Conflict Prevention and Peaceful Development:

Policies to Reduce Inequality and Exclusion

A CRISE Policy Conference
July 9-10, 2007
Queen Elizabeth House

Horizontal Inequalities and Conflict

Chapter 1: Horizontal inequalities: An introduction
and some hypotheses

Centre for Research on Inequality, Human Security and Ethnicity, CRISE

Department of International Development, Queen Elizabeth House, University of Oxford
3 Mansfield Road, Oxford OX1 3TB
Tel: +44 1865 281810; Fax +44 1865 281810
www.crise.ox.ac.uk
HORIZONTAL INEQUALITIES AND CONFLICT: UNDERSTANDING GROUP VIOLENCE IN

MULTIETHNIC SOCIETIES

Edited by Professor Frances Stewart

Chapter 1

Horizontal inequalities and conflict: an introduction and some hypotheses

Frances Stewart

1. Introduction

Violent conflict within multiethnic and multireligious countries is a major problem in the world

today – from the former Yugoslavia and USSR to Northern Ireland and the Basque country, from

Rwanda to Darfur, Indonesia to Fiji, numerous bitter and deadly conflicts are fought along ethnic

or religious lines. In addition to the direct injuries and loss of life both on and off the battlefield

that result, violent organized conflict is also a major cause of poverty often leading to economic

regress, with much the highest incidence of such conflict found in the poorest countries of the

world. Seeking a way of preventing these conflicts is thus of paramount importance.

Yet not all multiethnic or multireligious societies are violent. Indeed the vast majority are not

(Fearon and Laitin, 1996). So it is not the case that those with cultural, ethnic, religious or racial

differences cannot live together peacefully, as is suggested by the view that there is an

unavoidable ‘clash of civilisations’ (Huntington, 2002) . The question, then, is why ethnic or

religious conflict breaks out in some circumstances and not others. If we can answer that, we

may be able to identify ways of preventing such conflicts and their enormous costs in terms of

deaths, injuries, and economic and social collapse.

408
This book explores one important cause of such conflicts: the existence of major horizontal

inequalities (HIs). Horizontal inequalities are inequalities in economic, social or political

dimensions or cultural status between culturally defined groups. The book considers the role of

horizontal inequalities in causing conflict, and policies that would contribute to reducing HIs and

thereby reduce the likelihood of conflict.

This chapter provides the general framework for our discussion: the next section briefly indicates

the significance of the problem of conflict between identity groups, within and across nations;

Section 3 discusses the complex issue of how groups, which potentially provide the basis for

conflict, are formed and mobilized; Section 4 defines HIs in more detail and explains why they

may lead to violent group mobilization, drawing on the example of South Africa ; Section 5

develops the main hypotheses on the relationship between HIs and conflict to be considered in

the rest of the book; and the final section explains the organization of the book.

2. Groups in conflict

The incidence of violent conflict among poor countries is high – seven out of ten of the poorest

countries in the world are undergoing or have recently experienced some sort of civil war. These

conflicts involve very heavy costs. The immediate human costs in terms of deaths, injuries and

refugees are most obvious and well known. Deaths as a result of the fighting itself vary from a

few thousand, to an estimated 800,000 in the case of the Rwanda genocide (The History Place-

a). Refugee movements often amount to millions. ‘Indirect’ deaths as a result of war-induced

famine, following disruptions in production, marketing and purchasing power, are often far

greater than the direct deaths. The war in the Congo from 2000, for example, is estimated to

have accounted for nearly 4 million deaths, including both direct and indirect fatalities (Coghlan

et al., 2006). Moreover, there are many other heavy indirect costs as a result of the economic

409
and social disruption war causes. Violent conflict is one of the biggest obstacles to development,

reducing incomes and investment and undermining human development, as well as causing

immense suffering among the belligerents and the countries’ population more generally.

Regression analysis suggests an average loss in gross domestic product per capita of between

2.00 and 2.40 per cent per annum among countries experiencing conflict (Imai and Weinstein,

2000; Hoeffler and Reynal-Querol, 2003), while case studies show huge variability in costs, with

the worst conflicts leading to far greater losses. For example, one estimate suggested a

cumulative loss of half of GDP in the case of Iraq in the Iraq-Iran War, while the increase in

infant deaths during the Uganda conflicts amounted to 2 per cent of the population (Stewart et

al., 2001).

Violent conflict, of course, is not confined to poor countries, even though its incidence is greatest

among them. The conflict in Bosnia that accompanied the breakup of the former Yugoslavia led

to the deaths of over 200,000 Muslims and resulted in 2 million refugees (The History Place-b).

Similarly, many violent conflicts were associated with the disintegration of the USSR. Some,

notably that in Chechnya, continue to this day – again with heavy costs.

The incidence of violent conflicts rose quite steadily from the 1950s, accelerating following the

end of the Cold War, but then declining as the old Cold War-inspired conflicts (such as those in

Mozambique and Nicaragua) came to an end, while the often violent transition to new regimes in

the former Eastern block gradually subsided (Figure 1.1).

410
 Number of Refugees Worldwide (millions) Incidence of major conflict

0
2
4
6
8
10
12
14
16
18
20
19

0
10
20
30
40
50
60
45 70
1951 19
49
1953 19
1955 53
19
1957 57
19
1959 61
1961 19
65
1963 19
69
1965 19
73

Source: Calculated from Marshall (2006)

1967

Figure 1.2 Refugee trends 1951-2004

19
1969 77
19

Sources: UNHCR (2000), UNHCR (2006)

1971 81
1973 19
85
1975 19
89
1977 19
93
1979 19
1981 97
20
1983 01
1985
1987
0
20
40
60
80

1989
100
120
140
160
180
200

1991 Magnitude-weighted index of major conflict

1993
1995
1997
1999
Figure 1.1. Incidence and magnitude of major episodes of political violence, 1946-2004

2001
Incidence

2003
Magnitude

411
During the Cold War many conflicts presented themselves as wars about ideology or class,

following the ideological East/West division, with each side supported by the major powers along

ideological lines. The conflicts in Vietnam, Cambodia, Mozambique and El Salvador are

examples. But even then, some conflicts were overtly conducted along identity lines – for

example, the Catholic/Protestant troubles in Northern Ireland and a number of Middle Eastern

conflicts (between Christians and Muslims in Lebanon, Sunnis and Shias in Iraq/Iran and Jews

and Muslims in Palestine/Israel), as well as the recurrent ethnic conflicts in Burundi and Nigeria.

