Fragile and failed states have been with us since we’ve had a state-based international order. But the interest of policymakers in such states took on a new life after 9/11. The events of that day, and subsequent terrorist attacks, made devastatingly clear just how dangerous failed states such as Afghanistan could be, not only to their own people, but to communities around the world.

Afghanistan and the neighboring tribal areas of Pakistan continue to demonstrate the threat posed by ungoverned areas to their citizens, their neighbors and the broader international community. Somalia has been a failed state since the nineties, and has recaptured the international community’s attention in recent years – not because of the ongoing humanitarian catastrophe in the country, but because it has become a base and haven for pirates preying on international shipping. And an unsuccessful bomber who probably received training in Yemen has been the catalyst for a surge of international interest in that fragile state.

So what is a fragile state? There’s now a substantial body of literature on such states. And every academic or agency appears to have their own descriptor – so apart from fragile statues, you also have states that are “weak,” “failing,” “failed,” “collapsed,” “at risk,” “precarious”, “vulnerable” or “recovering”. Some of these are alternate descriptions to “fragile” and some are encompassed within it. Even the term “fragile state” has come under criticism in recent years. Some scholars now consider the term both pejorative and analytically imprecise. They claim that fragility is not an either/or concept but rather exists along a continuum, and that it is highly context-specific and comes in a variety of economic, political, and social forms. But regardless of the specific conceptual formulation, these descriptors and analyses all point to some type of significant state failure or dysfunction.

The World Bank provides a good working definition, observing such states “share a common fragility, in two particular respects: State policies and institutions are weak in these countries: making them vulnerable in their capacity to deliver services to their citizens, to control corruption, or to provide for sufficient voice and accountability. They face risks of conflict and political instability.”

Predicting conflict in fragile states

One consistent theme is the strong correlation between state fragility and conflict. Not all states experiencing conflict are fragile (India is a good example, with a number of internal conflicts, and the conflict in Kashmir), but most of them are; and not all fragile states are experiencing conflict, but almost all of them of them are or recently have. The World Bank identifies 37 fragile situations in 2010 – and all bar a small handful are post conflict or conflict affected.

It shouldn’t be surprising that there is such a strong correlation. Many of the indicators for conflict are indicators for state weakness. Low and declining growth are widely recognised indicators for conflict – and low income and weak growth typically translate into lack of state capacity. A lack of state capacity usually results in an inability to mediate between competing interests. Low income also lowers the cost of rebellion, making it more attractive to would-be rebels.

And the converse is also true. Conflict invariably has a negative impact on economic growth. Resources directed to conflict are diverted from development. Conflict destroys the infrastructure needed for economic activity. And without security, development efforts are unlikely to take hold and have the desired effects.

While the link between fragility and conflict is by now widely accepted there is certainly no similar consensus about how this link plays out in practice. Much academic work is now devoted to mapping the various causal flows between fragility and conflict. And it’s the kind of academic work that has direct relevance to policymakers, particularly when it comes to prevention and its necessary accompaniment, prediction.

Policymakers focus on the linkages in their efforts to obtain early warning of fragile states that may slip into conflict, and in an effort to ensure timely and cost-effective responses.

There is no shortage of early warning in this age of proliferating NGOs, instant and widespread internet and satellite communications, and an awareness of the threat posed by failing states. In fact, there may be too much of it – the challenge for policymakers can be to determine which of the barrage of warning they receive is credible, and requires action, and which can be ignored. The related challenge is to tie early warning to effective early action – early action here means policy response by governments and international and regional organisations. Analytical and advocacy NGOs have the luxury, and the frustration, of being able to warn, but not being able to respond.

So what kind of early warning do policymakers have access to?

Very broadly, there are two types of early warning – qualitative and quantitative warning.

