Full text of DOI:10.1111/1467-7717.00150

Disasters, 2000, 24(4): 291–313
What Happens to the State in Conflict?: Political
Analysis as a Tool for Planning Humanitarian
Assistance
Lionel Cliffe
Robin Luckham
Institute for Politics
Institute of
and International Studies Development Studies
University of Leeds
University of Sussex
It is now part of received wisdom that humanitarian assistance in conflict and post-
conflict situations may be ineffective or even counterproductive in the absence of an
informed understanding of the broader political context in which so-called ‘complex
political emergencies’(CPEs) occur. Though recognising that specific cases have to
be understood in their own terms, this article offers a framework for incorporating
political analysis in policy design. It is based on a programme of research on a
number of countries in Africa and Asia over the last four years. It argues that the
starting-point should be an analysis of crises of authority within contemporary
nation-states which convert conflict (a feature of all political systems) into violent
conflict; of how such conflict may in turn generate more problems for, or even
destroy, the state; of the deep-rooted political, institutional and developmental
legacies of political violence; and of the difficulties that complicate the restoration of
legitimate and effective systems of governance after the ‘termination’ of conflict. It
then lists a series of questions which such an analysis would need to ask — less in
order to provide a comprehensive check-list than to uncover underlying political
processes and links. It is hoped these may be used not only to understand the political
dynamics of emergencies, but also to identify what kinds of policy action should and
should not be given priority by practitioners.
Keywords: conflict, aid policy, politics, the state, governance.
Introduction
It is now part of received wisdom that peace-making and humanitarian assistance in
complex political emergencies (CPEs) require an informed understanding of their
political dimensions. All too often international agencies, donors and NGOs have
plunged in on the basis of an inadequate understanding of what are, after all,
profoundly political processes. The unintended, and even counter-productive,
consequences of such interventions have been amply documented in many studies,
some of which argue the thesis that humanitarian relief does (de Waal, 1997) or, at
least may (Anderson, 1996; Duffield, 1994), fuel conflict.
©
Overseas Development Institute, 2000.
Published by Blackwell Publishers, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.

292
Lionel Cliffe and Robin Luckham
The long experience in many conflict situations in Africa of UN and OAU
special envoy, Mohamed Sahnoun, leads him to see another advantage of analysing
context: that it is the premise for work with local people: ‘Effective relations with the
indigenous peoples of a country depend on understanding, ideally starting with a
better analysis of the political and social dimensions of conflict’ (1998: 88).
Mary Anderson (1998: 140) stresses that ‘emergency aid usually ignores the
causes of the emergency that prompted it’, and may ‘focus on "getting things back to
normal" … (when) it is the conditions of "normalcy" that give rise to the emergency
in the first place’. Others go on to argue that such an analysis of the roots, dynamics
and legacies of conflicts is just as essential during the phases of rehabilitation and
recovery if conflicts are not continually to recur. The boundary between states of war
and non-war is porous, and there are far too many cases of ‘return wars’ to allow for
complacency.
What is needed is an analysis that is both normatively grounded in humani-
tarian principles and based on an accurate appraisal of uncomfortable political
realities: a humanitarian Machiavellianism, so to speak. Specific case-by-case
analysis is essential since there are presently so few generalisable policy ‘lessons’, ‘no
calculus that can be counted on to work in different circumstances’ (Sahnoun, 1998:
92). The implication, which is certainly supported by the COPE work, is that what can
be learned and passed on is not a set of conclusions about the nature of conflicts and
ways of building peace or delivering assistance, but rather a method for analysis that
allows implications for policy and practice in a specific context to be generated. Thus
we endorse the conclusion from one recent collection of thoughts and experiences of
the great and the good in this field:
the need to understand as fully as possible what the realities are, the need to
avoid making untested assumptions or relying on superficial information but
to comprehend respectfully and seriously the interrelationships of the various
forces and actors in the play, and most of all to comprehend the situation of
the country that is being targetted to receive the tender mercies of
intervention (Moore, 1998: 6).
This article will strive to illustrate concretely how policy and practice can
and do benefit from such an awareness of the political context. These insights are
derived from the investigations carried out under the COPE programme, which was in
turn premised upon our initial conceptual framework for ‘analysing complex political
emergencies and the state’ (see Cliffe and Luckham, 1999). That framework has in
turn been further elaborated and amended in the course of the research, and is also
offered here in what is an improved (if abbreviated) form.
We started out with a check-list of different dimensions of the political
context that required to be taken into account, to that extent reflecting some earlier
attempts to develop practitioners’ handbooks for humanitarian action in war (see, for
example Minear and Weiss, 1991). A more up-to-date set of guidelines on ‘what relief
agencies need to know’ (Le Billon, 2000) usefully builds on those contributions to
analysis that have been labelled the ‘political economy of war’ and seeks to make that
perspective accessible to practitioners. Our approach has some similarities, although
we concentrate more on the political factors, whereas Le Billon dwells more on the
economic. Nevertheless, we are conscious of the inherent dangers of listing structural
components at the expense of an adequate understanding of process. Hence this article

What Happens to the State in Conflict? 293
is based on four linked themes, under each of which we explore the manner in which
conflicts unfold, transform the political realities which gave rise to them and create
new possibilities both for ongoing conflict and for building peace.
The first section of our analysis considers how the problematisation or de-
legitimation of the state generates violent conflict. The second explores the variety of
ways conflicts themselves reshape politics, the economy, social institutions and the
state, including (in some cases) the demise of the latter. The third outlines the
distinctive political and institutional legacies of conflicts, emphasising in particular
the need for an appraisal of these legacies to ensure that peace-building and
reconstruction strategies are sufficiently realistic. The final section concludes by
making the case for a strategic approach to peace-building, social and economic
reconstruction and reconstitution of the state itself, that seeks to make use of existing
balances of political and social forces and build on national and grass-roots initiatives
rather than working against them. In addressing each of these themes, we will seek to
offer insights from the COPE experience both in developing a methodology for
analysis and to point to some of the specific, and occasionally generalisable,
conclusions from our case studies.
The problematisation of the state and the political
origins of violent conflicts
In a previous article (Cliffe and Luckham, 1999) we argued that CPEs often arise
from failures of governance. But these failures in turn are symptoms of profound
authority crises in contemporary nation-states, with global, national and local
dimensions. Political analysis is needed at several different levels in order to develop
appropriate policy responses.
First, we contend — contrary to recent questioning of the relevance of ‘root
causes’ — that one should start from an adequate understanding of the historical and
structural antecedents of violent conflicts (see Figure 1).
The conflicts in the Horn of Africa, for instance, need to be contextualised in
relation to many centuries of state-formation, conquest, imposition of centralised
authority and surplus extraction, in the face of considerable ethnic, religious and
social diversity; being kept in motion by shifting balances of power and poverty, and
linked in turn to changing international alliances, long-distance trade, arms flows and
foreign assistance. The Ethiopian revolution of 1974, the rule of the military Derg and
the wars of the 1980s in Eritrea and Tigray represented both significant departures
from this historical pattern and important continuities. Similarly one may see the
current Ethiopian government’s experimentation with ethnic federalism as a creative
(if in certain respects problematic) attempt to overcome these historical legacies;
while the war with Eritrea stems from failure (on both sides) to come to terms with
the same legacies. Likewise the conflict in Sri Lanka can be traced back to the way
the colonial and post-colonial states reinforced exclusive conceptions of political
community and citizenship; the complex relationships between anti-colonial
nationalism, ethnic politics and state development; and the emergence of social
differentiation (and political violence) within the dominant Sinhala majority, as well
as between the latter and the country’s minorities.

