We need to talk about police dogs

Police dogs biting people cause more moderate and serious injuries than all other uses of force by police officers put together. It's time we had a national conversation about if the way police dogs are currently used does more harm than it's worth.

Since 1956,1 New Zealand Police have used working dogs as part of their operations. These working dogs are used for a variety of purposes, including search and rescue as well as tracking offenders.

For most of us, we only see police dogs portrayed in a wholly positive light. Sometimes that's through media articles about puppies beginning their training or graduating from the Police Dog Training Centre. You may also have seen some of NZ Police's social media accounts posting #FridayFloof photos of police dog puppies or promoting the Dog Squad: Puppy School TV show that features NZ Police's dog section.

Police dogs are heroes, we're told, and cute ones too!

Sometimes, our attention is drawn to the more dangerous side of police using working dogs. Like articles about the stab-proof vests for police dogs introduced in 2016, or the news in December 2020 that a man had shot a police dog in the head.

In that incident, the police dog's handler then shot the man three times. Both the dog and the man survived being shot, but many people took to social media to express their belief that the man shouldn't have been given medical attention, or that they hoped he would die.2345678910111213

With such a wholesome image being presented to the public, perhaps it's not surprising that many of us feel a sort of fierce loyalty towards New Zealand's police dogs. But there is a side to the use of dogs by NZ Police that is not often shown to us. One that's all too easy to look away from.

One of the ways police dogs are used in New Zealand is to attack people.

When using this tactical option, police dog handlers send their dog after someone to bite them, typically in order to immobilise and incapacitate that person so they can be arrested. This might sound quite reasonable on the face of it, but the truth is these attacks can have severe consequences. To my knowledge there has never been a meaningful review of this tactic, accounting for the harm it causes.

For any tactical option made available to police, it's important to consider whether it might do more harm than good. These conversations might happen readily for new initiatives, such as the Armed Response Teams which NZ Police discontinued in June 2020 following a trial, but it's a much harder conversation to start when it comes to established tactical options. NZ Police does not typically consider removing access to a tactical option once it has been implemented, let alone when it has been in place for decades.

Each year, NZ Police releases a summary report on their use of force, known as a Tactical Options Research Report. These summary reports are based on Tactical Options Reporting (TOR) data recorded by police officers. Each time a police officer uses force against a person, they fill out a TOR form recording information such as what tactical options they used, the context in which it was used, who it was used against, and what effects it had.

I was surprised, and impressed, to read on page 21 of NZ Police's 2019 use of force summary report that the author had considered that the availability of batons might cause more harm than good:

Baton use has been consistently low over the last five years and dropped even further in 2019—used at only 37 TOR events, compared to 54 in 2018.

Of interest, 38 staff members reported injuries resulting from baton training during 2019. Given it's [sic] low use, the required training time and associated injury risk, and the requirement that all staff carry a baton on their duty belt, it may be time for NZ Police to re-evaluate the costs and benefits of this tactical option to ensure that staff are equipped and enabled with the most appropriate equipment.

Tactical Options 2019 Annual Report | NZ Police

I hope this attitude could mean that NZ Police would be willing to examine the harms done by some of their current practices, and consider whether or not they are worth continuing.

These TOR reports, and the data on which they are based, give us insight into some of the harm inflicted by NZ Police's use of attack dogs.

In 2017, using TOR data released by NZ Police under the Official Information Act, I worked with Kirsty Johnston at the NZ Herald to produce an article that led with a story about a 12 year old girl who was bitten by a police dog in 2016. It was seeing this incident and the injury rate of police dogs — both revealed in this use of force data — that led me to question whether their use might do more harm than good.

Following the release of TOR data covering 2017, NZ Police began to refuse my requests for more of their use of force data. It took over two years of trying, including a comprehensive complaint to the Ombudsman that took almost nine months to resolve, before NZ Police finally agreed to release their use of force data covering 2018.

The TOR data allows us to compare how often police use various tactical options. There are some specific rules about when uses of force are recorded. For example, the "empty hand" tactical option is not recorded if a police officer pushes someone but they don't fall to the ground, and the use of handcuffs or other restraints is only meant to be recorded when used either with "pain compliance" or alongside another reportable tactical option.

