Small Wars Journal

There's no checklist for counterinsurgency

  1. Blog
Fri, 11/19/2010 - 11:35am
Joshua Rovner and Tim Hoyt; assistant professor and professor, respectively, of strategy and policy at the U.S. Naval War College; are skeptical about a recent RAND study Victory has a Thousand Fathers. They explain why in their Foreign Policy opinion piece There's No Checklist for Counterinsurgency. BLUF: "State-building usually includes a period of ruthless competition for power, and some "bad practices" are usually necessary to end it. Efforts to stop the process in midstream in the name of COIN doctrine may prove tragic if they end up prolonging the conflict without settling the underlying political issues."

Comments

Charles Knight (not verified)

Sat, 11/20/2010 - 5:24pm

Permalink

As I have just commented on the FP site, the Navy Professors review unfairly mischaracterises the method of the RAND report. It is a pity, becuse the reviewers perspective, as summarised by Dave above, is not actually at odds with the findings of the RAND work.

John T. Fishel

Sat, 11/20/2010 - 10:58am

Permalink

Having now read both the RAND study and the Rovner/Hoyt critique, I would offer the following comments.
1. The study breaks no new ground, except perhaps methodologically. It simply confirms the quantitative analysis that Max Manwaring and I first published in the journal Small Wars & Insurgencies (Winter 1992) and more recently in the Small Wars Journal as "The SWORD Model Of Counterinsurgency" as well as such qualitative studies as our Uncomfortable Wars Revisited and the Daniel Marston & Carter Malkesian edited volume of historical case studies, Counterinsurgency in Modern Warfare.
While I found the quantitative analysis in their QCA interesting, I believe that our use of Factor Analysis (not to be confused with their use of the term "factor") coupled with Probit Analysis was more compelling especially when amplified by qualitative analysis.
In the end, however, these authors add simply another set of supporting data and analysis to our general understanding of what works and doesn't work in COIN.
Rovner and Hoyt, on the other hand, attempt to make a case that the RAND group both fails in its elucidation of COIN success (and failure) but that their methodology is actually pernicious. Frankly, attempting to prove that a large scale analysis is wrong by introducing a few examples of where the analysis fails only argues for more research. That research is needed to explain the outlying cases. In our study, we had five outliers among 43 cases which, indeed, required additional explanation. So, too, does the RAND group have outliers in their larger data set (or among cases they did not consider as Rovner and Hoyt argue).
The most interesting case that Rovner and Hoyt introduce to "refute" the RAND study is that of Sri Lanka. Indeed, it does require additional analysis as to why it runs counter to more general population centric successes. However, I would start such analysis by proposing that the Sri Lanka government had relatively high levels of legitimacy among the Sinhalese majority population. And (after all) the Tamil Tigers revolt was that of an ethnic minority located in a relatively smaller area of Tamil concentration. This might be an excellent research project for Rovner and Hoyt to undertake.