By Paul Collier
Posted November 2009
Why has Africa had so much civil war? In all other regions of the world the incidence of civil war has been on a broadly declining trend over the past thirty years: but in Africa the long term trend has been upwards. Of course, every civil war has its ‘story’ – the personalities, the social cleavages, the triggering events, the inflammatory discourse, the atrocities. But is there anything more? Are there structural conditions – social, political or economic – which make a country prone to civil war? Might it be that the same inflammatory politician, playing on the same social cleavages, and with the same triggering events, might ‘cause’ war under one set of conditions and merely be an ugly irritant in another?
Although I am an Africanist, I like to set Africa in comparative perspective. If Africa is different – as it clearly seems to be in respect of civil war – there are two possible types of explanation. Africans could simply behave differently from others when faced with the same situation. Alternatively, for a given situation their behaviour could be much like anyone else’s, but the situations they face could be systematically different. To sort this out we need to look globally, not just regionally. Together with Anke Hoeffler, I have analyzed global data covering the last forty years, trying to see why most countries at most times have avoided civil war, while others have not. Our approach has been statistical – trying to see whether any characteristics of a society could account for a subsequent eruption into war. Within the limits of data availability, we have tried to include social factors (such as inequality, and the ethnic and religious composition of a society), history (such as the time since decolonization), and politics (such as the extent of democratic political rights) as well as economic characteristics. We find a pattern and we find that Africa fits that pattern.
Economic Roots of Civil War
Surprisingly, the dominant factors are economic. Three factors matter a lot for the risk of civil war: the level of income, its rate of growth, and its structure. If a country is poor, in economic decline, and is dependent upon natural resource exports, then it faces a substantial risk that sooner or later it will experience a civil war. Typically, such a country runs a risk of around one-in-seven every five years. Like Russian roulette, things might go well for a while, but then some conjunction of circumstances – the personalities, the triggering events – ignite violent conflict. Of course, when this happens, the media focus on the personalities and the triggering events. These are indeed the proximate ‘cause’ of the conflict. But the big brute fact is that civil war is heavily concentrated in countries with low income, in economic decline, and dependent upon natural resources.
Africa is the one region where such economic characteristics are the norm and this fully explains Africa’s distinctive incidence of civil war. Yes, Africa is riven by ethnic differences, so that where civil wars flare up they will invariably be fought along ethnic lines. But this does not mean that the ethnic differences are ‘causing’ the conflict. Globally, ethnically diverse societies are no more at risk of civil war than other societies. The only exception to this pattern is where the largest ethnic group is in a majority – that does indeed increase risks and we can think of examples in Africa. But Africa is so ethnically diverse that in most societies no group is in a majority. Fewer African societies have ethnic majorities than other regions.
In respect of the three economic characteristics, Africa is most distinctive in terms of natural resource dependence. Africa as a region still has a much higher ratio of land to population than other regions, and natural resources are basically randomly distributed under the land. Further, for various reasons, Africa has not managed to break into the global market for manufactures. Whereas twenty years ago the typical developing country was dependent upon natural resource exports, now 80% of developing countries’ exports are manufactures. Africa has thus got stuck in resource dependence while other developing regions have moved on. Why does this matter?
Why Natural Resources Fuel Conflict
Natural resources generate what economists term ‘rents’ – meaning profits that are much higher than the minimum level needed to keep the activity going. The trouble from natural resources stems from these rents. There are six routes by which natural resource rents increase the risk of violent conflict; four relate to political economy and two are straight economics. Let’s start with the political economy. The most obvious route is that natural resource rents are a ‘honey pot’. Politics comes to be about the contest for control of these revenues. This produces a politics of corruption – aided and abetted by foreign corporate behavior – and sometimes directly a politics of violence. The stakes are highest in low-income countries because the control of the state implies massive revenues relative to other income-earning opportunities. Further,this politics of rent-seeking diverts the public arena from its normal function of achieving the collective action that is necessary to supply public goods – the social and economic infrastructure that all societies need. The society thus loses out twice over: in the struggle for resource rents other resources are dissipated, and the supply of public goods declines. Nigeria provides a striking example of such a politics of contest for oil rents.
