The Myth of the Failed State: Political Order and Violence in Sub-Saharan Africa

In 2001, the term “state failure” became one of the structuring concepts of political thought and policy-making in the Western world. Although the term itself – along with the accompanying terminology of failing, troubled, and stressed states – was coined in the early 1990s to describe the situation in war-torn African countries and some of the post-communist societies, it was not until the attacks on the World Trade Centers on September 11, 2001 that it became popular. Any reflection on state failure must keep this relation between the concept and the historico-political arena in mind: since failed states could host terrorist organizations harmful to the national security of the United States, they could no longer be the focus of exclusively humanitarian organizations and political scientists, but now also merited discussion by American policy makers and security officials.

One of the most interesting consequences of state failure is violence. If for a great deal of modern sociology and political science the conception of the state is based upon Weber’s definition of the state as the entity that “upholds the claim to the monopoly of the legitimate use of physical force in the enforcement of its order,” the failure of the state would be defined as the loss of this monopoly, and it would have an empirical and direct increase in violence as its first corollary.

In this essay, I will analyze the links between “state failure” and violence in sub-Saharan Africa through case studies on Zambia, Sierra Leona and the Democratic Republic of Congo. In the first part, I will discuss different views of what constitutes a “failed state.” In the second part, I will apply the most suitable view, expanding it in order to apply it to the aforementioned case studies. Afterwards, I will study incidents of violence and death in each country, drawing on statistics of homicide, crime and a few other social indicators of material welfare. Ultimately, I argue that the success of modern states cannot only be defined by the monopoly of violence; states must be seen as institutional actors responsible for the provision of certain amounts of material welfare. State failure, then, cannot be defined simply as the loss of the monopoly of violence, but also as the inability of the state to fulfill these basic obligations, in which case the mortality rate could be seen as a primary indicator of state failure.

Failed State Theory

Scholars themselves have not come to an agreement on what constitutes a failed state, which means the same states can be classified as “failed” and “not failed” depending on the definition being used. The American political scientist Harvey Starr provides an overview of a number of definitions:

The careful reader may have been struck by perhaps the most problematic aspect of the study of state failure and of the existent literature—issues of conceptualization, definition, and measurement. The articles in this issue reflect the heterogeneity (confusion?) of definition and measurement regarding “state failure” or “failed states.” In brief, Goldstone defines state failure in terms of “stability,” while Bates is interested in “political order.” For Bates, “the mark of state failure is the government’s loss of its monopoly over the means of coercion.” Iqbal and Starr, after raising the issue of conceptual and definitional disarray, see state failure as “collapse,” using the POLITY code of “the complete collapse of central political authority.” Chauvet and Collier are interested in “failing” states, based on “inadequate performance of socio-economic functions.” Thus, they study low-income countries with extremely weak economic policies, institutions, and governance. Finally, Carment et al. are concerned with at-risk states or “state fragility,” which they admit is “an elusive concept.” For them, “fragile states lack the functional authority to provide basic security within their borders, the institutional capacity to provide basic social needs for their populations, and the political legitimacy to effectively represent their citizens at home and abroad.”

It is necessary to chose one of the aforementioned definitions, or in any case to come up with a set of criteria applicable to our three case studies. The definitions that seem to be the most complete are Bates’ and Chauvet & Collier’s. For the sake of simplicity, I will use Bates’ definition and set the threshold for state failure as the point at which “the government loses its monopoly over the means of coercion.” This theory means that a “successful” state is one in which all political agents within its territory have agreed to submit themselves to the authority of the government. In exchange for this submission, the state provides its citizens with a) personal security, b) basic health care and decent living standards, c) property security, and d) education. Defining the state’s responsibilities in this way furthermore necessitates a shift in terminology from “failed states” to “failing states,” as mentioned by Chauvet & Collier, since these elements can be partially provided, or provided only to a certain subset of the population. Weber, in Politics as a Vocation, stresses that it is the use of violence that allows for the formation of states, but admits that “force is certainly not the normal or the only means of the state (…) but force is a means specific to the state.” In this context, the four aforementioned elements constitute those other means of the state to preserve its power mentioned by Weber. One consequence of the adoption of an expanded Weberian definition of state and hence state failure combined with Chauvet & Collier’s focus on the performance of socio-economic functions is that several of the elements of the Failed States Index appear to be theoretically wrong: for instance, states can be deemed efficient even should uneven economic development along group lines and the rise of factional, aggressive and nationalistic elites persist. States exist in class-divided societies, and one of the main roles of violence is to maintain such a divide, hence revenue inequality cannot be a measure of state failure.

