On Feb. 11, 2021, the New Zealand Defence Force (NZDF) announced new rules on responding to civilian harm. Defence Force Order 35 (DFO 35) introduces NZDF-wide procedures for responding to reports of civilian of civilian harm arising from military activity in situations of armed conflict. The rules govern all members of NZDF, who can be disciplined in accordance with the Armed Forces Defence Act 1971 for non-compliance, and it also covers contractors and consultants, whose contracts might be terminated if they fail to abide by the rules outlined in the DFO. Announcing the new rules, NZDF Director of Legal Services Brigadier Lisa Ferris acknowledged that the complexities of modern warfare were exposing civilians to greater levels of risk, but insisted that “the protection of civilians is a strategic priority for NZDF.” DFO 35 is, according to the Defence Force, meant to “institutionalise a consistent and transparent approach… informed by past experiences of the NZDF, the values of New Zealand, and the developing norms of international best practice.”

In this effort, New Zealand has achieved something globally significant. It is the first national military policy that outlines both a commitment to and procedures for responding to potential and confirmed civilian harm. We are pleased on two counts: first, to see their deliberate thinking regarding this difficult and challenging process, and second, that they valued transparency enough to make this policy available to the public. As the first country to make this step, we can ask: how effective will this policy be in achieving its aims? This article looks at the policy, in light of the incident and inquiry that prompted it and finds there are areas where the policy falls short of its stated goals.

Although DFO 35 is commendable in terms of its broad aspiration, it is sorely lacking in terms of the details, and it risks repeating the common mistakes that other militaries have made in the past. We go through all five sections of the report, identifying what NZDF rules do well and where there is some scope for improvement. We also highlight some obvious gaps and omissions in the rules, recommending additional measures that could be introduced to enhance transparency and improve accountability, ensuring that civilians are better protected from the devastating consequences of armed conflict. We are particularly pleased to see that NZDF view DFO 35 as “a living document, [which…] will continue to be updated as required to that it remains as relevant and effective as possible.” At present, we feel that DFO 35 is a missed opportunity for NZDF.

And equally important, we look at this exemplar of a national policy in the light of what the U.N. Secretary General has described as “terrible suffering inflicted upon civilians” due to lack of concrete progress in the protection of civilians by the international community. Can this policy serve as an example for others to make this needed progress? There, too, we see that there are limitations in the policy and approach it takes, so that it falls short of what is needed to strengthen protection in concrete ways. Thus, DFO 35 is also a missed opportunity in a global sense.

Background: Operation Burnham

DFO 35 was a response to a particular incident and we see it is shaped significantly by the details of that incident and how it was handled. We begin, therefore, by outlining the context in which these rules emerged, as it helps to explain why the rules focus on certain aspects of civilian protection and not others. It also provides us with an obvious benchmark to assess whether the specific rules outlined in DFO 35 outline an appropriate response to reports of civilian harm.

In 2017, investigative journalists Nicky Hager and Jon Stephenson published a book entitled Hit and Run: The New Zealand SAS in Afghanistan and the Meaning of Honour, which alleged that NZDF committed war crimes during an attack known as Operation Burnham. Specifically, the book claimed that NZSAS attacked two villages in the Tirgiran Valley based on faulty intelligence that two insurgent leaders were based there. The book alleged that six civilians were killed during the attack, including a three-year-old girl named Fatima, and it claimed that another 15 civilians were injured. It also alleged that NZSAS and the American helicopter involved in the attack deliberately damaged or destroyed residential buildings, including a room containing religious books. The authors updated some of their original allegations as more information came to light.

Claims that civilians were killed during Operation Burnham had already surfaced (see here and here), but NZDF claimed that allegations of civilian harm had been investigated and determined to be “unfounded” (p.184). Following publication of Hit and Run, Chief of Defence Force Lieutenant-General Tim Keating held a press conference disputing the central premise of the book and claiming that the authors had made 105 separate errors. Keating acknowledged an NZSAS operation took place on that date, but argued the location of the village was two kilometers away from the coordinates in the book. He stated that helicopter video footage revealed that NZSAS did not deliberately target civilians and that air assets had not destroyed civilian homes, but conceded that some civilians might have been killed. The NZDF were subsequently criticized for being “too defensive” in their response, seizing on errors in the book rather than admitting their mistakes. Keating also accepted that NZDF could have been clearer from the outset that civilians might have been killed during the raid, even though NZSAS had not targeted them intentionally.

