Four theories on the causes of civil war are reviewed in the NAI Discussion Paper summarized in this article.
One theory emphasises feasibility factors over factors related to grievance; a second theory is built around the role of horizontal inequalities; a third theory highlights the different roles ‘need, creed and greed’ factors play in various phases of a conflict; and a fourth theory points out as a crucial factor ‘commitment’ problems leading to institutional failures.
The four theories lead to quite different policy conclusions. Their strengths and weaknesses, and their claimed empirical support, are discussed. Empirical shed additional light on some of the mechanisms that underpin these theories.
How are horizontal inequalities between ethnic groups linked to grievances? How do horizontal inequalities compare with vertical inequalities conflict indicators? No claims on directions of causality are made, but results do indicate that horizontal inequalities, in different dimensions, are important factors in grievances and violent conflicts.
Four different strands of the post-Cold War literature on the causes of civil war illustrate the variety of views on the roles of identity and inequality. Each of them has presented a ‘story’ – theories combined with empirical evidence – on the causes of civil war and the role played by identity and inequality. To keep it simple, we label them the Collier story, the Stewart story, the Zartman story and the Commitment story. These four stories also lead to quite different policy conclusions.
In essence, the Paul Collier story is one about civil war occurring when it is financially and militarily feasible, while downplaying the role of objective social grievances. A healthy economy and transparency of trade in extractive resources are some of the policy conclusions following from this story.
The Frances Stewart story is one about horizontal inequalities between identity groups being a main driver of conflict. Its key recommendation is for affirmative actions against horizontal inequality in its various manifestations.
The William Zartman story is one about a stylised sequencing of civil war, with different factors being decisive in different phases and greed factors of military entrepreneurs likely to become dominant over time. It underscores the importance of identifying the right moment for external actors to intervene in civil wars – not too late, when greed-related interests in perpetuating the conflict have built up.
The Commitment story, identified with the World Bank/World Development Report 2011, emphasises a lack of trust and failing institutions as an obstacle to agreement among conflicting parties on non-violent solutions. A prominent policy conclusion concerns the importance of investing in appropriate institutional frameworks and citizens’ trust.
When it comes to inequality, there is an ‘inequality between what’ question that may have different answers. What is often labelled “vertical inequality”relates to differences between individuals in a given population, while ‘horizontal inequality’ refers to differences between the averages within different groups. There is also an ‘inequality in what’ question, for instance in incomes, assets, political influence, social services, cultural status, etc. Finally there is an ‘inequality – how measured’ question, with a number of possible ways of mathematically transforming an array of numbers into a single indicator reflecting how ’unequal these numbers are.
The four stories reviewed could be seen as attempts to formulate grand theories about the causes of civil war, that is, to reveal patterns that repeat themselves over time and space. Case-study oriented scholars might question the approach and argue that studies of individual cases will reveal additional complexities and nuances that such grand stories are unable to capture. What additional contributions do these macro-models provide a person trying to make sense of ongoing conflicts in, say, Afghanistan or Eastern Congo?
I would defend all four of them in their attempts to paint the large picture (as long as it is not interpreted deterministically, applying always and everywhere). For those involved in formulating more general policies, or in prioritising investments in global public goods, the large picture is relevant.
Furthermore, the macro view may also inspire alternative interpretations of case-study findings, and vice versa. Each of the four stories reviewed carries a favoured set of policy recommendations.
Should, for instance, the focus be on reducing the feasibility of financing rebellions (Collier), on addressing horizontal inequalities (Stewart), on reinforcing state capabilities at an early stage of conflict (Zartman) or on restoring citizens’ trust in state institutions (World Bank)? The answers to these questions are by no means irrelevant to actors engaged in conflict-prevention and management.
The empirical explorations presented in the paper, based on Afrobarometer data, do not yield any final judgments on the stories. However, it appears that indicators of horizontal inequalities between ethnic groups have a larger potential than those for vertical inequality in predicting the level of grievance and possibly also the risk of civil war. A case where sharp shifts in indicators of horizontal inequalities took place before a violent conflict is Kenya 2008 (the electoral violence). Horizontal inequality is not captured in the empirical applications of the Collier-Hoeffler model, where only vertical inequality is used as a proxy for grievances. The downplaying of the role of factors related to grievances in the Collier story may thus be premature.
The explorations in the study have been able to identify correlations, but without establishing causal relations. Relative disadvantage and grievance may cause a conflict as well as be caused by it, and people may move in and out of different identities as a by-product of conflict and a group’s disadvantages. It is difficult to imagine how causal directions can be approached without access to more time-series data. Such data is, however, to be expected when future surveys deliver their results. An advantage with surveys such as the Afrobarometer is that it permits connections to be made between, on the one hand, changes in social and economic background variables and, on the other, perceptions and attitudes in relation to identities.