War’s evolution – Contemporary wars: trends and research

The decline of interstate conflict following the Second World War and the ‘long peace’ of the Cold War implied a movement towards global peace. However, an eruption of ethnic conflicts in the early 1990s brought to light a different kind of conflict that researchers had thus far largely overlooked: civil war. A research boom has now produced differing findings on the causes of and possible approaches to prevent civil war.

At the end of the Second World War, a combination of factors led to the obsolescence of interstate wars. These included bipolarity, the invention of nuclear weapons and the diffusion of nationalism on a global scale, all of which undermined the wars of conquest. In a world where ‘war’ was defined as interstate war, this trend was understood as a movement toward global peace – the paradoxical ‘long peace’ of the Cold War. For many years, academic research on war and conflict was centred on international relations. States were seen as the only actors capable of waging war, and research was focused on understanding the decline of interstate conflict. Various theories were formulated to explain this decline. The most influential ones were nuclear deterrence theory, whereby the fear of certain mutual destruction contained the superpowers, and democratic peace theory, whereby the successive waves of democratization reduced the likelihood of interstate conflict because democracies were unlikely to wage wars against one another.

Lurking below the surface, however, were non-state actors who engaged in a type of warfare that had many names, including civil war, intrastate war, low-intensity conflict, guerrilla war and wars of national liberation. It took some time for the academic community to realize that civil war was the one type of armed conflict that persisted in the post-Second World War period. This understanding was brought on by the spectacular eruption of ethnic conflicts that followed the end of the Cold War in the early 1990s. Indeed, the Cold War had sustained a flourishing industry of think-tanks and projects that allowed academic researchers to cling to the idea of war as interstate war, conveniently overlooking civil wars. Civil wars were always considered a messy form of conflict, deprived of armies with shiny uniforms, mass armour, pitched battles and clear frontlines – the elements that most people intuitively associate with the very essence of ‘war’. The only point of contact between the systematic study of civil wars and the policy world was the area of counter-insurgency studies.

Yet civil war is difficult to ignore, both for the damage it causes in the developing world and for its potential to produce negative effects for the developed world, primarily in the form of transnational terrorists who use failed states as their base, and population displacement that leads to mass migrations. The number of deaths in the civil wars that have taken place between 1945 and 1999 has been put at 143,883. In addition to direct fatalities, civil war causes many more indirect ones through mass displacement, epidemics, famines and the degradation of public infrastructure. Economic development is stalled or even reversed. This ‘conflict trap’ is now considered to be one of the main obstacles to economic development.

The realization that civil wars were real, relevant and worthy of systematic study led to a research boom that began in the mid-1990s. Development economists specializing in the study of African economies and initially funded by the World Bank sought to make the case that civil war was a major impediment of economic development. At the same time, the almost complete disappearance of interstate wars, combined with the end of the Cold War, led scholars of international relations and international security to shift their focus from interstate war and nuclear deterrence to civil war. And the resurgence of ethnic conflict during the early post-Cold War years led students of ethnic studies, including sociologists and political scientists, to turn to the study of civil wars.

There are three ‘sources’ of interest in civil war that correspond to three distinct styles of research and resulting ‘findings’. While recognizing the complexity of the issue and acknowledging that the causes of civil wars are multiple, economists have stressed the impact of natural resources; international relations scholars have pointed to ethnic antagonism; and comparativists and sociologists have focused on the role of the state.

Economics

Associated with the work of Paul Collier of the University of Oxford, development economics research has relied on econometric methods to explore the factors that differentiate countries that experience civil wars from those that do not. The main findings have centred on the importance of gross domestic product (GDP) per capita and the presence of natural resources. Countries that experience civil war tend to be poor and have an economy that is centred on the export of commodities. In contrast, characteristics such as regime type (whether a country is a democracy or an autocracy), economic inequality and ethnic fragmentation do not appear to be factors. Collier and his collaborators interpreted these results as indicating that civil wars were fundamentally wars of ‘greed’ rather than ‘grievance’. Rebels were in it to loot natural resources rather than to bring about social change or justice. Civil wars are a criminal enterprise rather than a political or social phenomenon. Sierra Leone’s conflict is one example of such a war.