In April 2018, Attorney-General David Parker announced an inquiry into the allegations described in the book, which would be headed by Sir Terrence Arnold (Supreme Court judge) and Sir Geoffrey Palmer (former prime minister). Parker stated that he had reviewed the video footage provided by Lieutenant-General Keating, arguing that it did not corroborate some key allegations outlined in Hit and Run, nor did it conclusively answer some of the questions raised by the author. Parker stated that, “in light of that, and bearing in mind the need for the public to have confidence in the NZDF, I have decided in the public interest that an inquiry is warranted.”

Over a period of two years, the Inquiry heard evidence from 159 individuals, received more than 4,431 documents and published 1,756 pages of public transcript. It was due to report by April 2019, requested several extensions due to difficulties getting the appropriate access to classified material, particularly material supplied by partners overseas. However, NZDF was criticized for failing to provide important material in a timely manner, with the Inquiry noting that “NZDF had a number of different systems, with the result that it was not able to find relevant material readily – nor could it be sure that the Inquiry had been provided with all relevant information.”

The Report of the Government Inquiry into Operation Burnham was eventually published in July 2020. The Inquiry found that the allegations made in Hit and Run were accurate in some respects, but inaccurate in others. It stated that NZDF had reliable intelligence about insurgent activity in the two villages and that NZDF personnel “acted professionally, although several miscalculations or errors may have been made.” It stated that Operation Burnham was not an attack on civilians and that video footage showed armed individuals emerging from a house. It stated that the helicopters did not open fire on residential houses, but confirmed that several rounds hit two or three houses, referencing an AR15-6 report that blamed a misaligned weapon on one of the helicopters. The Inquiry found that it is likely an unidentified girl aged between 8 and 10 was killed during the operation, but concluded that NZDF personnel “would not have known that there were civilians in close proximity to the man who was the target of the firing.” The Inquiry was unable to determine whether other individuals killed in the attack were civilians or insurgents, but found that their deaths were lawful. The Report also confirmed that at least six civilians were injured in the attack, including two children.

NZDF were cleared of any wrongdoing in terms of the civilian casualties, but the report concluded that “there were serious deficiencies in the way NZDF dealt with the allegations.” Although NZDF personnel were aware that civilians might have been killed in the attack, it concluded that officials “did not take any effective steps to investigate the allegations.” Moreover, the report assessed that the advice NZDF officials provided to ministers was “inaccurate because it misstated the position in relation to the possibility of civilian casualties [making…] a series of incorrect statements in both briefings to ministers and public releases, to the effect that the allegations of civilian casualties had been investigated and found to be ‘baseless’ or ‘unfounded.’” The Inquiry found that this misinformation stemmed from a “seriously misleading email sent by the Senior National Officer in Afghanistan, Lieutenant Colonel Chris Parsons.” In the email, Parsons stated that he had “sighted” a three-page report produced by the Joint Incident Assessment Team (JIAT) sent to look into the attack and that this assessment had concluded “there was no way that CIVCAS [civilian casualties] could have occurred”.

The JIAT actually determined the opposite. It concluded that “several rounds from coalition helicopters fell short, missing the intended target and instead striking two buildings, which may have resulted in civilian casualties.” Brigadier General Timothy Zadalis, who led the assessment team, also filed an addendum, noting that a review of the video footage had identified another possible occasion when civilian casualties could have occurred. During evidence, Parsons explained that he had mistaken an acronym in the report (“AF”) to mean “Air Force” when it was intended to mean “Assault Force”. He assumed the JIAT had cleared both the ground force and the air assets of causing civilian casualties, when in fact it cleared only the ground force. When Parsons realized his error, he did not take adequate steps to correct the position, something the Inquiry described as “a serious failure on his part, with significant and long-lasting ramifications.” However, the Report criticized senior NZDF personnel back in New Zealand, who seemed to accept his “email without question, even though there was substantial evidence contradicting it” (including ISAF media releases). Moreover, NZDF staff failed to correct the record after being sent the JIAT report on two separate occasions, continuing to claim that the report concluded allegations of civilian harm to be “unfounded.”