The early warning produced by my organization, the International Crisis Group, is a good example of qualitative early warning. The task of our analysts is to find out what is happening and why. They identify the underlying political, social and economic factors creating the conditions for conflict as well as the more immediate causes of tension. Our role is to warn, as early and effectively as possible, those who are able to influence a situation where the risk of new or renewed conflict has reached a dangerous threshold.

Crisis Group’s particular value-added in this respect is that all our reporting and analysis is field-based. At last count we had people on the ground from 50 different nationalities, speaking between them 49 different languages. They are steeped in local language and culture, getting dust on their boots, engaged in endless interaction with locals and internationals on the scene, and operating from 9 regional offices and 17 other locations in the field.

Crisis Group also produces the monthly CrisisWatch bulletin which summarises developments dur­ing the previous month in some 70 situations of current or potential conflict, assessing for each whether the overall situation has significantly deteriorated, improved, or on balance remained more or less unchanged. This is one of the few examples of very short term early warning in the public domain.

The challenge that this kind of qualitative early warning poses for policymakers is that its credibility and hence usability relies to a significant extent on the reputation of the external provider of such analysis. It is difficult, though not impossible, for governments to make big resource allocation decisions (ie whether to intervene to seek to prevent a looming conflict) on the basis of independent, non government, analysis. Of course, governments have their own analysts, but these often they won’t have the expertise, or the institutional freedom, of their non-governmental peers. There are ways to incorporate independent qualitative analysis into governments’ own analysis and planning, but the constraints will usually act to inhibit governments from using such external analysis as the predominant basis for their early response decision making.

This is where quantitative early warning comes into the picture. The advantage (in theory) behind quantitative analysis is that it relies on verifiable data, and hence provides an independent and transparent basis for making resource allocation decisions.

That’s the theory. The reality is more complex. Quantitative warning, relying as it does on statistical analysis, requires a model of conflict with quantifiable factors that can be measured, compared and analysed. But this conflict modelling is still more of an art than a science, despite rapid advances in the field over the last decade or so. The other challenge with quantitative warning is the timeframe of its predictive ability. It is much better adapted to highlight worrying trends than to identify with great specificity a likely tipping point into violence.

Quantitative theories used to be broadly, if simplistically divided into two camps – that of greed versus grievance – with the greed camp holding that economic factors were largely responsible for conflict, and the grievance camp blaming on inequality and political, ethnic and religious grievances.

The debate has been refined in recent years – concurrent with big improvements in the data – and now is more usefully characterized as one between feasibility and regime type, with the proponents of the feasibility thesis focusing on the conditions that determine the economic viability of rebellion, whereas the regime type proponents conclude that it is political institutions and not economic conditions that are the most powerful predictor of instability.

The doyen on the feasibility side of the debate is Oxford academic Paul Collier. His models have been developed and refined over the years, but the essence of his analysis is that the defining feature of civil war is the emergence and durability of a private rebel army, and under most conditions such organizations are likely to be neither financially nor militarily feasible. Civil war will only occur if a rebel organisation can build and sustain a private army. He and his co-authors go on to argue that “where insurrection is feasible it will occur, with the actual agenda of the rebel movement being indeterminate.” Their research shows that three factors in particular are important in demonstrating feasibility of conflict – namely low per capita income, slow economic growth, and large exports of natural resources. Further variables have been recently added to the model, namely whether a country is under the implicit French security umbrella and the proportion of its population who are males in the age range 15-29, and a weaker variable that mountainous countries are more conflict prone (see “Beyond Greed and Grievance: Feasibility and Civil War”, Paul Collier, Anke Hoeffler and Dominic Rohner, May 2008).

Low per capita income points to the incapacity of the state to maintain effective control over its territory. Both low income and slow growth can be interpreted as lowering the recruitment cost of rebel troops, and natural resources can provide rebel organizations with finance.

The attractiveness of this theory of conflict is that most of these factors can be quantified. And many civil conflicts over the past couple of decades can be readily explained by it – for example, Liberia, Sierra Leone, Angola, Congo, Sudan, Aceh, Niger Delta, and Cote d’Ivoire – but does not provide a robust explanation for all civil conflicts or for the lack of conflict in some states.