Looking specifically at how often police recorded having used each tactical option in 2018, excluding cases where they drew tasers or firearms but did not use them, the use of attack dogs was relatively rare:

Tactical options used by NZ Police
Jan 2018 – Dec 2018
  • 0
  • 500
  • 1,000
  • 1,500
  • 2,000
    • Tor Events Empty Hand: 1,748
    Empty Hand
    • Tor Events Handcuffs-Restraints: 1,571
    • Tor Events OC Spray: 1,260
    OC Spray
    • Tor Events Dog: 297
    • Tor Events Taser: 209
    • Tor Events Baton: 66
    • Tor Events Other: 19
    • Tor Events Firearm: 10

Out of the 4,324 incidents at which police recorded using force in 2018, attack dogs were used at 297 of them — 6.9%.14

In just a little over the three years, from 1 January 2016 to 15 March 2019, police dogs were used 5,284 times to apprehend people.15

NZ Police's annual use of force summary reports also reveal the number of incidents at which the use of attack dogs were recorded in 2016 (323)16 and 2017 (290).17 During those three years, police dog handlers only directed their dogs to attack people at approximately one in five incidents where they used a police dog to apprehend someone.

This reveals some of the benefits to NZ Police when it comes to their use of dogs. When police dogs are used to apprehend someone, most of the time they aren't directed to attack them, and the use of attack dogs makes up a relatively small portion of the force used by police each year.

But we haven't considered the harms yet.

In 2018, police recorded 903 "subject injuries" inflicted at incidents where they used force. About half of these injuries were classified as minor, either requiring no treatment or being able to be treated by the injured person. But 358 injuries were moderate — requiring medical treatment — and 89 required hospitalisation.18

Despite only being used at 6.9%14 of incidents where police used force in 2018, attack dogs caused more than half the moderate and serious injuries inflicted by police that year.20

Moderate and serious injuries caused by NZ Police use of force
Jan 2018 – Dec 2018
  • 0
  • 50
  • 100
  • 150
  • 200
  • 250
    • Moderate Empty Hand: 81
    • Serious Empty Hand: 32
    Empty Hand
    • Moderate Handcuffs-Restraints: 9
    • Serious Handcuffs-Restraints: 4
    • Moderate OC Spray: 10
    • Serious OC Spray: 1
    OC Spray
    • Moderate Dog: 218
    • Serious Dog: 24
    • Moderate Taser: 13
    • Serious Taser: 9
    • Moderate Baton: 4
    • Serious Baton: 2
    • Moderate Other: 2
    • Serious Other: 2
    • Moderate Firearm: 2
    • Serious Firearm: 4

On page 25 of NZ Police's use of force summary report for 2019, under the heading "Taking Every Opportunity to Prevent Harm", NZ Police published a table of injury rates for each tactical option available to their officers. One column is titled "Tactic Uses per 1 Injury (on average)", and the injury rate for attack dogs stands out: 1 to 1.

In that report, NZ Police downplayed the significance of the extremely high injury rate of attack dogs. Rather than treating it as a matter of concern, they dismissed it as an artefact of how the data is recorded:

Table 5 also shows the injury rate for each tactical option. Dog deployment had the highest injury rate, with an average of one injury resulting from every use.

However, dog deployment is only reported as a tactical option if the dog bites or injures someone (dogs are often used for tracking, which is not a use of force). Put another way—on average—for every dog bite (or injury), subjects sustained one injury.

Tactical Options 2019 Annual Report (PDF link) | NZ Police

This is not true. The Police Manual's chapter on "Use of force and Police dogs" explains more clearly when a police officer needs to report a TOR event because of how they used a police dog:

These tables detail the different dog bite reports and when that type of report is required.

Report Description
Tactical Options Report (TOR) Where in the execution of a duty a constable uses force on any person with a Police dog, regardless of whether or not that person has been arrested.
Use of force and Police dogs (PDF Link) | NZ Police

Though it is rare, with only 621 such incidents recorded in 2018, the use of police attack dogs is also recorded even if they don't bite or injure a person.

Rather than a statistical anomaly, this "1 to 1" injury rate is the probability that, when a police dog handler directs their dog to attack someone, the dog injures that person. Of course, "1 to 1" is an oversimplification. It's used in that summary report because all other uses of force have injury rates significantly below 50%.

