Principles and Practice of Risk Analysis and Risk Reduction

John N. Clarke

In this contribution, I will offer a personal view of the value of risk analysis and risk reduction, referred to here as John N. Clarke Principles and Practice of Risk Analysis and Risk Reduction. First, I will briefly discuss existing approaches and trends, in a manner that I hope complements what was already discussed in this morning’s session. Second, I will outline a number of principles that should be considered in developing any early warning system and finally, I shall offer a few preliminary thoughts on linking early warning systems to conflict prevention – which, of course, along with humanitarian preparedness represents the main reason for engaging in early warning analysis in the first place.

Existing Approaches

A recent analysis, published out of the Liu Institute at the University of British Columbia , The Human Security Report, has found that the number of armed conflicts around the world has declined by more than 40% since the early 1990s. The report goes on to suggest that the net decline can be explained by an increase in international activity focused on peacebuilding and conflict prevention.[i] This finding is supported by the results of a forthcoming book by Michael Doyle and Nicholas Sambanis, (forthcoming, Princeton University Press) in July 2006 of this year in which they describe a ‘peacebuilding triangle’ in which international assistance can support local capacities in establishing a stable peace.[ii]

The bad news is that of the remaining armed conflicts, the majority now take place in sub-Saharan Africa as do more conflict related deaths than in the rest of the world combined. The Human Security Report also highlights the fact that conflict exacerbates the conditions that gave rise to it in the first place – creating what has been described as a ‘conflict trap’.[iii] There is, therefore, still a need for early warning analysis at regional, national and, sometimes even at sub-national levels.

There is a vast literature on early warning, including statistical approaches, most notably, the Collier/Hoeffler model, work by Ted Gurr and Monty Marshall on peacebuilding capacities, and more case based approaches to the analysis of conflict.[iv]

Others have focused on the dynamics of particular cases and warning signs which either went either unrecognized or unheeded.

Still others have incorporated a comparative dimension to this work, looking, for example, at urban violence. Ashutosh Varshney, for example, compares three conflict prone and three non-conflict prone cities in Gujarat , India , to see what effects different forms of inter-communal association have on conflict. In particular he emphasizes the role of associative interaction – for example via trade or sporting associations in reducing the risk of conflict. This work is of interest both in the context of early warning and has implications for the development of preventive strategies, though the extent to which these findings can be generalized beyond Gujarat , remains open.[v]

In a basic sense, there are two types of indicators with which early warning analysis is concerned:

·        first, structural indicators, which very often consist of socio-economic factors such as GDP annual growth rate, infant mortality, dependency on primary exports et cetera;

·        second, what are commonly described as ‘proximate’ indicators, or those which are ‘closer’ to the onset of conflict), and;

·        finally, triggering events (or the match that lights the structural and proximate ‘fuel’).

By the time the level of a triggering event is reached, a country has likely moved beyond the point where prevention and preparedness are realistic policy options – rather conflict management will be the main (if not only) response possible. For this reason, early warning analysis tools should focus on structural and proximate indicators.[vi] There are a range of different ways in which one might devise an early warning system, and by and large, the format of the analysis depends upon the function for which it is designed. In the following section I therefore want to outline a number of principles that I think could usefully be kept in mind in your discussions on what sort of early warning system is needed and how best to go about creating it.

Principles for the Development of an Early Warning System

Clarity of Purpose. What is it you are trying to provide an early warning of and what is the target audience for that early warning? The answer to this question depends in large measure on the type of crisis one is trying to predict. Analyses must not only establish where crises are more likely to occur, but also, help shape strategies aimed at preventing and preparing for crises. Academic research certainly must inform this analysis, but our purpose as policy-makers is not academic but pragmatic – that is, to give us a good idea of where and why a crisis is likely and a preliminary idea of what measures could reduce the risk of conflict flaring. Ideally, early warning analysis provides information in a form that is easily digested by senior decision makers, whether governments, non-governmental organizations (NGOs), or international governmental organizations (IGOs).

A ‘good enough’ model will suffice: Broadly speaking, in analyzing social phenomenon, we rarely if ever achieve the level of predictability that is available in the natural sciences. Even the most robust statistical analyses of causes of conflict often leave a substantial amount of the variability in the dependent variable (conflict) unexplained. As a result, it is useful to think about early warning in probabilistic terms and aim only to establish where a humanitarian crisis is more likely, not inevitable. The standard by which early warning analyses are judged has often been unrealistically high – if a method does not yield the date, time and identify of the key actors in a crisis, it is often dismissed as useless. The point, of course, is that one does not need perfect predictions to formulate both preparedness and preventive strategies. A probabilistic account is sufficient, if the threshold for undertaking humanitarian preparedness and preventive measures is set at a sufficiently low level. We do not need an absolute date, place and time of a future crisis in order to know that the analysis should trigger an immediate policy response. The point here is that we need an early warning system that is ‘good enough’ not infallible.

Subsidiarity: Any centralized analytic and aggregating capacity needs to be balanced with engagement at a grassroots/field level. Analytic focus and responsibility can therefore usefully be ‘pushed down’ or devolved in a manner that ensures contact with events and analysts who are ‘on the ground’, allowing for both a macro and a micro focus.

