The limits and scope for the use of development assistance incentives and disincentives for influencing conflict situations - Case study: Rwanda




Paris, September 1999

By Anton Baare, David Shearer, Peter Uvin with contribution from Christian Scherrer

Sponsored by Denmark


1. There is no shared understanding of the dynamics of the Rwandan conflict amongst OECD members. This divergence inevitably limits the potential to use ODA as part of jointly agreed and systematically applied system of incentives and disincentives. In the current situation, only generic agreement on the direction in which to exert political pressure through ODA may be achieved within the donor community.

2. The findings and conclusions of this study should be seen in the context of the main conclusion of the JEEAR (1996) and supporting studies (Scherrer 1997, Uvin 1998): in the period leading up to the genocide the priorities of aid were largely unrelated to the challenges on increasing polarisation, inequity, hatred and violence that Rwanda was facing. In the post-genocide period, the scope for the use of incentives and disincentives (with or without specific conditions) in Rwanda is determined by a number of context specific factors. First, the genocide has had a profound impact on the social and political fabric of the country; this impact is hard to grasp for outsiders, who have never encountered such situation before. Second, the legitimacy of the donor community to exert influence is seriously hampered by its inaction before and during the genocide, as well as by some of its actions after the genocide. Third, security is such an overriding concern for the GoR, that, in the absence of a credible scenario to improve security through an international effort, the use of negative ODA conditionality is unlikely to be effective. Fourth, with remnants of the former government persisting in their genocidal attitudes and policies, the options and scenario's available are limited and there is little scope for a negotiated settlement to the conflict. Fifth, the polarisation and trauma in Rwandan society are such that any incentive strategy - even one that is not conditional or confrontational - is bound to require much time in order to have any visible impact.

3. Still too often, it is assumed that donors are somehow external to a conflict. However, whether through historical -, trade -or military relations, or even simply through the granting of ODA, all donors become part of a conflict. This study begins from the assumption that all ODA produce incentives and disincentives whether these are deliberate or recognised or not. This is probably a fortiori the case in the polarised circumstances of acute conflict. The issue is then not whether to create (dis)incentives or not, but rather how to be aware of the (dis)incentives impact of ODA interventions, and how to manage them in such a way as to promote conditions and dynamics that are propicious to peaceful conflict resolution.

Summary of recommendations

  • Act early. The longer one waits to use incentives and disincentives, the costlier their use will become, and the longer it will take to see results. Once the situation has come as far as Rwanda's, with an almost total destruction of the material and social fabric of society, it becomes extremely hard to change the dynamics of polarisation and violence. What this truly amounts to is that concerns of peace, justice, human rights, racism or inequality should not be on the agenda only after major violence has taken place, but always. This implies a rather fundamental reorientation of concepts and attitudes in the development community, for which the time is now ripe. The incentives/disincentives framework may be a useful tool for helping to operationalise these insights.

  • A policy of critical longer-term engagement is likely to be more effective than negative political conditionality. The use of political disincentives in Rwanda should not a priori be ruled out. But under the current circumstances their effectiveness is doubtful, squandering scarce leverage tools and political capital in the process. Positive incentives and dialogue may be more useful.

  • Promote 'transparency in diversity'. Rather than trying to develop unanimously agreed-upon - and thus often watered-down - positions and attempt to get all donors to behave the same way on each specific occasion. It is more important - and realistic - to agree on shared goals than on shared strategies, for the simple reason that there exists more than one path to the same goal. There are different manners in which donors can seek to produce incentives and disincentives. Countries adopt these tools in function of their assessment of the government, which itself is the result of a host of other factors. Donors may find it useful to think of simply being more candid and transparent about their assessments, concerns, and goals. The resulting clarity will benefit mutual understanding between the donors. It will also send clear signals to the recipient. These signals may not be co-ordinated, but they will be less ambiguous and watered-down.

  • Donors may wish to consider the important role that informal leadership plays in conflict situations. Some countries may be more willing and able than others to take certain risks, to innovate, or to engage parties to the conflict in dialogue.

  • Explicitly consider trade-offs and incentives and disincentives as part of a longer-term strategic framework for involvement. Such trade-offs can only be considered within a longer-term framework, preferably based on a GoR-donor dialogue.

  • Resist a tendency to compartmentalise conflicts to fit bureaucracies and funding mechanisms.

  • Be willing to support activities on the interface between development aid and security to more systematically address the need to establish viable national security frameworks in post-conflict situations, including professionalisation of the military and police.

  • International intervention needs to be directed to all sides in conflicts in a manner that is consistent with the objectives sought. In acute conflict situation, the scope for the use of ODA is often limited and primarily useful to create incentives or disincentives to governments only.

  • Strengthen and institutionalise mechanisms for reflection and self-criticism within donor agencies. Existing review and evaluation mechanisms have their part to play, but in conflict related situations , more flexible and rapid-feedback mechanism are required that can impact in earlier stages in the planning process.