The Report was particularly damning of NZDF record-keeping, describing a “surprising level of ineptitude and disorganisation within NZDF Headquarters.” Important documents were lost or misplaced – sometimes irretrievably – because they were stored on personal drives rather than in shared folders. NZDF was sent a copy of the JIAT report in 2011, but it was placed inside a safe and promptly forgotten. It was rediscovered – “essentially by chance” – several years later. The report was described as an “executive summary” so NZDF officials requested access to the “full report” only to be told that the executive summary was the full report. Despite being told this, NZDF officials made repeated requests to both NATO and ISAF for copies of the “full report”. The Inquiry described this as a “small illustration of a significant defect in the way that NZDF dealt with the civilian casualty issue, namely the lack of corporate memory within NZDF or, more precisely, NZDF’s failure to ensure that corporate memory on this matter was recorded, preserved and accessible.” This problem was exacerbated by the practice of moving people into new roles, which meant there was little coordination on this issue.

Although the Inquiry rejected allegations that NZDF had tried to conceal these casualties, it delivered a scathing attack on reporting procedures inside the organization. It argued that the no-casualty claim became the accepted NZDF narrative due to a “disappointing lack of commitment and rigour, both individual and collective, on the part of senior NZDF personnel.” It found that its “failure to obtain accurate information concerning the allegations meant that it advanced a false narrative, which misled ministers and the public.” Crucially, it stated that these institutional failings caused NZDF to undermine two core constitutional principles: namely, civilian control of the military and ministerial accountability to parliament. It recommended that the Ministry of Defence takes steps to ensure that the organizational structure of NZDF, together with its record-keeping and retrieval processes, are in accordance with “international best practice and are sufficient to remove or reduce the possibility of organisational and administrative failings of the type identified in this report.” It recommended that an office of the Independent Inspector-General of Defence be established – located outside the NZDF organizational structure – to investigate whether NZDF operations are being conducted lawfully and with propriety. Finally, it recommended that NZDF promulgate a Defence Force Order “setting out how allegations of civilian casualties should be dealt with, both in theatre and at New Zealand Defence Force Headquarters.”

DFO 35: An Overview

The new Chief of Defence Force, Air Marshall Kevin Short, issued an immediate apology, acknowledging that NZDF “let our frontline service people down through a series of organisational and administrative failings that saw incorrect information provided to Ministers and the New Zealand public.” Short promised to implement the recommendations in full, accepting that NZDF must do “better at the way we record, store and retrieve information, and then subsequently present that information.” Brigadier Ferris was charged with promulgating a Defence Force Order, consulting various experts during the development phase. On Oct. 5, 2020, a joint workshop was convened with the Ministry of Defence to discuss what obligations NZDF has to report instances of civilian harm, what possible outcome or amends should follow an investigation into civilian harm, how these investigations ought to be conducted including what evidence needs to be collated, and what steps NZDF should take to ensure the independence of any investigation into civilian harm (full disclosure: Thomas Gregory was a participant at this workshop).

It is important to note from the outset that DFO 35 is concerned with responding to civilian harm rather than avoiding it – a point that we will return to in a moment. In the introduction, it states that the “risk of harm to civilians is an unfortunate reality of armed conflict [and…] it is unrealistic to expect that civilian harm can be entirely eliminated.” It mentions in passing that there are processes that can be put in place to mitigate or minimize this harm, but DFO 35 does not outline any measures that could help to reduce the number of civilians injured or killed by NZDF personnel.

This  omission clearly reflects the problems encountered in the aftermath of Operation Burnham and the specific recommendations outlined in the Report. As the previous section explained, NZDF was cleared of any wrongdoing in terms of the harm caused to civilians, but was faulted for not reporting this harm in a timely or accurate manner. However, it is disappointing that NZDF has not attempted to build on lessons learned overseas, especially as DFO 35 notes that protecting the civilian population can be strategically advantageous to NZDF and that causing “civilian harm can be highly damaging to the reputation of a force and to the public support for the cause for which it fights.” Rather than responding to the last crisis, NZDF could have used this as an opportunity to avert future crises before they arise.