Moreover, in many of the ideological wars, the conflicts had an implicit ethnic or racial base. This

was the case, for example, in Mozambique, where the Xitsonga and the Ndau from the south

and centre fought northern ethnicities; or in Guatemala and Peru, where the conflicts were

presented as ideological and led by intellectuals of European origin, but the vast majority of

those killed were indigenous (Figueroa and Barron, 2005; Caumartin, 2005).

The identity basis of conflicts has become much more explicit since the end of the Cold War, as

ideological differences have diminished and socialism no longer seems a serious alternative, nor

using it as a banner a guarantee of external financial support. According to the Center for

Systematic Peace ‘ a virtual cornucopia of these seemingly intractable (and previously ‘invisible’)

social identity conflicts exploded onto the world scene and captured the public and policy eyes.’1

Data on conflict show a major increase in the proportion of all conflicts that are labelled as

‘ethnic’: from 15 per cent in 1953 to nearly 60 per cent by 2005 (see Figure 1.3)

412
Figure 1.3 Ethnic violence as a proportion of major political violence, 1946-2004
ethnic'

80%
Proportion of incidence/total magnitude classified as '

70%

60%

50%

Incidence
40%
Magnitude

30%

20%

10%

0%
1945

1949

1953

1957

1961

1965

1969

1973

1977

1981

1985

1989

1993

1997

2001

Source: Calculated from Marshall (2006)

These identity conflicts have also acquired global dimensions, with Islam versus the West

replacing the ideological divisions of the Cold War, with the wars in Afghanistan and Iraq and the

Israel/Palestine conflict as clear examples. Today, then, mobilization along group identity lines

has become the single most important source of violent conflict. The next section will consider

how such identities are formed.

3. Group formation and mobilization

People see themselves in many different ways; they have many identities. Some are fluid, short-

lived and insignificant (for example, being a member of an evening class), while others are more

permanent and more significant personally and socially (for example, gender, ethnicity and

413
religion). The importance people attribute to different aspects of their identity also varies

according to context and over time. Clearly, where violent conflicts are mobilized and organized

by identity, such identities must be sufficiently important to enough people to make them

prepared to fight, kill and even die in the name of that identity. Personal motivation also

obviously plays a part in causing people to fight (Collier and Hoeffler, 2000; Keen, 1998), but it is

our contention that in many conflicts people are primarily motivated by their group identity – their

religion or ethnicity –and consequently group motives are a vital driving force. For this to

happen, the group boundaries must be relatively clearly defined and have some continuity over

time.

People may be divided into groups in many ways – according to geography, behaviour,

language, physical characteristics and so on. Yet only those divisions that have or may acquire

strong social significance – that is, such meaning for their members and for others in society that

they influence behaviour and well-being in a significant way – are likely to form the basis of

identity conflicts. Group identities arise partly from individuals’ own perceptions of membership

of and identity with a particular group – that is, the self-perceptions of those ‘in’ the group – but

they are also determined by the perceptions of those outside the group about others2. The

important question, then, is why and when some differences are perceived as being socially

significant and others are not, both by group members themselves and by others.

Anthropologists have differed sharply on this question. At one extreme are the so-called

primordialists who argue that ‘ethnicity is a cultural given, a quasi-natural state of being

determined by one’s descent and with, in the extreme view, socio-biological determinants’

(Douglas, 1988).3 ‘Basic group identity consists of the ready-made set of endowments and

identifications which every individual shares with others from the moment of birth by chance of

the family into which he is born at that given time and given place’ (Isaacs, 1975: 31, quoted in

414
Banks, 1996). For primordialists, ethnic identity is etched deep in the subconscious of the

individual from birth.

The primordial view, however, doesn’t explain why ethnic groups change over time – why they

are of preeminent significance at some points and then the boundaries and characteristics of

groups change. For example, Cohen (1969) has shown how some rural people moving to towns

in Nigeria became ‘detribalized’, while tribal identity became more important for other urbanized

Nigerians. Furthermore, it is widely agreed that many tribal distinctions in Africa were invented

by the colonial powers: ‘Almost all recent studies of nineteenth century pre-colonial Africa have

emphasised that far from there being a single “tribal” identity, most Africans moved in and out of

multiple identities, defining themselves at one moment as subject to this chief, at another

moment as a member of that cult, at another moment as part of this clan, and at yet another

moment as an initiate in that professional guild’ (Ranger, 1983: 248). ‘Modern Central Africa

tribes are not so much survivals from a pre-colonial past but rather colonial creations by colonial

officers and African intellectuals’ (Wim van Binsbergen 1976, quoted in Ranger, 1983: 248). One

example is the distinction between Hutus and Tutsis, which some historians argue was largely

invented by the colonial powers for administrative convenience (Lemarchand, 1994).