294
Lionel Cliffe and Robin Luckham
Origins
(structure)
Political/policy
processes
(agency)
Precipitants of
violence
Logic of war itself
(perverse or
(historical
unanticipated
contingency)
consequences)
Global
• Global
• Changing nature of • International and • Perverse effects
economic
external policy
intervention:
regional military
intervention,
proxy wars
of humanitarian
intervention and
conflict resolution
• International
funding of military
activity (by TNCs,
diasporas, aid
etc.)
dislocations
• Global political
changes (e.g.
end of Cold
War)
conditionality,
democracy
• Arms
promotion
proliferation
(especially small
arms)
(replacing previous
support for author-
itarian regimes)
• Communi-
cations, media
• Technology
including
• Spilling of conflict
across national
boundaries
warfare
National state, national economy
• Economic
• Shrinking revenue
base and capacities
of state
• State repression • State promotion of
decline, uneven
development,
impoverishment
• Artificiality of
state; poor fit
with nation
of protest and
dissent
non-state
(paramilitary,
militia) violence
• ‘Asset transfers’:
both state elites
and their
• State crises and
policy failures
• State fracture or
collapse
• Declining legitimacy • State loss of
of predatory,
authoritarian
regimes
monopoly of
legitimate
violence
opponents acquire
vested interests in
war
(‘criminalisation’)
Civil and uncivil society
• Economic and
social injustice
• Marginalisation
of minorities
• Identity politics:
• Privatisation,
criminalisation of
violence
• Parallel war
economies
politicisation of
ethnicity by the
state as well as
against the state
• Mobilisation of
popular discontent:
war as drama of
social exclusion
• Strategies and
vested interests of
warlords
• Ethnic cleansing
• Arming of revolt
from below (by
marginalised
• Legacies of
earlier violence
• Livelihood and
survival strategies
of combatants
groups, child
soldiers, etc.)
Figure 1 The production and reproduction of armed conflict
The nature of these authority crises and how they articulate conflict varies
greatly from case to case. They do not invariably result in state ‘failure’ or ‘collapse’.
Sometimes to the contrary the problem may be functioning but repressive or
exploitative states. But in either case what is crucial is how the state’s legitimacy
comes in question. Hence we take issue with recent analyses which contend that
failing states and political disorder constitute rational or adaptive responses to
changing global conditions (Reno, 1998; Bayart et al., 1999). They may ‘work’ (as
Chabal and Deloz, 1999, put it) for predatory political elites, but usually not for those
who bear the brunt of impoverishment and violence. Moreover one should resist the
temptation to read history backwards by starting with the conflicts, and then list all
their possible ‘causes’. Nevertheless, one may provide a rough and ready diagnostic

What Happens to the State in Conflict? 295
check-list of factors (see also Figure 1 which categorises these somewhat differently,
in terms of whether they are structural, arise from political or policy decisions or
emerge from particular historical contingencies). These include:
•
Colonial and post-colonial state formation and its impact on underlying social
diversities, especially where there is little correspondence between state and
nation, or where colonial powers have reinforced ethnic or other differences in
order to govern and create reliable instruments of state coercion (for example,
through the recruitment of ‘martial races’ into colonial armies, often with
disastrous consequences for the post-colonial state, as in Uganda).
•
•
Global economic transformations and their impact (via deteriorating terms of
trade, debt burdens) on economically vulnerable states: reducing their capacity to
provide their citizens with the benefits of ‘development’, to fund redistributive
policies which might alleviate discontent, or even to assure basic law and order.
The ‘silent violence’ of underdevelopment, poverty and inequality, although the
crucial issue remains how and when this is transformed into overt violence.
Certainly all the conflicts considered in our study have arisen in part from strong
grievances about past or present deprivations, even if war-makers have
sometimes capitalised on the latter for their own more self-interested goals. In the
parlance of current debates about whether conflicts are driven by greed or
grievance (Berman and Malone, 2000), it is the interaction of both which is
critical.
•
•
•
The adoption of exclusive rather than inclusive conceptions of national
citizenship; along with differential access to state, military and economic power
on an ethnic, religious, regional or class basis.
Patterns of social exclusion (of ethnic and religious minorities, less-developed
regions, the poor, women, subaltern classes, underemployed youth); and above
all how such exclusion has become politicised.
The structure and institutional design of the state itself and whether it sharpens
political polarisation (like Sri Lanka’s post-independence winner-takes-all
electoral system, or the majoritarian and one-party versions of democracy
practised by the Obote governments in Uganda) or diminishes it.
•
•
•
Authoritarian or despotic governance and absence of non-violent mechanisms for
expressing dissent, making political and economic demands and resolving
disputes (democracy deficits may even contribute to conflict in some democratic
systems as well, like Sri Lanka or Colombia). In short, excluded groups turn to
violence because they lack more legitimate means of securing their rights.
The de-institutionalisation, patrimonialisation or indeed ‘criminalisation’ of the
state, for instance in some sub-Saharan African (SSA) countries, like the Congo,
Liberia or Sierra Leone, where predatory political and economic elites turned the
state into a mechanism for their own private enrichment, and in so doing left it
hollowed out or destroyed.
The fracturing and incapacitation of states through divisions and conflicts within
the political elite and within the military (coups and mutinies have often been
precursors of conflicts, as in post-independence Uganda or Sierra Leone) —
above all when this has encouraged challenges to the state’s monopoly of
violence.

296
Lionel Cliffe and Robin Luckham
•
The historical entrenchment within the polity and social formation of political
violence itself: one of the more robust findings of the statistical literature is that
the best predictor of conflicts is previous conflicts, as can be seen in the recent
history of countries such as Colombia, Northern Ireland, Ethiopia or Angola.
The political goals of armed groups and the circumstances under which they
organise, mobilise support, acquire arms, and turn to violence rather than other,
non-violent forms of protest are clearly crucial, yet are surprisingly neglected in
some recent analyses of conflicts. There are, for example, differences between
guerrilla forces with political goals like national liberation or social justice,
armed groups with exclusionary ethnic agendas and purely opportunistic or
predatory warlord forces stepping into the political vacuum opened up by the
disappearance of the state (see also the typology in Cliffe and Luckham, 1999:
35–7). ‘New wars’ or ‘postmodern conflicts’ increasingly blur such variations,
for instance when groups like the SPLA in the Sudan lose their initial idealism
and begin competing for resources and political space with tribal militias and
warlords. Yet the differences continue in many cases to be important as we shall
show below.
•
•
There are often complex regional and sub-regional considerations, such as the
porousness of national boundaries, the funding and arming of revolt by
neighbouring states and the failure of regional mechanisms for conflict-resolution
— though these are perhaps even more in sustaining conflict (see below) than in
initiating it.
For our purposes what matters is how these structural conditions become
authority crises and lead to violence. One may trace the state’s and the regime’s
descent into predation and violence: as the economy declines, the state’s revenue base
shrinks; its legitimacy fades; its reliance on coercion increases; and at the same time
its monopoly of the means of violence comes under challenge. The crucial turning-
point is often when state violence starts to make the state weaker and less secure, not
stronger. At the same time, state decline is usually reinforced by the disengagement of
citizens from states they no longer regard as legitimate or feel they can count on for
protection from political or criminal violence. Another turning-point comes when
disaffected groups are able take up arms and begin to challenge the state’s monopoly
of violence.
Yet none of this is inevitable. Of course it is not easy to reverse the legacies
of history, and there may not be much national or international policymakers can do
about structural constraints like vulnerability in international markets. Nevertheless
political and policy choices have to be made, and can make a difference: both to the
political and economic conditions which breed violence; and to the chain of events
which bring it about. In Somalia, for instance, it was the Siad Barre government’s
mismanagement of the economy which eventually produced an acute fiscal crisis of
the state; and it was its manipulation initially of pan-Somali nationalism and
subsequently of clan and sub-clan rivalries, which led the country to war with
Ethiopia, and later to factionalisation and disintegration of the state. The policy
failures were international as well as national, including Great Power intervention in
the conflicts in the Horn of Africa; Western political and economic support for the
Barre regime, as well as its withdrawal after the end of the Cold War; the failure of
international financial institutions to use their leverage to secure economic reforms;
the lack of international or African support for democratic transition during the