Collier has more recently looked at the links between democracy, development and conflict. He has argued that income level is the key factor in preventing violence during transitions to democracy. He believes that while the international community often favours promoting democracy as a solution to overcoming violent conflict, democracy also constrains the technical possibilities of government repression, and that this makes rebellion easier. Although the net effect of democracy is therefore ambiguous, he suggests that the higher is income the more likely is it to be favourable. He finds that whereas in rich countries democracy makes countries safer, below an income threshold democracy increases proneness to political violence.[i]

The regime type advocates take a different approach. The most sophisticated and sustained research on factors influencing motivation has been done by the Political Instability Task Force (formerly known as the State Failure Task Force), a panel of scholars that has worked since 1994 to collect and analyze data on political regimes and conflict around the world. Its work has focused particularly on regime type and quality.[ii]

The Task Force’s research postulates that regime type is overwhelmingly the dominant factor behind revolutions, ethnic wars, and adverse regime changes. However, the effect of regime type is not a simple function of the degree of democracy or autocracy. The starting point is that strong autocracies are rarely prone to conflict. Strong democracies are also not prone to conflict. It is certain kinds of partial autocracies and partial democracies that are much more vulnerable than other regime types, with the vulnerability depending on the patterns of executive recruitment and political participation under those regimes. A particularly strong contributor to instability is that of factionalism within the political process.

The taskforce model has four independent variables: regime type, infant mortality (as a proxy for poverty), a “bad neighborhood” indicator flagging cases with four or more bordering states embroiled in armed civil or ethnic conflict, and the presence or absence of state-led discrimination. They claim for this model an 80% success rate in identifying likely instability within a period of two years.

Interesting recent research has also looked at the linkage between climate and conflict. Studies by scholars such as Edward Miguel at Berkeley attempt to establish clear causal links between climate factors and conflict. While most studies claim that poverty (which can be intensified or induced by climate change) has suffered from questions of reverse causality – namely, whether conflict leads to poverty or vice versa – Miguel finds that drops in rainfall in Africa, clearly an exogenous factor not affected by conflict, produce drops in income, increasing the likelihood of conflict the following year by nearly half. He recommends pre-empting violence by targeting foreign aid to shore up incomes in regions where livelihood is affected by rainfall, thereby removing a short term trigger of violence.[iii]

The latest work in this field has moved beyond precipitation to looking at warming and finds strong historical linkages between civil war and temperature in Africa, with warmer years leading to significant increases in the likelihood of war. One recent research paper suggested a roughly 54% increase in armed conflict incidence by 2030 if current climate model projections are correct.[iv]

So that is a very quick run though of the current models for predicting conflict and instability. As mentioned earlier, the effectiveness of this type of quantitative analysis depends on the robustness of the models, and the quality of the data – and these are both continuing to evolve.

For policymakers seeking to establishes processes in which early response is less ad hoc and more systematic, perhaps the best course of action is to use quantitative analysis to identify a small group of fragile states at risk of violent conflict within a two year timeframe, and then incorporate external qualitative analysis to refine that list and determine the most appropriate intervention.

Policy approaches to fragile states

Another difference between quantitative and qualitative analysis is that while the former may tell policymakers when to intervene, it doesn’t give them much guidance on how. Good qualitative analysis is much better geared to inform policymakers’ interventions in particular fragile states.

So how should policymakers engage with such states? Given the variations between states, their problems, the tools available to interveners, the political will to intervene and all the other permutations, I won’t attempt to set out a menu of policy options here. However there are a number of guidelines that could usefully inform interventions in all fragile states.

1. Understand the problem.

This is perhaps a statement of the obvious – but it is salutary to understand how often the obvious is ignored when the international community intervenes. Far too often lessons painfully learned in earlier interventions are forgotten or ignored.