The exact injury rate for attack dogs is not published explicitly in NZ Police's summary reports, but it can be calculated from the TOR data they have released under the OIA.

In 2018, the injury rate for police attack dogs was 86%.22 This is comparable to the injury rates in 2016 (89%),23 2017 (88%),24 and 2019 (86%).25

When a police dog handler directs their dog to attack someone, they know with near certainty that it will injure that person.

The tactical options with the next highest injury rates in 2018 were "firearm" (ignoring events when firearms were drawn but not fired) and "other". These were both used far less than any other tactical option (10 and 19 times, respectively). After them, the "empty hand" tactical option had the highest injury rate, at 24%.26

The majority of injuries inflicted by the "empty hand" tactic — 73%27 — were classified as minor, whereas only 5.5%22 of injuries inflicted by attack dogs were minor.

When speaking to RNZ in May 2020 about the high proportion of mental health or emotional distress in incidents where people were hospitalised by attack dog injuries, NZ Police defended their use of this extremely dangerous tactical option:

our dog handlers are probably dealing with the most violent, dangerous and drug-fuelled members of our society on a day-to-day basis and at times we're going to have to use force and we have to use our police dogs - that's just the reality of the situation.

Mental health calls make up quarter of serious police dog bites | NZ Police dogs national co-ordinator Inspector Todd Southall

NZ Police recorded drugs as being a relevant factor in 42%29 of incidents where officers set police attack dogs on people in 2018. Whether or not it's reasonable to describe those people as "drug-fuelled", drugs of some kind were somehow relevant at almost half these incidents. But are these people violent and dangerous?

81%30 of the time police set attack dogs on people in 2018, those people were unarmed. They were no more or less likely to be armed than anyone else police used force against in 2018 — 81%30 of all people subjected to any kind of police force in 2018 were unarmed.

To categorise people's behaviour when recording use of force, police officers use a framework called Perceived Cumulative Assessment (PCA). The PCA framework has five levels, which are described on page 15 of the "Use of force" chapter of the Police Manual, released under the Official Information Act in 2017:

Cooperative Willingly responds when approached.
Passive resistant Refuses verbally or with physical inactivity.
Active resistant Pushes away, pulls away, or runs away.
Assaultive Intent to cause harm, expressed verbally, through body language/physical action.
GBH or death Shows action intended to or likely to cause grievous bodily harm or death to any person.
Willingly responds when approached.
Passive resistant
Refuses verbally or with physical inactivity.
Active resistant
Pushes away, pulls away, or runs away.
Intent to cause harm, expressed verbally, through body language/physical action.
GBH or death
Shows action intended to or likely to cause grievous bodily harm or death to any person.
Use of force (PDF link) | NZ Police

In 76%32 of incidents when police officers set attack dogs on people in 2018, those people were below the "assaultive" threshold. That is, they were not expressing any intent to cause harm, either verbally or through their body language or actions.

Almost every time that happened — 73%32 of all attack dogs uses in 2018 — NZ Police recorded those people as being "active resistant". They were pushing away, pulling away, or running away.

The TOR data doesn't clarify how many of these people classified as "active resistant" were running away, as opposed to pulling or pushing away. However, it seems clear that police use attack dogs most of the time not because they are avoiding putting themselves in the path of violence, but instead because they don't want to let someone escape.

The Independent Police Conduct Authority (IPCA) recently released its findings from its investigation into an incident in which a police dog handler fatally shot a man in Auckland. At the time, the police dog handler and his dog were both fairly new to their roles.

Apparently this dog handler was accustomed to only setting his attack dog on people who were unarmed and running away:

The officer finished his training as a dog handler about four months before the incident.

The officer's dog was also relatively junior. The officer says that he had deployed the dog no more than six times before this incident and on all of those occasions the offender was unarmed and running away from them.

Investigation into fatal shooting of Hitesh Lal in Auckland | IPCA

I don't know if this is representative of police dog handlers more generally, but it would be consistent with the data.

The people police set attack dogs on are overwhelmingly unarmed and just trying to get away. But that's not all, they are also mostly Māori.

