 was even lower at 37% (dropping from 72% in the 2008 GE). The Mana BE with a 54% turnout was exceptionally high for a BE. If you measure the total vote in the Christchurch East BE as a percentage of the total vote in the preceding GE, it does represent the worst result (47% versus 49% for Botany, 61% for Mt Albert and 68% for Mana).

Labour's vote

Labour (and left friendly commentators) have made much of Poto Williams getting 61% of the total vote being a 7% increase from the 54% that Leanne Dalziel got in 2011. However the Christchurch East BE vote for Labour was only 52% of its 2011 GE vote. This was a smaller percentage than the Labour vote achieved in the previous three Bes. In Botany, Labour's BE vote was a hefty 67% of its 2008 GE vote, in Mana in was 61% of its 2008 GE vote and in Mt Albert it was 65% of its 2008 GE vote.

National's vote

In the Christchurch East BE National's share of the vote dropped from 36% to 26% of the vote or 10%. Doocey only managed 34% of the vote that Gilmore got in the 2011 GE. However the Christchurch East result was little different to how National fared in the Mt Albert BE in 2009. This too was a safe Labour seat vacated by a prominent long standing MP. National's percentage of the vote in the Mt Albert BE dropped to a paltry 17% from its 28% share in the 2008 GE – at 11% a bigger drop than it suffered in Christchurch East. Similarly if you look at the BE vote for National's Melissa Lee, it was only 36% of the vote compared to what Ravi Musuku achieved against the then sitting Minister Helen Clark in the 2008 GE. Given that Botany was a safe National seat, the comparison with Christchurch East is more difficult to make as the seat profiles are entirely different.

What happened in Mana

Whilst Mana has historically been a safe seat for Labour, Kris Fafoi only managed to eke out a paltry 1,400 vote majority for Labour in the November 2010 BE. Hekia Parata (already a list MP) managed to keep National's vote at a staggering 80% of its vote in the 2008 GE and increased National's share of the vote by 6% from 35% to 41% – something quite rare in BE history. Turnout was also very high at 54% of registered voters and 67% of the number who voted in the 2008 GE!


The result in the 2013 Christchurch East BE was very similar to what happened in the 2009 Mt Albert BE. A popular long standing sitting MP vacates a very safe Labour seat. National chose candidates who fought what could only be classified as very average campaigns in each BE. Matthew Doocey is not a rising star in the mould of say Simon Bridges and Melissa Lee, whilst already a List MP, made a number of pratfalls in the Mt Albert campaign. Given the historical trend of the Labour vote in both of these safe Labour seats, National's head office were not going to devote much in the way of funds or personnel to fighting a hard campaign in seats they were never going to win. Thus the respective GOTV operations of the Doocey and Lee campaigns were largely down to the candidate's capacity to motivate local National Party members. The importance of the local team the candidate puts together cannot be underestimated. Having fought in a few BEs, the quality and size of the team can make a huge impact on the eventual vote as turnout in BEs is crucial. Hekia Parata's outstanding result could be put down to her putting together an aggressive and well organized team. Plus, whatever people might think of her skills as Minister of Education, she is a much more formidable campaigner than Doocey or Lee.

On the Labour side, Cunliffe had much to lose if the Christchurch East result was a repeat of the Mana result. Whereas in Mt Albert, Clark's tenure there attracted some good electorate workers AND the candidate (David Shearer) was a handpicked high riser with an intriguing background with the UN, the Labour team in Mana were poorly organized and it would be fair to say that had taken the Labour vote for granted in what had been a long standing safe Labour seat. Fafoi thought his media experience and name recognition would be enough but he was an entirely inexperienced candidate so you add a poorly staffed local electorate who had never really had to work hard to elect their MP with a novice campaigner against a crack local team behind an excellent campaigner and you can see how the Mana BE result happened.

Labour's team needed to lift its game from even the Mt Albert BE because of the need to avoid an embarrassment to its new leader and because National had won the party vote (narrowly) in Christchurch East in 2011. Thus it was that much time and effort was put into Labour's Christchurch East BE campaign. This began with deliberately siting the 2013 Labour Annual Conference in Christchurch and sending the 500 odd delegates out to canvas for a whole afternoon. As an organizer of canvassing efforts, I know that this would be enough people and time to be able to identify all Labour leaning voters in the electorate (which had 5,000 fewer enrolled voters than in the 2011 GE which in turn was down almost 10,000 enrolled voters on the 2008 GE due to earthquake related out-migration). Enter Jim Anderton, a veteran operative of the left who many may have forgotten was the Labour Party President from 1979 until he took over the Sydenham seat from the retiring and lackluster John Kirk in 1984. Anderton made sure that Labour had an excellent election day GOTV and that was off the back of now accurate canvassing records – something that was difficult to do given the upheaval and movement of people in the electorate since the earthquakes.