The purpose of DFO 35 is to implement NZDF-wide procedures for responding to reports of civilian harm and is guided by the following five principles:

1. Independence – NZDF will conduct itself without interference from persons or bodies not assigned to the matter, without fear of reprisal, and without expectation of favour for finding any particular outcome;

2. Impartiality – there is to be no undeclared conflict of interest or personal or professional bias for any individual assigned to investigate a report of civilian harm, including legal advisors whose advice may have contributed to a particular behaviour or course of action;

3. Thoroughness – resources are to be adequate to allow for quality record keeping and to ensure all feasible steps to source, analyse, store, and retrieve relevant information can be met;

4. Promptness – NZDF is to act promptly to record, assess, report, and resolve reports of civilian harm (delay can affect the rights of affected persons, implicated NZDF personnel, and witnesses), however, a prompt response should not be prioritized over a thorough approach; and

5. Transparency – NZDF’s response to reports of civilian harm should offer those involved a high level of assurance in the processes engaged (achieved through regular communication, openness, and accountability).

Subsequent sections attempt to convert these broad principles into more concrete processes and procedures. An eight-step processes for responding to civilian harm is outlined, beginning with Incident Awareness. DFO 35 states that “it is a mandatory requirement for all NZDF personnel to report all civilian harm, regardless of source, to the relevant SNO [Senior National Officer] or equivalent.” The SNO is expected to produce an Initial Notification Report, which is sent to the Strategic Commitments and Engagements Branch at Headquarters Joint Forces New Zealand (HQJFNZ SCE Branch). Working with the Ministry of Defence, HQJFNZ SCE Branch are required to produce a note for the Minister of Defence (with referral to the Prime Minister and Minister of Foreign Affairs). This note should be signed out by the Chief of Defence Force and the Secretary of Defence. In the meantime, the SNO is expected to conduct an Initial Assessment of the incident to determine whether the report is possibly credible or not credible. After conducting their initial assessment, the SNO is required to produce an Incident Report (even in cases where the allegation is determined not to be credible), which should be sent to HQJFNZ SCE Branch. An Investigation should be initiated in cases where civilian harm is determined to be both possibly credible and critical (such as death or serious physical injury or significant damage to civilian property).

DFO 35 states that investigations ought to be conducted in accordance with the Armed Forces Discipline Act 1971, which allows for a Court of Inquiry (COI) to be established (appointments to the COI should be independent from the individuals and units under investigation). The purpose of the COI is to establish the context and facts, determine whether any individual or systemic factors contributed to the civilian harm, and to decide whether any further action needs to be taken (e.g., amends to the victims). COI Reports are to be filed with HQJFNZ SCE Branch, who will produce a cabinet paper for the Minister of Defence. Before the report is finalized, however, the COI are expected to Share Findings with those affected (in the appropriate language), so that they have an opportunity to comment on the report. Amends should be provided to the victims, but will vary depending on the context (see below). Finally, the Closing Authority is responsible for the archiving these reports. In cases where the report was deemed to be non-credible, the closing authority is HQJFNZ. And in cases where the report was deemed to be possibly credible, the closing authority is the COI Office.

The fact that the SNO is still required to submit an incident report, even in cases where the allegation is deemed not to be credible, is a commendable development because it will enable officials to reopen case files if new information emerges and it will be easier for the media to petition for information under the Official Information Act of 1982. However, we believe that every initial assessment ought to be made public, or at least a statement indicating the conclusion and justification, not least because external complainants might otherwise never learn what happened to their case.

Some Positive Steps

After the embarrassment surrounding the Burnham Inquiry, NZDF has attempted to create a comprehensive framework for responding to civilian harm, eliminating the organizational and administrative failings that allowed them to pass on inaccurate information about allegations of civilian harm. Although there are some serious shortcomings with DFO 35, which we will discuss in the following section, it is important that we recognize some of the positives. As this section explains, the order establishes an important precedent when it comes to multinational operations, which is something that other militaries ought to replicate in their own policies on civilian harm. The section on amends also raises some interesting considerations, although it is unclear how NZDF intends to implement these in practice.