International relations

For scholars of international relations, civil wars were seen as interstate wars writ small. Like states fighting against each other, ethnic groups in multi-ethnic states faced off seeking control at the centre or secession. Initially, a popular theory was that of ethnic animosities associated with the work of Donald Horowitz of Duke University, which was based on the social psychology of group identity. Political scientist Samuel Huntington extrapolated this theory on a global scale with his ‘clash of civilizations’ thesis. Barry Posen of the Massachusetts Institute of Technology (MIT) supplied the crucial mechanism of ‘the security dilemma’, whereby preparations for defensive action by an actor are interpreted by its rival as an intention of aggression, thus leading to pre-emption in a self-fulfilling kind of way. An example is the conflict in Bosnia.

Comparative politics and sociology

Another set of findings from scholars of both comparative politics and sociology emphasizes the role of state capacity in civil wars. According to an influential cross-national study, the best predictors of civil war were, again, GDP per capita and the proportion of a country that is composed of ‘rough terrain’. This finding implied that states that have difficulty controlling their periphery are vulnerable to rebels. An example is El Salvador. In contrast, sociologists who take a macro-historical view have pointed out that ethnic wars are closely related to processes of dissolution of empires, such as the Ottoman and Habsburg empires in the past and the Soviet empire more recently.

Overall, this research boom has been very valuable. The most significant and robust statistical finding across almost every study is the positive correlation between GDP per capita and the incidence of civil war. All other correlations tend to be weak. It is possible, therefore, to say with a great degree of certainty that poor countries are more at risk of civil wars compared with wealthy countries. However, we still cannot tell which particular causal pathway links poverty to civil war: lack of opportunities, greedy rebels, weakened states? Furthermore, we can only say that poor countries face a higher risk of civil war. We cannot really tell why only some poor countries experience civil war when most do not – or which ones among the poor countries are most at risk.

Ethnicity

The fact that ethnic fragmentation does not appear to be statistically significant does not mean that ethnicity is irrelevant. Ethnic fragmentation is notoriously difficult to map, as it may combine linguistic and religious differences. Some researchers have suggested that one should look only for politicized ethnic differences, while others have suggested measures of polarization rather than fragmentation. In a recent paper, Andreas Wimmer, Lars-Erik Cederman and Brian Min of the University of California, Los Angeles, argue that it is not ethnic diversity per se that is associated with violent conflict, but rather three ethno-political configurations of power that lead to different types of violent conflict. States that exclude large portions of their population on the basis of ethnic background are likely to be challenged by armed rebellions. Segmented states in which power is shared across a large number of competing ethnic elites risk violent infighting, mainly in the form of coups. And cohesive states with a short history of direct rule are more likely to experience secessionist conflicts.

One of the most negative findings is that regime type has little impact on the risk of civil war. Poor democracies and poor autocracies are equally vulnerable. A less robust finding is that so-called anocracies – regimes that are neither democracies nor autocracies – tend to be at a higher risk of civil war, either because they lack both the legitimating capacity of democracy and the repressive capacity of autocracy, or because transitions to democracy increase the risk of conflict. In fact, there is some historical evidence suggesting that emerging democracies with weak political institutions are also at risk of civil war, because their leaders are likely to manipulate nationalist feelings and invoke external threats to scare their populations into supporting wars. The main implication here is that building strong institutions should precede running elections.

Despite these valuable contributions, this body of research also has problems. Civil wars are particularly challenging phenomena to study: their actors are often obscure, the countries in which they erupt are poor and they often lack major defining events such as decisive battles. As a result, data tend to be of questionable quality.

