Drilling Deeper: A Systematic, Context-Sensitive Investigation of Causal Mechanisms in the Oil–Conflict Link

Abstract

This article employs an innovative methodology to study causal mechanisms in the oil–conflict link by combining a systematic approach with taking into account country details. Engaging in a deductive test of causal mechanisms in a controlled comparison of four major oil exporters, results show that no oil-related causal mechanism can fully explain the differences in violence. A more inductive analysis complemented by process tracing suggests a more comprehensive mechanism: oil contributes to the formation of opposition through grievances. Levels of violence, however, vary largely according to non-oil conditions – particularly the cohesiveness of opposition and the reaction of government to challengers.

Acknowledgements

The authors would like to thank the German Research Foundation (DFG) for generously funding our research.

Notes

1. Unless otherwise indicated, ‘conflict’ refers to ‘internal violent conflict’.

2. For an overview of the literature, see also Le Billon (2012) and Ross (2012).

3. In a recent publication, Bazzi and Blattman (2011)– in contrast to earlier studies – find very weak evidence for the widespread hypothesis that income shocks provoke conflict.

4. The study finds, moreover, that oil (and gemstone) reserves double the duration of conflict if the resources are located inside the specific zone of conflict.

5. Ross (2004a) develops a number of indicators per mechanisms, but does not quantify or list them (and their evidence); Humphreys (2005) uses single indicators per mechanism only.

6. Our questionnaire contains 155 variables systematically subdivided into four clusters: ‘Resource-specific’; ‘Non-resource-specific’; ‘Domestic’; and ‘International’ conditions. This provides the opportunity to test all variables either as clusters or individually, and also without explicit reference to a causal mechanism – thus allowing for inductive control. It also means that we took the formula ‘few cases, many variables’ seriously (Przeworski and Teune, 1970).

7. The questionnaires were completed by country experts: two researchers within our research team have studied the countries in detail and have carried out extensive field work in all four cases. By a dual review process within the project group we checked the validity and reliability of their assessments.

8. For instance, the question ‘Do nationwide parallels of cultural and social boundaries also exist in the resource region?’ was answered with ‘yes’ for the case of Nigeria and was complemented by the following explanation: ‘To some degree ethnic boundaries and class boundaries run parallel in the Niger Delta. While under military rule there was a high level of political and economic discrimination against some ethnic groups, especially Ogoni and Ijaw, there have been improvements since 1999. Nevertheless, according to Minority at Risk the economic exclusion of Ijaw people persist.’ If there was an ordinal scale with three options, the risk thresholds commonly corresponded to the worst assessment on our scales.

9. A number of variables were measured at only one point in time. These included those capturing events before 1989 and some for which changes are unlikely.

10. See Table IV in the Online Appendix, our selection is based on the descriptive information of the conflicts provided by UCDP.

11. The questionnaire is based on the matrix in Table II in the Online Appendix, the detailed version of the questionnaire, including values, is available on request.

12. FIS did split up in the 1990s and these divisions weakened the movement (Heristchi, 2004, p. 117). However, this does not affect its power prior to the civil war onset.

13. Divisions within the political opposition have diminished since 2010, rendering the oppositional forces more powerful.

14. The historical dynamics underlying the violent conflicts in Venezuela are only partly captured by UCDP data. For more details, see Mähler (2011).

15. Especially harsh government responses may explain why, during the ‘Arab Spring’, countries like Libya and Syria lapsed into full scale civil wars, while others (Egypt, Tunisia) have not.