Instrumentalists, in contrast, see ethnicity as being developed instrumentally, to be used by

groups and their leaders in order to achieve political or economic goals. Cohen, cited above,

explained the development of Hausa consciousness and customs in this way.4 Similarly, Glazer

and Moynihan (1975) argued that ethnicity was maintained and enhanced by migrant groups in

the US in order to promote their economic interests. The colonial inventions, according to the

instrumentalist view, served administrative purposes. Further, the use of ethnic symbols and the

enhancement of ethnic identities, often through the reworking of historical memories, is

frequently used as a powerful mechanism for the mobilization of support for conflict. Numerous

415
examples, presented in Alexander et al. (2000), as well as by Cohen, Turton and others, have

shown how ethnicity has been used by political and intellectual elites prior to, and in the course

of, wars. In international wars, this takes the shape of enhancing national consciousness, with

flag waving, historical references, military parades, and so on. In civil wars, it is a matter of

raising ethnic or religious consciousness. An example here would be the radio broadcasts by the

extremist Hutus before the 1994 massacre in Rwanda, in which the Tutsis were repeatedly

depicted as subhuman, like rats to be eliminated, echoing Nazi anti-Jewish propaganda of the

1930s. Similarly, Osama Bin Laden has appealed to Moslem consciousness, arguing that the

conflict he and his followers are waging against the west is ‘in essence a religious war’ (The

Observer, 4 November 2001).

A third perspective, with much in common with the instrumental one, is that of ‘social

constructivists’. Constructivists too believe that ethnicities are frequently used instrumentally for

political purposes, but their emphasis is on the ‘making’ and ‘remaking’ of ethnic boundaries that

must occur to make such instrumentalism possible. Differences are emphasized, even invented,

by leaders in order to construct social groups. Such construction is an ongoing process which

may reinforce existing group boundaries or develop new ones following the political and social

motivation of the leaders responsible for such construction. Group boundaries, it is argued, are

made and remade (Anderson, 1983). For example, Akindès (forthcoming) has analyzed how

identities are ‘marketed’ by the media as well as by political leaders in the case of Côte d’Ivoire.

Yet both instrumentalists and constructivists generally recognize that there need to be some felt

differences in behaviour, customs, ideology or religion to make it possible to raise ethnic or other

consciousness in an instrumental way. For example, Glazer and Moynihan (1975: 379) state

that, ‘For there to be the possibility for an ethnic community at all, there will normally exist some

visible cultural differences or ‘markers’ which might help to divide communities into fairly well

416
defined groupings or ethnic categories.’ Thus some shared circumstances are needed for group

construction – for example, speaking the same language, sharing cultural traditions, living in the

same place or facing similar sources of hardship or exploitation. Past group formation, although

possibly constructed for political or economic purposes at the time, also contributes to present

differences. Whether the origins of a group are instrumental or not, the effect is to change

perceptions and make the differences seem real to group members – this is why group identities

are so powerful as sources of action. As Turton (1997: 82) puts it, the power of ethnicity or ‘its

very effectiveness as a means of advancing group interests depends upon its being seen as

“primordial” by those who make claims in its name.’ Hence what was a dependent variable at

one point in history can act as an independent variable in contributing to current perceptions.5

Construction of a cohesive identity can be effected by leaders of a group, or others,including the

state, who classify or categorize people into groups (Barth, 1969; Brubaker and Cooper, 2000).

Such categorization can be quite arbitrary, as is argued to have been in the case of the Belgian

colonial classification of Hutus and Tutsis, or, despite some arbitrariness, it may follow some

visible markers. The Nazi classification of Jews and the classification of ‘blacks’ in the U.S.

historically are examples of such categorization, combining arbitrary decisions on boundaries

with some common ‘markers‘ of ancestry. Where categorization by others is the source of group

boundaries, what people themselves feel about their own identities may not be important at all:

what matters is what others think they are.

Group identities can potentially provide the basis for violent conflict mobilization whether they

are primarily own-constructed identities – when they may be used to motivate people to demand

rights and to rebel – or identities constructed by the state or other groups – when they may be

used to discriminate against (and sometimes fight against) particular groups. Whether group

boundaries emerge out of the felt identities of the group itself or through categorization by

417
others, groups which mobilise in a way that threatens social stability generally have some

shared characteristics which usually makes it quite easy to identify members; and they also have

some continuity. Yet in almost every case, there is also some fluidity and uncertainty about

precise group boundaries, which evolve over time in response to circumstances – for example,

during the Biafran war, the Iwerri in Nigeria chose to reject their prior Ibo identity; the Telegu-

speaking people, who were an apparently homogeneous group seeking autonomy from the state

of Madras, became quite sharply divided once they had gained this autonomy (Horowitz, 1985:

66). Similar developments are occurring in Aceh now that it has achieved the autonomy for

which it fought.

This book argues that group mobilization along lines of identity is a central feature of many

conflicts, taking a broadly social constructivist view of group formation. The salience of particular

identities is increased by political action – by political leaders, media or the education system –

sometimes in order to raise consciousness of own identities, sometimes of that of others. Yet

though we take a social constructivist line, we also argue that people themselves can become

strongly convinced about the essential nature of their identities and that of others – which is why

mobilization by identity can work. Moreover, while people can choose which identities are

important to them, for the more enduring aspects of their identity they are not free to choose any

identity, as it were, ‘off a shelf’, shifting to whatever seems most convenient at a given moment.

Thus while someone can readily choose to change their social club or to abandon it altogether,

Kenyans without mixed parentage cannot choose to stop being Kikuyu and become Luo, though

they can choose to downplay their ‘Kikuyuness’. In any particular case, history and social

context will determine the possibilities. For example, in Europe today a change in religion is

relatively easy, but this was much less so in earlier centuries, when religious divisions were a

major cause of conflict; and in some contexts it is almost impossible in some contemporary

developing countries.6 In Peru today, someone who is of indigenous origin can choose to define

418
themselves as mestizo but they still cannot choose to avoid all racism. It is where there is limited

freedom to switch group that groups boundaries are particularly important in terms of creating

potential group grievances, and hence in terms of political mobilization. Where people can shift

groups in an instantaneous and costless way, then group distinctions and boundaries matter

much less.