Additionality: As indicated above, there is a vast literature on early warning. Given the number of systems, principles and sources already available, anything new must serve a function not already addressed by those that already exist. What will a new system accomplish that is not already achievable within the scope of what already exists? How will the system envisaged link with existing analyses/approaches? Any new system must build on that which exists already and be tailored to the realities of the particular region.

Simplicity: The key to constructing a workable, policy relevant methodology is balancing the level of detail in the analysis with the need for simplicity in its use. In order to ensure balanced analysis, a wide range of analysts at differing levels (local, regional, national and different backgrounds/disciplines) should be engaged. Early warning templates must therefore simplify complex concepts without undermining the integrity of the analysis.

Linking Early Warning to Programming that reduces Risk

The first point to be made about the link between early warning and prevention is that 43.6 percent of countries emerging from conflict, relapse into conflict within five years.[vii] For these countries, conflict is effectively cyclical and therefore early warning or risk analysis is of great value in many post-conflict settings. Risk analysis should therefore take place in both pre and post-conflict settings, otherwise we are not going to deal with those countries that – statistically – we know are most likely to relapse into conflict. This carries with it a second advantage and that is that in the immediate aftermath of conflict, many of the political barriers to preventive programming that exist in a pre-crisis country are often eliminated. This, arguably, can create a window of opportunity to undertake programming that would otherwise be politically impossible.

There is, I think, a danger of over-intellectualizing peace-building/ prevention/conflict transformation et cetera. One can forever debate the meaning of these terms but from a policy making point of view the question is whether such fine distinctions help us to make better policy and programming at the field level. The measure of the value of such debates is the extent to which they improve the caliber of the policy developed and programming undertaken.[viii]

Preventive or risk reducing responses can be grouped into two categories:

·        The first has commonly been described as a ‘do no harm’ approach – or trying to ensure that in our programming and policy we avoid making a bad situation worse.[ix] This leads, for example, to particular food distribution practices, and particular camp structures and management practices.

·        The second form of prevention is more aggressive, whereby in addition to ‘doing no harm’ we try proactively to reduce risk. Here there are no shortage of programs and approaches that seek actively to reduce risk – for example, political mediation – as in the Axworthy/Gaviria mission to Peru in 2000 or Norway ’s mediation between the parties in Sri Lanka .[x]

This second type of prevention requires a holistic approach to ensure that different responses are coordinated in a manner that ensures that at a minimum, different programmes are not working at cross purposes and ideally, that when combined the total impact of all programming is greater than the sum of the individual parts. This was reflected some five years ago in the Secretary General’s June 2001 Report on the Prevention of Armed Conflict, wherein he emphasized that preventive strategies should not only be initiated ‘at the earliest possible stage of a conflict cycle in order to be most effective’[xi] but also that they should incorporate:

“short-term and long-term political, diplomatic, humanitarian, human rights, developmental, institutional and other measures taken by the international community in cooperation with national and regional actors.”[xii]

Central to this process of linking risk analysis and risk reduction programming is its institutionalization: The quality of the analysis produced depends ultimately on both the expertise of the analysts and their ability to work with other experts in developing the analysis, but institutional arrangements also matter. Early warning systems must be anchored within existing institutions with the mandate to not only identify risks but with clear mechanisms to address them. Independent early warning systems can easily become marginalized from the responsive processes of other parts of an organization. Equally, however, if one seeks to integrate early warning analysis and programming into all areas within existing institutions, there is a danger that the analysis is never aggregated and assembled in a manner that looks at the whole picture. There is of course a middle option that is to try and balance the two approaches – a small dedicated group that seek to mobilize and interact with other parts of any given institution. This middle way can help to strike the balance between ensuring someone is responsible (and that something gets done) and that what gets done, is integrated into the broader structures of an institution or organization. The role of such an analytic unit is therefore as a fulcrum, assembling risk analysis with a view to leveraging institutional action in response to that analysis.

Fifth and finally, programming must – to the greatest extent possible - be entrenched within domestic social and political structures, and sustained by domestic actors both in planning and implementation.

About the author

John N. Clarke (PhD, Cambridge University ) is Post-doctoral Fellow, Yale University , and Field Coordinator, Office of the Resident Coordinator, Sudan . He has previously served as Team Leader for the 2006 UN and Partners Workplan for Sudan and in both the Bureau for Crisis Prevention and Recovery (UNDP, New York), and in the Early Warning and Contingency Planning Unit (OCHA, New York). From 1999-2000, he served as a Policy Advisor to Canada's then Minister of Foreign Affairs, the Hon. Lloyd Axworthy. Along with Geoffrey R. Edwards, he is co-editor of Global Governance in the Twenty-First Century (Palgrave, 2004).

For quotation, refer to this article as follows:

John N. Clarke Principles and Practice of Risk Analysis and Risk Reduction, in An Early Warning System for Sudan, edited by Dominique Wisler and Eltayeb Haj Ateya, Khartoum, Coginta, 2007

| Terms of Use |  Privacy Policy |  © Coginta - All rights reserved