The Government says violent crime is down. Is it right?

Monday, 24 November 2025

There were 38,000 fewer victims of violent crime in the year to August than when the Government was elected, Justice Minister Paul Goldsmith said on Friday.

Every day Kiwis come to Stuff to see what’s happening in their world. We’re committed to telling you quickly and accurately. But we’re also committed to providing nuance and context, to explain how and why things happen. Our explainer articles are all about getting to the heart of complex events and giving you reliable information. If you’ve a subject you want us to explain or fact-check, fill in the form at the bottom of this story.

On Friday morning, the Government released its latest figures on violent crime.

“The latest New Zealand Crime and Victims Survey shows there were 147,000 victims of serious violent crime in the 12 months to August,” Justice Minister Paul Goldsmith said. “That’s 38,000 fewer victims than when we came into government, and 9000 fewer than our last update.”

Based on these statistics, the Government has surpassed its target of 20,000 fewer victims of violent crime by 2029. That required an annual total of 165,000.

But according to Victoria University criminologist Trevor Bradley, the claims should be taken with a pinch of salt.

“Political claims that crime is going down are usually made by the Government, and then claims that crime is going up are usually made by those in opposition,” he said.

“Statistics can be made to say lots of different things and different political parties can have a very different interpretation of the very same statistics.”

With that in mind, there are a few things to unpack here - what is considered serious violent crime? Where does this data come from? Is it reliable?

Let’s get into it.

Figures are based on a Ministry of Justice survey

Whenever the Government publishes violent crime data, it is always based on the New Zealand Crime and Victims Survey (NZCVS). Conducted annually by the Ministry of Justice, this survey defines violent crimes as sexual assault, non-sexual physical assault, and robbery.

The Government says violent crime is down, but are the figures reliable?

According to Bradley, the NZCVS is the most comprehensive measure available for household and personal crime.

“It gives us really useful insights into the volume of crime, particularly compared to the police statistics. But at the same time, it doesn't give us a complete picture,” he said.

The survey works by interviewing 7000-8000 adults every year and measuring the number of victims and incidents of crime for the 12 months to October.

According to the Ministry of Justice’s general manager of sector insights, Rebecca Parrish, it is an important source of information for measuring victimisation because it measures both reported and unreported crime.

You may have noticed the survey measures the number of victims and incidents of crime for the 12 months to October. Yet the Government’s announcement covers incidents of violent crime in the year to August.

That’s because the Government introduced quarterly reporting for the nine targets it announced last year - one of which was the reduction in serious violent crime. To accommodate this, the ministry has started collating data on a quarterly basis. While useful, these results are less reliable than the annual estimates, Parrish acknowledged.

“In short … the quarterly figures are best seen as indicative. This is because the annual estimates fully match the survey design,” she said.

That doesn’t render the quarterly data useless, but it is a bit less reliable.

There are a few other limitations with the NZCVS as well. Because the survey uses a household-based sampling method, it does not include anyone in aged care facilities, prisons, Oranga Tamariki residences, or hospitals.

Claims made by all politicians that crime is up or down should be taken with a pinch of salt, Victoria University criminologist Trevor Bradley said.

It also relies on victims’ honesty.

The people involved in conducting the survey are skilled, Bradley said. They don’t simply ask participants if they have been the victim of a crime.

“Instead they’ll talk to them about scenarios, you know, ‘Have you lost anything or had anything stolen?’ They're not leading them to a yes or no type of response. But once the victims share their stories, the researchers then categorise that into different offence types.”

Still, some participants would struggle to be forthcoming.

“There are certain types of offences which victims find really, really difficult to speak about - particularly violent offences,” Bradley said. “Sexual offending is one of those. Domestic violence is another.”

So it has a few limitations… but is it the only source of violent crime data?

The NZCVS is one source, another is the Reported Crime and Victimisation Series produced by NZ Police.

As the name suggests, this series records crime that is reported to police. The data includes details like a victim’s age, gender, ethnicity and relationship to the offender. Victims can be people or organisations.

“Neither the NZCVS nor the RCVS is more reliable than the other, they are simply different data sources used for different purposes,” Parrish said.

“NZCVS is best for tracking long-term victimisation trends, while RCVS is best for guiding day-to-day crime response decisions.”

According to the RCVS data, there were 37,538 unique victims of violent crime (assault, sexual assault and robbery) in the year to August 2025.

That was up from 37,080 in the year to August 2024, and 36,380 in the year to August 2023 - a couple of months before the current Government took office.

There are two points of interest here: One is that the number is significantly lower than those reported in the NZCVS. The other is that the numbers are trending up, not down.

It’s important to note that Police RCVS data only records crime when it was reported - so it excludes violent crime that police never hear about. That’s part of what the NZCVS tries to capture.

“We know that the crime statistics collected and reported by police only represent about 25% of all crime reported to the survey,” Bradley said. “The other 75% is what we call the dark figure of crime. It’s the volume of crime that's never reported to police and therefore never recorded by them.”

This explains the vast difference between NZCVS data and RCVS data. But which is more reliable?

Well, that’s not a question that can really be answered.

“There is no way to measure crime in such a way that we do get the complete picture, because there will always be a dark figure of crime,” Bradley said.

“So to make claims that crime overall is reducing, and violent crime in particular is reducing, can be quite spurious.”

In a statement to Stuff, Goldsmith noted the Government has been consistent in its use of NZCVS for crime data, which provides insights into victimisation trends over time.

“I would note that there were very few objections to this approach when the data was showing an increase in the numbers of victims last year,” he said.

Bradley acknowledged that this is helpful - to an extent.

“If we compare this six-month period to the last, there seems to be a reduction in the violent crime that's reported. But that doesn't mean that's reflected across our entire society, because there are dark spots.”