Issue 13: Editorial – Codes of Conduct
by Nick Leader, Research Fellow, HPG June 2003

[This issue is also available in French: Échange Humanitaire No. 13]

Codes of conduct have become quite a phenomenon in the humanitarian system in the past few years. This edition of the newsletter carries pieces on the Code of Conduct of the Australian Council for Overseas Aid (ACFOA), the Code of Conduct for Humanitarian Agencies in Sierra Leone, the Principles of Engagement for Humanitarian Assistance in the Democratic Republic of Congo (DRC), and the NGO Field Cooperation Protocol. Other codes in recent years include the Red Cross/NGO Code of Conduct (perhaps the ‘mother of all codes’ in the humanitarian system), the People in Aid Code in the UK, the Sphere Project Minimum Standards, the Joint Policy of Operation in Liberia (JPO), and the Ground Rules Agreement in south Sudan.

Superficially these codes share a common idea: they are public statements of principles or standards of performance to which a number of agencies voluntarily sign up and against which each agency states it is willing to be judged. The Humanitarian Charter of the Sphere Project expresses this philosophy thus: ‘The Humanitarian Charter expresses agencies’ commitment to these principles and to achieving the minimum standards’. Or, as the Ground Rules declares in its preamble, it ‘seeks to define the minimum acceptable standards of conduct for the activities of OLS agencies and SRRA. [1]


There are many origins to the current emergence of codes. The following three, however, are perhaps the most important:

  1. The huge proliferation of humanitarian agencies working in conflict since the late 1980s. The original Red Cross/NGO Code of Conduct arose specifically from the concern of the established agencies in Sudan in the late 1980s – notably IFRC and Oxfam – that much of the work being done there was of low quality and that some kind of mechanism was needed to raise standards. The response to the Rwanda crisis and the subsequent Joint Evaluation reinforced understanding of the problem of over-proliferation of agencies and low standards.
  2. Growing awareness of the problems of working in the midst of internal war and the growing feeling by many, both within and without the humanitarian system, that aid could be ‘fuelling conflict’. Again the Goma camps were seminal in this respect.
  3. The ‘vacuum of regulation’ which confronted the hugely increased numbers of NGOs working in the midst of conflict and often in areas where the state had collapsed. There was no regulatory or protective environment either for the agencies or, more importantly, the civilian population. Indeed the flagrant abuse of the rules of war by parties to a conflict was the greatest part of the problem.

In this context accountability became a watchword, both among critics of humanitarianism and its supporters.

Accountability and NGOs

Accountability is by no means a straightforward concept. To complicate matters, the ‘accountability trail’ for NGOs is bewildering. Agencies are accountable to the people they are intending to serve. They are also accountable to their donors. They are also accountable to their own organisation’s charter or mandate and legislation governing charitable organisations, both in their home country and the country of operation.

This is complicated further in that they are also, in a sense, accountable to international humanitarian law. This is a grey but important area. Agencies are not signatories to the Geneva Conventions but there are parts of the conventions that cover humanitarian assistance, notably that it should be impartial and neutral. Authorities are indeed entitled to bar the passage of aid if they think it is not neutral. It could be argued that agencies are thus accountable to both principles of humanitarian action, such as neutrality, and to the governing authorities of an area to ensure the aid they provide is neutral and impartial.

Thus what agencies are accountable to is not an easy concept to define and there are potentially significant ‘conflicts of accountability’ for an agency between the various ‘objects of accountability’. This is a source of ongoing confusion. One of the favourite mechanisms for enhancing accountability has been codes, and this confusion has affected them too.

Typologies of codes

Since the original Red Cross/NGO Code of Conduct, codes have mushroomed. It is useful to contrast them across various axes. One contrast is country specific codes, such as the Ground Rules, the JPO, and the Sierra Leone code, versus universal codes such as Sphere or the Red Cross/NGO code. The country-specific codes are generally concerned with dealing with the specific problems of a particular operation; for example, the Sierra Leone code deals with the problem of armed convoys. Often they go through a process of inter-agency development during which only a few agencies (interestingly often the same ones) take the lead. They usually assume that agencies need to confront ethical problems and dilemmas with a united front as a way of avoiding manipulation. All draw heavily on the Red Cross/NGO code and all, like the Red Cross/NGO code, have little in terms of detailed protocols or guidelines as to, for example, what neutrality means in a specific context.

A second distinction is between general statements of principle and more detailed statements of performance standards. The best examples of detailed statements is the Sphere Project and the People in Aid code, both of which commit agencies to detailed standards of performance in a number of technical areas. Conversely, the Red Cross/NGO code is more a statement of general principle. And while the country specific codes may be quite detailed about a particular problem – for instance, armed escorts in Sierra Leone – they say little about anything else. An exception is the Ground Rules which commits Operation Lifeline Sudan and the relief wings of the factions to a detailed list of mutual responsibilities. How, or indeed if, general principles such as neutrality can be converted into practice in a specific context is an important problem. The Sphere Project, for instance, despite its technical detail, says very little about neutrality.