The most significant feature of DFO 35 is its scope, with the policy establishing a framework for responding to reports of civilian harm resulting from military activity that NZDF has participated in, whether directly or indirectly. This has obvious implications when it comes to coalitions, with DFO 35 stipulating that “civilian protection, including response to reports of civilian harm, must be considered when NZDF is planning to participate in a multinational operation.” Whenever it considers deploying as part of a multinational force, NZDF is required to undertake an assessment to determine whether the multinational force has the capacity to respond to reports of civilian harm. NZDF is required to actively encourage the codification of civilian harm response procedures in coalition arrangement, including clear expectations for information sharing. During multinational operations, NZDF is expected to “support the prioritisation of civilian protection during the operational planning process, including encouraging early dialogue with humanitarian actors.” The order also commits NZDF to ensuring that local forces trained by NZDF personnel are able to respond to instances of civilian harm, from addressing the immediate medical needs of those impacted to the appropriate “management of the dead, [and] missing persons.”

This commitment to promoting civilian protection in multinational agreements will ensure that coalition partners take the issue seriously, encouraging them to develop their own policies and procedures. We hope that NZDF takes this commitment serious enough that it will withdraw from multinational operations if there is evidence to suggest that civilian protection is not being prioritized. This is a useful notion in an era of combined operations where it is possible to diffuse responsibility and accountability under the mantle of a coalition. Later we discuss ways NZDF can explore the ramifications and potential implementation responsibilities of this policy commitment in the light of its recent role in several multinational operations.

The section on amends is also significant as it addresses some of the problems that have hampered effective amends programs. Rather than simply focusing on monetary payments, DFO 35 suggests that amends can be categorized into three broad areas:

1. Restitution – designed to restore affected persons to their situation prior to harm, including monetary payments, livelihood assistance, contribution to community infrastructure.

2. Rehabilitation – including medical, psychological, legal, and social services including resources to search for the missing.

3. Satisfaction – including pledges of non-repetition, communicating changes in policy to prevent further civilian harm, official acknowledgement of the facts, public or private apologies, commemorations to victims, and acceptance of responsibility.

We are pleased that NZDF is looking beyond just financial restitution when it comes to amends, acknowledging that it might be necessary for them to provide specific services to the victims. It is particularly significant that NZDF might be willing to review procedures in certain circumstances, communicate any changes to the victims, and promise to avoid repeating the same mistakes again. This represents a significant departure from existing practices, which have tended to focus on ex gratia payments. DFO 35 is less clear on the decision process determining when NZDF should consider alternatives to financial restitution and who is responsible for delivering them. Indeed, the specific processes and procedures focus almost entirely on ex gratia payments, such as the amount that should be offered, who is responsible for delivering these payments, and when it might be appropriate for payments to exceed the established threshold.

Some Missed Opportunities

Although NZDF should be praised for these developments, there are also some significant omissions that need to be flagged and some confusion that needs to be cleared up. This section outlines three problems that need to be corrected: the failure to include any specific provisions on minimizing civilian harm, the emphasis that has been placed on the initial assessment, and the uncertainty around specific roles and resources. We are concerned that NZDF is too focused on responding to its last crisis that it has failed to anticipate or avert future crises. At the same time, we are not convinced that the policies and procedures outlined in DFO 35 actually address the concerns raised in the Burnham Report. Put simply, there is still a risk that reports will be overlooked and incidents not properly assessed, allowing inaccurate information to circulate without correction.

