The Use of Force, Its Authorisation and Consequences

Operational Experience from Bosnia to Iraq

ROE: The Use of Force, Its Authorisation and Consequences

RUSI Journal 2006

I have been asked to give a few examples of ROE from my own experience, with some indication of how these played out. I will do that, and then give a few conclusions on how ROE affect a field commander, and indeed how a field commander affects ROE. I must admit to some difficulties in finding the historical examples of ROE: MoD, PJHQ and its predecessors have not been good at archiving, and so I am relying on my own notes, diaries, and some press lines.

Bosnia under the UN

Turning first to UNPROFOR. In 1995, I was commander Goražde force, responsible for an isolated enclave separated from the main UNPROFOR AOR by both geography and Mandate. In Central Bosnia, the Mandate, and therefore the ROE, focused on supporting UNHCR’s humanitarian operation, chiefly the delivery of aid. Here the ROE were all about force protection: defensive, neutral as far as the warring factions were concerned, and with no offensive operations permitted. This was because at least in theory, the mandate was consent-based. Of course, things changed over time in response to Serb behaviour.

Gorazde force ROE were also formulated on a consent-based mandate, but were stronger from the beginning. They were drawn up for the protection of a UN designated Safe Area. This brought some difficulties in itself. The Safe Area, in General Rupert Smith’s words, existed “to safeguard the civilian population, so far as it is possible to do so in a war.” However the area was never geographically defined other than a by a 3-km circle around the city centre from which the Serbs had withdrawn. Nor was it a Safe Haven in the legally accepted sense, since security was not guaranteed by a proper protection force, nor was demilitarization imposed. The ROE were still defensive, but allowed for the use of force in protecting the civil population as well as UN troops.

For most of the time, though, the UN force was always over-faced by the Serbs in terms of weapon types and ranges. There was a disconnection therefore between the required ends, the mandate, and the available means – both in terms of hardware and ROE. When the enclave was attacked in May 1995, the UN force was able to fight for long enough to delay the attack until the Bosnian Army division in the enclave was in a position to take over the defence. This came at the cost of thirty-three hostages.

The one effective tool in the bag was air power, but this was heavily circumscribed, with permissions needed from UNPROFOR command in Zagreb. It was also under separate, ie NATO, command which further complicated matters. I never succeeded in getting it until the final evacuation when it was clear that Operation Deliberate Force was imminent, and the operation was conducted with close air support overhead and on call.

Bosnia under NATO

Two years later I was first Chief of Staff and then Deputy Commanding General of NATO Multinational Division South-West, based in Banja Luka, responsible for territory covered by all three factions. Here the Mandate was derived from the Dayton Agreement, and it was therefore compliance based, not consent based. The factions had to have positive clearance by us for any military movement or activity.  The ROE specified that minimum necessary force, including deadly force, could be used to protect NATO troops and their mission. It was also authorized for use to protect persons with protected status (i.e. diplomats and aid workers), and designated property. Pre-emptive, offensive, operations were authorized to enforce compliance with the Dayton Agreement. Deadly force was authorised against anyone committing or intending to commit a hostile act, i.e. an act designed to kill or damage a soldier, a person with protected status, or designated property

The available means were also considerable: armoured infantry fighting vehicles, tanks, artillery, and significant NATO air power all at hand. Its test was the Kosovo War, when our task was to ensure that the Serb forces did not react to NATO action against their brothers over the border, and that the Croat and Bosnian factions did not decide to take advantage of the situation. The authorization for pre-emptive operations was invoked here in order to prevent the movement of personnel or hardware. There were also specific operations mounted against individuals like Vojslav Seselj which had the potential to stir up a response and which had to be conducted using considerable and very obvious force in order to prevent or deter any hostile act.

The ROE called for all reasonable precautions to be taken to prevent injury to third parties, and the destruction of property. Warnings were to be issued when the operational situation permitted but were not mandatory. The use of force as punishment was specifically excluded, and there were specific directions to respect local social and religious customs. The use of warning shots was also authorized which, for the British Army, was a first.

Mandate, ROE and available means were therefore in my view appropriate and in balance, and as British officers and men we never had any particular difficulties arising from them. But this was, of course, a Multinational operation and the degree to which the contributing nations would sign up to the top end ROE varied. Only the British and Czech contingents would take part in all operations including offensive ops; only the British could be deployed anywhere in Bosnia. The Canadian, Dutch and Belgian troops all had specific exclusions and restrictions. This to a considerable degree governed where these contingents were deployed, what they would do, and whether or not they could be deployed outside their own boundaries. This should not come as a surprise, since national governments often have to secure parliamentary approval for deployments, and these will therefore come with specific political caveats. In the case of the Netherlands’ contingent, for example, what the troops could do was governed by the political fall-out from the Srebrenica experience.

