Is there racial bias with Police charging decisions?
Radio NZ reported in July:
On 2 April, then-police commissioner Mike Bush told the Pandemic Response Select Committee that while historically police had not used discretion fairly across all communities, this had changed.
“We’ve moved a very long way and we now have data to say that we now are in a place where we apply that discretion evenly across communities.”
According to information supplied to RNZ under the Official Information Act, that statement was based on data collected from police’s strategy to reduce Māori offending, Te Huringa o te Tai, which showed the number of pre-charge warnings for first-time adult offenders was the same for Māori and non-Māori, at 73 percent.
The discovery of “an analytical error” has however corrected the number of pre-charge warnings given to Māori as just 53 percent, compared to 59 percent for non-Māori.
Police Deputy Commissioner Wally Haumaha acknowledged there was bias in the police – but said he was confident it was changing.
This got me interested as a long time analyst of crime statistics. The first thing I know is that overall rates often mean little as they treat all crimes the same – a homicide and a drink driving.
Likewise the difference in pre-charge warnings might be related to the nature of the crimes, rather than racial bias.
This is not to say racial bias might not be a component. It is quite possible there is some. But the overall topline data is not a good evidence base on which to decide. So I asked the Police to break down the pre-charge warnings by offence type and ethnicity. Their OIA response is here.
| Māori | Non-Māori | Difference | |
| Acts intended to injure | 32% | 49% | 17% |
| Abduction/Harassment | 17% | 23% | 6% |
| Theft | 33% | 29% | -4% |
| Drugs | 44% | 43% | -1% |
| Weapons | 54% | 47% | -7% |
| Property | 5% | 15% | 10% |
| Public order | 89% | 90% | 1% |
| Offences against justice | 57% | 53% | -4% |
| Overall | 54% | 57% | 3% |
The biggest difference is in violent offences where 32% of Maori get a warning and 49% of non-Maori. That does suggest racial bias, unless the makeup of offences within that category is very different for each ethnicity. You don’t caution someone for GBH but you might for a pub fight. If I have time I’d be keen to break down that category more.
Fewer Maori get warnings for harassment offences and property offence.
But in some categories it goes the other way. Non-Maori are less likely to get a warning for theft offences, for weapon offences and for offences against justice.
Worth noting that the overall difference in warning rates of 3% in no way can explain the huge gap between Maori and non-Maori offending. If the Police are slightly less likely to give a first time offender a warning if they are Maori, this won’t explain why overall offending is around 400% higher.
But nevertheless any racial disparity in who gets a warning is of concern. Even this data is somewhat inconclusive. Why are non-Maori less likely to get a warning for theft offences and Maori less likely to get a warning for violent offences?
To really answer that you’d want to get access to say a 10 year dataset and examine it down to the individual offence level.
My thanks to the Police for the OIA response.