In many respects given how much Labour threw at the BE in Christchurch East up against a novice National candidate with zero name recognition in an electorate probably the most hostile to the government than any in the country given the anger and frustration of voters over the insurance related rebuilding delays and infrastructure chaos, this is actually not that bad a result for National. It's true that Shearer was a ‘sexier' candidate in Mt Albert than Poto Williams but really in so many ways the Christchurch East BE result was very close to the Mt Albert BE result.  Labour and its cheerleaders have already read too much into the result. National do have cause to be concerned in that its strong party vote showing in Christchurch in 2011 was an integral part of its nationwide winning margin.

In by-elections, getting out your vote is everything. National's turnout in the 2011 GE was lower than all the polling showed partly because of the teapot tape kerfuffle but also because its soft and low information voters heard the ‘govern alone' mantra and figured John Key was all good and didn't need their vote. If National invests in a thorough and extensive nationwide Party vote GOTV effort, they will be able to weather almost any campaign storm that inevitably the left leaning media will throw at them, especially if the economy continues to gather steam and the Christchurch rebuild sees more residential homes targeted.

 Below the break

 be2 be1


  1. Only General Electorates were compared so the results of the Maori electorate BEs (Ikaroa Rawhiti on 29 June 2013 and the Te Tau Tokerau on 25 June 2011) were not included in this analysis. The reason is that Maori electorates have a number of features that make them quite different from General Electorates not the least of which is that only Maori can be registered to vote. These differences include vastly larger geographical territories, the existence of parties specifically targeted at Maori that do not run candidates in General Electorates, the lower levels of voter registration and the generally lower levels of voter turnout. These factors combine to make comparison with General Electorates less meaningful.
  2. The comparison is with all General electorate BEs held in the 48th Parliament (2008 to 2011). Given that there were three General Electorate BEs in that period which is an above average number, it enables valid comparisons to be made.
  3. The BE results are only compared with the previous GE especially with respect to the change in the vote for candidates of the respective parties.
  4. Only parties that stood candidates in both the preceding GE, the BE and the subsequent GE (for BEs held in the 48th Parliamentary term) are included in the comparison. The percentage of the vote indicated by each party in each electorate by electorate comparison is of the total electorate votes cast in that particular election. The minor party results are just not analyzed but are part of the total vote.
  5. The comparison is with the ELECTORATE vote in the GE analyzed and not the PARTY vote. The publishes the total votes for each electorate in GEs in two separate categories (Party then Electorate) whereas in a BE, there is no Party vote to count and so only electorate votes count.
  6. The turnout percentages for GEs are calculated from the total number of Electorate votes cast in the GE (please note that in almost all electorates the total number of Party votes cast can differ from the total number of Electorate votes cast – this is because some voters only exercise one of their entitled votes in the GE). This means that the by-election comparison is an apples to apples comparison. Please also note that for ease of comparison of respective party performance, only the party names are recorded whereas GE Electorate and a BE ballots lists the candidates' names more prominently than the party affiliation. Likewise the reporting of results at for the Electorate or BE vote, they also features the candidate names prominently.

In order to accurately determine the voter turnout percentage to be calculated, for all the GE results, the total number of enrolled voters for each electorate came from the Electoral Commission's published number of enrolled voters eligible to vote in that electorate at that election. For the BEs, the total number of enrolled voters for the BE was taken from media kits released for the specific BE by the Electoral Commission. The exception was the Mt Albert BE where no BE total enrolment figure was released in the media kit. In order to approximate that figure, the increase in enrolled voters in Mt Albert between the 2008 and 2011 GEs was halved and then added to the 2008 enrolment total on the premise that the voting population of the electorate was rising slightly over time and that the size of the roll at the time of the BE was somewhat between the roll of the GEs on either side. Given that the roll increased by only 772 voters, this approximation does not mean the voter turnout percentage for the Mt Albert BE  is very far off being 100% accurate.

Thanks to Kiwi in America for the analysis.

Comments (14)

Login to comment or vote

Add a Comment