P1: JZP

9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

I n late-century Africa, things fell apart. By way of illustra-

tion, consider Figure 1.1, which lists civil wars in African
countries from 1970 to 1995, as judged by the World Bank.
As time passes, the list grows. Angola, Chad, Namibia,
Nigeria, and Sudan enter the 1970s war-torn; in the mid-1970s,
Sudan exits the list, but Equatorial Guinea and Zimbabwe join
it; by 1980, Zimbabwe departs from the ranks of the war-torn,
but is replaced by Mozambique, Nigeria, and Uganda. The
pattern – a few dropping off, a larger number entering in –
continues into the early 1990s. Only one country that was con-
flict ridden in 1990 becomes peaceful by 1992, while eleven
others crowd into the ranks of Africa’s failed states.
Humanitarians, policymakers, and scholars: Each de-
mands to know why political order gave way to political con-
flict in late-century Africa. Stunned by the images and realities
of political disorder, I join them in search of answers. In so
doing, I – a political scientist – turn to theories of the state and

3
 P1: JZP
9780521887359c01

year 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95

Burundi

Chad
CUNY1261/Bates

Congo
Djibouti
Ethiopia

Kenya
Liberia

4
Mali

Mozambique
Namibia
978 0 521 88735 9

Nigeria
Rwanda
Senegal
Sierra Leone
Somalia
Sudan

Uganda
Congo
Zimbabwe

Figure 1.1. Civil wars, Africa 1970–1995. Source: World Bank (Sambanis 2002).
November 13, 2007
18:1
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

locate the sources of political disorder midst the factors that

lead states to break down.
I anchor this book in the work of Weber (1958) and view
coercion as the distinctive property of politics. As will become
clear in the next chapter, I depart from Weber – and his “struc-
turalist” descendants1 – by turning to the theory of games.
Driven by the realities of Africa, I view political order as
problematic: In light of the evidence Africa offers, political
order cannot be treated as a given. Rather, I argue, it results
when rulers – whom I characterize as “specialists in violence” –
choose to employ the means of coercion to protect the creation
of wealth rather than to prey upon it and when private citizens
choose to set weapons aside and to devote their time instead
to the production of wealth and to the enjoyment of leisure.2
When these choices constitute an equilibrium, then, I say,
political order forms a state.3
To address the collapse of political order in late-century
Africa, I therefore return to theory – the theory of the state – and
to theorizing – the theory of games. I do so because proceeding
in this fashion points out the conditions under which political
order can persist – or fail. I devote Chapter 2 to an informal

1
Evans, P., T. Skocpol, and D. Rueschmeyer (1985), Bringing the State Back
In, Cambridge, U.K.: Cambridge University Press provides perhaps the
best-known example.
2
I am drawing on Bates, R. H., A. Greif, et al. (2002), “Organizing Violence,”
Journal of Conflict Resolution 46(5): 599–628.
3
The ambiguous phrasing is intended.

5
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

derivation of those conditions. In the remaining chapters, I

turn from deduction to empirics and explore the extent to
which these conditions were to be found, or were absent, in
late-century Africa. The evidence leads me to conclude that
in the 1980s and 1990s, each of three key variables departed
from the levels necessary to induce governments and citizens
to choose in ways that would yield political order.

The Literature

Following the outbreak of conflict in Serbia, Somalia, Rwanda,

and elsewhere, the study of political violence has once again
become central to the study of politics. Familiar to many, for
example, would be the attempts by Collier and Hoeffler (2004)
and Fearon and Laitin (2003) to comprehend the origins of civil
wars. Also familiar would be studies of the impact of ethnic-
ity (Fearon and Laitin 2003), democracy (Hegre, Gates et al.
2001; Hegre 2003), and natural-resource endowments (e.g.,
Ross 2004). In my attempts to comprehend why things fell
apart in late-century Africa, I draw upon these writings. But I
also take issue with them, for virtually all share common prop-
erties from which I seek to depart.
Consider, for example, the assumption that civil war can be
best treated as the outcome of an insurgency. When thinking
about the origins of political disorder in Africa, I can find no
way of analyzing the origins of insurrection without starting

