Why violent crime is up 7% and down 21% – Derek Cheng

There are obviously multiple ways to measure and define violent crime, allowing politicians to pick the data trend that suits them.

When Prime Minister Christopher Luxon talks about violent crime rising 33% under the previous Labour Government, for example, he’s referring to police victimisation data for “acts intended to cause injury” (since reclassified to “assault”).

When he lauds the Government’s success on its goal to reduce violent crime – by 20,000 victims by 2029, and which has already been surpassed – he is referencing different data.

This is from the NZ Crime and Victims Survey (NZCVS), which is collected from interviews with people about their experiences of crime. The survey estimates the level of all crime, including the estimated 74% that is never reported to police.

In this way, Luxon paints a convincing picture of violent crime skyrocketing under Labour, and then dropping dramatically under his Government.

And it appears to resonate: “crime/law and order” was the only subject in the latest Ipsos Issues Monitor survey, in November, where National was ahead of Labour.

But the annual number of violent crime victims, according to the NZCVS, also dropped in most of the years Labour was in power. This was also a period (from 2018 to 2022) when the prison population was falling.

This doesn’t fit Luxon’s narrative: that Labour let prisoners out of jail and, consequently, violent crime rose.

So what does comparing apples with apples look like?

Using the NZCVS, violent crime fell between 2018 and 2022, when Labour was in office, and again since 2024, under Luxon’s watch.

Violent crime also rose in the last year of Labour’s tenure and for the first year of the present Government’s term, though not as steeply.

Overall, under Labour, there was a 4.5% increase in the number of annual victims in its final year of office, compared with its first year.

The reduction under Luxon’s watch so far is 21%, to August 2025, though it’s an even bigger drop (32%) if we take the peak – in mid-2024, when the annual estimate was 215,000 – as the baseline.

Looking at police victimisation data for assault, the number of victims fluctuates seasonally. For example, there were about 7000 reported assault victims in November 2022, and again in November 2023, but fewer than 6000 in May of each of those years.

Annual figures provide a better picture of the overall trend. To include the latest monthly data (September 2025), the Herald has taken figures from October to September, starting with October 2017 to September 2018.

Violent crime under this measure fell between 2017-18 and 2018-19, under Labour’s watch, and then rose steadily, year on year, until September 2023.

Overall, for Labour’s two terms, the annual number of reported assault victims rose 42%.

Under Luxon, the number dipped by a tiny fraction in his first year, compared with the previous year. It then rose 2% for the year to September 2025, a fraction of the rise under Labour.

One of the complexities with the overall picture is the multitude of definitions for violent crime.

The NZCVS definition is sexual assault, non-sexual assault and robbery. Information is collated from thousands of face-to-face interviews with people about their experiences over the previous 12 months.

Applied nationally for the adult population (15 and over), the survey then provides an estimate of the annual number of victims.

Luxon’s definition for Labour’s time in office – police victimisation data for “assault” – covers assault, serious assault and assault of a prescribed officer. The data is the number of people who have reported being a victim of such a crime to police, regardless of whether any charges followed.

There is probably some crossover with the NZCVS and the police data; however, because the vast majority of crime is unreported, the NZCVS picture is considered much more comprehensive.

For another snapshot, the Justice Ministry has its definition for violent crime.

This captures several offences: murder, attempted murder, manslaughter (excluding driving causing death), assault, sexual assault, child procurement or grooming, child abuse, intimate image offences, sexual offences, human trafficking or slavery, abduction or kidnapping, false imprisonment, robbery, and blackmail or other extortion.

The ministry holds this data across the justice pipeline:

The ministry publishes this data, except for the police victimisations. This is, however, routinely shared with justice sector ministers for their regular meetings.

One of the ministry briefings for such a meeting, released to the Herald under the Official Information Act, reveals police victimisations for violent crime are increasing.

“The number of unique victims for violent offences reported to police in May 2025 increased by 7% compared to May 2024,” the July briefing said.

“The overall decrease may not be evenly distributed across all levels of severity.”

The “overall decrease” here is the dramatic fall in violent crime according to the NZCVS, though it has its own definition for violent crime that includes unreported crime.

The unevenness is the rise in violent crime in the police data bucking the NZCVS trend.

If a more severe experience of crime is generally more likely to be reported to police, then the suggestion in the briefing is that lower-level violent crime is decreasing more than higher-level violent crime.

As any diligent statistician will tell you, none of the data is definitive and it’s best to look for trends.

The trend here is that the number of victims – using police data with the ministry’s definition of violent crime – rose more sharply under the previous Government than under Luxon’s tenure so far.

This is also observed in the data for police victimisations for “assault”: an increase under both governments, but a bigger annual increase on average under Labour.

It’s also observed in the NZCVS data, comparing Labour’s final year and Luxon’s first year, before the rising trend across those two years dramatically reversed.

It’s also apparent in the Justice Ministry’s violent crime data further down the justice pipeline, in court.

The annual numbers for finalised charges, convictions, people charged, and people convicted have all been rising since Luxon became Prime Minister, but less sharply than in the final year of the previous Government.

There was also an across-the-board drop for all these data points from 2020-22. This was during a period when victimisation reports to police for violent crime were rising.

This discrepancy can be explained by Auckland spending much of the latter part of 2021 in Covid lockdown, which froze court proceedings and exacerbated the backlog of Auckland court cases.

This has taken years to work through and may also be a contributing factor to the across-the-board increases in subsequent years – under both Labour and the present Government.

Politicians generally get more cut-through with simple messages: “Crime is down, yay us,” or “Crime was up, boo them.”

But crime is invariably complex, with many moving parts in the justice pipeline and independent of it: health, housing, poverty, education and employment, Oranga Tamariki interventions and intergenerational trauma, especially when it comes to gangs.

This means even if justice policies are putting downward pressure on the level of violent crime, it might still be rising, or vice versa.

The Justice Ministry briefings to ministers provide as complete a picture of violent crime as possible, looking primarily at the justice pipeline.

This includes regional breakdowns of the NZCVS data: while violent crime is falling, it was trending up in the Waikato and Canterbury regions for the year to May, compared with the previous quarter.

Police numbers were 2% higher in May 2025 compared with a year earlier, while court backlogs fell 12% and the number of serious recidivist youth offenders dropped 15%.

The latter alludes to another of the Government’s law-and-order targets that is also well ahead of schedule.

Other parts of the overall puzzle include:

The retail picture has been influenced by the introduction of Auror, which has made reporting a retail crime much easier.

“Increase in overall retail crime reports were in line with the increase in use of the Auror tool,” the briefing said.

Then there’s the important matter of the impact of government policy.

The drop in the number of victims of violent crime under Luxon’s watch started well before any of his key law-and-order policies – Three Strikes, legislation cracking down on gangs, sentencing reforms – took effect.

There has been an increase in police on the beat despite slow progress towards the coalition commitment of 500 more officers.

The drop in the number of serious youth offenders also started to happen before the Government’s legislative move to enable bootcamps (separate to the bootcamp pilot), the removal of section 7AA from Oranga Tamariki legislation, and the expansion of the previous Government’s fast-track programme.

This led child advocates to speculate that the drop in serious youth crime was a reversion to pre-Covid trends.

Justice Minister Paul Goldsmith has claimed the Government’s tough talk on law and order has contributed to the drop in violent crime.

There is, understandably, no data available on the severity or impact of tough talk.

Derek Cheng is a senior journalist who started at the Herald in 2004. He has worked several stints in the press gallery team and is a former deputy political editor.