We are particularly concerned that DFO 35 contains no procedures on preventing or reducing civilian casualties, not least because the two processes – preventing civilian harm and responding to civilian harm – are interdependent. Striving to achieve a more accurate understanding of potential civilian harm before an engagement, an essential part of mitigating harm, can better inform an assessment of estimated civilian harm post-engagement. Likewise, we have seen in real world operations that there is a propensity for militaries to fail to adequately anticipate civilian harm, especially for strikes on structures. A rigorous post-engagement assessment process should identify these areas of risk and enable learning and improvement in these areas of risk. The failure to actively link these two processes can lead to a vicious cycle—a weakness in detecting civilian harm followed by a cognitive bias that civilian harm was unlikely to occur since it was not detected. As well as representing good military practice to learn and improve, such a process could also be seen as a way to fulfill Additional Protocol I’s obligations to take all feasible precautions to minimize loss of civilian life.

Perhaps NZDF officials reasoned such a focus on mitigation of civilian harm was not necessary since the inquiry found that potential civilian casualties during Operation Burnham were likely only from United States air assets, not the NZ ground force. However, that does not mean there were not important lessons to be learned. For example, in the last engagement in Operation Burnham, a combined attack by an AC-130 aircraft and an AH-64 Apache helicopter on four individuals south of the individual objective, the engagement was cleared by the NZ Ground Force Commander (GFC). The inquiry found that the GFC was unaware of whether the individuals were armed or not, and in fact believed they were moving north—representing an imminent risk that justified an engagement in self-defense—when in fact they were moving south. The inquiry decided in the end that there was no evidence that the four individuals were civilians, though that could not be ruled out either.

The point of the inquiry was that this engagement was in compliance with international law and ROE. But equally as important, mistakes were made that can be corrected in the future. From United States analysis of civilian harm incidents, the misunderstanding of the tactical situation seen in that last engagement, a disconnect between the ground force and supporting air assets, is a regular feature of civilian harm incidents involving air assets. For example, the 2010 civilian casualty incident in Uruzgan, Afghanistan, where a United States airstrike killed dozens of civilians, and the 2015 United States airstrike on an MSF hospital in Kunduz both had this same disconnect between ground and air forces. While militaries should strive to learn lessons from past mistakes and improve practical elements that increase risk to civilians, neither the inquiry nor the DFO represent a strong commitment to that end.

Relatedly, the Burnham Inquiry recommended that NZDF prioritize its response to civilian harm, but it is disappointing that it has not seized the initiative and used this as an opportunity to incorporate lessons from overseas. As mentioned before, the United States analyzed hundreds of civilian harm incidents from ISAF operations and the lessons and recommendations were shared with all ISAF contributing nations. But there is no mention of this in the inquiry or the DFO. If NZDF had referred to these lessons, they would be aware that there are ways that NZDF can proactively shape the operating environment to reduce the likelihood that civilians will be harmed, thus reducing the likelihood that NZDF will have to respond to this harm. This is not something that can be implemented retrospectively. Troops need to be trained in advance on the importance of reducing civilian casualties, and the methods by which civilian casualties can be reduced. They need to be equipped with the appropriate weapons, including nonlethal options, to ensure they have the right tools at their disposal. Operations need to be planned in ways that put civilian protection at the center of considerations, so that commanders have the opportunity to assess the impact on civilians and consider alternative courses of action. Specific procedures can be put in place to strengthen air-to-ground coordination to avoid the miscommunications observed in Operation Burnham and in many historical air-to-ground civilian harm incidents. None of these measures are mentioned in DFO 35, let alone developed in satisfactory detail. Rather than trying to pre-empt future problems by looking at measures introduced overseas, NZDF has fallen into the old trap of fighting the last war – in this case, the problems with its reporting procedures.

Our second concern focuses on how much hangs on the initial assessment, which DFO 35 defines as an “informal means to ascertain basic facts that will allow NZDF to triage civilian harm reports and filter out non-credible reports.” The SNO is responsible for conducting an initial assessment, and their reports are expected to contain “sufficient information to both demonstrate the rationale for the conclusion of the assessment and capture the fullness of the event for future review.” There are several problems with this process. It is unclear, for example, whether the SNO is able to conduct this assessment in an unfair and impartial manner. One of the concerns raised in the Burnham Report is that Lieutenant Colonel Parsons – the SNO who misread the ISAF assessment – demonstrated what the report described as “confirmation bias.” As the Report explains, he read the assessment “in a way that supported his preconceptions – he saw what he wanted to see and discarded what he otherwise knew.” This was obviously a serious error on his part, but NZDF needs to ensure that it has installed appropriate institutional safeguards to prevent others making the same mistake. Unfortunately, the procedures outlined in DFO 35 would allow for this to happen again because so much hinges on this initial assessment and the individuals responsible for making this assessment are too closely connected to the events they are meant to assess.