Sierra Leone 2000 – 2001

In West Africa, I commanded the British Joint Task Force, which was not part of the UN force, and as Military Adviser to the Government of Sierra Leone I also exercised de facto command of the Sierra Leonian armed forces through embedded British staff officers and advisers throughout the chain of command. The Mandate was based on an MOU between the British Government – the former colonial power – and the Government of Sierra Leone.

I was however not permitted to undertake combined operations with the Sierra Leone Armed forces, and this made the position of embedded officers and NCOs difficult; for they could only do their job if they accompanied SLA units to the front, making sure that what we had taught was put into practice. And, by the way, demonstrating that we meant business in sharing the dangers of the campaign. Nor did I undertake operations with the UN: my specified task was to make sure that the UN “did not fail” – a rather different matter from making sure that it succeeded. A further factor was the enemy, the RUF, which certainly did not respect the Geneva Conventions and was not treated as a combatant force under it; on the other hand its members had the option of joining the UN run DDR programme.

There was thus. as it was put to me at my initial briefing, strategic confusion and a degree of operational pause. Strong ROE were needed, and my notes from the time and the press lines are clear that we had very robust ROE for the British-only element of the force. Necessary force, including deadly force, was authorized to protect the officers and men and our mission; and the means available were considerable, including a large over-the-horizon embarked reaction force and a destroyer or two always close off shore. I was constrained greatly however in conducting offensive or pre-emptive operations, and had to refer to London to do these. It never ceases to amaze me, how much more people several thousand miles away know about local conditions and the solutions to problems, than do those on the ground. I only received authorisation three times, and on one occasion – an operation to secure the town of Kenema in eastern Sierra Leone at Christmas 2000, this only came through an hour before the operation was to be launched.

This operation was one in which, in my view, the Mandate, Mission, ROE and available means were not in balance and it was as well that the enemy was the RUF, and not something more capable. London never really made up its mind whether it wanted to support the government and defeat the RUF, or to underpin the UN. In fact a combination of both was required, but the British and SLA should have been the stick, to the UN’s carrot. This is what in effect we achieved on the ground, but only in the face of considerable inertia at home. I think much had to do with an over-estimation of the power of the opposition: this was not exactly the Waffen SS, although to hear some people talk you might have thought it was. The RUF had gone on the rampage when faced with tribesmen or local forces, but whenever they were given a serious smack by European forces, they caved in at once.

Iraq, 2003 – 2005

I first went to Iraq in 2003 with the US Army, and subsequently commanded British Forces Iraq and Multinational Division (South-East). During the campaign in Iraq up to late 2005, three separate sets of ROE were employed. The first set were those used during the invasion, which were geared towards offensive operations governed by the law of armed conflict, citing UNSCRs from 1990, 1991 and 2002 relating to WMD. Military operations were limited to what was necessary to enforce compliance with these SCRs, and to remove the Iraqi regime. It was stipulated that as areas were declared free from hostile forces, more restrictive ROE would be imposed.

The ROE were designed to direct military action against necessary military objectives or secure a military advantage; and stipulated that military objectives should not be attacked if such an attack was likely to cause civilian casualties that would be excessive in relation to the military advantage gained. In general they were very permissive ROE, with powers of release for the use of artillery, close air support etc delegated to theatre commanders. Interestingly they also authorized the use of force against non-Iraqi military forces demonstrating hostile intent.

These ROE were revised in October 2003 and became much more restrictive, being based not only on the general provisions of the law of armed conflict, but also on the duties of an occupying power under the 4th Geneva Convention. As other nations joined Britain to form MND SE, national differences also became a factor, more of which later. These ROE recognized that the environment was one of low consent, in which close air support and the use of heavy weapons might be required. But commanders had to use force to maintain security and stability, or demonstrate resolve in the face of a threat to the force, using minimum force – rather than to gain a military advantage. The principles on which these ROE were based were therefore those of proportionality, distinction, and humanity. Respect for social religious factors, private property, courtesy and impartiality were also stressed. Although the inherent right of self defence was upheld, lethal force could only be justified where there was an imminent threat to human life. These ROE, in my view, created a good deal of uncertainty in the minds of both commanders and soldiers on the ground, about when they could and could not fire their weapons. Many, I know, feared that if they did kill an insurgent under these ROE, they would end up facing a murder charge in court at home.