What Happens to the State in Conflict? 297
crucial period when the Barre regime was falling apart; in addition, of course, to the
well-documented perverse effects of international intervention during the 1990s.
It is for this reason, we argue, that CPEs result in part from failures of
governance, not only by national governments, but also by international financial
institutions, aid agencies, NGOs and others who influence policy choices (see Figure
1). For it is often the case that different political and policy choices could have
resulted in different, less violent resolutions of the same structural problems.
Countries like Tanzania with greater ethnic diversity than Somalia, and a similarly
weak state and poor economy have largely avoided violent conflict, in part we would
argue, because of the greater legitimacy of the post-independence leadership and the
political choices made by it. Of course major mistakes were made in Tanzania, above
all in economic management, but what has so far distinguished it from many other
states has been the leadership’s ability to learn from these mistakes and alter policies
accordingly.
In a previous article (Cliffe and Luckham 1999: 33) a distinction was made
between three broad types of governance failure:
•
•
•
development policy failures;
absence of or defects in democratic process; and
failures in conflict management.
To these we might add, fourth, failure to assure the basic security of citizens
from human rights abuses, crime and physical violence (what O’Donnell calls ‘the un-
rule of law’), as well as from the silent violence of poverty and inequality. We will
not repeat that analysis here, but will reiterate the need for donors and NGOs to
consider the potential impact of their own policies and programmes on actual and
potential conflicts. Under what circumstances, for instance, have donor-sponsored
economic-adjustment programmes contributed to the social tensions breeding conflict
(as arguably in Sri Lanka and Somalia — although in the latter it was the Barre
regime’s ability to subvert the economic reform process that was damaging), and
when has it not? When is external pressure for political reform and human rights
actually effective, and when has it merely goaded political elites into raising the
political stakes through violent action against their opponents, as in the sequence of
political events leading to the Rwandan genocide? This approach of identifying
contextual preconditions for violent conflict is also explored in White and Cliffe in
this issue.
Policy choices are never made in a political vacuum. Power and interests
matter, and need to be explicitly taken into account if proposed reforms are not to be
purely symbolic or indeed counterproductive. Predatory or repressive regimes like the
Barre government in Somalia, the Abacha government in Nigeria or the Mobutu and
Kabila governments in the Congo either resist reforms or find limitless ways of
subverting them to their own advantage (namely, the manner in which some African
governments have managed to convert privatisation and market-oriented economic
reform into new channels for patronage and private enrichment through the state
(Reno, 1998)).
Even governments like the current administrations in Sri Lanka or Colombia,
with some real commitment to reform or to ending conflicts, may be hamstrung by the
political process (in Sri Lanka political polarisation within the majority political
community has been a recurrent obstacle to constitutional and political reform), or by

298
Lionel Cliffe and Robin Luckham
the structure of vested interests (in Colombia those of the military and paramilitaries,
of the drug cartels and, arguably, of the United States). The two countries also
illustrate our earlier point, that the historical and structural antecedents of conflicts
cannot be ignored. Any negotiated settlement of the conflict in Sri Lanka would have
to recognise and find ways of reconciling the interests of the majority and minority
communities; while in Colombia no peace process would be sustainable that did not
seriously address the problems of rural poverty and social exclusion — even if, as we
argue below, the conflicts themselves have created new interests which also need to
be factored into political calculations.
Neither can one assume that international promotion of political and
economic reform is politically neutral. This was glaringly obvious during the Cold
War, when international assistance and military support directly promoted the foreign
policy interests of major powers. Either they directly fuelled conflicts, as in the Horn
of Africa or Central America, or they contributed indirectly to incipient CPEs by
propping up corrupt, inefficient and repressive governments, as in Liberia, Somalia
and the Congo.
During the post-Cold War era there has developed a subtler political
economy of international assistance and ‘humanitarian’ intervention. Many studies,
including our own, have documented the tensions between humanitarianism and
national interest, the distortions created by the bureaucratic and fund-raising
imperatives of aid agencies and NGOs and the lack of co-ordination between donors
with conflicting political priorities (for example, de Waal, 1997). Moreover, old-
fashioned international realpolitik is by no means dead, although it assumes multiple
forms in a multipolar world. In addition, regional interventions by neighbouring
countries have played an increasingly important role, both in triggering conflicts and
in sustaining them afterwards, as in northern Uganda, Sierra Leone, the Horn of
Africa (Cliffe, 1999) or Kashmir.
Time and again the international community has failed to respond in a
timely, co-ordinated or politically informed manner to impending or actual
emergencies. Partly this may reflect the inadequacy of early warning systems, as
several recent international reports and analyses have suggested. But even were early
warning systems based on the analysis we propose above more widely available, they
would have two major drawbacks.
In the first place, it is not very difficult to list the relevant precursors of
conflicts, or to come up with lists of countries which are vulnerable; indeed political-
risk analysis is widely used by foreign investors for just this purpose. But it is far
more difficult to predict the onset of particular conflicts. The Congo, for instance, was
a CPE waiting to happen for most of its post-independence history according to
almost all the criteria considered above, and the real issue perhaps is why the country
did not fracture sooner. Misguided foreign support for a central government which
was already known to be corrupt, inefficient, repressive and highly resistant to reform
— and which had already ceased to administer the greater part of its territory or
control its own armed forces well before war actually broke out — bears a major
responsibility both for its state of suppressed crisis and for the subsequent upheaval.
Second, some hard questions need to be asked therefore, about the
international community’s own role and ability to react appropriately to looming
emergencies. The problems lie partly with the structure of the international
community itself, the difficulties of reconciling the conflicting interests of different
donors and international agencies and the potentially perverse effects of foreign

What Happens to the State in Conflict? 299
interventions. But they also reflect the difficulties of defining appropriate forms of
international pressure where national governments are determined to resist reform,
manipulate external assistance or deliberately pursue policies which increase the risk
of conflict, not because they are irrational but because their political priorities
(namely, regime survival at all costs) are different from those both of the international
community, as well as (all too often) from those of the majority of their own citizens.
The political and economic dynamics of conflict
Recent analyses of conflict concur on two central propositions. First, violent conflicts
almost invariably transform the political, economic and social realities which gave
rise to them. Second, the nature of conflicts has changed significantly, especially
during the past decade or two: in certain respects they may be considered ‘new wars’
(Kaldor, 1998) or ‘post-modern conflicts’ (Duffield, 1998).
Our own studies document the transforming potential of conflicts, even if we
do not believe this consigns analysis of their political origins to irrelevance. In Sri
Lanka, for instance, 20 years of political violence have made the divisions within and
between the majority and minority communities so deep that constitutional and
political reforms which might have been sufficient to protect the rights and satisfy the
political aspirations of minorities two or three decades ago, may no longer be
adequate. In Rwanda, war and genocide not only redefined and polarised Tutsi–Hutu
relations, they physically liquidated most political moderates who might have
reinforced the political middle ground, brought members of the diaspora Tutsi
community to power, made it extraordinarily difficult to construct workable
democratic institutions and set off further conflicts throughout the Great Lakes region
and the Congo. In Somalia, conflict splintered what had previously been regarded as,
ethnically and linguistically, the most homogeneous country in sub-Saharan Africa. It
sharpened and reconfigured clan and sub-clan rivalries, liquidated the state and its
armed forces, and brought powerful groups into being who had a vested interest in
war, or at least in a state of non-peace. Later in Somaliland a quite different dynamic
of reconstructing political institutions from below was also to emerge.
The consequences sometimes ascribed to conflict may also be partly the
product of antecedent conditions. For instance the collapse of the state and
informalisation of the economy in the Congo and Sierra Leone, or the politicisation of
the clans in Somalia began well before the outbreak of open hostilities. Indeed, the
distinction between ‘war’ and ‘not-war’ can be just as elusive at the start of conflicts
as at their end.
How fundamentally the nature of contemporary conflicts has changed, is,
however, rather less certain. It is certainly true that how we understand and analyse
conflicts has been transformed by recent theoretical advances. In addition, the
international community’s policy responses to conflicts have changed significantly.
The openly partisan Major Power funding, arming and intervention in Third World
wars of the Cold War era has apparently ceased, and been replaced by the theory and
practice of ‘humanitarian’ assistance and intervention — and this in turn has altered
the political dynamics of conflict and conflict resolution.