In fairness, NZDF have introduced some additional safeguards in the reporting process, but these are unlikely to prevent past problems from occurring again. DFO 35 stipulates that initial determinations will need to be reassessed in cases when it is “later brought into question by the emergence of new information.” Unfortunately, this instruction is buried in the footnotes, even though Lieutenant Colonel Parsons was criticized in the Burnham Report for “fail[ing] to take adequate steps to rectify the position once he appreciated that he may have misinterpreted the paragraph.” DFO 35 states that HQNZDF SCE Branch should be provided with a copy of the initial assessment even in cases where reports are determined to be not credible. However, it does not specify what steps (if any) HQNZDF SCE Branch are expected to take to confirm the accuracy of this assessment, such as cross-referencing them against assessments undertaken by coalition partners or non-governmental organizations. Recall that the Burnham Inquiry was particularly critical of senior officers at HQNZDF for uncritically accepting reports from the SNO despite it being “inconsistent with the significant amount of other credible information they held about the operation, which came from a variety of sources.” Finally, the Chief of Defence Force is required to make an annual statement summarizing, among other things, reports of civilian harm involving NZDF troops that were determined not to be credible. This additional layer of transparency is welcome, but so much hinges on what information is actually included in these summaries.

Here we also note an unexploited opportunity to learn from others. Besides the United States assessments of ISAF-caused civilian harm, the United States also conducted assessments of the civilian harm reporting and assessment process used in Operation Inherent Resolve (counter-ISIS coalition operations) in Iraq and Syria (covering operations from 2015-2017), as well as in Operation Resolute Support in Afghanistan (covering operations from 2017-2019). Similar failings were observed in the reporting and assessment process, but additional analysis examined trends and precisely where the process tends to fail and why. These findings could have helped NZ learn from these failings and provided additional specificity to help learn from them.

We also note that NZ was a coalition member of both Inherent Resolve and Resolute Support. Though we earlier commend NZDF for the DFO’s attention to multinational operations, we think a good way to prepare for meeting this commitment to ensuring that adequate civilian casualty reporting mechanisms are established would be a review of its recent multinational operations and what measures the coalition had in place for reviewing and assessing civilian harm. How did those multinational operations measure up? And what specific standards does NZ wish to have? Are these standards procedurally related to reporting and assessment only? Or do they also involve steps to mitigate civilian harm? If the latter, has NZDF considered the increased risk of civilian harm over time observed in both operations, according to coalition operational data? These are questions that are better to answer now before NZDF is involved in another multinational operation.

Our final concern relates to the allocation of roles and responsibilities, as it is unclear who – if anyone – actually has institutional ownership of this issue. It is important to note that NZDF has gone to considerable lengths to detail who is responsible for the component parts, with more than two pages dedicated to dividing up responsibilities between the Office of the Chief of Defence Force, the Commander of Joint Forces New Zealand, HQNZDF Strategic Commitments and Engagements Branch, Senior National Officers (SNOs), along with the responsibilities that fall on every NZDF member. The trouble is that no single individual or team seems to be responsible for ensuring that DFO 35 is implemented consistently. We are concerned that far too much is left to the discretion of individuals in the chain of command. Whilst some will undoubtedly appreciate the importance of protecting civilians, there is no guarantee that everyone will prioritize to the same extent, particularly if there are other pressures on their time and resources. It is worth recalling that Lieutenant Colonel Parsons blamed his mistake on the fact that “it was sandwiched between multiple tasks that resulted in him working for ‘20 hours straight.’” Responding to civilian harm seems to be just another task that has been added to everyone’s workload without anyone having ownership of the issue.