  • Promote institutional memory. Important documents, often funded by and available in the same agency, are not known or used. This is a problem for all development aid, but its consequences can be especially dramatic under conditions of conflict. Such institutional memory can also provide an infrastructure for reflection and self-criticism.

  • Many donors have created conflict units since the Rwandan genocide (and to no small degree as a result of it); however, there is a need to mainstream conflict impact analysis and integrate it in human resource development and objective-oriented planning and management tools.

  • Develop multilateral rapid response financing mechanisms that allow for fast action to intervene in conflict. In Rwanda, a first step would be to initiate a joint review and evaluation of the UNDP Trust Fund mechanism. An evaluation would have a two-pronged focus: (i) technical and managerial performance; (ii) lesson learned (strength, weaknesses and opportunities) on the conceptual aspects of the trust fund mechanism in the context of rapid response in (post) conflict situation.

  • Realise the inevitability of initial donor-driven support in key sectors such as justice and security framework reform. If donors are convinced that certain actions are important and the government does not prioritise these actions, but does also not oppose donor investment in them, donors must be willing to pay for them.


4. This study examines the way donors, after a severe violent conflict, use their assistance to create incentives and disincentives to reduce violent conflict and build a durable peace. Donors may well use different terminology than incentives and disincentives - or even not examine their aid in these terms. The task of this analysis is to promote fresh understanding and discussion of these matters.

5. The impetus for this study came from the Joint Evaluation of Emergency Assistance to Rwanda (JEEAR). Published in 1996, it examined the international response to the Rwandan genocide and its immediate aftermath. It concluded that while human rights conditionality was "preached" it was not "practised." (p. 32) This point is supported by other studies (Scherrer 1997; Uvin 1998) that the priorities of aid in the early 1990s were largely unrelated to the challenges of increasing polarisation, inequality, hatred, and violence Rwanda was facing at the time. Thus important opportunities to use aid to induce a response away from increasingly violent conflict through the strategic use of incentives and disincentives were missed.

6. Many study team discussions began by clarifying what was meant by the terms incentives and disincentives. The study team found that the term 'conditionality' was more readily understood and used by practitioners and policy makers alike. Conditionality, especially in its usually discussed form of negative conditionality, consists of the threat of withholding aid (and, if necessary, the execution of this threat) in order to force recipients to modify their behaviour. As such, conditionality is only one possible tool among many others within an incentives/disincentives framework. Incentives are inducement processes involving the offer of a reward by a sender in order to promote a particular action or response by a recipient; disincentives seek the same goal by increasing the costs of non-desired behaviors. (Cortright 1997). Both incentives and disincentives then may be conceived in a conditional or in an unconditional manner (i.e. with or without reciprocity requirements).

7. We also found that in speaking about 'influencing conflict' in our interviews, a range of underlying assumption needed to be made explicit: was the goal to 'resolve' conflict, 'manage conflict' or 'transform conflict' (Lederach, 1995). The terminology reflects quite different ideas on issues such as the role of outsiders, the desired outcomes, or the degree to which one is prepared to reflect on the unintended and unforeseen negative effect of one's actions. In some cases, we will refer to these unintended signals and outcomes as disincentives. Generally, the role of outsiders in transforming or resolving the conflict in Rwanda is limited; only Rwandans themselves can create the desired outcomes.

8. At some points in this study, we will briefly return to the period before the genocide, for donors have played important roles in Rwanda for thirty years already. Yet, the focus of the study is on the period from July 1994 to the present. It has been argued - among others by some members of this study team - that this time limitation undercuts the relevance of this study, for it excludes analysis of the way the aid system has been an important part of, if not complicit to, the processes that eventually led to genocide. Rwanda's crisis also had an impact beyond its borders and has taken on a major regional dimension. As a result, this study encompasses donor responses that affect Rwandan populations inside and outside the country in the period July 1994 to December 1998.

9. The conflict in Rwanda remains uniquely violent, intense, and destructive. Many argue that civil war and genocide are still ongoing (although they have abated somewhat in recent months). Human rights violations continue to be widespread, bringing daily suffering to hundreds of thousands. The racist and extremist rhetoric that caused the genocide is still alive; the challenges of justice are stil enormous; and reconciliation has hardly begun. Levels of poverty are still significantly higher than they were even in.1994, when they had already increased sharply. This dismal situation has to be compared to the truly gigantic nature of the challenge faced by aid agencies - and, foremost, by the people and the government of Rwanda. It can be argued that, given this context, the progress achieved is substantial.

10. The paper is organised the following way. Section 3 presents a brief overview of the major events during the period under study. Section 4 examines how bilateral donors have sought to employ incentives and disincentives so as to have different impacts on the conflict. It begins with a general overview of aid and then focuses on the more specific sector and project levels. Sections 5 and 6 discuss the two cross-cutting issues of donor co-ordination and donor coherence. Section 7, finally, synthesises the insights of the study and proposes some general lessons.

=A9 OECD 1999