300
Lionel Cliffe and Robin Luckham
Yet whether and how conflicts themselves have been transformed requires
rather more consideration. ‘New’ wars, so it is argued (Kaldor, 1998), have four main
characteristics:
•
•
•
•
They are simultaneously a response to the dislocations of globalisation and a
reflection of global interconnectedness (through flows of arms, finance, diasporas
and refugees).
They have developed their own distinctive political economies and incentive
systems linked to parallel war economies, which transcend existing states and
national economies.
The boundaries between the public and the private spheres, state armies and non-
state military formations, warfare and criminal activity have become increasingly
blurred.
Wars are no longer fought to achieve substantive political goals such as national
liberation or social justice; more and more they are driven by narrower forms of
ethnic mobilisation and identity politics.
All these assertions have validity as generalisations and each has important political
and policy implications. For instance, an obvious priority is concerted international
action to break the global connections (be they small arms flows, international
diamond markets or more controversially diaspora support of insurgents) which
sustain warfare. Conflicts cannot be dealt with in isolation, but require co-ordinated
regional approaches, as in Central America or West Africa. Conflict resolution and
demobilisation of former combatants should be based on realistic appraisal of the
incentives making people fight. More attention should be given to the real politics of
ethnic mobilisation, including how, by whom and for what purposes ethnic and other
identities are mobilised in conflicts.
Yet not all conflicts match this characterisation of new wars in every respect.
For example, in the wars fought by the EPLF and TPLF/EPRDF in Ethiopia and
Eritrea the armed movements had well-defined political goals, were highly disciplined
and sought to build support among civilians rather than terrorising them. It is arguable
that reform or liberation wars are becoming less frequent, and that war itself has
corrupted and divided many a liberation movement such as the SPLA in the Sudan
(but see Johnson, 1998, for a less-negative view).
Policy prescriptions require detailed understanding of the dynamics and
impact of individual conflicts, including any departures from the generalised patterns
mentioned above. These remain extremely varied in their political origins, economic
ramifications, military organisation, dynamics and impacts. Below we summarise
some of the more important variations.
•
The political and other goals of the contending groups and how far they have
changed in the course of conflict itself. Is the goal the seizure of power (as in
Taylor’s ‘coup from the periphery’ in Liberia); the reform or reconstitution of the
state; or secession from it? To what extent has there been a shift, as in the Sudan,
from initial ideological political goals (liberation, reform, minority rights) to
narrower agendas like ethnic mobilisation, control of oil or personal enrichment?
The mode of warfare (conventional, guerrilla, warlord, terror). How far does it
match recent characterisations of ‘new’ or ‘post-modern’ conflicts in terms of the
•

What Happens to the State in Conflict? 301
targeting of civilians, criminalisation of violence, or blurring of local, national
and international distinctions?
The nature and organisation of the various armed groups, both those contesting
the state (liberation movements, guerrillas, ethnic or clan militias, warlord
followings, criminal gangs) and those supporting it (official armed forces, police
and intelligence bureaucracies, paramilitaries, mercenaries, state-sponsored
militias and gangs).
The fate of the state during conflict. Is it challenged in all national territory or
only part? Has it been fractured by conflict, or is it relatively uncompromised?
How far has it ceased to carry out its normal functions (law and order, welfare,
economic management)? There is clearly an important difference between states
able to carry out their normal development and other functions outside conflict
zones as in Sri Lanka or Uganda, and those where war has consumed the entire
state, as in Bosnia, Somalia or Sierra Leone.
•
•
•
•
How far have power relations within the state been transformed, for instance
through the entrenchment of extremist factions or parties as in Bosnia or the
Sudan; through the enhanced influence of the military, security bureaucrats,
intelligence networks, paramilitaries and other groups with access to the means of
violence, as in Colombia; or through the spread of corruption and criminal
activity as in Liberia.
There are crucial variations in the privatisation or informalisation of violence and
the fracturing of the state’s own military and security bureaucracies. Breakdown
of military discipline, military and police harassment of and extortion from
civilians, the blurring of distinctions between government and rebels (as with
Sierra Leone’s sobels) are both a frequent consequence of war and further reduce
the state’s ability to deal with CPEs, to maintain law and order or to defeat rebel
forces.
•
•
CPEs also typically transform power relations in political and civil society as
well as within the state, by enhancing the influence of groups controlling the
means of violence or enjoying access to emergency aid, diaspora funds, scarce
commodities and other assets created by conflict.
An important aspect of these societal transformations is the impact on gender
relations, including the differential impacts of conflicts on women and men and
the varying political agency they enjoy as combatants or peacemakers;
depending, for instance, on the character of the armed movements and whether
they have a transforming political agenda, like EPLF in Eritrea, are patriarchal
like the mujihadeen in Afghanistan, or are simply predatory as in Liberia.
CPEs tend to give rise to and be sustained by parallel war economies, usually
transcending national boundaries, their scope varying from conflict to conflict.
These undermine the state by creating assets not subject to its control, removing
its revenue base and hence reducing its ability to provide services, including
basic physical security. In the most extreme cases like Somalia, the informal
economy has entirely replaced the national economy. In others like the Congo,
the fragmentation of the latter into a series of regional informal economies
actually preceded conflict, of which it was an important cause.
•
•
In a number of recent conflicts — those in the Congo, Bosnia, the Horn and
Sierra Leone are good examples — the state has in effect been subsumed in wider
regional conflict formations transcending national boundaries. These have been

