To what extent does empirical evidence confirm or question the value of conflict early warning and response for effective practice by regional organizations? This article presents a brief overview of existing key EWR mechanisms and analyzes if, and under what conditions, these mechanisms might be a useful peace and security promotion tool for regional organizations. It looks at three regional and subregional organizations--the African Union, the Economic Community of West African States/Economic Community of West African States Monitoring Group in West Africa, and the Intergovernmental Authority on Development in East Africa that have established such conflict EWR mechanisms. Until now, these tools have not been adequately implemented or fully used. The principal reason for this is not a lack of sufficient EWR data. Instead, regional organizations often fail to respond in time to prevent an emerging violent conflict because of weaknesses of the organization and political disagreements within the organization. Keywords: early warning and response, regional organizations, African Union, Economic Community of West African States, Inter-governmental Authority on Development, peacebuilding.
AMONG THE MANY ASPECTS IN THE DEBATE ON BROADER CONCEPTS OF global governance, two largely unrelated desires can be identified: First, regional organizations are increasingly requested to provide security by engaging in the prevention of violent conflict and in peacebuilding. (1) Second, since the mid-1990s, conflict early warning and response (EWR) has been conceived as a means of preventing violent conflict in order to protect people's lives. (2) Partly on the insistence by and with the assistance of donor organizations, some regional organizations, especially in Africa, are now beginning to use EWR as a peace and security instrument to prevent crises. This comes at a time when the methodologies of EWR have improved. After a decade and a half of experience, we raise the question in this article whether both of these trends--to implement EWR and for regional organizations to use this tool to prevent conflict--have improved the security of the people. Our research questions are the following:
What are the strengths and weaknesses of the existing EWR mechanisms?
Have they been put to appropriate use in predicting and preventing violent conflict by regional organizations?
What is the experience of regional organizations in implementing EWR mechanisms?
Can regional organizations capitalize on the most recent progress in EWR research?
Our analysis enables us to present preliminary results on two separate fields of inquiry and offer conclusions on their value if combined as in the case of regional and subregional organizations in Africa.
Early Warning and Response Mechanisms: How Do They Work?
Our hypothesis is that the predictive capacities of conflict EWR mechanisms have greatly improved over the past two decades. (3) However, they still suffer from two deficiencies: First, the underlying theories (or, at least, hypotheses) about causal chains toward violence and the role of small events are not always spelled out in EWR models, which are either based on simplified rational choice models or on statistical findings from large-n analysis. The lack of focus on small events is additionally due to a disconnect between the local level (where the majority of violent conflicts take place and where monitoring systems vary a lot or have not systematically been established) and the center of attention of EWR models on global or macrodata. Furthermore, empirical evidence suggests that the link between warning and response remains weak. Response is often lacking, despite clear warning signals. The current conflict in Darfur, for example, was not acted on in a timely fashion, but not because of a lack of information on the emergence of the conflict. All the indications of a major conflict were known. Similarly, the dangers of violent conflicts and wars in the former Yugoslavia, in Rwanda, or in the Iraqi invasion of Kuwait were recognized before the killing started. As a preliminary conclusion we argue that the improvement of early warning is needed, but this alone will not result in closing the warning-response gap. Already in 1997 Alexander L. George and Jane E. Holl argued in their contribution to the Report of the Carnegie Commission on Preventing Deadly Conflict that there exists a warning-response gap and that "the design and management of early warning systems should be intimately connected with the task of responding to warning. (4) This warning-response gap has not changed fundamentally since then. In this article, we analyze why this remains to be the case with regard to the three regional and subregional organizations. We argue that this is largely due to several structural deficiencies, among them the still existing disconnect between decisionmakers and the EWR that is partly due to a "cognitive distance" with regard to the localities of the conflict. (5) Furthermore, even now decisionmakers are often unwilling or unable to act before the emergence of a crisis. In the concluding section, we detail several deficiencies and identify areas for improvement in the establishment and management of EWR as well as in the regional organizations that use EWR.
Conceptualizing Early Warning and Response
Early warning and response can be defined as "systematic data collection, analysis and/or formulation of recommendations, including risk assessment and information sharing," and early response normally "occurs in the latent stages of a perceived potential armed conflict with the aim at reduction, resolution or transformation." (6) EWR tries to estimate the magnitude and timing of risks of emerging threats; it analyzes the nature of these threats and communicates warning analyses to decisionmakers. (7) Early warning systems for the prevention of violent conflict are latecomers compared with their application in such fields as intelligence, military reconnaissance, or humanitarian emergencies. Early response mechanisms are even more recent efforts to close the gap between early warning and early action.
A crucial, yet so far underreflected, issue is the question of who is going to be warned and who is supposed to act on this warning. Is a "recognized authority" (e.g., a regional organization) ideally situated to be the primary addressee of such a warning? Little thought is given to warn those who are about to be attacked. (8) The underlying assumption of most early warning systems is that international actors will take over responsibility as protectors as soon as adequate information is being processed. (9) This assumption, however, has so far not been confirmed in practice. Accordingly, Casey Barrs proposed "to focus more effort on a warning capacity within the killing grounds." (10) Such an approach could assist in overcoming the gap between early warning and early response.
Efficient EWR systems can tackle various threats to human security such as: (1) wars and armed conflict; (2) state failure; (3) genocide and politicide; (4) other gross human rights violations; and (5) humanitarian emergencies caused by natural disasters. In this article, we focus primarily on the related threats to the first three categories for which we use, mainly in accordance with the Political Instability Task Force (PITF), the umbrella term "political instability." These are also the types of conflict in which a prevention role is expected from regional organizations. EWR mechanisms in this sense are a part of an overall crisis prevention architecture and are "intended to detect rising tensions headed towards violent conflict." (11)
Categorizing EWR Systems
Despite growing scepticism in the policy and donor communities, research and publications on EWR have experienced an upsurge within the past few years. (12) Review studies use different ways of categorizing the broad spectrum of EWR models. We find Monty G. Marshall's taxonomy most useful because it focuses on the aims of the models, making them more comparable. He classifies twenty-one early warning models into three types:
Conditional and causal models deal with empirical evidence for causal interference between independent variables and violent conflict or political instability;
Predictive models try to forecast the outbreak of violence in a time span of one to five years. They focus on selected variables and process indicators or event-based information;
General risk and capacity models are used to rank countries from weak to strong related to social problems, political conflict, and poor state performance. (13)
In our compilation (Table 1) we take up Marshall's categories, albeit with two specifications: (1) in order to keep the vast amount of general risk and assessment models under control, we divide them between those that aim at rankings or performance ratings and those that aim at targeted intervention; (14) and (2) we include in-depth investigative research and intelligence as an important additional category--a qualitative component that is regarded as highly valuable by practitioners as well as country and area specialists. In a review of EWR mechanisms, we list five major models: (1) five predictive models (mainly run by the PITF); (2) a dozen institutions that rank and rate states according to their risks and capacities; (3) about twenty efforts to integrate risk and capacity assessments into early response models; (4) several private companies, nongovernmental organizations (NGOs), and government agencies that offer or use investigative case study research; and (5) intelligence for early warning. (15)
Selected EWR Models, Tools, and Mechanisms
What are the assumptions underlying the different early warning projects? Which methodologies are used? Within these five categories, we have singled out one prominent example each in order to illustrate how these models, tools...