The second route by which natural resource rents increase the risk of war is through the detachment of government. Because resource-rich governments do not need significant other tax revenues they become detached from their electorates. In most societies, because electors have to pay high taxes, they scrutinize the government to see how it uses their money. This was indeed how democracy developed in the West. The campaigning slogan ‘no taxation without representation’ can be inverted to the depressing reality of ‘no representation without taxation’. In many resource-rich societies the resource rents are not seen as belonging to ordinary people in the same way as income taken from them in taxes – hence the detachment. The government is able to ignore the concerns of the population. Mobuto’s Zaire was a classic example of such detachment.
Even when neither of these first two effects directly leads to a civil war, they produce between them a dangerous cocktail of a dysfunctional politics of self-interest and a governing elite that is detached from the concerns of ordinary people. These two characteristics then combine with a third dangerous factor: natural resources are usually found in only part of the country, often in a peripheral area. The people who live in this area are ready prey for secessionist political movements. To the usual romantic propaganda of identity politics, secessionist leaders can add the powerful language of economic self-interest: ‘our’ resources are being squandered by a corrupt and alien elite. Large natural resource rents not only make civil war more likely, they make it more likely that a civil war will be secessionist. Biafra, Katanga, Cabinda: Africa’s secessionist wars have usually been related to natural resources.
The final political economy route by which natural resources increase the risk of civil war is that they provide an obvious source of finance for rebel groups. Even if the rebellion is not motivated by these rents it is greatly facilitated by them: from the proceeds leaders can purchase arms and pay recruits. Warfare is a costly business: whereas thirty years ago rebel groups largely had to depend upon a friendly government for finance and armaments, now rebellion has been privatized’ – markets in natural resources and armaments have developed to the extent that rebel groups can be self-sufficient. Rebel groups gain access to natural resource rents in several ways. One is to run protection rackets against the companies or people who are the exporters. Another is directly to operate extractive businesses. Yet another is to sell concessions to mineral rights in anticipation of subsequent control of the territory. The prolonged viability of UNITA in Angola and the RUF in Sierra Leone; the violent gangs of the Nigerian Delta; and the successful rebellions of Laurent Kabila in Zaire and of Denis Sassou-Nguesso in Congo Brazaville, were all assisted by one or the other of these methods of natural resource financing.
In addition to these political economy effects there are two economic effects that increase the risk of civil war. Resource rents appreciate the real exchange rate causing ‘Dutch disease’, whereby the rest of the export economy contracts. Usually in Africa the non-resource export economy is based on agriculture, so that small farmers in some areas face sharply declining income despite the influx of wealth into the economy. Finally, the prices of natural resources are usually highly volatile, so that the economy becomes subject to booms and busts. This pattern typically depresses the long term growth rate, and it also implies periods of severe contraction. Recall that economic decline is itself a risk factor for civil war.
Does It Have to be Like This?
Through these six routes, to date natural resources have largely been bad news for Africa. But it doesn’t have to be like that. Revenues from natural resources are an enormous opportunity for low-income African economies. Especially for landlocked countries with hostile climatic conditions, such as Chad, natural resources probably offer the only option for significant poverty reduction. This is the dilemma: resource rents have the potential for good as well as for bad. The strategy of saying ‘just leave the resources in the ground’ sacrifices the potential for good as the price of avoiding the bad. Historically, such a strategy would usually have been an improvement on what actually happened. But, as well as being a counsel of despair, it is unrealistic. As a result of the geo-politics of oil, there is a rush of new discoveries in small, poor African countries – Equatorial Guinea, Mauritania, Sao Tome and Principe, Gambia. These resources are not going to be left in the ground. The challenge for both Africa and the international community is to change the political and economic governance of such resources so that the future is not a repetition of the past.