Despite making up only 17%34 of the population of Aotearoa, 57%35 of the time police used attack dogs in 2018 they used them against Māori. Most of the moderate and serious injuries inflicted by police attack dogs were inflicted on Māori, too.35

At every level, the harms of the justice system have a disproportionate impact on Māori. Māori are more likely to be sentenced to prison than Pākehā, when convicted of the same crime.37 More than 50% of our prison population is Māori.38 Depending on which year you look at, Māori are 7-9x more likely than Pākehā to be on the receiving end of police force.3940

The disproportionate use of attack dogs against Māori is just one of many aspects of our society that disproportionately harms Māori. It affects a relatively small number of people, a few hundred each year, but for them its effect is particularly brutal.

As well as recording information about the behaviour of people during a TOR event, NZ Police also recorded information about what offences they charged those people with.

These charges are split into a number of categories, but of course there can be a huge difference between two different charges within the same category. For example, both homocide and "intimidation/threats" fall under the category of "Violence".

However, looking just at those broad categories, there were some clear differences. For the 297 TOR events where police officers used attack dogs in 2018, the subject was charged with a violent offence 28%41 of the time. But for those events where dogs weren't used, subjects were charged with violent offences 38%41 of the time.

Offences under the "Drugs & Antisocial" category were also less common for TOR events were dogs were used, and it was incredibly rare for any charges of sexual offences to be laid following any TOR event.

People who had police attack dogs set on them were much more likely to be charged with dishonesty and traffic offences. Dishonesty charges were laid following 43%41 of TOR events where attack dogs were used, compared with 7%41 of other TOR events.

Traffic offences were laid 33%41 of the time for TOR events involving attack dogs, but just 13%41 of the time for other TOR events.

When attack dogs were used, the most common charges were "traffic offences" (97),41 "car conversion etc." (81),41 and burglary (60).41 Both "car conversion etc." and burglary are considered "Dishonesty" offences.

The most common charge following TOR events where police dogs weren't used, by a large margin, was "disorder". These charges were laid following 27%41 of TOR events where attack dogs weren't used, but only after 6%41 of TOR events where attack dogs were used.

On some occasions each year, NZ Police sets attack dogs on children and young people. In 2018, NZ Police recorded 452 incidents at which they set attack dogs on children, under 14 years old. At a further 3352 incidents, police dog handlers set attack dogs on young people aged 14-16.

When asked in 2017 about another group of cases in which police set attack dogs on children aged 12 and 13, Children's Commissioner Andrew Becroft said:

I would expect there would be a clear policy for the police as to how police dogs — and for that matter tasers and pepper spray — are deployed when it is known to the police officer the person is a child.

Children's Commissioner Andrew Becroft | NZ Herald

As far as I am aware, no such policy exists. When I asked NZ Police in 2019 for all documents regarding the use of force and police dogs, only one of the documents they released contained a single mention of what to do when setting attack dogs on children or young people:

If a person [bitten or injured by a Police dog] refuses treatment, make a note book entry of the refusal and have that witnessed by either the person, medical person or another officer.

In the case of a child or young person their parent or guardian must be informed as soon as possible and they should be advised that they should seek medical assistance as soon as possible and every attempt should be made to provide that treatment.

Use of force and police dogs (PDF link) | NZ Police

Police officers are protected from criminal responsibility when they use force to prevent someone from escaping when an officer is trying to arrest them by section 40(1) of the Crimes Act 1961. This section has a couple of caveats: the officer must not have been reasonably able to prevent the escape in a less violent manner, and the force they use cannot be intended or likely to cause death or grievous bodily harm.

Police attack dogs cause serious injuries 8.1%22 of the time they are used, but currently this law does not appear to be limiting police officers using attack dogs to stop people escaping arrest. I expect that relying on it for that purpose would require NZ Police to lay criminal charges against their own dog handlers for doing what they've been told to do, which seems very unlikely.

But the law isn't the only thing that can restrict the use of force by police. NZ Police policy can also place restrictions on how they use force.

For example, NZ Police's taser policy prohibits police officers from ever discharging a taser on a person below this "assaultive" threshold, because that use would not be proportionate.

Always use a TASER in a manner consistent with the "Tactical Options Framework" and never 'use' in situations below the assaultive range, e.g. active or passive resistant.