There is no checklist of appropriate policies for fragile states. What may have worked in Iraq for instance – such as international support for tribally based militias – is unlikely to work in Afghanistan. In fact in the latter country, the last five years have seen failed incarnations of the same policy on militias – first in the form of arbakai (tribal militias), then the Afghan National Auxiliary Police, then the Afghan Public Protection Force, and now the Coalition appears determined to repeat the failures of those initiatives with its latest effort, the Local Defense Initiative. So perhaps an exhortation to understand the problem is not quite as obvious as it seems.

[ad#Google Adsense (336×280)]

Policies have to be evidence-based. They have to build on a field-based understanding of the history, culture, political dynamics and region. This is where the work of organizations like Crisis Group is so important. We produced field-based policy reports. Our analysts are stationed in or near the countries they cover. They are steeped in an understanding of the country, its culture, politics, and the interests of the key players. They can and do travel around countries much more freely than embassy staff can. They usually have better access too. All of which is reflected in our analysis.

2. Recognise that prevention is better than cure, and that prevention does work.

There was an excellent report published in 2005 – the Human Security Report – which documented the trends in conflicts since the Second World War. (The 2009 issue of this report is forthcoming.)

Its headline statistic is an encouraging one, and perhaps counterintuitive – namely that there has been a 40% reduction in the number of state-based armed conflicts since the early 1990s. And, while there has been a small uptick in the number of state-based conflicts since 2003, when non-state conflicts are included (i.e. conflicts in which all both parties are non-state actors, such as rebel groups) there has been a continuing decline in the overall number of conflicts since 2003. There has also been a longer term trend decline in battle deaths (ie military personal and civilians killed in fighting.) There were only some 12,000 reported battle deaths in 2005 – less than any year since 1946.

Why has there been a decline in conflict, and what lessons can we learn from these trends? The Human Security Report posits a number of causes, such as end of the Cold War leading to a reduction in proxy conflicts, and the growth in the number of democracies. But it attributes much of the reduction to a surge in international conflict prevention and resolution activities in the 1990s, led by a reinvigorated United Nations. Between 1987 and 2008, the number of Special Representatives of the Secretary- General increased six-fold. UN peacekeeping missions – which play a key role in preventing renewed conflict – increased from four in 1990 to 15 currently.[v] The international financial institutions and donor governments and civil society have played a significant role with their efforts to address the root causes of conflict. The key message is that conflict prevention efforts, for all their failings and inadequacies, can make a real difference.

There are some examples of where these advances – seen, for example, in the rise in resources devoted to peacekeeping and conflict prevention work among national governments, stronger regional peace and security response mechanisms, the evolution of vibrant civil society engagement in conflict resolution and reconciliation initiatives, and advances at the level of international law – have paid dividends. In some cases, coordinated international engagement has been instrumental in shifting states affected by devastating civil wars onto the fragile road to transition. Countries such as Liberia, Sierra Leone, Burundi and Nepal continue to face considerable challenges, in economic development, institution building and professionalization of their public services in the wake of devastating civil wars. The painful and long-term task of reconciling societies damaged by war will remain relevant for some years. But these states also managed to execute a transition to post-war recovery that would have seemed inconceivable in the early 2000s.

Concerted efforts by regional organisations, such as the Economic Community of West African States (ECOWAS), have also played a decisive role in preventing violent outbreak in fragile states and in bringing states back to the fold. We saw some encouraging evidence of the impact of coordinated regional and international engagement in Guinea over the past few months. Effective regional and international actions in response to 2009’s military takeover have been instrumental in encouraging the country’s shift back to civilian rule from January this year.

Of course, there is no room for complacency. Climate change, the fallout from the global economic crisis (including falling commodity prices and reduced remittances from diasporas), the aftermath of recent fuel and food price shocks, and a likely fall in the aid and development budgets of rich countries are all likely to increase the likelihood of conflict in fragile states in the coming years.