6
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

with the behavior of governments. The conditions that led

to the breakdown of order in Africa include the authoritarian
nature of its states and their rulers’ penchant for predation. By
rendering their people insecure, they provoked insurgencies.
While both insurrectionaries and incumbents must necessar-
ily feature in the analysis of political disorder, in this instance it
makes sense not to focus exclusively on the rebels but to stress
as well the behavior of those whom they seek to drive from
power.
Recent contributions exhibit a second common feature:
the methods that they employ. Utilizing cross-national data,
they apply statistical procedures to isolate and measure the
relationship of particular variables with the onset and duration
of civil wars. I, too, make use of cross-national data; but rather
than collecting data for all countries in the globe, I restrict my
efforts to Africa. I do so in part because Africa provides an
unsettling range of opportunities to explore state failure and
because political disorder is so important a determinant of the
welfare of the continent. I also do so because I find it necessary
to draw upon my intuition. To employ that intuition, I need
first to inform it, be it by immersing myself in the field or in
qualitative accounts set down by observers. I have therefore
made use of a selected set of cases – those from the continent
of Africa – and my knowledge of their politics.4
4
The use of a subset of countries also eases the search for exogenous vari-
ables, and thus causal analysis. For example, given the small size of Africa’s

7
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

Lastly, if only because they are based on the analysis of

cross-national data, contemporary studies exhibit a third
property: Their conclusions take the form of “findings.” These
findings are based upon relationships between a selection of
key variables and the outbreak or duration of civil wars. Collier
and Hoeffler (2004), for example, stress the importance of
“opportunities,” that is, chances to secure economic rewards
and to finance political organizations. Noting that the magni-
tude of primary product exports, the costs of recruiting, and
access to funding from diasporas relate to the likelihood of
civil war, they conclude that “economic viability appears to be
the predominant systematic explanation of rebellion” (p. 563).
Fearon and Laitin (2003), by contrast, conclude that “capa-
bilities” play the major role: “We agree that financing is one
determinant of the viability of insurgency,” they write (p. 76).
But they place major emphasis on “state administrative, mil-
itary, and police capabilities” (p. 76), measures of which bear
significant relationships to the outbreak of civil wars in their
global set of data.
In this work, I proceed in a different fashion. I start by
first capturing the logic that gives rise to political order. While
I, too, test hypotheses about the origins of disorder, I derive

economies, I can treat global economic shocks as exogenous – something

that yields inferential leverage when seeking to measure the impact of
economic forces on state failure.

8
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

these hypotheses from a theory. By adopting a more deductive

approach, I depart from the work of my predecessors.

Key Topics

Energized by such works as Kaplan’s “The Coming Anarchy”

(1994), students of Africa have focused on the relationship
between ethnic diversity and political conflict. At least since
the time that William Easterly and Ross Levine penned “Africa’s
Growth Tragedy” (1997), empirically minded social scientists
have sought to capture the impact of ethnicity on the eco-
nomic performance of Africa’s states. Interestingly, however,
they have found it difficult to uncover systematic evidence of
the relationship between measures of ethnicity and the likeli-
hood of political disorder.5
In this study I, too, find little evidence of a systematic rela-
tionship. And yet, the qualitative accounts – be they of the
killing fields of Darfur or of the tenuous peace in Nigeria – con-
tinue to stress the central importance of ethnicity to political
life in Africa. In response, I argue that ethnic diversity does
not cause violence; rather, ethnicity and violence are joint

5
For a discussion, see Bates, R. H., and I. Yackolev (2002), Ethnicity in Africa,
in The Role of Social Capital In Development, edited by C. Grootaert and T.
van Bastelaer, New York: Cambridge University Press; and Fearon, J., and
D. Laitin (2003), “Ethnicity, Insurgency and Civil War,” American Political
Science Review 97(1): 75–90.

9
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

products of state failure. Their relationship is contingent: It

occurs when political order erodes and politicians forge polit-
ical organizations in the midst of political conflict.
The political significance of resource wealth has also
attracted much attention. Analyzing their data on civil wars,
Collier and Hoeffler (2004) report that “dependence upon pri-
mary commodity exports” constituted “a particularly power-
ful risk factor” for the outbreak of civil war (p. 593). Africa
is, of course, noted for its bounteous natural endowments of
petroleum, timber, metals, and gemstones. And scholars and
policymakers have documented the close ties between the dia-
mond industry and UNITA (National Union for the Total Inde-
pendence of Angola) in Angola (Fowler 2000), the smuggling
of gemstones and the financing of rebels in Sierra Leone (Reno
2000), and the mining of coltan and the sites of rebellion in
eastern Zaire (present-day Democratic Republic of the Congo)
(Kakwenzire and Kamukama 2000).
And yet, using Collier and Hoeffler’s (2004) own data,
Fearon (2005) has demonstrated that their findings are frag-
ile, depending in part on decisions about how to measure
and classify cases. In this study, too, I fail to find a signifi-
cant relationship between the value of natural resources and
the likelihood of state failure.6 Once again, then, there arises

6
For both Fearon (2005) and myself (this work), only the value of petroleum
deposts is related to political disorder. Even here the relationship is fragile,
however.

