It is often said that a country that has experienced civil war has a 50 per cent chance of sliding back within five years. Indeed, the UN has used this figure in preparation for its Peacebuilding Commission. However, the authors of this figure have since revised their original estimate to 23 per cent. This paper examines the process whereby academic findings become accepted wisdom and warns against the wholesale adoption of such figures by policymakers.
In the UN system, forces favouring a stronger international peacebuilding regime seized upon the work of Paul Collier and Anke Hoeffler. In 2002, they published an article which investigated ‘peace duration’, defined as ‘the number of months since the end of the previous conflict, or since 1945’. It concluded that ‘shortly after a conflict, on average, countries face a 50 per cent risk of renewed conflict during the next five years’. However, changes in definition and methodology can produce radically different estimates. In such cases, we need to understand why this is so, what the changes imply and which estimates, and in what form, are presented to policymakers.
If the figure expresses the average risk of recurrent war with the civil-war group within a given period then a significantly lower figure is produced. Taking this definition and using the same data set as Collier and his team covering the period 1960-99, the results are as follows:
- of the 49 countries listed, civil war did not recur at all in about half the cases within the entire 40-year period
- of the remaining cases, a second civil war did occur in 12 countries but after the first five-year period of peace. For example, after 1962, Algeria did not experience another war until 1991
- in only 13 cases did a second civil war recur within the first five-year period giving a recidivism rate of around 26 percent.
Similar work by Barbara Walter using the same data set but for a longer period (1945-96) also estimated a lower recurrence rate. She concluded that single wars rather than recurrent conflict was the norm. Collier and his associates have continued to work on civil war recurrence. Based on the changes in methodology and data in a 2006 study, they estimated a recurrence rate of 20.6 per cent for the first four years after the previous war had ended. In the revised version of the same paper two months later, this figure came out at 23 per cent. Yet, these new figures have received little comment even though changes in conclusions of this kind have important ethical and policy implications.
- Policymakers may not wish to know about the finer points of methodology, but they do need proper caveats about the soundness and durability of a given formula.
- Findings should be interpreted with great caution in areas where little work has been done and where a particular set of authors has dominated the agenda.
- Researchers have a responsibility to present their methods and data for evaluation and scrutiny by others. They also need to explain any divergence between previous figures.
- The lower figures produced by the revised data would point towards a less intrusive model of international peacebuilding.
