A 2022 long-term insights briefing has been pointed to as evidence of the violent nature of our prison population. But it’s a flawed document, says Roger Brooking.

Only 3.2% of The Spinoff’s readership supports us financially. We need to grow that to 4% this year to keep creating the work you love. Sign up to be a member today.

A couple of months ago, a minor furore was sparked by Green MP Tamatha Paul saying “the vast majority of people who are in prisons are there for non-violent offences” in a video posted on social media. The comments were picked up by the media and widely criticised, with the police minister Mark Mitchell calling them “total nonsense” and “an insult”. 

The Ministry of Justice’s “long-term insights briefing into imprisonment in New Zealand”, a document submitted to parliament in 2022, was relied on to back up the idea that Paul was wildly wrong. The briefing made statements like “most people in prison have committed serious offences”, “most people serving sentences in prison have been convicted of serious sexual or violent offending”, and “more than 54% of the New Zealand prison population are either serving sentences or facing charges for sexual or violent offending”.

But there are some problems with this document. It relies on one-day “snapshot” data, which looks at who is in prison at a particular point in time, in this case as of December 31, 2021. At this point, the total prison population was 7,702. As of March 31, 2025, it had risen to 10,680. But being out of date isn’t the only problem with the report. For something that is supposed to provide long-term insights, surely it’s better to take a long-term view and to analyse prisoner data over a 12-month period, instead of a single day. In any given year, around 20,000 New Zealanders spend time in prison, and about 15,000 are released. The numbers are this high because the vast majority of those sentenced to prison are given short sentences.

The reason these offenders receive short sentences is because their offending is not serious enough to warrant a long-term sentence. It includes offences such as drug dealing, shoplifting, theft, driving while disqualified, breaching court conditions – none of which necessarily involves interpersonal violence. Offenders given a sentence of two years or less are automatically released halfway through the sentence. That’s why there’s a high turnover.

Statistics NZ has the definitive data on 12-month criminal profiles. It shows that in the fiscal year 2023-2024, 1.7% of the sentenced prisoner population were there for homicide, 22% for acts intended to cause injury, 6.7% for sexual assault, 4.5% for abduction, harassment, and 4% for robbery. That comes to 39.1% for what are considered violent offences. So Tamatha Paul was absolutely right – the majority (around 60%) of those who are sentenced to prison every year are there for non-violent crimes. Does this represent the “vast” majority, as Paul claimed? Well, 60% is substantially bigger than 39%, so she was more right than wrong. Does it include prisoners on remand? No it doesn’t – because they’re not sentenced and may not end up in prison. So let’s leave them out of it and focus on the percentage of violent criminals actually sentenced to prison – 39%.

Mark Twain had it sorted. He said “there are lies, damned lies, and then there are statistics”. This pithy observation points to the potential for statistical data to be misinterpreted to support a weak argument. And it’s pretty weak to claim that 39% represents a majority, or that a one-day snapshot presents an accurate picture when presenting a long-term analysis.

Here’s another insight that Mitchell and the media missed. The briefing has a whole chapter addressing the question “What works to keep people out of prison?” The ministry’s conclusion was that “prison-based rehabilitation programmes have consistently delivered better results than community-based programmes in New Zealand”.

This is not true either. Annual reports published by Corrections show that almost all its rehabilitation programmes reduce reoffending by less than 5% in the 12 months after release from prison. The best-performing programme that Corrections has ever run in prison reduces reoffending by between 8% and 13% in the year after release. However, in 2019, the Ministry of Justice published an 83-page evaluation of the drug court in Auckland, the AODTC, which it said reduced reoffending of graduates by 86% in the first year of release. It operates in the community – and that makes it nearly 20 times more effective than prison-based programmes. 

This remarkable result is probably the most important insight that the 124-page long-term insights report should have documented. But it isn’t even mentioned. And yet, this briefing was presented to parliament, and has since been quoted by politicians and the media to beat up on Tamatha Paul. It is seriously short on insight and relies on one-day snapshots of the prison population to replace long-term analysis. Its accuracy cannot be relied on.