While many conflicts have a cultural dimension, that is the groups involved perceive themselves

as belonging to a common culture (ethnicity or religion) and are partly fighting for cultural

autonomy, it is evident that cultural differences are not a sufficient explanation for conflict, since

the peoples of many multicultural societies live together relatively peacefully. Indeed, Fearon

and Laitin (1996) have estimated that from 1960-1979, of all the potential ethnic conflicts in

Africa (defined as occurring where different ethnic groups live side by side) only 0.01 per cent

turned into actual violent conflict. In some cases, groups may live together peacefully for

decades and then conflict erupts.

We need, therefore, to go beyond cultural explanations of conflict to economic and political

explanations. As Abner Cohen argued:

Men may and do certainly joke about or ridicule the strange and bizarre customs

of men from other ethnic groups, because these customs are different from their

own. But they do not fight over such differences alone. When men do …fight

across ethnic lines it is nearly always the case that they fight over some

fundamental issues concerning the distribution and exercise of power, whether

economic, political, or both.

(Cohen, 1974: 94, italics added)

419
In other words, cultural differences do not lead to violent conflict unless there are also major

economic and/or political causes.

The motivation of the participants is clearly at the root of any violent situation. Many

contemporary economists emphasize the pursuit of individual economic advantages as the

prime force driving conflicts (see, for example, Keen, 1998; Collier and Hoeffler, 2000). But the

majority of internal conflicts are organized group conflicts – they are neither exclusively nor

primarily a matter of individuals committing acts of violence against others. What is most often

involved is group mobilization of people with particular shared identities or goals to attack others

in the name of the group. While young men may fight because they are unemployed,

uneducated and have few other opportunities, they also generally fight out of loyalty to a group

(or sometimes to an ideology or a cause). Examples include the militia in Najaf in Iraq, the Hutus

in Rwanda, the Tamils in Sri Lanka, and the Catholics in Northern Ireland, amongst others.

Sometimes, indeed, the power of their beliefs or loyalties is so strong that they are prepared to

sacrifice their own interests – in the extreme case their own lives – for the wider objectives of the

group. Often governments, dominated by a particular identity group, are involved: sometimes

instigating attacks against other groups, and sometimes under attack. In fact Holsti (2000)

argues that state violence has more often than not been the initiating cause of recent conflicts.

In contemporary conflicts, group affiliations occur along a variety of different lines. In some

cases it is religious affiliation that provides the relevant binding and categorizing identity for the

groups involved (see, for example, the conflicts in Northern Ireland; Muslim/Hindu conflicts in

India; and the Muslim/Christian conflicts in the Philippines). In other cases the salient cleavage

seems to be racial (for example, in Fiji). Ethnicity is a binding factor in some conflicts (as in

Rwanda and Sri Lanka) while in other cases clans are the main source of affiliation (for example,

in Somalia). There are also many overlapping distinctions: in some situations, for example, both

420
ethnic and religious affiliations are pertinent (such as in Jos in Nigeria, in Poso in Indonesia and

in the Balkans), while class and ethnicity overlap in Central America, and caste and ethnicity are

intertwined in Nepal.

Large-scale group mobilization – particularly for violent actions – is unlikely to occur in the

absence of serious grievances at both leadership and mass level. The role of leaders is

important in political mobilization, in choosing the grounds for mobilization (whether, for

example, religion, class or ethnicity) and in ‘selling’ the importance of the chosen identity to the

people being mobilized. At the leadership level, the main motivation may be political ambition

and hence such mobilization is particularly likely to occur where there is political exclusion of the

group’s leaders; while the followers may also be concerned with the political representation of

the group as a whole, their primary motivation is more likely to be grievance concerning the

economic and social position of their group relative to others. Both leaders and followers may

become strongly motivated where there are severe and consistent economic, social, and political

differences between culturally defined groups, termed here multidimensional horizontal

inequalities.

It should be noted that it is not necessarily the relatively deprived who instigate violence. The

privileged may also do so, fearing a loss of power and position. The prospect of the possible

loss of political power can act as a powerful motive for state-sponsored violence, which occurs

with the aim of suppressing opposition and maintaining power.

4. Horizontal inequalities and mobilization

As noted above, horizontal inequalities are inequalities between culturally defined groups or

groups with shared identities. They are called horizontal to distinguish them from inequalities

among individuals, which we refer to as vertical inequalities (VI). These identities may be formed

421
by religion, ethnic ties or racial affiliations, or other salient factors which bind groups of people

together, following the discussion in the previous section.

For simplification, we can categorize HIs into four areas: political participation; economic

aspects; social aspects; and cultural status. Each of these contain a number of elements. For

example, HIs in political participation can occur at the level of the cabinet, the parliament, the

bureaucracy, local government or the army, amongst others. HIs in economic aspects

encompass access to and ownership of assets (financial, land, livestock and human and social

capital), employment opportunities and incomes. HIs in social aspects encompass access to

various services (education, health, water, sanitation and housing), and human outcome

indicators (such as measures of health and educational achievements). HIs in cultural status

include the extent to which a society recognizes (or fails to recognize) a group’s cultural

practices (for example, in matters of dress, holidays and so on).

While the four broad categories are relevant to every society, the elements that are relevant in a

particular case depend on the nature of the society, its political system, its economy, and its

social structure. For example, land may be irrelevant in modern urban societies but is clearly of

paramount importance in many developing rural economies such as in Zimbabwe, while

employment seems to be important in most countries. In natural-resource-rich economies, the

control over such resources, either directly or via the state, is an important source of group

competition. Access to housing is of critical importance in more developed economies, such as

Northern Ireland, but is less important where people mostly construct their own housing (as in

many African countries), where access to public sector employment is particularly important as a

way out of poverty.