10
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

a disparity between the evidence from cross-national regres-

sions and that from qualitative accounts. I shall argue that the
disparity suggests that the exploitation of natural resources
for war finance is a correlate rather than a cause of political
disorder.
A third factor plays a major role in the literature: democ-
ratization. Qualitative accounts, such as those of Mansfield
and Snyder (Mansfield and Snyder 1995; Snyder 2000) sug-
gest that democratization produces political instability and
leads to the mobilization of what Zakaria (1997) calls “illib-
eral” political forces. Careful empirical researchers, such as
Hegre (Hegre, Gates et al. 2001; Hegre 2004), confirm that new
democracies and intermediate regimes – those lying some-
where between stable authoritarian and consolidated demo-
cratic governments7 – exhibit significantly higher rates of civil
war. As demonstrated by Geddes (2003), many of these inter-
mediate regimes are the product of the “third wave” of democ-
ratization (Huntington 1991) and the collapse of communist
regimes and are therefore themselves new and vulnerable to
disorder.
In the 1980s and 1990s, many of Africa’s governments
reformed. Regimes that once had banned the formation
of political parties now faced challenges at the polls from

7
Using Polity coding. Available online at: http://www.cidcm.umd.edu/
polity/.

11
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

candidates backed by an organized political opposition. And

in the late 1980s and early 1990s, militias assembled, states
failed, and Africa faced rising levels of political disorder. The
experience of Africa thus appears to conform to what the liter-
ature has recorded: Electoral competition and state failure go
together.
In analyzing the impact of political reform, I employ two
measures: the movement from military to civilian rule and the
shift from no- or one- to multiparty systems. In discussions of
democracy, the followers of Schumpeter (1950) argue for the
sufficiency of party competition; those of Dahl (1971) contend
that party competition is necessary but not sufficient. Without
an accompanying bundle of political and civil rights, the latter
argue, contested elections are not of themselves evidence of
democratic politics. In debates over the relationship between
party systems and democracy, I concur with the followers of
Dahl. When addressing political reform, I pay no attention to
the number of political parties, their relative vote shares, or
the conditions under which the opposition is allowed to cam-
paign. I therefore address not the relationship between democ-
racy and political conflict but rather the relationship between
political reform and political disorder.
Lastly, there are those who emphasize the impact of pov-
erty. That poverty and conflict should go together is treated
as noncontroversial, as if disorder were simply an expected

12
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

corollary of the lack of economic development.8 But consider:

If, as many argue, lower per capita incomes imply lower wages
and therefore lower costs of rebellion, so too do they imply
fewer gains from predation; income thus cancels out the ratio
between the costs and benefits. From the theoretical point of
view, moreover, there is simply little that can be said about the
relationship between the average level of income – or, for that
matter, poverty – and incentives for violence. As I will argue
in Chapter 2, for our purposes, discussions of private income
can be set aside; for the logic of political order suggests that
the focus be placed not on private income but rather on public
revenues. Economic shocks will indeed play a major role in this
analysis, but the focus will be on their impact on the revenues
of states, not on the incomes of individuals.9 In this work, when
I measure the impact of income per capita, I treat it as a control
variable, rather than as a variable of theoretical interest.
In Chapter 2, I parse the logic of political order. I recount the
theory informally, portraying the interaction between govern-
ments and citizens and among citizens as well. Presented as a

8
Indeed, see Sambanis, N., and H. Hegre (2006), “Sensitivity Analysis of
Empirical Results on Civil War Onset,” The Journal of Conflict Resolution
50(4): 508–35. The authors point to per capita income as one of the very
few variables that bears a robust relationship with civic violence.
9
See the arguments in Hirshleifer, J. (1995), Theorizing About Conflict, in
Handbook of Defense Economics, edited by K. Hartley and T. Sandler, New
York: Elsevier.

13
P1: JZP
9780521887359c01 CUNY1261/Bates 978 0 521 88735 9 November 13, 2007 18:1

Introduction

fable, the argument is based upon rigorous foundations and

points to the conditions under which governments choose to
engage in predation and citizens choose to take up arms.10
Chapters 3 through 5 set out the conditions that prevailed
prior to the collapse of political order. They document the
social and political configurations that were in place at the
time of the impact of the economic and political shocks that
dismantled the state in Africa. In Chapter 6, states fracture
and political disorder engulfs nations in Africa. Chapter 7
concludes.

10
The informed reader will note the parallels between my analysis and that
of Azam, J.-P., and A. Mesnard (2003), “Civil War and the Social Contract,”
Public Choice 115(3–4): 455–75; Snyder, R., and R. Bhavani (2005),
“Diamonds, Blood and Taxes: A Revenue-Centered Framework for Ex-
plaining Political Order,” The Journal of Conflict Resolution 49(4): 563–
97; and Magaloni, B. (2006), Voting for Autocracy, New York: Cambridge
University Press.

14