What Hager's Dirty Work claims about the Taxpayers' Union and David Farrar
Nicky Hager has released Dirty Work: The Taxpayers' Union and New Zealand's Hard Right, launching the book at Unity Books on 22 September 2026.
The launch was reported by RNZ. The publisher is Potton & Burton, according to The Post, and the launch venue was Unity Books, according to The Press. The book follows a nine-year gap since Hager's previous book, according to the NZ Herald.
The early debate has focused on David Farrar. Farrar runs market research firm Curia, which produces the Taxpayers' Union-Curia political poll and the National Party's internal polling. Farrar and Jordan Williams are co-founders of the Taxpayers' Union. Farrar retired from the Taxpayers' Union board in 2023. Hobson's Pledge was set up by former National leader Don Brash. Those details were set out in RNZ reporting on the book.
Hager alleged Farrar used privileged access to a National Party election database to target mailouts for the anti-capital gains tax campaign before the 2020 election. The allegation concerns the handling of party data and its use in an outside campaign.
Farrar told RNZ he did not access the National Party database. He called the allegation "totally, 1000 percent wrong". Farrar also told RNZ he had not read Dirty Work but agreed to respond to specific allegations put to him.
A second allegation concerned the Taxpayers' Union's Stop Three Waters campaign. Hager alleged Farrar pushed the campaign to lean more heavily on Māori issues. Farrar denied that account.
For readers who follow polling, the working arrangement matters more than the launch-day exchange. Curia works in two roles at once. It is a public pollster through the Taxpayers' Union-Curia poll. It is also the National Party's internal pollster. That dual role is common in Wellington, where a small pool of pollsters, strategists and advocates move between party, commercial and third-party work. It also means any claim about data flows or message coordination gets close attention.
The broader context here is how these claims will be tested. Hager's method relies on leaked material and on-the-record reconstruction. Farrar's response relies on direct denial before reading the full text. Neither settles the question of what records exist for the mailouts, who authorised them, and what lists were used. In my view, those are the points select committee staff, party officials and campaign lawyers will look for next, rather than the competing characterisations on launch day.
A narrower question for people who use polls is about trust. Public polling and internal polling serve different functions. One informs voters. The other informs party strategy. When the same firm does both, disclosure of clients and separation of data become central to trust. The denials on the database allegation leave that issue live, not closed.