422
There are causal connections between different HIs. For example, inequalities in political power

often lead to similar social and economic inequalities. A biased distribution of government jobs

and provisions of infrastructure is common, with the group in power discriminating in its favour.

For example, in Burundi in the 1990s, half of government investment went to the Bujumbura

region and its vicinity, which is the home of the elite Tutsi group (Gaffney, 2000). In some

countries, the president and his coterie have taken a massive share of state resources for their

private use, such as the Duvaliers in Haiti and the Somoza family in Nicaragua (Lundahl, 2000;

Pastor and Boyce, 2000). Moreover, there are connections between economic and social

elements. Lack of access to education leads to poor economic opportunities, while low incomes

tend to result in poor educational access and achievements in a vicious cycle of deprivation.

There are also reinforcing cycles of privilege and deprivation because of the way that one type

of capital requires others to be productive. (These cycles, which help explain the persistence of

horizontal inequalities, are explored in Chapter 4.)

The presence of sharp HIs provides a general motive for political mobilization. If governments

fail to respond to demands – or indeed repress them violently – this political mobilization may

become violent, with the power of identities binding people together. Where there are sharp

political HIs, then group leaders may find that violence is the only way to secure political power.

Similarly, where groups are denied high-level jobs (for example, in the bureaucracy) educated

people have a strong source of resentment. For the masses, in contrast, resentments can be

caused by lack of access to land and employment, as well as social services. Lack of cultural

recognition can be a running source of resentment, while particular attacks on cultural symbols

can be a trigger for conflict (for example, through the desecration of a holy place, as in the

destruction of the Babri Mosque in Ayodhya in India).

423
HIs may be spatially distributed: that is, particular regions of a country may be deprived (or

privileged) compared to other regions. In such cases, HIs can lead to separatist claims where

resource-rich provinces seek autonomy, resenting the redistribution of local resources to other

parts of the country (for example, Biafra in Nigeria or Aceh in Indonesia). Yet, sometimes it is

poorer regions which feel exploited by the richer areas (for example, in Bangladesh and Eritrea).

Different types of conflict emerge, however, where people from competing groups live in the

same geographic area, as in Rwanda or Burundi. In such cases, the deprived may seek political

and economic rights, or control over government institutions. There may also be attacks on

particular groups and pressure for ethnic cleansing without direct government involvement.

The four categories of HI, and some major elements within each, are presented in Table 1.1,

together with examples of where particular HIs appear to have been instrumental in provoking

conflict.

424
Table 1.1: Some examples of horizontal inequality in conflict situations

Political participation Economic Social access and situation Cultural status recognition

Assets Employment and incomes

Participation in government Land Incomes Education National holidays

Fiji, Burundi, Fiji, Cambodia, El Malaysia, South Africa, Fiji, Rwanda, Burundi, Haiti South
Bosnia and Herzegovina, Salvador, Haiti Chiapas Africa
Uganda Northern Uganda, Kosovo
Sri Lanka
Privately owned capital Government employment Health services Respect for cultural sites
Malaysia, South Africa, Sri Lanka, Fiji Burundi,
Burundi India,
Northern Uganda, Chiapas
Northern Ireland

Government Infrastructure Private employment Safe water Respect for cultural behaviour
Chiapas, Mexico, Burundi Fiji, Uganda, Malaysia Northern Uganda, Chiapas Latin America, historically
Army/police Aid ‘Elite’ employment Housing

Fiji Afghanistan, Sudan, South Africa, Northern Ireland

Northern Ireland, Rwanda Fiji,
Burundi, Kosova
Northern Ireland

Natural resources Unemployment Poverty

Liberia, Sierra Leone, South Africa, Northern Ireland Chiapas,
Indonesia Uganda,
South Africa
4.1 Horizontal inequalities in South Africa

The case of South Africa is illustrative of the role of HIs, their pervasive and multidimensional

nature and how they can be politicized, leading to political mobilization, protest and eventually

violence.

Historically, during the Apartheid era, the South African case was an extreme example

of very sharp HIs of every type. HIs between blacks (77 per cent of the population in

1996) and whites (10.9 per cent) were entrenched by a white political elite (initially

colonial) over the centuries preceding the democratic transition in 1993.

As one scholar observed:

The history of South Africa’s polity is dominated by the use of political power to attain

and maintain socio-economic ends. A white minority inherited political power in 1910

and during the next eight decades used this power to entrench itself politically and to

enhance its economic, cultural and social interests.

(Schrire, 1996: 59-50)

The large and consistent HIs that existed under the white-dominated government are

well known (see Figure 1.4).

• Economic HIs: Real per capita GDP of blacks in 1980 was 8 per cent of that of whites;

this had risen to 10 per cent by 1990 and just over 12 per cent by 2000 (van der Berg

and Louw, 2004). Whites still owned 90 per cent of the land in 2007.7

• Social HIs: State education expenditure on each white student was 14 times that on each

black student in 1980 and adult literacy among blacks was two-thirds that of whites. In

1993, there was only one formal-sector brick house for every 43 Africans compared to
 one for every 3.5 whites.(Knight, 2001). Infant mortality rates among blacks were six

times those of whites in 1980 and life expectancy was 56 among blacks and 70 among

whites (UNDP, 2003). The ratio of the Human Development Index of blacks to whites

was 0.64 in 1980, and 0.60 in 1990 (UNDP, 2003).

• Political HIs: throughout the period, the whites had all the cabinet posts,

accounted for 94 per cent of the higher echelons of the civil service, and all the

senior positions in the police and army (UNDP, 2000).

Figure 1.4 Ratio of performance of black South Africans to white South Africans

10

6
1980
1995
ratio

5
2002
4

0
u/e rate* Labour HDI Service HIV/AIDS Civil service Real per
participation* deprivation management capita
income
dimension
Source: UNDP 2003. * ratio of Black to other.