Two contrasting examples help to bring the issues into focus. Thirty years ago Botswana and Sierra Leone had the same level of per capita income. Then they both received enormous diamond income. The government of Botswana succeeded brilliantly in harnessing these revenues for economic growth: for many years Botswana was not just the fastest growing economy in Africa, it was the fastest growing economy in the world. As a landlocked desert, it is easy to imagine Botswana’s fate in the absence of diamonds. Sierra Leone had a dramatically different experience. The diamond revenues fomented violent political contests which destroyed the society. The economy collapsed, and now the country is at the bottom of the Human Development Index. The differential between the two countries in terms of per capita income is now an astonishing ten-to-one. The economic and political governance of natural resource revenues was evidently absolutely vital in producing this massive divergence in outcome. In short, although policies and governance always matter, they matter much more where there are large natural resource rents. Africa needs more Botswanas and fewer Sierra Leones: which of them will Equatorial Guinea, Sao Tome and Principe, Mauritania, and Gambia resemble two decades hence?
How African Societies Can Help Themselves?
If I was a citizen of an African natural resource economy I would want to know how to become Botswana and to avoid the fate of Sierra Leone. I think that the magic ingredient that makes the difference is scrutiny of government by the country’s citizens. Unfortunately, scrutiny is a ‘public good’ – that is, if it is provided, the whole society benefits. The incentives for individual action are thus all wrong – basically, the smart thing to do is to sit back and hope that someone else goes to the trouble of providing public goods such as scrutiny. Societies need ‘collective action’ to overcome the public goods problem and because Africa’s societies are so highly diverse –more ethnically diverse than anywhere else in the world – they find it unusually difficult to supply public goods at the national level.
Of course, people and groups lobby the government, but overwhelmingly this lobbying is not for the national interest but for individual or group advantage. But there are ways around this problem. In an ethnically diverse society it is probably much easier to organize scrutiny at the local or regional level than at the national level – at the local level ethnicity is likely to unite people in collective action, just as at the national level it is likely to divide them and frustrate collective action. If the rents from natural resources could be transparently and fairly distributed to sub-national levels of government there is some hope that such governments would come to face serious citizen scrutiny. The challenge is to get to this stage where rents accruing at the national level are seen to be fairly distributed to the regions.
The Right Agenda for Outsiders
This is where the rest of us come in – those of us who are not African citizens and so have little basis to tell African governments what they should and shouldn’t be doing. What we can legitimately do is to make it easier for African citizens to get to the stage at which they can overcome their collective action problem and scrutinize how resource rents are used at the local level. Specifically, we can help to make natural resource rents transparent at the national level. This has been the agenda of NGOs such as Global Witness – now picked up by the British government’s Extractive Industries Transparency Initiative – and I think that it is the right agenda. At least, it is the right agenda for us. Transparency in reporting revenues is itself only an input into scrutiny – it makes domestic scrutiny easier. It doesn’t make it happen automatically, but without transparency in revenues there can be no scrutiny of how they are used.
Another key area for international action is that banks should be required to cooperate in tracking down misappropriated natural resource rents. For example, the Nigerian government has recently abandoned the attempt to repatriate the vast Abacha wealth from London banks because the process was proving to be an unending legal nightmare. What is the incentive for African societies to scrutinize their leaders if corrupt wealth is so well-defended by Western legal systems?
A further area for international action is the acquisition of natural resource contracts. Too often Western corporations have connived with African political leaders to reach deals that were mutually profitable at the expense of the country. Transparent competitive tendering must become the norm. When NorthSea oil concessions were awarded we would not have tolerated an oil company concluding a secret private deal with a minister; we should not tolerate such a practice in Africa.
This, to my mind, is the agenda for corporate social responsibility in Africa: transparency in bidding for resource concessions; transparency in revenue payments to governments; and cooperation by banks in tracking misappropriation of rents. Sadly, it is far from the currently dominant agenda. International resource extraction companies live in terror of two powerful forces – Western consumers who may boycott their products; and the local people living around their installations, who may kidnap employees and damage equipment. They have responded to Western consumer pressure – itself based on a lazy, teenage misdiagnosis of Africa’s ills – by trying to look like good employers and good environmentalists. They have responded to local extortion rackets by providing health and education facilities in the neighborhood of their installations. Frankly, both of these are at best irrelevant. High wages mess up the labour market and so cost jobs; it is governments, not companies, that should be supplying basic social services. What has got lost is the legitimate, indeed essential role that companies can play in helping African societies to scrutinize their governments. Corporate social responsibility in Africa must be radically redefined.
* This article originally appeared on the Crimes of War Project’s website on
October 2004 (www.crimesofwar.org)