Given this outlook, policymakers need to develop smarter and more cost effective interventions. And that being the case, we need to recognise that prevention is not only more effective than intervention after the bullets have started flying – but (and this should be music to policymakers’ ears) it is also much cheaper.

A 2004 study estimated that on average one euro spent on conflict prevention generates over 4 euros in savings to the international community.[vi] As with all such studies, there are a number of heroic assumptions involved – but not so heroic to render the key finding redundant, namely that the cost of properly targeted prevention is a lot less than the cost of conflict.

To give some concrete figures, though again, very much context specific: the former UK Foreign Secretary Jack Straw has estimated that the small preventive military deployment in Macedonia, stopping the slide to war there, cost the British taxpayer £14 million (24 m euros), while fighting the war in Kosovo, by contrast, cost Britain £200 million (350 m euros), and in Bosnia over £1.5 billion (2.6 billion euros). And those are just the military costs, not the costs of reconstruction.

3. Understand the regional context

Fragile state policies are generally geared towards individual states, and too often ignore the regional context. But as the experience of Africa and the Balkans and South Asia have amply demonstrated — conflict and state-failure usually have very strong regional dynamics.

Neighboring states can contribute to conflict in a number of ways. The most obvious is by being a party to the conflict. They can be active participants, for instance by being at war with the state in question, but more usually they are indirect participants, preferring to support proxy militias or fund rebel groups. Rwanda and Uganda were both in the Congo. Neighboring states can also provide a safe haven for rebels or spoilers, as Pakistan is doing with the Taliban, and Chad and Sudan are doing for each other’s rebels. They can funnel arms and supplies to governments or rebels (as some 11 countries are alleged to have been doing in Somalia, many of them being neighboring countries); or they can be more subtle in their destabilisation – as one could perhaps characterise Ethiopia’s role in Somalia, as it pulled its troops after its intervention, leaving a security vacuum and likely ensuring Somalia remains a failed state for many years yet.

As noted above, the Political Instability Task Force has identified “bad neighborhood” as a statistically significant risk factor for conflict, with bad neighborhood here being defined as four or more bordering states embroiled in armed civil or ethnic conflict.

So when it comes to strengthening or rebuilding fragile states, failing to address these regional dynamics will probably consign even the best-designed and most well-intentioned peacekeeping mission or development assistance package to failure as soon as the troops leave or the donor community’s generosity runs dry.

4. Commit the necessary political, financial and security resources

The key political resource is a commitment to stay the course. It takes many years to rebuild a state, and premature disengagement can very quickly destroy all the progress and the billions invested in rebuilding. Just look at Timor Leste, where premature disengagement allowed that country to fall back into conflict, with the result that in 2006 it was almost back to where it started after its violent rebirth in 1999.

Doubt among Afghans about the international community’s commitment to stay feeds insecurity there, and feeds patronage-based politics, and a willingness to do deals with the insurgent leadership – driven in part by their fear that the internationals will abandon them, as they have done in the past.

The sad fact is that the international community isn’t good at staying the course. Too often we adopt a formulaic approach – particularly to those states requiring large scale international intervention. The standard response is a four or five year commitment, in the form of largish peacekeeping missions to back up internationally mediated peace agreements or Security Council resolutions, some DDR (disarmament, demobilization and reintegration of combatants) and SSR (security system reform) followed by a rush of indigestible aid dollars, a flurry of institution building, and premature elections – followed too often and too soon by a withdrawal of peacekeeping troops and a drying up of aid money just when it is most needed and the absorptive capacity is able to beginning coping with it.

A rush to early elections can be particularly problematic. Holding elections – although good for political show business (and, in post-conflict peacebuilding contexts, much loved as an exit benchmark for governments anxious to meet their commitments and go home) – quite often has nothing much to do with democracy. Crisis Group was one of the first organisations to really make this point loudly and clearly, opposing as we did a rush to an early election in Bosnia in 1996 because we feared this would consolidate ethnic divisions which hadn’t had the chance to be counterbalanced by the development national secular political forces, or at least strong civil society institutions. The recent presidential election in Afghanistan, held when there evidently wasn’t the domestic institutional capacity to manage a credible and legitimate process, is another example.