The extent of failure of a certain state depends on its ability to provide the aforementioned elements of material welfare, but the mark to determine if it has failed or not depends on the emergence of a new political actor who uses force and the consequent inability of the state to stop it. In this way, states fall along a continuum in which an ultimately successful state, i.e. one in which there is no other violent political actor, may be haunted by its relative inability to provide its citizens with personal and property security, education, and decent living standards. Chauvet & Collier’s definition, then, establishes this continuum; Bates’ sets the threshold for a state to be considered a failed one. However, Bates’ threshold sees state failure as contingent on the appearance of armed guerrillas. While this variable may be theoretically correct, as the existence of armed guerrillas indicates the failure of the state’s monopoly on violence, it should be discarded. First of all, though the emergence of guerrillas constitutes violence, if one defines a failed state as that one in which non-institutional violence takes place, one cannot simultaneously see violence as a possible consequence of state failure, nor can one adequately address cases in which violence persists not in the form of armed guerillas but in the form of petit-criminals. Furthermore, the mere existence of a guerilla group does not imply the inability of the state to eventually destroy it.

Before turning to the case studies, one must also examine the extent of violence in Sub-Saharan Africa more generally, by comparing statistics on the increase and decrease in violence in different African states to world averages. The Geneva Declaration calculates that 740,000 people die violently every year, but between three quarters and ninety percent of these deaths occur in non-conflict settings, meaning that crime is much more likely to kill an average citizen than armed conflict. Murder, furthermore, does not rank among the top ten causes of death: one is one hundred times more likely to die out of natural causes than to be killed by someone, and only one out of four assassinations has to do with armed conflicts. Even in light of this, however, the state’s role is crucial. From here onwards, I will take the world mortality rate of -.87 percent as the “natural” one, and the gap between it and the mortality rate of an individual state as due to the state’s lack of intervention or lack of efficiency.

Three Case Studies

Zambia has been a fairly peaceful country for the past few decades, with neither civil wars nor guerilla movements since independence from Britain in 1974. After several years of a single party state, a wave of protests in 1991 pushed the long-time president Kenneth Kaunda to allow elections and a multi-party system. Following the victory of his rival Frederik Chiluba, the country has been democratized to a great extent, and it seemed an oasis of liberal democracy and economic growth amidst the desert of civil wars and poverty that affected Africa during the 1990s. Yet Zambia remains one of the most violent countries in the world, with a life expectancy of only 48 years. The child mortality rate was .038% in 1988 – three years before the end of the dictatorship – but .074% in 2002. Democracy did not improve the general living standards of the population, and, according to some indicators, it has diminished. Zambia’s death rate is 1.6%, significantly more than the continent’s rate of 1.17%, and almost twice the world’s rate of 0.87%. Zambia’s homicide rate is also one of the five highest in the world, at .038%, or approximately 4,500 people killed every year.

In sum, it is clear that the Zambian state does not currently provide its citizens with the basic services that modern states should: criminality is rampant, people die young and children die often. Simultaneously, however, the literacy rate is not one of the highest of the region, but neither it is one of the lowest. Zambia’s economy is a fast growing one: since 2006, it has grown at a rate of more than 5%, meaning that the high rates of criminality and death cannot be explained by poor economic resources.

It is very easy to see that in Zambia the state has the monopoly on violence: despite the country’s low social standards, there are no armed groups aside from the army and police force. In this sense, Zambia cannot be considered to be a failed state. However, it is not an efficient one, or at least not for the bulk of the population. The wealth produced by the country in the past years has not been distributed amongst the majority of the population, which still lives in poverty and resorts to violent crime. The fact that these homicides are not perpetrated by a political opponent of the state but by simple criminals does not change the underlying analysis: the Zambian state is unable to provide its citizens with basic physical security and health services.