At the same time, there is little guidance in the DFO to help direct what are both difficult and by nature subjective decisions regarding credible civilian harm. The United States uses the guidance “more likely than not” to help direct such decisions, but previously mentioned assessments found that this general guidance combined with cognitive biases meant that personnel could tend to inappropriately reject the possibility of civilian harm—just as was seen with LtCol Parsons. This is a difficult task, demanding more specific guidance as well as personnel trained in its implementation.

Accordingly, NZDF needs to create a dedicated civilian protection team within HQNZDF to coordinate and monitor the implementation of DFO 35. This is specialist work and it requires a specialist team – with the appropriate institutional support – to ensure that it is completed to a satisfactory standard. According to the current allocation of responsibilities, for example, Commander Joint Forces New Zealand is responsible for considering civilian protection in relation to NZDF participation in multinational forces, developing reporting and amends protocols for specific theaters, ensuring that pre-deployment training and operational orders comply with this order, proactively fostering dialogue with relevant non-governmental organizations, ensuring that HQNZDF SCE Branch has the appropriate reporting mechanisms in place, receiving reports of civilian harm, ensuring that records of civilian harm are stored in line with privacy and data protection regulation, ensuring that civilian harm records are monitored for lessons learned, ensuring that any lessons learned are disseminated appropriately, and ensuring that the Chief of Defence Force is briefed regularly. This is an extraordinary amount of work, but no provisions are made to establish a dedicated civilian protection team within NZDF. Not only will this negatively affect the ability to respond to allegations of civilian harm, it also means that NZDF is unlikely to develop the “corporate memory” recommended in the Burnham Report. And once more we also note the NZDF’s omission in consulting existing lessons and best practices by coalition partners when creating the DFO.


We are pleased that NZDF seem to be taking the issue of civilian protection seriously, but there are a number of missed opportunities in DFO 35 that need to be addressed. There are some obvious omissions that need to be revisited, not least the absence of any provisions aimed at reducing civilian harm in the first place. Even though the recommendations outlined in the Burnham Report focused on responding to instances of civilian harm, NZDF could have used this as an opportunity to address operational deficiencies seen in Operation Burnham that increase risk to civilians. We also describe the missed opportunity to incorporate the lessons learned for civilian harm that were shared with all ISAF contributing nations. Unfortunately, it seems that NZDF is primarily concerned in this particular effort with managing the problem of civilian harm, rather than preventing it.

Moreover, it is unlikely that the processes outlined in the DFO would be able to prevent another Operation Burnham. When it comes to reporting civilian harm, far too much hangs on the initial assessment, which is conducted by the SNO. From analysis of the reporting and assessment process in several multinational operations where NZ was a coalition partner, it is very possible that, under the DFO, credible reports of civilian harm could be dismissed as non-credible without anyone noticing until it is too late. Publishing or summarizing all initial assessments – including those deemed non-credible – would help to improve transparency and ensure that complainants know what has happened with their case.

At the same time, there is no guarantee that NZDF will be able to develop the appropriate reporting mechanisms and corporate memory to ensure that it learns from its mistakes. Both the Burnham inquiry report and the DFO are notable in the absence of lessons and analysis regarding civilian harm, showing the NZDF did not learn tragic lessons from its coalition partners. NZDF will need to establish a dedicated civilian protection team within HQNZDF to ensure that it has the appropriate infrastructure in place, both for DFO implementation and for learning, before the next civilian is killed. And in an ideal world, NZDF might consider introducing some measures to reduce the likelihood that civilians will be killed as a consequence of NZDF operations.

In addition to representing an opportunity for New Zealand to sharpen its national response, we also see here factors that continue to hinder the global effort to better protect civilians. In armed conflict, two factors will help protect civilians from harm by militaries: the practice of compliance with international humanitarian law and a commitment to maintaining and improving operational practices and procedures to reduce risk to civilians. For militaries that have a strong commitment to international humanitarian law, the latter represents the most promising way to better protect civilians and reduce the “terrible suffering” of civilians in conflict. Unfortunately, we have yet to see clear leadership in this area. New Zealand’s failure to identify operational areas of risk to civilians in Operation Burnham exemplifies this continuing global failure. We hope that New Zealand and others will recognize this and pick up the mantle of leadership for the protection of civilians.

Photo: New Zealand Defence Force Website