During the election campaign RNZ is tracking polls, Facebook ads, campaign financing, data from the debates and advance voting numbers. The latest figures are published right here on this page.
The party vote decides the allocation of seats in Parliament. If a party doesn't reach the 5 percent threshold to get an allocation of seats, but does win an electorate seat, it will make it into Parliament.
The seat projection below reflects the current averaged party vote based on polls, as well an assumption Te Pāti Māori wins two electorate seats.
The amount of money gained from donations and loans is submitted each May. The latest available full data covers 2022. Data from 2023 only includes donations of more than $20,000, these donations must be reported to the Electoral Commission within 10 days of receipt.
Each election year the Electoral Commission allocates money to parties to be used for the broadcasting of election programmes and election advertising.
A total of $4,145,750 has been allocated for the 2023 election.
Television and radio advertising can only occur during the official campaign period which starts 10 September and ends at midnight 13 October. Parties can only use the funds allocated by the Electoral Commission for advertising on these mediums, not other party funds.
Advertising on the internet is treated differently. Parties can use the Electoral Commission funding to pay for internet advertising during the official campaign period but can also advertise outside of this period if they pay for it themselves. They can also spend money other than the broadcasting allocation for online advertising during the campaigning period as long as they don't breach the election advertising spending limit. For 2023, this is $1,388,000 per party.
There is no spending limit on online advertising which occurred before the regulated period of 14 July.
Groups or individuals who aren’t a political party or a candidate can run election advertising. If they expect to spend more than $15,700 they must register with the Electoral Commission as a ‘registered promoter’. All advertisements must include a statement about who is responsible for it.
If a registered promoter encourages people to vote for a particular party or candidate, permission must be sought from the party or candidate. This spending counts toward a political party or candidate’s spending limits as well as the registered promoter’s spending limit.
‘Attack advertisements’ - running down a party or candidate - also count as election spending by a registered promotor.
There is a cap on spending during the election regulated period of $367,000 for registered promoters. If they spend more than $100,000 a promoter must submit a report of expenses. These reports are published on the Electoral Commission’s website.
RNZ has asked all registered promoters what they plan to promote this election.
Facebook's ad library opens a window into how political parties are spending money to get their ads featured in Facebook and Instagram feeds.
The following charts track advertising via the Facebook pages of political parties. They don't track candidate pages and do not include posts which aren't ads. The exact amount spent on each advertisement isn't shared in the ad library, instead an upper and lower spending estimate is given.
Political parties can pay to have their ads on Facebook targeted to different demographic groups, including by age and gender.
The sentiment of the text in ads can also be calculated. RNZ's analysis shows the proportion of ads for each party that had text containing more positive or negative words or terms.
Hot topics for New Zealanders
The Ipsos New Zealand Issues Monitor tracks what people think are the three most important issues facing the country today. The below chart tracks issues which have featured in the top five since October 2021.
Based on the text of ads on Facebook, this chart shows the topics parties have been spending money on.
How we analysed Facebook advertisements:
Data on Facebook advertising is publicly available through Facebook's transparency ad library.
Topics were determined based on keywords. If an advertisement had keyword matches across multiple topics, the topic with the highest number of matches was selected. Where there was an even number of matches for multiple topics, the advertisement's topic was set to 'unknown'. Ads identified as 'generic' were ads which promoted reading policy or voting, but where no words related to topics were identified. Ads where there was an image or video and no text were manually assigned to a topic. Keywords assigned to topics are updated as the campaign continues.
Sentiment analysis was based on ad text and used the Lexicoder Sentiment Dictionary 2015. Advertisements with no text were classified as neutral. Additional terms were added to the negative and positive dictionary based on what appeared in advertisments to improve sentiment identification.