We know that the period of transition to democracy is in many ways one of the most dangerous and fragile of all. This doesn’t mean that we should retreat from democratisation, but that we should rethink our priorities in the way we pursue it. The most important of all things to prioritise is the rule of law – often a very difficult challenge, but essential if democratic institutions are to take root and flourish.

When it comes to financial resources – the temptation is to do it on the cheap. This is the ultimate false economy. Conflict imposes horrendous financial costs, not to mention the devastating human toll. And failure to get it right after conflict significantly increases the risk of a return to conflict.

It has been estimated (in 2004) that civil war in a low-income country costs that country and its neighbors on average 42 billion euros in direct and indirect costs. That is for a single conflict.[vii] To put that figure in perspective, the worldwide aid budget in 2004 was 60 billion euros.

So we should properly fund effective prevention, and thereby reduce the costs spent on peacebuilding post-conflict reconstruction. And when we do have to fund peacebuilding, we shouldn’t do it on the cheap – as the likelihood is that this will increase the risk of the country falling back into conflict, thereby requiring yet further expenditure on peacekeeping and stabilisation forces and the follow-up peacebuilding.

Finally there are the security resources. A secure environment is a necessary but not sufficient condition to strengthening a fragile state. It’s sobering to realise that rebel groups and militias responsible for terrorizing a country or large parts of it are often very weak and brittle. They often survive and prosper because there are no capable forces to oppose them. When confronted with effective forces, they will often collapse – as happened with the rebels in Sierra Leone (confronted first by the mercenaries of Executive Outcomes, and subsequently routed by a few hundred British special forces); militias in Bunia, DRC (confronted by EU/French forces in Operation Artemis); or rebelling soldiers recently in Timor Leste (confronted by Australian troops and police).

More importantly, security guarantees can be critical in deterring future spoilers. The ability to rapidly deploy an effective military force can be sufficient to ensure that the need to deploy will not arise. The commitment can be an over-the-horizon one – as the UK provided to Sierra Leone – but it needs to be real and credible.

But sometimes a small security commitment will not be sufficient, and large scale sustained commitments are required. Even then, an effective and timely commitment may repay the investment many fold in terms of the avoidance of future costs. Such commitments are expensive and politically difficult – particularly where there is a likelihood of casualties – but as always the cost of failing has to be weighed against the cost of the commitment. Back in March 2002, when there were some 4,500 NATO peacekeepers stationed in Kabul, Crisis Group called for the peacekeeping force to be expanded to 25,000-30,000 and deployed around the country – a call subject to much criticism and ridicule from some Coalition governments, and NATO itself, as being greatly in excess of what was needed or feasible. Some eight years later, NATO is on track to have 150,000 troops deployed by September 2010.


Fragile states need lots of security and lots of development if they are to become viable and effective states in future. And it’s important to remember that in a globalised world, it’s not just capital and trade that travel the world – terrorism and extremism can also be exported, or nurtured in fragile states, and ill-gotten funds laundered by them.

But while that provides an incentive for international engagement, it’s important that it not become the sole justification for engagement. Fragile states inflict untold misery on their citizens and on neighboring countries.

When it comes to addressing state fragility we increasingly understand what works and what doesn’t – even if we are a long way from having all the answers yet. We certainly know enough to know that we aren’t doing enough to assist such states. Earlier and better targetted assistance to fragile states would dramatically improve the quality of life for hundreds of millions of people, which makes it a tremendously worthy goal in its own right.

Speech by Nick Grono, Deputy President of the International Crisis Group, to Institut Royal Supérieur de Défense, Brussels, 27 March 201

Source: International Crisis Group