302
Lionel Cliffe and Robin Luckham
sustained in different ways by mutual interventions by states into each others’
conflicts, by a variety of non-state military formations operating across national
boundaries, through regional arms economies and through regional peace-
keeping efforts (as in West Africa).
A variety of interventions by international organisations, donor agencies, relief
bodies, NGOs and peace-keeping forces shape the course and political impact of
conflict. In some CPEs (like Eritrea), local political actors have been better able
control the flow and impact of relief than in others. In some cases (like Somalia)
external intervention has become part of the problem, in others (like the Rwandan
genocide) it has been flawed by failures of political will, and in others (like
Mozambique) it has contributed to the resolution of conflict. But it is never
politically neutral, and donors and relief agencies cannot escape the responsibility
to make informed political choices.
•
Political, institutional and developmental legacies of
conflicts
The point is often made that conflicts are liable to re-ignite — as in southern Sudan
— since the distinction between war and peace is often blurred. Conflicts sometimes
live on as what Outram terms ‘violence following conflict’ (1999). Even when peace
is apparently restored, there are many traumas to be healed, as well as institutions
destroyed by conflict needing to be rebuilt if reconstruction is to be made sustainable.
Hence building peace and restoring effective and accountable governance
should be based on as accurate an assessment of the political, institutional and
developmental legacies of conflict as possible. Some of these legacies are immediate
and concrete, including destroyed infrastructure, displaced populations, dispossessed
peasants, victims of rape and torture, and are obvious priorities for relief agencies.
But here we focus on the more subtle and corrosive impact of conflict on states and
political communities, on the rules and practices of politics and upon institutions, or
what others term ‘social capital’.1 Among these are:
•
‘Governance voids’ resulting from the problematisation or collapse of the state:
the absence of an effective government presence in all or part of national
territory; breakdown of law and order; limited or no government capacity to
provide public goods (water, health, education). These vary both within and
between countries, and sometimes there is an effective absence of governance in
weak states without or prior to conflict, as in the Congo. Or the voids may be
filled by ‘shadow’ administrations built by liberation movements (as in Eritrea),
or on the basis of traditional structures (as in Somaliland).
Despotic governance tends to be reinforced by CPEs, both within the framework
of existing authoritarian systems, as in the Sudan, and within formal democracies
like Sri Lanka or Colombia, where conflict may corrode existing protection of
democratic rights and freedoms. It is not necessarily inconsistent with hollowing
of the state. Indeed as Mann has argued in his classic work on the sociology of
militarism, the despotic power of the state (its reliance on force) often diminishes
its infrastructural power (its capacity to raise resources and govern effectively).
•

What Happens to the State in Conflict? 303
•
Serious ‘accountability dilemmas’ also tend to arise from the weakening of the
state and transfer of significant proportions of its power, revenues and
responsibilities: for example to guerrillas, militias and warlords; and on the other
to peace-keeping forces, international agencies, NGOs or diaspora communities.
The basic problem is that the former may not be accountable to anyone at all
except those in charge of the gun; while the latter are responsible to remote
international stakeholders like donor governments or charity organisations — not
national electorates or local communities. This not only tends to leave the latter
with little or no effective means of influencing policies or securing redress, it also
furthers the shrinkage of the public domain and the atrophy of national political
institutions and processes.
•
•
Such accountability dilemmas tend to be compounded by the regionalisation of
conflicts: flows of arms, combatants and refugees across national boundaries,
parallel markets, interventions by neighbouring countries all blur lines of
responsibility and bring in new political protagonists.
The ‘rule of un-law’ is another crucial aspect of the breakdown of governance
during conflicts, often with an enduring impact long after their end — including
breakdown of judicial and policing systems, massive violations of human rights,
widespread impunity for such violations and great difficulties in naming or
punishing the perpetrators (Pankhurst, 1999).
•
•
CPEs can have a deep impact on political culture by creating ‘societies of fear’
(Koonings and Kruijt, 1999), in which violence and rights abuses become
normalised, gender relations are brutalised and predation and corruption are
endemic, as in the conflicts in West Africa and Central America.
There is often a systematic redistribution of power, wealth and status, toward
military factions, warlords, ethnic or clan militias, traders in arms, drugs and
other scarce commodities, sometimes directly encouraged by the state, as in
Sudan, and sometimes not. This redistribution creates new political realities,
which may be exceedingly hard to reverse, as in ‘post-war’ Liberia or Bosnia,
where the current balance of political forces makes long-term reconstruction
highly problematic.
Disempowerment and political and social exclusion — of minorities, war-
affected regions (such as northern Uganda), refugees and internally displaced
people, women or indeed some categories of former combatants, including child
soldiers — are also in differing degrees aggravated by CPEs in ways that may be
difficult to reverse.
•
Yet there are important exceptions to the above. First, guerrilla movements
with well-defined political goals may be able, as in Eritrea, to establish parallel
administrations in the governance voids left by the retreat of the state, based on new
forms of political mobilisation and social capital creation, as well as co-operative (and
relatively equal) relationships with donor consortia. A second exception may be
where the destruction of the state creates political spaces in which new forms of
politics can emerge from below, as in Somaliland, where clan elders and women’s
groups have played different roles in peace-building and in the construction of more
legitimate forms of public authority.
The CPEs which our research has examined in greatest detail indeed
illustrate some of these exceptions, as well as the need to consider some of the

304
Lionel Cliffe and Robin Luckham
potentially transformational legacies of CPEs, which programmes of reconstruction
might build upon. Nevertheless, even in Eritrea and Ethiopia, the liberation
movements developed commanding styles of politics which carried over to post-war
period, influenced their approaches to reconstruction and democracy, and were
arguably a factor in their mutual failure to negotiate their border conflict during the
late 1990s. The LTTE in Sri Lanka, though quite different from the predatory militias
of Liberia or Sierra Leone, has an exclusive political agenda, is extremely violent, is
suspicious of negotiation, and would find it very difficult to engage with democratic
politics in any post-conflict settlement. Though the liquidation of the state may have
opened political spaces for reconstruction in Somaliland, it did not do so in the
remainder of former Somalia.
Peace-building, state reconstruction and socio-economic
recovery
A thorough understanding of the political background and processes is useful in
identifying and prioritising humanitarian and peacemaking tasks after as well as
during an ‘emergency’. In particular should the state be reconstructed, and if so how
and by whom? It follows from our preceding analysis that a narrow focus on
rebuilding the state as such can be highly problematic, all the more if promoted by the
donor community from the outside, rather than based on the construction of political
alliances and political dialogue from within. Moreover it is clearly not enough simply
to restore dysfunctional state (including military) structures, which may have been
responsible for conflict in the first place. On the one hand reconstruction will require
rebuilding public authority, including the state’s capacity to organise socio-economic
recovery and provide public goods, including basic physical security — and the latter
will necessitate the restoration of the state’s monopoly of force. On the other hand
such public authority will remain on shaky foundations if the state is not also re-
legitimised and rebuilt as far as possible upon the basis of the consent of its citizens
rather than violence. Hence democratisation and a more accountable state are
necessary dimensions of peace-building and reconstruction (although the wholesale
importation of Western models of democracy may also be fraught with dangers).
But to illuminate what needs to be tackled in the longer term during a ‘post-
conflict’ phase — including whether such a neat turning-point is likely to occur —
itself is a task of two stages. The first is predictive: using the analysis of the context
and ongoing processes during conflict that we have outlined in sections above to
begin to identify a future agenda when it becomes possible to look beyond mere relief
towards rehabilitation. Just as an understanding of the origins and progress of a
dramatic flood or famine immediately begins to make clear what will eventually have
to be done in terms of resettling people, re-establishing homes and livelihoods and
what early warning systems, flood control or food security measures can help avoid a
repeat performance, so it is with complex political emergencies — but with
complications. Not only is there the complexity, but the political character, in terms of
causes and consequences, means that reconstruction and recovery are political tasks.
The reconstitution of the state itself is at once prerequisite, means and target of
rehabilitation. Moreover, the very act of identifying and discussing such agenda in
advance can be a route into talks about ending conflict where these have been difficult