TASER (Electronic Control Devices) (PDF link) | NZ Police

Compared with attack dogs' 86%22 injury rate, tasers cause injuries only 12%26 of the time they are used.

Of course, NZ Police policy isn't always followed perfectly. In 2018, police officers recorded that they discharged their tasers on people who they decided were only "active resistant" at 757 incidents.

As well as prohibiting the use of tasers on people below the "assaultive" threshold, NZ Police's taser policy also says they must not use a taser on a person in a vehicle or in control of machinery, if there's a risk that vehicle or machinery could injure them. But in 2018 a police officer tasered a 15 year old boy who was not assaultive and who was driving a stolen tractor.

Despite the IPCA finding that the officer's use of their taser "was an unreasonable and excessive use of force", and noting that "anyone who is authorised by law to use force is criminally responsible for any excessive use of force", the officer faced no consequences for their actions. It is extremely rare for NZ Police to charge its own officers in cases where they are found to have used excessive force.

The Independent Police Conduct Authority's reports offer some more insights into just how badly things can go when police dog handlers misuse attack dogs. Here are four incidents from 2017, which came to light through IPCA reports:

1. After police forced a car they were chasing to crash, a police dog handler set his attack dog on the young driver in a manner that IPCA chair Judge Colin Doherty characterised as "unnecessary and unprofessional".

The IPCA ruled that it was unnecessary to use an attack dog at all, but what was most concerning was that the dog handler intentionally yanked on his attack dog while it was biting the young person. This yanking exacerbated their injuries, and the police dog handler tried to justify his behaviour by saying they did it to help train his dog.

I believe Judge Colin Doherty's response speaks for us all:

No member of the public should ever be subjected to a dog bite for training purposes.

Inappropriate force used on young people after a pursuit through Christchurch | IPCA

In their response, NZ Police said "The actions of the dog handler were addressed by a supervisor."

2. After a pursuit that the IPCA ruled should not have happened, a police dog handler set his attack dog on two youths aged between 13 and 15. The attack dog seriously injured one of the youths, who then required hospitalisation for four days. The IPCA ruled that the length of time the police dog handler allowed his dog to continue biting was excessive and an unreasonable use of force:

Officer P ordered his Police dog to engage (bite) Mr V who then surrendered. The second passenger was arrested without incident and the third passenger, Mr W, jumped over a fence. Officer P ordered his Police dog to engage him and the injury caused by the dog bite required surgery and a four days' recovery in hospital.

Police should not have pursued youths on Auckland motorway | IPCA

In their response, NZ Police refused to accept the IPCA's finding, and the dog handler faced no consequences for his actions:

Police do not agree with the IPCA's findings that the time the dog handler allowed the dog to bite was excessive, and was unreasonable use of force.

Police acknowledge IPCA findings relating to pursuit of stolen vehicle | NZ Police

3. In another incident, a police dog handler set his attack dog on a woman who had already surrendered. She received serious injuries, and was hospitalised for three days requiring surgery to repair the damage to her arm. The IPCA ruled that the use of an attack dog in that incident was unnecessary and excessive.

In their response, NZ Police acknowledged that officers' use of force "must be considered, timely, proportionate and appropriate", and that their dog handler should have used less violent tactical options before releasing his attack dog. They did not, however, mention any form of consequences for the dog handler's unnecessary and excessive use of force.

4. Sometimes, the mere presence of a police attack dog can escalate a situation. In the worst incident I am aware of, a police dog handler dropped his attack dog's lead and lost track of his dog.

The handler only found his dog again when he followed the sound of a woman screaming. Having been trained to attack people, the dog had entered someone's property and attacked a woman who was hanging up her washing. She was seriously injured, and had to be taken to hospital to treat injuries to her leg.

Both the attack dog and its handler underwent re-certification training, then continued working.

This was not the only incident in which police attack dogs have attacked innocent bystanders.

Nor is it the only IPCA investigation which reveals how the simple presence of a police attack dog can escalate a situation.

In 2014, police took a dog with them while unlawfully searching a person's home. The presence of their dog resulted in a fight between their dog and a dog that lived in the house, which police officers then pepper sprayed.

In 2017, a passenger in a car yelled an obscenity at a pair of police officers, one of whom happened to be a dog handler. The officers then followed the car to a petrol station, where they attempted to unlawfully arrest the person who had yelled at them. After he resisted, the dog handler set his attack dog on the man.

