Issue 24 - Article 10

Public expenditure and humanitarian response

October 7, 2003
Karin Christiansen

This article outlines the familiar problems of ‘off-budget’ donor expenditure, and the impact of emergency response practice, on public expenditure systems. In particular, it looks at the benefits that could be had by thinking about Public Expenditure Management (PEM) when setting up humanitarian operations, so as to make them ‘budget compatible’ in the longer term, once the emergency phase has passed. The idea is not to subsume the humanitarian sector into development approaches: there are distinct functions and acknowledged roles for these different types of work. However, there are ways of building practical links between them that maintain the integrity of each, but from which both can benefit. This article sets out a way of ensuring that the financial issues around humanitarian work do not inhibit core government functions around public expenditure and budgets.

Screen Shot 2012-11-02 at 11.35.42 AM

The off-budget problem

Many governments in low-income countries that are receiving large amounts of aid have no clear picture of how much is coming in, or where it is going, let alone what is or is not working. The vast majority of this donor expenditure is delivered through projects, some of which are on-budget, most of which are not. Partly for reasons of independence, impartiality and neutrality, humanitarian aid is largely delivered through ‘off-budget projects’.

Evaluations of ‘project aid’, including humanitarian assistance, have drawn the following main conclusions:

  • Administrative costs are extremely high because of the multiplicity of different reporting and accounting requirements, including tied aid.
  • Spending is inefficient because it is dictated by donors’ priorities and procurement arrangements, not local priorities or accurate needs assessments.
  • Funding levels are extremely unpredictable, and there is a discrepancy between pledges, commitments and disbursements.
  • State systems are undermined or unsupported through special staffing arrangements and parallel structures.
  • Accountability mechanisms are designed to satisfy donors’ requirements, rather than those of domestic/beneficiary constituencies.
  • Emphasis is placed on the micro and technical aspects of interventions, with a failure to engage in the macro issues, including at times the political or military aspects of a situation.
  • It is difficult to sustain a positive impact beyond the short term, with high levels of reliance on donor funds.
  • The delivery of assistance can encourage corruption, fraud and rent-seeking, even at the extreme the development of parallel economies and interests. Independence from government does not obviate these problems.
  • Systems are absent by which to monitor and evaluate information in such a way as to feed back in real time and change practice. There is a lack of learning.

Getting ‘off-budget’ assistance ‘on-budget’, and trying to get donors and NGOs to develop systems that synchronise activities with government systems, has become a major challenge for governments (and donors). As projectised systems become entrenched, so too do the incentives to remain separate, both for donors and for those involved in projects, within and outside of government. Bringing financial flows ‘on-budget’ becomes increasingly difficult. In turn, fiscal discipline is limited, policy-making remains ad hoc and driven by external resource flows, and it is difficult to set up any systematic planning and implementation across government. The longer this continues, the harder it is to change.

Screen Shot 2012-11-02 at 11.35.18 AM

The idea

The suggestion that public expenditure issues need to be incorporated into humanitarian assistance seems counter-intuitive at first. The idea being put forward here is simple: all agencies should use a budget-compatible information-gathering format from the beginning of the flow of resources into a country. This is not the same as starting a ‘budget process’, nor does it imply a gain or loss of control by any party. Rather, it would facilitate the development of such a process in the future.

In practice this might mean:

  1. Adapting a standard budget classification to include likely humanitarian activities. Ideally, this would be based on the budget structure previously used in the country. If the existing classifications are problematic, one could start with an International Monetary Fund (IMF) classification. It would probably require some revision and modification as events unfolded.
  2. This way of classifying data could be used from the beginning of humanitarian operations as the standardised format in which all players present their activities.
  3. The coordination mechanisms and structures that are set up in such situations would then have a standardised format, which would greatly facilitate information collection and exchange between humanitarian actors. It could be seen as building on the work done by, for example, ReliefWeb, and could greatly assist reporting, accountability and learning.The choice of where to locate the responsibility for this function would probably have to depend on the situation. It would however be important to ensure that the gathering of such information does not bypass government.
  4. It would probably be useful if development aid to support public expenditure management was also started up earlier than is currently the case. This would facilitate continuous and iterative liaison between the developing management structures and humanitarian actors, to make sure that this method of gathering and sharing ‘resource data’ is serving humanitarian relief operations effectively.
  5. To enable this, all actors need to have a clear understanding of the distinct principles and objectives of humanitarian aid and development assistance in order to protect their different purposes.

Why is this important for humanitarian actors?

This kind of approach would offer several benefits for humanitarian actors.

  • Using budget-compatible information-sharing mechanisms within humanitarian relief operations should facilitate the move through the ‘grey zone’ before developmental aid flows emerge to replace relief.
  • The information that would be in this ‘budget’ is being collected and shared anyway, so why not use something standardised that will help for the longer term too?
  • Standardising classifications of information would make it much easier to work out what different actors are doing, avoid duplication of efforts and so deploy resources more effectively.
  • Such an approach would provide a potentially more ‘neutral’ vehicle for information on expenditure. This could be of particular value in contested environments and situations of conflict. Relief operations cannot afford to ‘avoid’ the national government once that government possesses even a vague sense of legitimacy. National government systems need to be part of the solution to humanitarian problems, even in the short term. Such an approach would provide a transparent and neutral tool for starting up such collaboration, and would help to reduce co-option and the political manipulation of funds.
  • The lack of capacity or engagement of crisis or post-conflict states is frequently either the direct cause of a humanitarian crisis, or results in a failure to respond and mitigate that crisis. While there is a consensus that the state is the only long-term mechanism for fulfilling humanitarian needs and obligations, state-building needs also to be seen as a crucial short-term activity. This is partly because the costs of badly coordinated, inefficient humanitarian aid delivery are high, and also because there are likely to be important lost opportunities in terms of better governance and domestic accountability. The benefits of incorporating financial information and developing budget-compatible systems early in the response to a crisis are likely to be high.
  • This problem would seem to be more acute in post-conflict situations in more developed states, like Iraq or Bosnia. The level of urbanisation, technology and complexity required to get water, sanitation food supply and health systems up and running means that ‘reviving’ such systems is not only the aim of a longer-term developmental process, but also a short-term humanitarian necessity.

Karin Christiansen is a researcher in the Centre for Aid and Public Expenditure (CAPE) in the Poverty and Public Policy Group (PPPG) at ODI.

The ideas in this article have been developed in consultation with members of the Humanitarian Policy Group at ODI.

Thoughts and comments on the practicability of such an approach by those with experience of operating in these environments would be greatly welcome, as would any expressions of interest in taking the approach forward. Please contact Karin at

References and further reading

Rob van den Berg and Niels Dabelstein, Iraq and Rehabilitation: Lessons from Previous Evaluations, OECD/DAC Network on Development Evaluation, 16 May 2003,

Lawson et al., General Budget Support Evaluability Study, ODI/OPM, 30 December 2002,

David Booth, ‘Poverty Reduction and the National Policy Process’, in Colin Kirkpatrick, Ron Clarke and Charles Polidano (eds), Handbook on Development Policy and Management (Cheltenham: Edward Elgar, 2002).

The Transitional Government Of Afghanistan, Analysis of Aid Flows to Afghanistan,


Comments are available for logged in members only.