What Happens to the State in Conflict? 305
to initiate as a result of warring factions imposing preconditions or disagreeing about
agenda. A ‘Beyond the Conflict’ agenda (see for instance the Horn Conference under
this name in 1990 (Doornbos et al., 1991) or the War-torn Society Programme of
UNRISD) may be less contentious and also allow identification of non-partisan
elements or those on each side with commitment to peace.
A second and continuing process of analysis begins once some post-conflict
transition is under way. That of course involves monitoring the actual processes of
ending conflict, of transition to a sustainable peace, and of social, economic and
political reconstruction, including consideration of the necessary interconnection of
all these challenges and their phasing.
Difficult political choices will often be required, since conflicts are never
brought to an end or public authority rebuilt upon a level playing-field. The choices
are especially problematic where those responsible for violence and human rights
abuses either remain armed and able to wreck any prospective settlement (like the
RUF and other militias in Sierra Leone), or can manipulate the peacemaking process
so as to emerge in control of the state (like Charles Taylor in Liberia). To what extent
is it desirable or even feasible to make Faustian bargains with groups who may have
little real interest in building peace or an inclusive political process? Should one
concentrate on reforming the formal instruments of state power and democracy in the
hope that they can eventually displace the corrupt and violent ‘shadow states’
sustained by conflict? Current donor attempts to promote ‘security sector reform’, for
instance, draw on elements of both approaches in seeking to integrate former
combatants into reconstituted national armies and police forces. Yet they are unlikely
to succeed without a genuine domestic constituency for reform, which is precisely
what seems to be absent in some post-conflict situations.
But equally one should not fall into the trap of assuming transformations are
impossible. As we have already argued, conflicts sometimes open up opportunities for
change, as well as legacies to be overcome. But to make use of these opportunities
requires accurate diagnosis of the real politics of transition and of the political forces
and alliances and groups which can be mobilised to create effective constituencies for
change — and of how to open the political spaces within which these constituencies
can organise freely.
Whether in anticipatory prediction or actual observation of these processes, it
is possible to identify what dimensions are crucial, in the sense of carrying great
weight in themselves or being strategic in their impact on other problem areas. These
dimensions and a framework for analysing them, with some findings from cases to
illustrate how important it is to understand these elements of the processes, are
presented in this section.
The ‘post-conflict’ circumstances
The end-game
The outcome of the fighting and the circumstances under which it is ended or reduced
in scale vary along several dimensions, which in turn will shape what it is possible to
do in any actual or predicted ‘post-conflict’ stage and who will do it.
Was violent conflict brought to an end through outright victory, imposed
truce or genuine agreement? Interestingly, almost all of the several conflicts in the

306
Lionel Cliffe and Robin Luckham
Greater Horn have ended in an outright victory, or at least the outright defeat of
existing regimes, which marks this region off from many others, where some form of
power sharing was the outcome. The circumstances under which open conflict came
to an end did, as we shall trace through, have a major impact on subsequent patterns
of demilitarisation, recovery and state reconstitution, but not necessarily in an over-
determined way. Other factors and other turning-points came into play.
Were terms of settlement dictated by victors, agreed by factions or imposed
from outside; and were they clear-cut or ambiguous? In the Greater Horn, ‘victory’
would suggest the first alternative was most likely, but terms and subsequent
processes were not always clear and unambiguous. Eritrea was the most clear-cut
case; there was an outright victor: a coherent, disciplined and self-confident
movement with a regular army capable of conventional warfare not just guerrilla
methods — itself shaped by the character of the conflict through which it had
evolved. It was that movement that articulated the immediate settlement, and the
bases for demobilisation, recovery and state building, even if these increasingly
involved participatory processes. In Ethiopia, Somalia, Uganda, Rwanda, it is more
accurate to talk about the outright defeat of the former regimes, for who exactly were
the victors was still somewhat in dispute, and thus other actors were involved in what
was more of a negotiated process. Of course in the case of Somalia, these issues
remained unresolved for a decade after Siad Barre’s regime was deposed.
Balance of forces
Three dimensions of the transitional political arrangements were decisive in dictating
the terms of such settlements, as well as what emerged in future developments, like
the kind of new political dispensation, and whether there was a clear-cut end to the
internal wars and whether there was a prospect of conflict petering out or re-erupting:
•
•
•
the various forces and factions that were or should have been incorporated;
whether the regime was newly formed or a reconstituted version of the old; and
its social and political base: whether relying entirely on military means or with
some degree of popular support or legitimacy.
Again, Eritrea represents one pole. The EPLF had the stage to itself and so
did not have to accommodate or deal with any other significant aspirant for power.
The Ethiopian army and administration had been defeated and their personnel
excluded from the territory; the only other political grouplets were divided, in exile
and without arms. Thus, strategies for all aspects of transition could be worked out
and implemented, but also political compromise was not a practice in which the new
leadership was skilled. The new bureaucratic structures were an amalgam of the
movement’s own and some incorporated elements of the old administration of what
had been run as a region. It had widespread popular support for its image and
programme of independence.
Although achieved at the same moment and as part of the same set of
circumstances, the main movement in Ethiopia proper, the EPRDF, that had defeated
the military regime in 1991, was not the sole contender for political power. Other
regionally based movements had also taken on the former military government. A
bureaucracy with a degree of independence and experience was more or less intact,

What Happens to the State in Conflict? 307
and there were social forces at work in what was left of civil society — especially
those associated with the former ruling nationality, the Amhara, on which the empire
and the military regime had been based. The different balance of forces led to a
different pattern of transition. In the process of working out a new dispensation, these
elements were assigned minor roles within a formalised ethnic federalism — a role
some grudgingly accepted, others did not, with implications for the future of the
reconstituted state and for sustainable peace.
Uganda offers another example of where the former regime was routed and
where a main contender took control of the transition. The latter sought to incorporate
all previously existing forces and factions into a new dispensation, though on its own
terms, rather than banishing or suppressing them. An innovative form of state
eventually emerged from these circumstances, as will be discussed below. The
incorporation formula did bring peace and consolidated power in the most populous
southern and central areas, but did not extend, for whatever reason, fully to the
political forces in several of the peripheral areas of the country — with predictable
effects on security in those areas.
In Rwanda, too, a single main contender was left alone on the political stage
with the defeat of the regime and its army, but one which was largely made up of a
movement from the near diaspora, and which was left an empty stage when the rump
of the genocidal regime decamped across the Congo border with half the population.
In Somalia, although one phase of internal war ended in 1991 with the
overthrow and flight of the Siad Barre regime, there were so many contenders for a
power that was so fiercely contested that the country became an extreme case. Not
only has fighting continued, in a war among various warlords, but the tasks of
demobilisation and of reconstructing both the economy and the state have been
postponed for a decade.
Somaliland, with consequences we shall explore below, was an exception to
this pattern. The period 1988–90 had been marked by all-out war between the
garrisons of the Siad Barre regime’s army and the guerrillas of the Somali National
Movement in the north-western part of what was then Somalia. Like in Ethiopia, the
military outcome was a complete rout of the regime’s army, which fled the scene or
broke up. But unlike Eritrea where a well-disciplined movement and its army took
over the reins of power, the guerrilla forces of the SNM decided they had had enough
of fighting and melted away, leaving the field to small militia bands and to indigenous
institutions. These circumstances formed the backdrop to what was required and also
what was possible in humanitarian activities, rehabilitation and sustaining peace.
These tasks were tackled through complex and long-drawn-out deliberations.
Circumstances dictated that all of them had to be carried out simultaneously, while
they interacted one with the other, but where patience and lengthy consultation were
crucial requirements of humanitarian or rehabilitation activities.
The manner in which the end-game was played out will affect the prospects
for peace in both the short and longer term:
•
•
Will hostilities be terminated and herald a distinct post-conflict phase or drag on?
What is the risk of re-igniting major conflict?