Another very important distinction could be described as ‘internal’ versus ‘external’. A number of the country specific codes have an explicit objective to influence the behaviour not only of humanitarian agencies but also the warring parties. The Ground Rules were signed by the SPLA and were intended to promote the SPLA’s respect for international humanitarian and human rights law. The Sierra Leone code workshop had as one of its two objectives ‘To ensure that the parties to the conflict recognise and observe the impartiality and the inviolability of humanitarian principles.’ The DRC code likewise is addressed to the warring parties and has been signed by them.

This is probably the most difficult area for codes and here they move away from self-regulation into the area of conditionality: both the JPO and the Ground Rules have elements of conditionality in them, though fall short of ‘full’ humanitarian conditionality. Research by ODI shows that this is the least effective area for codes. In short, humanitarian codes cannot fill the vacuum of regulation or impose regulation on warring parties; that can only be done by the great powers, if at all. (An upcoming RRN Network Paper will document this research in more detail.)

Codes in practice

Development and Opt-in

The development of, and opt-in to, codes are closely related. For country specific codes development is often hurried, pushed by a few agencies, and lacking in detail. The current code for the DRC has been pushed by some donors, notably ECHO, and one large agency, MSF, has refused to opt-in. On the other hand the Sphere Project learned from the experience of the Red Cross/NGO code and its project team went through a process of widespread and lengthy consultation with many agencies. Opt-in to Sphere is consequently widespread, though not so much with southern agencies. (For a more detailed analysis of Sphere, see Newsletter 12.) The Ground Rules too took six months to negotiate between OLS and the SPLA and as a consequence addressed many concerns of both organisations.

Development does not stop with the production of a code and many codes continue to develop after they are signed. The JPO, for instance, went through several versions after first being signed in 1996 (see RRN Network Paper 22). And the Sphere Project too will continue to develop after field testing.

Although codes are voluntary, there is often an element of compulsion to opt-in. As illustrated by the ACFOA piece, agency motivation in signing up can be partly defensive in that agencies are concerned with their reputation. Donors too are increasingly using signatory of a code as a criterion for disbursing money – for example, the Disasters Emergency Committee (DEC) in the UK with the Red Cross/NGO code, and ECHO with the DRC code.

This raises important questions about the independence of humanitarian agencies and the extent to which codes used in this way could have a negative impact. Indeed under such circumstances they might reduce innovation and experimentation and make agencies refuse to work if they fear they cannot uphold standards for good reasons beyond their control.


One problem with agencies signing up to codes is that there is too little thought as to what changes will have to be made in agency procedures as a result. This is particularly the case with shorter, vague and more general codes. More detail makes agencies think harder about what it means to sign up. A recent study of British agencies reported that the Red Cross/NGO code, a short and general code, is a statement which has ‘not been internalised by organisations and remains unused as a means of guiding and auditing their work.’ Agencies need to examine carefully what it means for them to sign up to a code, and develop a strategy for its implementation; implementation also needs to be properly linked with other initiatives concerned with increasing professionalisation, such as training.

So far it appears that evaluations have hardly made reference to codes of conduct. There is room for development here.


Probably the weakest area of codes is what happens when a signatory breaks them. If the essential nature of a code is that it is a public statement of principles or performance objectives against which an agency commits itself to be judged, then the consequences of breaking the code are important. One of the problems is that all of the compliance mechanisms are, as the ACFOA piece puts it, ‘complaints driven’. Signing up to a code is voluntary and the nature of the business means that agencies are loathe to report one another for transgression. This is reflected by the fact that compliance mechanisms are usually under-developed, and if they do exist are under-used.

The Red Cross/NGO code, for example, has only had a handful of complaints in its five year history and as yet has no formal complaint or compliance mechanism. The idea of an Ombudsman has been floated by British agencies as one way of addressing this problem. Again, formal external evaluations could have a role here.


As argued above, perhaps the most important motivation behind codes is that agencies are struggling with a ‘vacuum of regulation’, not just for themselves but for the way in which war is fought.

However, although codes have considerable promise in some areas they are problematic in others. Any ‘real’ code needs wide participation in its development and opt-in and so depends on a constituency with shared values and objectives. With a shared set of values, vagueness in a code can be acceptable; without it vagueness is a weakness and leads to confusion.

Codes also require individual agencies to develop strategies for implementation, mechanisms for reporting on the strategy, independent evaluation of the success of implementation, an independent compliance mechanism, and the threat of publicity if codes are broken. Given these pre-requisites codes will probably be most successful in improving technical performance.

Codes are, unfortunately though not surprisingly, probably least successful in terms of imposing regulation on warring parties. This is simply not something agencies can do. What they can do, however, is establish their own ethical position and it is in terms of developing common ethical or principled positions in specific circumstances that success is most mixed.

On the one hand it would seem that often the ‘conflicts of accountability’ are too great across the range of agencies involved in humanitarian work for them to share common positions. However, developing a common approach to the principles of humanitarian action in a specific context is very important if agencies are to reduce the likelihood that they will be manipulated by anti-humanitarian forces. To achieve this will require considerably more work in terms of thinking through how to put principles, as opposed to standards, into practice. And codes are only part of the answer.

[1] Sudan Relief and Rehabilitation Association, the humanitarian wing of the Sudan People’s Liberation Movement.