Because of this uncertainty, and because of the high level of violence during August 2004, these ROE were again revised in December 2004. It was recognized that a state of non-international armed conflict existed, and British commanders were authorized to use offensive action in accordance with LOAC in response to emergent threats or for force protection. Particular armed groups were named in a supporting targeting directive. Under these ROE, the force used under LOAC was to be as a last resort or where the risk to life was unacceptable. The amount of force used was to be “reasonable and proportionate, and the minimum necessary to accomplish the mission.” Damage to civilian property was to be avoided if possible, and commanders were urged to be aware of the risk of aggravating local tensions. However military necessity was the driver.

These ROE were much more in accord with the circumstances on the ground; unfortunately the damage had already been done in the minds of the soldiers, who remain wary of firing their weapons. I had to write personally to all COs, as GOC British Forces, spelling out what was required and explaining that as long as force was reasonable in the circumstances, no-one should be afraid of finding themselves in court. This was not, of course, the same thing as committing crimes, including murder. We should be clear about the difference, despite emotional reporting in the press. Robust ROE have never and can never be at odds with the responsibility to obey international law. While the military system must protect people who operate within the ROE when things turn ugly, it must equally prosecute misconduct. The Adjutant General’s paper “The Military Criminal Justice System” is quite clear about this.

By now there were of course considerable differences in the ROE of the different nations in MND SE. The British were permitted to do anything and everything from CIMIC operations up to full scale attack and pursuit operations. The Danes, Italians, Czechs and Australians were almost as robust, excluding only a few specifics such as crossing brigade boundaries or handling detainees. The Italians and Czechs were also not permitted to conduct offensive operations. At the other end of the spectrum, the Japanese contingent was permitted only to defend itself – nothing else. Knowing this, I and the staff had to give missions or tasks that were appropriate; and make sure we applied for exemptions or modifications in advance. For example, cross-boundary authorization for specific capabilities during the Election period. On the other hand, there was enough flexibility to allow, for example, British troops to be placed under Danish command and vice-versa, Romanians under Italian command, and Australian troops under British command.


  • ROE are commanders’ business, not lawyers business. General Abizaid, Commander US Central Command, was fond of saying that “lawyers advise, commanders decide”; and “commanders” includes the whole chain of command. If they are not right, then commanders have a duty to represent that. This is because the commander on the ground is both responsible for the execution of the mission, and accountable for the success or failure. He must have the delegated resources and authority to carry it out. Separating these three will cause trouble – and by the way, failure to delegate is the same as failure to resource.

  • Commanders at every level must ensure, personally, that the staff has a method of disseminating ROE, including changes. As important as this is the method of ensuring that the ROE are explained and understood. This must start as part of generic training, and not left until arrival in a theatre of operations; in NI days, for example, we were very good at training and educating on the basis of the yellow card. ROE are an aspect of discipline, and as Napoleon remarked, discipline is the soul of an army.

  • ROE are supposed to be, fundamentally, permissive because they have a direct connection with the fulfilment of the mission. They must of course lie within the limitations of what is necessary, proportional, and legal; but the means – which include the ROE as much as hardware, money, people and so on – must be in balance with the ends. They must therefore be based on an analysis of the whole problem at issue: that is, our own mission, and the object of that mission. In most cases this will be the enemy, and the ROE must be appropriate to the enemy’s aims, objectives and capabilities. But in some operations the object of the problem may be ourselves, our public opinion, our political expediency; or in a complex PSO, the dispute between parties.

  • ROE must be flexible where the use of force for mission accomplishment and protection is concerned, because situations vary over time, and from place to place. But they must not be flexible where the defence of our soldiers’ lives is concerned. ROE that result in our own people being killed or injured, or that put our own people in court, are bad ROE. I do not mean to condone illegal behaviour here – a crime is a crime – but we must be clear about the need to secure and protect ourselves as the essential precondition to taking the fight to the enemy.

  • When dealing with a multinational force, the commander must understand what the partners will and will not do – understand it, because he will not be able to change it. I kept a chart of all the nations in MND SW in Bosnia, and in MND SE in Iraq, to remind me constantly who could do what. However by preparing the ground carefully in advance with national capitals, specific things can be negotiated on a case-by-case basis. As a rule of thumb, though, a coalition commander should never ask for anything from his partner contingents unless he knows the answer will be yes.


© Jonathon Riley, 2006. This is the text of a speech given at the RUSI, which later appeared as an article in the RUSI Journal