308
Lionel Cliffe and Robin Luckham
Demilitarisation and demobilisation
Four important post-conflict issues include: the prospect of down-sizing armed forces
without jeopardising security; demobilising ex-combatants and ensuring livelihoods
for them but also ensuring the decoupling of armed groups from the informal
economy; the prospects for removal of weaponry from society at large and from
future disaffected groups; and the reconstitution of security forces and the basis for
that — what mix of former fighters, regular army, new recruits — so as to ensure
composition does not itself generate conflict (as it did in Zimbabwe in 1982–3).
Let us explore the differing experiences with respect to continuation or
termination of conflict, and demilitarisation. The sharply clear-cut transition in Eritrea
can be taken as one pole. There was no opposing armed force to threaten security, a
new national army had to be created, but the base was there in the experienced and
well-disciplined EPLF fighters and that army could be down-sized. The remaining
fighters could be disarmed and demobilised without any leak of weaponry or
professional killers into society. The provision of livelihoods for demobilised fighters
and also returning refugees could in turn be planned, even though it had to be
undertaken in a context shaped by the effects of a long war: extreme economic
destruction and lack of indigenous means for generating much revenue, foreign
exchange earnings or employment.
Ethiopia in 1991 had parallels and differences. The EPRDF forces were
similarly well disciplined and were disarmed and demobilised efficiently, with
relatively few, however, in the new army, to make way for new recruits and to some
extent old soldiers from regions other than Tigray, and for former fighters from the
several other armed factions. The EPRDF fighters found a place as militia members in
Tigray, and were later catalysts in the escalation of war with Eritrea. Some former
soldiers of the military regime and armed groups not under the EPRDF umbrella
remained at large in society and were able to fuel banditry as well as peripheral armed
opposition to the new regime.
The new Ugandan regime in 1986, too, was able partly to control the end of
hostilities and demobilisation, but unable to eliminate resistance to central authority
from peripheral areas, nor to stop the further escalation of the use of weaponry in
rustling and other competition for resources among pastoralists and their neighbours.
This pattern was repeated in southern Ethiopia.
The victorious RPF forces in Rwanda were in a position to manage the
demilitarisation within its borders, but only contain, for the moment, the threat from
outside of forces associated with the old regime — circumstances that almost
inevitably involved them in incursions into the complex conflicts in the Congo
Democratic Republic from 1996. The creation of a new army was also affected by the
post-genocide situation of flight and mistrust; it absorbed mainly RPF forces and very
few soldiers of the old regime’s army.
Somalia represents the antithesis: a case of almost complete failure to
prevent the emergence of new violently contested divisions and more widespread
anarchy, characterised by the absence of any demobilisation or demilitarisation.
Current estimates suggest that the number of automatic weapons abroad in society is
enough for members of every family to have their own. Together with other past and
present conflicts in the Horn, the ready market in weapons poses a threat even to
those neighbouring countries that have successfully undertaken some demobilisation

What Happens to the State in Conflict? 309
and disarmament. But exceptions to these dominant trends should also be noted.
Heinrich (1998) provides a useful reminder of the many examples throughout Somalia
where mediation mechanisms, ‘traditional’ and newly forged, played important roles
in dampening the violence of conflict. In Somaliland too, society was awash with
small arms but also some more substantial weapons. Militia groups competed for
resources and terrain in the vacuum after the defeat of the regime and the dissipation
of the SNM — the one organised force, even if one associated with one clan
essentially. But unlike other areas of Somalia proper, clan elders and other informal
groups set in train processes of traditional conflict resolution and dialogue on other
issues. These processes gradually marginalised warlords, painstakingly arranged a
disarmament of the militias which had of necessity to be a negotiated process,
resolved inter-clan and other clashes, re-absorbed the former SNM fighters into a
broader framework and came up with a formula for reconstituting the state.
Options and prescriptions
In the many and varied longer-term tasks of reconstructing the state, economy and a
sustainable peace, actors are making policy choices which have strategic implications,
whether intended or not. Our arguments are twofold: that the options are constrained
by the patterns of the earlier — before and during — conflict stages and by the
circumstances of the immediate transition following the diminution of conflict which
we have outlined above; and that such choices should be informed by a systematic
analysis of that context. We will explore how options have been constrained, and
some of the lessons from choices made, again using cases from the areas mentioned.
The focus will be limited to two specifically political agendas as follows.
The manner and form of the reconstitution of the state
State reconstitution encompasses a wide field. It includes the size of the task of
rebuilding a collapsed state and how to achieve such essentials as a revenue base, law
and order. There is the choice of whether to change regime or personnel only or opt
for a different form, or incorporate both state and ‘parallel’ institutions. There is also
the politics of the process of bringing about that change: what process of constitution
making or review; how far top-down or from below; and whether constitutional
structure should be federal, decentralised or unitary; a decision on the extent of power
sharing; the role of the military; and how far should accountability be built in, for past
and future abuses of power, for policy and performance?
In all our main cases from the Greater Horn, a state had been overthrown and
had to be reconstituted but still the demands of this task have varied depending on the
prior degree of collapse and on the end-game and circumstances of transition.
Rwanda faced one of the most challenging immediate tasks. As a result of
the genocide, defeat of the extremist Hutu regime and flight of its personnel, the
former administration was either non-existent or discredited and had to be replaced
comprehensively. Moreover, the genocide had involved the slaughter of many Hutu
politicians who would have been prepared to engage in coalition politics and so
severely complicated the politics of state reconstruction and of national rebuilding.
Those processes relied on the RFP personnel who had been in exile for much of their

310
Lionel Cliffe and Robin Luckham
lives and had little accumulated experience or detailed knowledge of local
circumstances.
In Somaliland, the overthrow of the regime, flight of most of its personnel
and absence of a significant organised alternative led to a protracted period of
statelessness. The technically and politically challenging task of reconstituting a state
was pursued through the lengthy but ultimately effective process of traditional
negotiation and mediation, with clan-based representation which provided a voice for
peace-brokers rather than warlords, and which simultanously acted as a constituent
assembly, an electoral college for a two-chamber legislature as well as a peace
conference. The practical tasks of generating a revenue base, a civil service and a
court and police system were, however, made immense given previous history. Those
challenges have also been faced in Puntland in the north-east, and on a local scale
elsewhere in Somalia. But the process of relying on externally based conferences with
the faction leaders and warlords centre stage has so far not succeeded in reconstituting
a central state. The latest attempt, in 2000 in Djibouti, may have gone further than
others and allowed involvement of a broader array of forces but still excluded key
actors.
In Uganda, Ethiopia and Eritrea there have been formal processes of
constitutional review, involving open discussion and forms of participation among the
population at large, even though these have been initiated and to some degree
orchestrated by the main contenders. Those who occupied power in Ethiopia
promoted the most original formula, the institution of a system of ethnic federalism as
a means of curbing factionalism based on exclusion — an understandable (though
risky) reaction to recent and past chapters of history. Uganda coped with the same
centripetal forces by allowing indirect representation of people at four levels up to the
national while excluding formal political parties, with their history of separatism,
from that electoral process.
All these cases differ from those in other experiences as varied as Sierra
Leone, El Salvador, Honduras and Mozambique, where negotiated endings of conflict
allowed for more continuity in the existence of the state, and where implementation of
political reform was often problematic and itself the subject of continuing negotiation
— or of a retreat into open conflict.
Rethinking the state’s role in recovery and long-term development
Only one example will be cited here — and again using the illustrative capacity of an
extreme case — to show the predictive potential of political analysis. The emergence
through a long struggle in Eritrea of a strong, self-confident new regime, without
much need to share power or compromise what it regarded as principles, also
determined the role the state set itself in rehabilitation. Although wedded to some
privatisation and a market economy, this was to be a directed market economy. The
stance toward all outsider donors and international agencies was to accord them only
a supportive role, welcoming support providing it bolstered nationally determined
priorities. This stance also generated a predictable attitude to NGOs:
Theories that view popular participation as a weapon opposing, rather than
co-operating with national policies, assume the existence of fundamental
conflicts between those policies and popular demands. The whole argument