Even if the data were more reliable, the same statistical findings are typically interpreted in different ways. For Collier and his collaborators, GDP per capita relates to the opportunity costs of fighting (poor people have nothing to lose), whereas James Fearon and David Laitin of Stanford University interpret it as state capacity (poor countries are weak countries). Likewise, the gap between proxy variables and the ideas and concepts they are supposed to represent (known as ‘internal validity’) tends to be large. For example, the index of ethno-linguistic fractionalization (ELF) is a very problematic indicator of ethnic divisions. Moreover, causal mechanisms tend to be absent from the analysis. For example, it may be that the mechanism underlying the link between natural resources and conflict captures the looting intentions of ‘greedy rebels’ as posited by Collier and his associates; the looting intentions of ‘greedy outsiders’, such as neighbouring states who have an interest in fomenting civil conflict; popular grievances caused by inequality and corruption that characterize natural resource-based economies; the possibility to finance an armed rebellion (rather than any motivations per se); the weakening of the state or of social networks. Obviously, policy intervention requires the identification of the proper causal mechanism.

Moreover, this research suffers from its twin reliance on the assumption of ‘unit homogeneity’ (the notion that participants, or experimental units, are identical) and static analysis. Civil wars are treated as if they were all manifestations of the same fundamental and constant underlying phenomenon. Although this is an acceptable assumption when the goal is to observe broad patterns, it becomes more problematic when the objective is to unearth causal mechanisms or interpret recent trends. In fact, an unexpected development has recently taken place: after an initial spike, following the end of the Cold War, the rate of civil wars began to decline significantly. Relaxing the assumption of unit homogeneity would allow a better grasp of the evolution of civil wars.

The end of the Cold War

The end of the Cold War was welcomed by optimists as a blessing that would bring global openness, democracy and peace. At the same time, pessimists warned of a coming global anarchy, an era of ethnic conflict and instability. As we near the 20th anniversary of the end of the Cold War, it is worth taking stock of the consequences of this momentous event for violent conflict.

Civil war spiked immediately after the end of the Cold War. Observers and analysts alike were initially swayed by this spike. Many thought that the end of the Cold War spelled a ‘coming anarchy’, through the eruption of ‘new wars’ which ‘shattered the dreams of the post-Cold War’. Following this wave of doomsday predictions, and after the rate of civil war onsets had returned to its Cold War average, many researchers concluded, as Fearon and Laitin did, that ‘the prevalence of civil war in the 1990s was not due to the end of the Cold War and associated changes in the international system’. More recently, however, the observation of a declining trend in civil wars has produced renewed sensitivity about the end of the Cold War – and rightly so. For example, researchers associated with the Human Security Centre at the University of British Columbia have characterized this decline as an ‘extraordinary and counterintuitive improvement in global security’. They note that by 2003 there were 40% fewer conflicts than in 1992, and that the deadliest conflicts (those with 1000 or more battle deaths) had fallen by some 80%. They also added that the end of the Cold War was the single most critical factor in this decline, and identified international intervention as the key mechanism. Because the two superpowers ended their interest in ‘proxy wars’ in the developing world, the United Nations and other international agencies, donor governments and NGOs were free to play a new global security role that entailed active diplomacy, peacekeeping and peacemaking, thus preventing new conflicts from taking place and brokering peace agreements to end those that had already erupted. Obviously, the divergence between these interpretations is largely a result of observations taken at different points in time. The end of the Cold War era seemed a disaster in 1992, appeared unimportant in 2001, but struck observers as having led to a clear improvement by 2005. Evidently, descriptive inferences based on highly sensitive short-term trends are treacherous.

Three types of civil war

Drawing from a new wave of research that systematically examines the dynamics of recruitment and violence within civil wars, I have suggested a distinction between three types of civil war. Widely considered to be the most common type is irregular (or ‘guerrilla’) war, in which the strategically weaker side refuses to match the stronger side’s expectations in terms of the conventionally accepted basic rules of warfare. These wars of ambush and surprise are typically wars of attrition, with the rebels seeking to win by not losing, while imposing unbearable costs on their opponent. A common manifestation of these wars is the absence of clear frontlines. Examples are the conflicts in Chechnya or Kashmir.

Conventional civil wars, on the other hand, are thought to be much less common. This type of warfare requires a commonly shared perception of a balance of power between the two sides, in the sense that they are both willing to face each other conventionally, across clearly defined frontlines. The distinction between conventional and irregular wars is intuitive: whereas the former is a symmetric conflict, the latter is an asymmetric one. Examples of conventional wars include the Spanish civil war, the Biafran civil war in Nigeria, the Bosnian war and the war in Azerbaijan.