408
Following unsuccessful peaceful protests, the sharp HIs in every dimension led to

armed rebellion from 1976, until the transfer of power in 1993. Over this period there

was some diminution in HIs, partly for economic reasons and in a very partial and

unsuccessful effort to secure peace without transferring power.

An overriding objective of the black majority post-1993 government has been to reduce

black/white differentials, but these efforts have been constrained by limits on

government expenditure and by the economic liberalization agenda. There was a

complete reversal in political inequalities at the top, and a major reduction in HIs in

managerial posts in the civil service, where the black share had risen to 63 per cent by

1996. There was a substantial improvement in service access for the black population,

and the ratio of blacks/whites on an index of service deprivation fell from nine in 1995

to 2.4 in 2002 (UNDP, 2003). The poverty rate among blacks, which had been over 40

times that of whites in 1995, fell to eight times in 2002. Differentials in infant mortality

rates, life expectancy and adult literacy initially narrowed but are rising again with the

unequal incidence of HIV/AIDS. The ratio of the HDI of blacks to whites rose from 0.60

in 1990 to 0.73 in 2000 (UNDP, 2003).

Efforts to reduce economic inequalities were less successful. While the blacks’ share

of national income rose from under half in 1985 to three-quarters in 1995, their share of

population also increased and there was only a small decrease in the differential of real

adjusted GNP per capita. Efforts to ‘empower’ black business by increasing their role in

private capital ownership, with a target of 30 per cent ownership, appear to have

faltered with the share of quoted companies with significant black influence at just 5%

in 2007 (Financial Times, 29 June 2007), while Sherer (2000) finds evidence of

persistent labour-market discrimination in the post-apartheid era. In general,

409
differentials are diminishing, but remain extremely high. According to a 2005 report,

‘Vast racial and gender inequalities in the distribution of and access to wealth, income,

skills and employment persist’ (DTI, Republic of South Africa, 2005: para. 1.5).

Nonetheless, with political power and increasing socioeconomic opportunities, the

black population gives broad political support to the government, while the white

population remain the gainers economically. There seems little danger of renewed

conflict, but the extremely high level of criminality testifies to the unsatisfactory nature

of the economic inequalities.

5. HIs and conflict: the main hypotheses

This section considers the main hypotheses concerning the relationship between HIs and violent

conflict, to be explored in the rest of the book.

In this book when we talk of violent conflict we are referring to serious political violence, i.e.

violence that primarily has political objectives. Thus we are not referring to criminality, domestic

violence, or relatively minor episodes of political violence, of which there are many. Such violent

political conflicts can be of different kinds. One important distinction is between conflicts that

involve the state, as one significant actor, and those that do not. Within the former category,

there are conflicts in which state repression, sometimes spontaneous and sometimes in

response to rebellion, is an important source of violence. A second type of violence in which the

state is integrally involved is separatist conflict. There are also communal conflicts in which the

state is not one of the parties to the conflict, though it might intervene on one side or the other.

The violent conflicts discussed in this book include examples of each of these kinds.

The hypotheses concern the political outcomes likely to be associated with HIs, where HIs are

observed inequalities between groups. We are aware that people take actions on the basis of

410
their perceptions of others and of their relative position rather than actual inequality. However,

in much of this book we will investigate the relationship between externally measured

inequalities, rather than self-perceived HIs, and conflict. The validity of this approach rests on

the assumption that perceptions broadly reflect the observed reality. This assumption is being

investigated by surveys of perceptions in seven countries8. Chapter 10 reports on the findings of

two of these surveys – in Ghana and Nigeria – permitting some insights into the relationship

between observed and perceived HIs.

1. Our first hypothesis is that conflict is more likely where there are significant political or

economic HIs, or both.

Both political and socioeconomic inequalities are of major relevance to political outcomes: strong

political HIs mean that leaders of groups feel politically excluded and are thus more likely to lead

opposition and possibly rebellion; while socioeconomic inequalities mean that the people as a

whole have strong grievances on ethnic lines and are thus likely to be more readily mobilized.

2. Our second hypothesis is that political mobilization is especially likely where there are

consistent HIs, i.e. both political and economic HIs run in the same direction.

Where HIs are inconsistent (that is, there are political HIs but not economic, or there are

economic HIs but not political), it seems probable that they are less conflict-promoting than

where they are consistent. This is because where there are political HIs but not economic, that

is, the economically privileged are politically excluded, the group may mind less about their

political exclusion. This seems to have been the case, for example, among the Kikuyu under Moi

in Kenya. Similarly, where the economically underprivileged are politically included, the stimulus

to opposition will be less, because the potential leaders gain from their political inclusion, and

the mass of people may get some satisfaction from that and from the prospect that political

411
inclusion will confer economic benefits. This describes the situation in South Africa post-

apartheid.

3. Our third hypothesis is that lack of cultural recognition and equity, or cultural status HIs, will be

provocative, while cultural inclusion will help sustain peace. A change in either direction may be

particularly relevant to group mobilization.

As we proceeded with our research into HIs, it became apparent that as well as political and

socioeconomic HIs, cultural events and mores were highly relevant to group grievances. These

include issues such as language and education policy, recognition given to particular groups in

terms of national holidays, formal and informal attitudes to dress and other types of group

behaviour. While it is difficult to measure ‘cultural HIs’, they do form a component of the overall

picture, and one which Brown and Langer explore further in Chapter 3.

4. Our fourth hypothesis is that political mobilization and possibly conflict will become more

likely where HIs are widening.

The direction of change in HIs is also likely to be relevant to political outcomes, since if the

situation is worsening it seems probable that this is provocative, and if it is improving it will be

ameliorative.

Apart from these formal hypotheses, the context in which HIs occur also matters. In particular,

political conditions (including the nature of the state), cultural demographic conditions and

economic conditions all affect the likelihood of conflict for any given state of horizontal

inequalities.