In 2018, during an arrest, a police dog handler set his attack dog on a man who was face-down on the ground with his arms pinned beneath him. The report notes that the dog handler discounted the possibility of using his hands or his taser, which would have been a less violent use of force, because they wouldn't have been able to control his dog.

Officer A said assisting Officers B and C with restraining Mr X using his hands, or using a Taser to 'contact stun' him, were not options available to him because of the need to keep control of his dog.

Police Dog bite in Upper Hutt (PDF link) | IPCA

This same reasoning was also used by the dog handler who shot and killed Hitesh Lal:

We accept that the officer needed to control his dog with one hand, therefore he had to decide between arming himself with the Taser or the firearm in his other hand.

Investigation into fatal shooting of Hitesh Lal in Auckland | IPCA

I asked NZ Police if it believes its use of dogs to bite people is worth the harm caused through the injuries they inflict.

I also asked them if, in light of the harm caused by police dogs compared with other tactical options available to police officers, they would consider changing their policy around the use of force and police dogs.

NZ Police didn't answer my questions directly, but they did provide this statement:

The purpose of using Police dogs is not to bite/injure people, but to help apprehend a person when other tactical options are not considered practical or effective, and sometimes a bite/injury occurs.

In 2018 there were around 88,000 jobs attended by dog teams, however there were only 297 TOR events which involved a Police dog bite or injury (Note that there may be multiple TOR events relating to the same incident, for example an incident involving multiple offenders). In the vast majority of situations where an offender is apprehended with the use of dogs there has been no bite/injury, but those non-bite/injury successes are not seen in the TOR ‘use of force’ data.

Nearly all dog deployments occur when people have committed offences and run away, or are in possession of weapons and confronting police, and in most cases the use of OC spray or taser is impractical.

Had a dog not been used for these events it may well have resulted in the offender evading apprehension and continuing to pose a risk to the community.

All dog handlers must issue a loud clear challenge to an offender advising the dog will be released if they do not comply.

We are acutely aware of the impact our dogs can have when biting someone. Our dogs and dog handlers receive very robust training and are regularly validated throughout their working life.

All dog bites are reviewed by a dog handler’s supervisor and a commissioned officer of police. Any serious bites are referred to the IPCA.

Of the people bitten in 2018, the vast majority did not require hospital treatment, and less than 10% met the requirement for notification to the IPCA.

To clarify the term ‘actively resisting’ - an actively resisting offender who has run off could have just committed a serious offence such as homicide, aggravated robbery, rape or burglary and police are duty bound to apprehend this person to prevent further harm.

The structured TOR data that you received does not include the contextual detail around the events. The full circumstances of why apprehension using a dog was necessary are explained in the narrative description of a TOR and these are not well reflected in the structured data fields.

Inspector Nic Brown, Acting Director: Capability | NZ Police

There is recent precedent for NZ Police changing their policies to prevent harm, even though it would also mean apprehending fewer people.

In December 2020, NZ Police reworded their pursuits policy in order to clarify it so officers would be less likely to initiate pursuits when the risk of harm means they wouldn't be justified. In an internal email leaked to Stuff, Wellington road policing manager Inspector Wade Jennings told police officers:

A very basic question to ask yourself is ‘is what I am chasing worth the risk?’.

Wellington road policing manager Inspector Wade Jennings | Stuff

Pursuits can be lethal — 7458 people were killed in police pursuits from 2010 to 2019 — and because there are thousands of pursuits each year they cause more serious injuries overall than attack dogs. But any individual pursuit is significantly less likely to cause serious injuries than setting an attack dog on someone.

On average, 100 pursuits cause 1.458 serious injuries, compared with 8.122 serious injuries for every 100 times police dog handlers set an attack dog on someone.

Ideally, I think NZ Police should discontinue their use of attack dogs. I believe their use simply does more harm than it is worth. But until that happens, I would like NZ Police to tell their dog handlers to ask themselves the same very basic question:

Is what you are chasing worth the risk?

If the answer is no, then they should not command their dogs to attack anyone.

If police dog handlers won't ask themselves that question, then we should ask them. This is a conversation that we need to have.