What Happens to the State in Conflict? 311
of civil society getting organised to form an alternative to government is
based on such premises. There would [then] be … as many civil society
institutions and NGOs as there would be issues. … Eritrea does not allow
this. The argument would be that there is a strong government in this country
determined to make its own policies, commit its own mistakes, correct them
in its own way and run the country according to what it perceives to be its
needs (Alemseged Tesfai, 1998: 353–4).
Yet these very predictable stances proved difficult for outside agencies to
comprehend and adjust to, and led to many misunderstandings which hindered
support for policies in such areas as refugee repatriation, demobilisation of fighters
and revenue raising.
Contrast this central role for the state with that in post-conflict situations like
Somalia, the opposite extreme, where the virtual absence of effective state-like
institutions means any strategy for state reconstitution, including defining the
economic role of the state, and for rehabilitation, is dependent on the role of external
actors, whether UN agencies, great powers, regional organisations or NGOs. Their
record — in Somalia, Liberia, Sudan — has not been notably successful in putting the
humpty-dumpty of the state back together.
Conclusions
This article has added its voice to the many that have stressed how inadequate or
misinformed political analysis has contributed to inappropriate and counter-
productive interventions in seeking to provide humanitarian assistance, promoting
peace or rebuilding war-torn societies. Its particular argument has been that ‘complex
political emergencies’ deserve that descriptive title if only because of the multiplicity
of forces shaping outcomes. Each CPE develops a dynamic of its own; once violent
conflict emerges, it transforms itself and all around it — the state, livelihoods,
national economy, social relations. Consequently any intervention by outside or even
internal actors is but one of those forces and cannot be the decisive factor in
determining outcomes. Any actions that aim to make a strategic impact must then be
opportunistic, taking advantage of particular forces and conjunctures offered. But we
have also argued that, although complex and varied, CPEs are not beyond all
understanding. We have offered one perspective on the dynamic of CPEs,
concentrating on three dimensions or stages (not linear or even necessary): the origins
and in particular the typically growing inadequacies and fissures in the state; the form
of the gathering momentum of violent conflict, concentrating again on what happens
to the state; and the challenges offered by a context where any let-up in violence
allows sustainable reconstruction of society and reconstitution of the state. This
outline is offered as both a conceptualising of some of the processes and differing
patterns of such dynamics and a possible methodology for making policy-relevant
analysis.
That approach was abbreviated enough so no attempt will be made here to
précis it further. Two brief issues can be briefly addressed by way of conclusion: to
give some idea of what difference such informed analysis could have made or might

312
Lionel Cliffe and Robin Luckham
make to actual policies; and what is involved in making operational such applied
research.
At a general level, one of the conclusions of our analysis of several countries
of the Greater Horn — namely that those that share the same violent overthrow of
existing regimes and states have gone on to generate conditions so varied as to cover
the whole range from the most stable to the most anarchic — reinforces the view that
there is no single template that can give policy guidance. One more specific
conclusion about the appropriateness of strategy can be taken from among the many
errors identified with hindsight from interventions in Somalia. The adoption of a
mandate to limit armed intervention to ensuring relief aid delivery and the methods
pursued was implicitly premised upon the view that the warlords would be indifferent
to actions which would so seriously affect their economic base — a blind spot that
any analysis of the political economy of the conflict would have revealed. A second
example of strategic shortcomings on the part of both international and non-
governmental agencies is illustrated in Eritrea where a better-informed conception of
the nature of the liberation movement and the new embryonic state might have
avoided the long delay in implementing a joint refugee repatriation programme with
UNHCR. Such misperceptions were also instrumental in the often tense relations
between the government and NGOs.
Note
1. Our focus is somewhat different from the literature dealing with how conflict destroys
social capital, and more narrowly political.
References
Alemseged Tesfai (1998) The Challenges of Post-War Development: A Review of WSP
Research Findings. In M. Doornbos and Alemseged Tesfai (eds.) Post-Conflict Eritrea:
Prospects for Reconstruction and Development. Red Sea Press, Lawrenceville.
—— (n.d.) Eritrea: Post-War Challenges of Development: Analysis of War-Torn Societies-
Eritrea Project. Unpublished thesis, Asmara.
Anderson, M. (1996) Do No Harm: Supporting Local Capacities for Peace through Aid.
Collaborative for Development Action, Cambridge.
—— (1998) ‘You Save My Life Today, But for What Tomorrow?’ Some Moral Dilemmas of
Humanitarian Aid. In J. Moore (ed.) op. cit.
Bayart, J.-F., S. Ellis and B. Hibou (1999) The Criminalisation of the State in Africa. James
Currey, Oxford.
Berdal, M. and D.M. Malone (eds.) (2000) Greed and Grievance: Economic Agendas in Civil
Wars. Lynne Rienner, Boulder.
Chabal, Patrick and Jean-Pascal Deloz (1999) Africa Works: Disorder as a Political
Instrument. James Currey, Oxford.
Cliffe, L. and R. Luckham (1999) Complex Political Emergencies and the State: Failure and
the Fate of the State. Third World Quarterly 20(1): 27–50.
De Waal, A. (1997) Famine Crimes: Politics and the Disaster Relief Industry in Africa.
James Currey, Oxford.
Doornbos, M., L. Cliffe, A. Ahmed and J. Markakis (eds.) (1992) Beyond Conflict in the
Horn: Prospects for Peace, Recovery and Development in Ethiopia, Somalia, Eritrea
and Sudan. James Currey, Oxford.

What Happens to the State in Conflict? 313
Duffield, M. (1994) Complex Political Emergencies and the Crisis of Developmentalism.
IDS Bulletin 25(4): 37–45.
—— (1998) Post-modern Conflict, Warlords, Post-adjustment States and Private Protection.
Journal of Civil Wars 1(1): 65–102.
Heinrich, W. (1997) Community-based Peace Processes in Somalia, 1991–96. Life and
Peace Institute, Uppsala.
Johnson, M. (1998) The SPLA. In C. Clapham (ed.) African Guerrillas. James Currey,
Oxford.
Kaldor, M. (1998) New Old Wars: Organised Violence in a Globalised Era. Polity,
Cambridge.
Koonings, K. and D. Kruijt (eds.) (1999) Societies of Fear: The Legacy of Civil War,
Violence and Terror in Latin America. Zed Press, London.
Le Billon, P. (2000) The Political Economy of War: What Relief Agencies Need to Know.
Humanitarian Practice Network Paper 33, Overseas Development Institute, London.
Mann, M. (1988) States, War and Capitalism. Studies in Political Sociology. Cambridge
University Press, Cambridge.
Minear, L. and T. Weiss (eds.) (1993) Humanitarian Action in Times of War: A
Practitioners’ Handbook. Lynne Rienner, Boulder.
Moore, J. (ed.) (1998) Hard Choices: Moral Dilemmas in Humanitarian Intervention.
Rowman and Littlefield, Oxford.
Outram, Q. (1999) Liberia: Roots and Fruits of the Emergency. Third World Quarterly 20(1):
163–74.
Pankhurst, D. (1999) Issues of Justice and Reconciliation in Complex Political Emergencies:
Conceptualising Reconciliation, Justice and Peace. Third World Quarterly 20(1): 239–
56.
Reno, W. (1998) Warlord Politics and African States. Lynne Rienner, Boulder.
Sahnoun, M. (1998) Mixed Interventions in Somalia and the Great Lakes. In J. Moore (ed.)
op. cit.
Zack-Williams, A. (1999) Sierra Leone: The Political Economy of Civil War, 1991–98.
Third World Quarterly 20(1): 143–62.
Address for correspondence: Lionel Cliffe, Institute for Politics and
International Studies, University of Leeds, Leeds LS2 9JT. E-mail:
<<l.r.cliffe@leeds.ac.uk>>