Counterinsurgency study is much more than a checklist

On 18 November Joshua Rovner and Tim Hoyt posted a critical but incomplete review of Victory Has a Thousand Fathers: Sources of Success in Counterinsurgency, a RAND study for which I am the lead author. I feel compelled to reply.

Our report isn’t perfect; little in social science is. I think we are fairly candid about the limitations of our study, candor which Mr. Rovner and Mr. Hoyt exploit in their review while at the same time overlooking the strengths of the work. Victory Has a Thousand Fathers reports findings from detailed case studies of 30 systematically selected insurgencies.  It is distinctive in the extent of its qualitative and quantitative detail over a significant number of cases, and it is far more rigorous than this critique suggests. 

The 30 cases cover all of the insurgencies started and completed between 1978 and 2008 that also meet a relatively modest set of criteria regarding casualty thresholds and the nature of the conflict in question.  The five specific insurgencies mentioned by Rovner and Hoyt weren’t "left out" due to "puzzling" choices on our part – they just didn’t fit the clearly specified historical bounds or satisfy the case selection criteria.

We didn’t have the resources to study all of the insurgencies of the 20^th and 21^st centuries, and we didn’t cherry pick the ones we did study. We did what rigorous historical comparativists do: we defined and bounded a set of cases, and we studied all of them. The utility to policymakers of findings based on all insurgencies begun and completed between 1978 and 2008 should be self-evident.

While there is some risk of tautology in the relationships between some of the factors we evaluated and case outcomes, the examples chosen by Rovner and Hoyt are poor ones. For example, "The majority of the population in the area of conflict supported or favored the [counterinsurgent] force" is not perfectly associated with the outcome of the cases; in two of our 30 cases (Turkey, 1984-1999 and Croatia, 1992-1995) the counterinsurgent (COIN) force prevailed without having the support of the majority of the population.  Our study further avoids tautology by dividing each of the cases into between three and five phases, and assessing all of our 77 factors for each phase of each case, to include progressive change in factors from one phase to another.  In several cases the COIN force lacked the support of the majority of the population in early phases of the conflict, but earned it in later phases, and (as described in the detailed case narratives for those cases) it clearly contributes to the government’s subsequent success.  These findings are far from "meaningless," as these critics contend, but have important implications for population-centric counterinsurgency, especially when considered alongside our findings about the importance of what we refer to as tangible support. 

Briefly, we distinguish "tangible" support from "expressed" support and construct it to include the provision of manpower, funding, materiel, sanctuary, and intelligence to insurgents.  Most insurgencies draw (or ultimately fail to draw) their tangible support from a supporting population, but some draw significant support from outside the area of conflict, perhaps from external state supporters or Diaspora networks.  While tangible support is strongly correlated with popular support in the 30 cases we studied, when the two differ (that is, when insurgents lack popular support but meet tangible support needs from sources outside the area of conflict, or when the majority of the population in the area of conflict expresses support for the insurgents but government forces are still able to prevent insurgents from meeting their tangible support needs) victory always follows tangible support.

Although Rover and Hoyt take issue with it, repression is, based on the facts, a bad COIN practice.  Of the 30 insurgencies we studied, the vast majority in which the COIN force employed repression in any phase (be it early, middle, or late) were losses for the COIN force. We acknowledge that two of 30 counterinsurgent forces (again, in Turkey and Croatia) managed to prevail while engaged in collective punishment and escalating repression, but offset the negative effects with a host of better practices, including effectively reducing the flow of tangible support across borders, establishing and then expanding secure areas, and ensuring the provision of basic services in areas controlled by the government.

And we do not suggest that all applications of force are bad; that would be the kind of simplification that is attractive to critics. Proportionate and discriminate use of force against insurgents and not against the surrounding population plays a prominent role in government success in many of the cases as described in the detailed case narratives (the study’s detailed narratives are wholly ignored by Rovner and Hoyt).

The recommendations of the report follow closely and logically from the quantitative and qualitative results. I would encourage those who wish to consider the study’s conclusions to decide for themselves.   

Dr. Christopher Paul is a social scientist at the RAND corporation.