Sierra Leone, in contrast, seems like the textbook example of a failed state. The country went through an eleven-year-long war, from 1991 to 2002, in which some 50,000 people are estimated to have died as a direct consequence of the conflict. Almost a decade after the war, in 2010, the country had one of the lowest levels of development: the life expectancy is below 50 years, the literacy rate under 60%, the infant mortality rate 0.185%. Even if there was an improvement from the wartime situation, it is fair to say that the state is not able to perform several of its basic duties.

The degree of violence, however, also merits further scrutiny. If 50,000 people died during an eleven-year-long conflict, slightly fewer than 5,000 people died each year, which for a population of 5,000,000 people means a homicide rate of .1%, 2.5 times higher than the Zambian rate. A few years after the war, in 2004, the homicide rate dropped to .034%, meaning that the recovery of the monopoly of violence by a single political agent – the state – reduced homicide by 60 percent. Thus, while war-torn Sierra Leone had a yearly average of almost 5,000 violent deaths, three years after the end of the war, only slightly more than 2,000 people were murdered.

2,000 assassinations, however, is still a high figure, and thus merits an explanation. Furthermore, why is post-war Sierra Leone as underdeveloped and violent as the politically stable and economically dynamic Zambia? If our initial model based upon Bates’ definition were true, there should be a direct correlation between the existence of armed groups outside of the state on the one hand and a very high homicide rate on the other. Additionally, the existence of armed groups does not explain the tremendous gap in death rates between countries that underwent civil wars. While an eleven-year-long conflict in Sierra Leone killed 50,000 people, a five-year-long conflict in Congo killed about 5.4 million people. This fact points to the existence of another cause or set of causes explaining the incidence of death and violence, i.e. the monopoly of violence is not the sole explanation for its occurrence.

The Second Congo War started a year after the peace agreement that followed the deposition of United States-supported dictator Mobutu Sese Seko, who ruled the country from 1965 to 1997. Laurent Kabila’s government quickly proved itself incapable of ruling the country. Following Kabila’s political failure and an end to what had been a strong alliance with the Rwandan government, the population found itself divided along ethnic lines, in a bloody fight between some twenty-five to thirty armed groups and eight different nations for the control of its extremely rich mineral resources. According to Reuters and several non-governmental organizations, some 500,000 people were killed during battle, 10 times more than in Sierra Leone.

It is clear that the fight for mineral resources was at the heart of the Congolese conflict, a possible cause of the high death ratio. Perhaps, in Congo, political order and stability were only necessary for the different warring groups insofar as it allowed them to exploit resources, which in practice meant that it was enough to control the main mining spots and the links between them and ports or borders, while extending political stability to the rest of the country was simply beyond the groups’ economic interests. A humanitarian crisis that under any other circumstances would hinder the foundations of the warring groups’ claim to legitimacy, therefore, was not a problem in this case, since the groups simply did not work under the logic of a state-building political project, but under a predatory, resource-exhausting one.

Consequences and Findings

The case studies reveal the increase in the incidence of violence as the direct consequence of the state’s loss of monopoly over violence, with the homicide ratio up to .1% of the total population, 2.5 times higher than that of non war-torn countries. On the other hand, as we have seen, Zambia’s death ratio is 1.6%; twice the world’s ratio of .87%. In Congo, however, some 5,500,000 people died as a consequence of the war, which gives us a rather similar yearly death rate of 1.66%.

Thus, a narrow definition of state failure as simply the loss of the state monopoly over violence cannot account by itself for all the violence that has taken place in these three case studies: the deadliest conflicts in Africa have killed around .1% of the population, doubtless a very high figure, but still low when compared to pre-modern wars and to the amount of death provoked by curable sicknesses or by the complete disorganization of the economic system in war-torn countries.