Although asymmetry is predominantly expressed in irregular war, the converse is not necessarily the case. Symmetry (or parity) is not synonymous with conventional war. Rather, it is possible to point to a third type of civil war, ‘symmetric non-conventional war’. This type of war is often described as ‘primitive’ or ‘criminal’ war, and entails irregular armies on both sides in a pattern resembling pre-modern war. Whereas in conventional civil wars the rebels are able to generate a military capacity that rises up to the state’s capacity, in symmetric non-conventional wars, the state’s capacity has fallen so much as to resemble that of the rebels; the state, in other words, is just another militia. Think of the conflicts in Lebanon, Somalia or Sierra Leone. A key difference between these types of conflict and large-scale criminal violence is that in the former the state has ceased to be a military superior actor and often has completely collapsed.

Recent research finds that, although irregular war is the dominant type of warfare in the post-Second World War period, it is much less common than is often thought. These wars constitute slightly more than half of all civil wars. Conventional civil wars are much more common than thought (about 35% of all civil wars), and symmetric non-conventional civil wars account for the balance. What is very interesting, however, is that the dominance of irregular war is completely associated with the Cold War. Since the end of the Cold War, conventional and symmetric non-conventional wars have emerged as the dominant types of war, while irregular war has declined significantly. Indeed, the irregular wars being fought in Iraq and Afghanistan appear to be anomalies.

This analysis suggests that irregular war is not a method available at all times to all actors, but a historically contingent phenomenon. During the Cold War, three factors led to the preponderance of irregular wars: the financial and military support of the Soviet Union and its allies (and to a smaller degree of the US for ‘anti-Marxist’ insurgencies); the widespread belief that social change could be achieved through irregular war bolstered by the creation and operation of transnational networks of activists that gained experience by participating in several wars before launching their own; and the development and popularization of the doctrine of revolutionary people’s war.

These irregular wars of the Cold War had three characteristics: they required high levels of skill (and hence ‘high-quality’ rebels), they tended to be long-lasting and they were often won by the rebels.

The end of the Cold War led to a decisive degradation of rebel military capacity. The civil wars that continued in the post-Cold War period were either conventional, associated with the implosion of multi-ethnic empires and states, or symmetric non-conventional, associated with state failure. These types of war are neither new, nor are they more deadly compared with irregular wars; they became prevalent, however, by default.

Simply put, the end of the Cold War proved to be a tremendous blow to potential rebels. By depriving them of plentiful superpower support, of the belief that social change could come out of the barrel of a gun and the transnational networks that sustained them, and of the means to implement the doctrine of revolutionary war, the end of the Cold War selected, so to speak, these rebels out of history. Now, relatively weak states that had been previously vulnerable to irregular war could easily deter rebels or defeat them before they could pose a serious challenge. In contrast, the weakest states, deprived of any superpower support, became much more exposed to opportunistic rebels. In a way, Collier and his collaborators were not necessarily wrong when they characterized rebels as greedy looters rather than justice seekers; they just had in mind a subset of civil wars that happened to become dominant in the post-Cold War era with a concentration in sub-Saharan Africa.

What are the implications of this research for policy makers? Obviously, there is little we can do if the causes of civil wars are poverty or low state capacity. Both are slow, long-term processes that cannot easily be affected by outside intervention. In contrast, if the decline of civil wars proves lasting, and if it is indeed due to the secular decline of irregular wars, the news have a positive spin. Unlike irregular wars, rebels in both conventional and symmetric non-conventional wars are easier to defeat, or at least contain, by a well organized international force. These conflicts are no guerrilla quagmires and the rebels are no organized forces capable of putting up long-term resistance. The worst outcome is the persistence of lingering, low-intensity conflicts at the peripheries of states experiencing conflict, as opposed to higher-intensity wars of attrition that threaten power at the centre. This is not to minimize this type of instability, but it is one that is easier to manage than the irregular civil wars of the Cold War. State-building skills and long-term investment in institution building do require the cover of security operations, but the security challenge is more manageable than in the past.

References

Footnotes