5.1 Political conditions

412
The nature of the state is of fundamental importance in determining both the cause and the

duration of conflict. Several aspects are of relevance:

• The structure of the state and of power. Some political systems are highly concentrated,

lacking division of powers or geographic decentralization of power. In general, one might

expect conflict to be more likely with highly centralized systems than in less decentralized

systems (such as federations, or where power is diffused across institutions, or there is

substantial decentralization), because in centralized systems more is at stake when

determining who controls the state (Suberu, 2001; Bakke and Wibbels, 2006).

• How accommodating the state is. People may feel grievances and mobilize politically, but

such mobilization need not take a violent form. Those with grievances may acquire

power, or be accommodated, peacefully. Violent mobilization is more likely where there

is little accommodation. In general, one would expect democratic systems to be more

accommodating, especially where the economically deprived group is in a majority.

However, fragile democracies can be conflict-provoking so there is no simple equation

between democracy and conflict (Snyder, 2000; Stewart and O'

Sullivan, 1999). Relevant

to this is the nature of the democratic system. A first-past-the-post, winner-takes-all

system (the Westminster model) is likely to be much more conflict-prone than one

involving more power-sharing of a formal or informal nature. Thus proportional

representation systems have been found to be less conflict-prone. This may be because

the presence of PR and other power-sharing mechanisms tend to reduce political HIs, as

well, probably, as reducing the likelihood of conflict for any given political inequalities.

Whether democratic or not, governments can be accommodating and inclusive making

violent opposition less likely. In contrast, governments (especially nondemocratic ones)

413
 may make no attempt to meet people’s demands, and can react to opposition with harsh

repression, which may provoke a further violent reaction.

5.2 Cultural demography

The potential strength of any violent movement depends also on the geographic location and

relative size of the population in different groups. Where groups are concentrated geographically

(or largely so), separatism or irredentism often becomes a political aim. But this is not a

possibility where members of each group are dispersed across the country. In the latter case,

groups may mobilize in order to gain control over the state, or improve their political and

economic rights, without separatist ambitions.

As far as relative size is concerned, at one extreme many small groups may find it difficult to

mobilize collectively. At the other, if there is a homogeneous population violence may be less

likely, though it can still occur as new and different identities emerge from what appears to be a

homogeneous whole – for example, Somalia and Haiti, both relatively homogeneous in some

ways, have been subject to endemic conflict (Bangura, 2006; Collier and Hoeffler, 2000; Fearon

and Laitin, 2003). In between there are many possibilities. If the privileged group forms the vast

majority and the underprivileged group a small minority then any violent protest is likely to be

readily suppressed (as in the northern part of Kenya), although foreign support can prolong the

conflict (as in northern Uganda). Where there is a small privileged group and a large

underprivileged one, the deprived can be mobilized against the privileged – the Jews, the

Chinese and the Lebanese, for example, have been subject to such attacks periodically over

many centuries.9 The presence of a few large groups may be conducive to conflict, as each

fights for political supremacy (an example is Fiji; but such conflict is not always violent, as shown

in the case of Guyana)10.

414
A further relevant feature is the cohesion of the group; strong cohesion and hence mobilization

potential may be the outcome of cultural unity, or of political leadership emphasizing group unity.

As we have argued, leadership, the media and treatment by others are important in determining

which identities are felt so strongly that people mobilize for political violence behind them.

However, cohesion is inherently easier in some contexts. Thus, where there is cultural

fragmentation (for example, many languages being spoken), mobilization behind some

overarching identity may be less likely, and geographic dispersion may also make cohesion less

likely. For example, in Peru, one possible overarching identity is an ‘indigenous’ identity, which

would bring together about half the population – a powerful political entity. Yet differences within

the indigenous population as to language, history and location may mean that such an

overarching identity is not strongly felt and hence political action on the basis of indigeneity may

be unlikely. Yet it seems that powerful leadership, plus discriminatory treatment, can get people

to overcome such differences, as we observe among the Orang Asli in Malaysia, where a pan-

Asli identity has emerged. This group contains at least 17 distinct subgroups, yet they came

together to protest colonial and later postcolonial discrimination against them (Nicholas, 2000).

Two features seem to determine a group’s potential for mobilization. One is whether others

categorize people as belonging to a single group, thereby stimulating unity through external

pressure; the other is the presence of leadership (normally, but not always, from within the

group) which is effective in mobilizing people as a group. In some situations it seems deprived

groups can be mobilized because of their felt grievances, yet not in the name of the group but

under some other (often ideological) banner. This seems to have been common in the Cold War

era.

5.3 Economic conditions

415
The nature of the economy is another type of conditioning factor which helps determine the

outbreak (or not) of violent conflict. Countries with low incomes per head are more prone to

violence according to econometric investigations (Auvinen and Nafziger, 1999). This may reflect

lack of viable occupations among poor populations, who therefore find war an attractive

proposition; or lack of state resources, so that the state offers too little to its citizens to persuade

them to respect their civic obligations; or it could be that low incomes are associated with weak

states which are unable to repress violence effectively.

Low-growth economies seem to be more violence-prone than high-growth. This is probably

because with higher growth often all groups benefit to some extent, and so inequalities seem to

matter less. If the economy is growing then peaceful economic opportunities expand, so people

may be less likely to want to disturb the economy and become involved in conflict.

Many econometric investigations also show that the presence of natural resources raises the

risk of conflict (Ross, 2004). There are a number of possible mechanisms which explain this

(Humphreys (2005)). In many contexts, the presence of high value natural resources is an

important source of HIs, both regionally, as the resource rich areas are richer than the rest of the

country, and locally, as particular groups within a locality gain control over the resources.

