Patterns in the dynamics of insurgencies

Development Policy27 Feb 2011Frauke de Weijer

The second day of the working group zoomed in on the identification of patterns in the dynamics of insurgencies. Wikileaks data on violent interactions between insurgents and counter-insurgency forces (CIGACTs) had been scrutinized for patterns in scale and location of events. Such type of analysis could prove interesting, but suffers from a number of weaknesses which limit its usefulness in practice.

Firstly, the way data is being collected limits its usefulness. The number of CIGACts is highly dependent on the presence of foreign troops and the story they tell is more illuminating about their movements rather than the insurgents’ behaviors. Secondly, patterns require a narrative to become explanatory and natural scientists are not well placed to provide these. This stresses the importance of combining scientific analysis with local knowledge. Slicing the data in different ways, based on known events or potential explanatory stories, can then verify or falsify existing assumptions of policy makers and provide fascinating new insights.

Another scientist from the Santa Fe institute went beyond the Afghan insurgency and had analyzed a large dataset covering a range of insurgencies with the purview of detecting possible generic patterns of insurgencies. The results were fascinating and, to me at least, highly unexpected. There does seem to be a generic pattern to insurgencies! Across the board, attack rates accelerate with increased experience of the organization and are proportional to organization size. This implies that the binding constrain to insurgency is labour, and policies should be aimed at restricting recruitment and encouraging desertion. This is surely an interesting finding, but translation to policy is still not straightforward.

At first glance the conclusion may be that the current policy of incentivizing Taliban fighters to desert and reintegrate is sensible. This was however rapidly challenged by the development actors around the table. There still is little evidence that individual monetary incentives and jobs reduce recruitment rates. A multitude of factors contribute to the decision-making process, of which monetary rewards are just one and arguably not a major one. An analytical approach based on game theory that can capture these factors would be extremely useful, but in reality still seems quite far off.

The working group ended with a much deeper understanding of the scientists and the foreign policy and development community of the importance, constraints and potential of each other’s work. We were excited about the possibilities to collaborate and devise new ways of complementing each other’s work, and realized that we were breaking completely new ground. The sub-title of the workshop was “frontiers of data analysis”, which it surely was. But even more strongly it was the frontiers of collaboration and communication between highly distant professional fields.

Our brains were all buzzing with the potentiality of these worlds coinciding, but at the same time cognizant of the bridges that need to be built. Differences in culture and language need to be bridged, and data collection methods adapted to the needs of analysts. The most important obstacle however, we all felt, is the institutional incentives in both worlds. But the energy and excitement that we all felt for this new frontier of evidence-based policy making was tremendous. A number of initiatives for collaboration are already being spawned, and hopefully many others will follow when we all go back to our respective fields of work. This is surely one of the most exciting events I have attended in a very long time!