Is hard to understand why a country like Zambia, that has not been through any sort of war lately and with a rapidly growing economy, has such a high death and homicide rate, and, if the average citizen of the developed world lives 75 to 80 years, the fact that the average Zambian lives no more than 48 cannot be explained by simply natural causes. The explanation, then, must be related to the state: while the Zambian state has had, for the past decades, a very stable political situation and a growing economy, it has been unable to provide basic relief for its citizens. It is clearly not efficient and it does not honour its duties; yet its power is uncontested. Can it be said that the Zambian state has failed? Certainly not, since no other political actor has ever tried to stand against it. Is it, then, a successful state? That also does not seem to be the case: a country with relatively stable levels of economic development with such poor performance on social indicators cannot be deemed to be successful by modern standards.

This observation, then, provides the grounding for our preference of the concept of “failing state” over “failed state.” The Zambian state is clearly failing; even if it has managed to maintain political order. The same definition can be applied, mutatis mutandis, to the Sierra Leone case: once the war was over, in 2002, the state recovered the monopoly of violence and there was an immediate diminution of the death ratio. The state stopped being a “failed” one. Yet, for international standards, the level of living of the population remained dangerously low, and the death ratio, though considerably lower than during wartime, remained very high. Did the Sierra Leone state suddenly jump from being a “failed state” to a completely successful and efficient one? Clearly not. Even if all the armed groups outside of the state were defeated, the state was nonetheless unable to provide basic health care, education or security for its citizens, which again places it into the category of “failing states.”

In this sense, the Congolese experience is an aggregated example of the aforementioned ones that allows us to measure the combined influence of civil war and lack of state efficiency. During the war, the state lost its monopoly of violence, which by itself killed around half a million people, but the fact that the different armed groups or the state were unable to provide with basic relief assistance for the population was what produced the overwhelming majority of casualties. This points to another element that lies at the heart of the difference in death ratios between Congo and Sierra Leona: the fact that a nation state loses control over part of its territory due to foreign invasion or regional secession (losing monopoly of violence; i.e., becoming a failed state according to Bates) does not mean that it loses its power over those regions that it still controls, nor that the seized regions sink into a state of chaos, looting and anarchy. On the contrary, if the guerilla group that seizes power in that region is a legitimacy-seeking, organized group, it will be better for them to establish political order and to become a state within the contested official jurisdiction of the enemy’s territory. This is what seems to have happened in Sierra Leone, in which, for most of the conflict the country was split into two parts: the south and the east controlled by the Revolutionary United Front rebels, which based its popularity upon the idea of a more equitable distribution of the state revenues. In contrast, the volatility and number of guerilla factions and foreign powers intervening in Congo made the situation much more deadly for the average civilian since no political group had an interest in administrating regions of the territory or aspired to produce any kind of political order within them. The consequence of this volatility is that the country as a whole sank into an abyss of chaos and anarchy.

Finally, then, the failed state theory does explain the levels of relative violence in these countries, but only when this theory is expanded to view the state as a set of political institutions that has the duty of guaranteeing a certain level of life and security to its citizens, as opposed to only as a violence-specialist holding a nation-wide monopoly. The failed state theory is problematic not because it is inherently wrong, but simply because it is used too liberally. Its future utility depends on the ability of the academic community to reach a consensus about its definition and methodology.


1) Call, Charles, The Fallacy of the Failed State, Third World Quarterly, Vol. 29, No. 8, 2008, pp 1491-1507

2) Weber, Max, Economy and Society, University of California Press, 1978

3) Starr, H., ed. 2006. Approaches, levels, and methods of analysis in international politics: Crossing boundaries. New York: Palgrave Macmillan

4) Starr, Harvey Introduction to the CMPS Special Issue on Failed States, Conflict Management and Peace Science, Volume 25, Issue 4, 2008.

5) Chauvet & Collier, Aid & Reform in Failing States, Asian‐Pacific Economic Literature, ISSN 0818-9935, 05/2008, Volume 22, Numéro 1, pp. 15 – 24

6) Bates, Robert The Logic of State Failure, Conflict Management and Peace Science, Volume 25, Issue 4, 2008.

7) Weber, Max Politics as a Vocation, at

8) “Failed States,”,

9) The Geneva Declaration Leaflet, in

10) The CIA World Fact book.






Leave a Comment

YRIS is a student publication, and Yale University is not responsible for its content.