Sometimes, moreover, it’s a matter of ‘aspiration to inequality’ (Tadjoeddin, 2007) when the

populations of regions rich in natural resources find themselves relatively poor in terms of levels

of living, and resent the redistribution of ’their’ resources’ to the rest of the country. There are

also mechanisms which may explain the connection between natural resources and conflict

which have nothing to do with HIs. For example, natural resources offer finance for the

prosecution of conflict; it may enrich those fighting (or offer them the prospects of

enrichment);and natural resource availability can weaken state capacity and the state’s

relationship with its citizens because of the high levels of corruption that often result. It is

416
plausible that more than one mechanism is in play in any particular case. What this suggests for

our work is that the presence of natural resources may generate an additional propensity to

conflict, at times operating by causing sharp HIs, and at times independently of them.

The presence of particular political, demographic and economic conditions of the sort just

discussed will mediate the relationship between HIs and conflict. For example, conflict may be

less likely to break out, given sharp HIs, where there is a growing economy, or where a very

strong state represses any conflict, or where political accommodation occurs. Moreover, HIs are

not, of course, the sole source of conflict.11 Other economic explanations of conflict include:

private incentives where war provides opportunities of enrichment and where alternative

opportunities are very poor (with low incomes, high unemployment), fighting may offer an

attractive alternative; environmental pressures have also been argued to lead to conflicts over

resources, especially land; and conflict may arise from a failure of the social contract, where the

state fails to deliver public services, security and incomes. Each of these explanations has some

statistical and case-study support. Moreover, different conflicts may have different explanations,

with more than one often appearing to be relevant. For example, the north/south conflict in

Sudan was both an example of horizontal inequality (with the south being heavily deprived), and

one of powerful private gains that perpetuated the struggle (Keen, 1994); and while it has been

argued plausibly that environmental poverty was a significant element in the conflict in Rwanda

(André and Platteau, 1996), horizontal inequalities were clearly also important.

It thus follows that we should not expect an automatic relationship between HIs and conflict,

not only because particular factors intervene to make conflict more or less likely, but also

because there are other factors which may cause conflict even where HIs are limited.

Nonetheless, if we conclude that HIs can be an important source of violent conflict, this is of

major significance since it points to a range of policies that could help prevent conflict by

417
reducing HIs. The rest of this book is devoted to exploring whether HIs indeed do cause

conflict, and under what conditions, following our four hypotheses above.

6. Plan of this book

The book is divided into four parts. The first part deals with general and conceptual issues

pertaining to horizontal inequalities. First, Gibney analyzes an important aspect of political

inequalities – notably how the concept of citizenship can be used as a mechanism of

exclusion; secondly, Brown and Langer take up the issue of cultural status inequality,

analyzing the concept and showing its relationship to group mobilization. One of the reasons

horizontal inequalities are so significant for individual well-being is that they are often very

persistent, lasting generations and sometimes even centuries, so that they trap individuals in

situations of inequality. The question of why such inequalities are so persistent in some

contexts but not in others is dealt with by Stewart and Langer in Chapter 4.

Part II of the book deals with issues of measurement and statistical analysis. It starts with a

general discussion of how to measure horizontal inequalities (,Mancini, Stewart and Brown).

This is followed by two chapters that use econometric analysis to investigate the relationship

between HIs and violent conflict: Mancini does this across districts within Indonesia in the

1990s and Østby uses cross-country data and investigates how political as well as

socioeconomic inequalities affect conflict.

Part III presents material from in-depth studies, including Côte d’Ivoire, Ghana and Nigeria in

West Africa; Bolivia, Guatemala and Peru in Latin America and secessionist movements in

Southeast Asia. This includes an analysis of political inequalities in Ghana and Côte d’Ivoire

by Langer (Chapter 8), followed by a comparison of two cities in Nigeria, by Ukiwo. To

418
complete the analysis of the West African case, Chapter 10 reports on surveys of perceptions

of identity and inequality in Ghana and Nigeria. Chapter 11 presents a comparative analysis

of Bolivia, Guatemala and Peru by Caumartin, Gray Molina and Thorp; and Brown provides an

analysis of secessionist movements in Southeast Asia in Chapter 12. ,

The concluding section (Part IV) first presents findings from the book as a whole. (Chapter

13). The final chapter then analyzes a wide range of policies that would contribute to reducing

HIs, and considers some political problems and constraints in introducing them.

1
CSP, Global; Conflict Trends, http://members.aol.com/CSPmgm/conflict.htm
2
Summed up in advice given in 1858 by a Jew in Prussia, to his son: ‘Always remember that

you are a proud citizen of Prussia, entitled to equal rights. And never forget that you are a Jew. If

you do, there will always be others to remind you of your origins’ (Frister, 2002: 58).
3
A view associated with Smith (1986; 1991) and Soviet ethnobiologists – for example, Bromley

(1974).
4
He argued that ‘Hausa identity and Hausa ethnic exclusiveness in Ibadan are the expressions

not so much of a particularly strong “tribalistic” sentiment as of vested economic interests’

(Cohen, 1969: 14).

5
Smith has argued that ‘the [past] acts as a constraint on invention. Though the past can be

“read” in different ways, it is not any past’ (1991:357-358, quoted in Turton, 1997).

6
Reynal-Querol (2001: 2) argues for developing countries that ‘religious identity is fixed and

non-negotiable.’

419
7
S. Tsiko. ‘South Africa: Whites Own 90% Land’. The Black Star News, March 16, 2007.
8
The Centre for Research on Inequality, Human Security and Ethnicity has carried out surveys

of perceptions in Ghana, Nigeria, Peru, Bolivia, Guatemala, Malaysia and Indonesia.

9
The type of violence analyzed by Chua (2003).

10
See the case studies in Bangura (2006).
11
See Collier and Hoeffler (2000) who emphasize private incentives; Nafziger and Auvinen

(2002) who attribute conflict to individual grievances and a failed social contract; and Homer-

Dixon (1994) who points to environmental pressures.

420