The debate over Afghanistan strategy since Obama's troop increase last
year may not have produced any solutions yet, but it has produced plenty of
think tank reports purporting to have them. One of the most recent is a new RAND
Corporation study
that makes bold claims about victory in counterinsurgency. The authors of the study argue that
debates over COIN are usually "based on common sense, a general sense of
history, or but one or two detailed historical cases." Policymakers and
military officers are desperate for solid research that can help them evaluate
the menu of strategic options, but the best they can expect is advice based on
analogies or selective readings of history. To remedy this situation, the authors set out
to perform a thorough analysis based on "extensive data collection, rigorous
analysis, and empirical testing."
It's a laudable enough goal -- but for all their
claims to superior rigor, the authors fail to live up to it. They make a series
of basic methodological mistakes that throw doubt on their conclusions. Most
importantly, they confuse cause and effect.
The authors identify fifteen
"good" practices and twelve "bad" ones and conclude that success will occur as
long as COIN forces implement more good practices than bad. In other words,
there is a universally applicable checklist for victory. The authors are
unequivocal about the meaning of their analysis of 30 past conflicts: "These
data show that, regardless of distinctiveness in the narrative and without
exception, COIN forces that realize preponderantly more good than bad practices
win, and those that do not lose."
Unfortunately,
several of the good practices they identify are tautologies. In victories, for
example, "The majority of the population in the area of conflict supported or
favored the COIN force," and "The government/state was competent." Such
statements are true; they're also meaningless. A more useful discussion of best
practices would focus on how one creates a situation in which the majority of
the population supports the government, and how the government could
effectively rule.
The authors acknowledge some of these problems. Remarkably,
they concede that they cannot "disentangle the causal order" between practices
and outcomes, meaning that they cannot tell which practices led to victory and
which ones were simply evidence of success. Nonetheless, they defend their
results by claiming that they "do not make strong causal arguments, in part
because so many good COIN practices occur together that we cannot arbitrate
between their many possible causal contributions."
This bit of methodological
legerdemain lets them off the hook for actually demonstrating the influence of
the factors they believe are important, while simultaneously implying that implementing
their laundry list of good practices will in fact lead to victory. In any case,
the caveat is not consistent with their unequivocal conclusions throughout the
report. Frankly, it is hard to take their disclaimer seriously given the
subtitle of the study: "Sources of Success in Counterinsurgency."
The authors also defend their approach by arguing that
it is difficult if not impossible to measure the causal impact of any given
practice because "effective COIN practices tend to run in packs." In other
words, teasing out the influence of particular factors is tricky because
outcomes are overdetermined. This claim is highly dubious, however, because
many of the factors that supposedly correlate with success are in fact
redundant and overlapping.
In addition to these methodological problems, the authors
make puzzling choices about case selection. Again aspiring to rigor, they
justify their decisions on several grounds. They cast the net widely for contemporary
insurgencies by including no fewer than thirty recently concluded cases. As a
result, the sample is "perfectly representative of the recent history of
insurgency," controlling for geographic and cultural variation, as well as
important variation in the military capabilities of COIN and insurgent forces.
Closer
inspection, however, raises significant concerns of bias. The authors chose
their sample based on the conflicts' start date, only including those that
began after 1978. This means that
the study misses some of the most important recent cases which have actually
taken much longer to resolve: Northern Ireland, Basque separatism in Spain, the
Philippines (MILF), Angola (UNITA), and Sri Lanka are all left out. Because
these wars lasted much longer than many of the cases in the sample, they're
arguably more relevant than some of the cases that do make the cut, such as a
one-year conflict in Congo.
A second problem is that
the cases focus on places that are, in terms of both U.S. national security and
international stability, relatively unimportant. There isn't a single case from
Western Europe, the Middle East, or South Asia (excluding Afghanistan). Indian
COIN efforts in Punjab and Kashmir (largely concluded in terms of organized
terrorist violence and militant activity), or Pakistan's operations in Sindh
and Baluchistan merit analysis. What about success in Northern Ireland and
Spain? Can any study really ignore the Middle East and still claim to be
relevant for policymakers today?
The authors claim a
"perfect correlation" between good practices and COIN success, but important
case studies that cut against this conclusion are conspicuously absent. No
example more clearly demonstrates the flaws in case selection than Sri Lanka,
which succeeded through the ruthless suppression of human rights. This omission
is particularly inexcusable because the
government successfully applied repression in the last phase of the conflict,
which contradicts the study's contention
that such tactics only temporarily subdue insurgencies.
The authors note that
in rare cases, countries can use repression in the decisive phase as long as
they also undertake a series of good practices that dampen the negative effects
of coercive violence. But it might also be the case that repression in early
phases helps separate insurgents from civilians, thereby enabling different
approaches later in the war. Or it might be that good practices in the middle
phases make it easier to annihilate insurgent forces at the end through a
combination of repression and mass force, which is one way of looking at the
Sri Lankan success. In short, it is may be how
one uses repression that is a "good" or "bad" practice, rather than
repression always representing a "bad" practice.
The RAND study concludes
with a series of recommendations for counterinsurgent forces. Not surprisingly,
some of these echo the current emphasis on population-centric COIN, warning
against repression and other approaches that are likely to alienate civilians. Others
are surprisingly banal, as in the exhortation to "keep a scorecard" of good and
bad practices and "make changes" if the balance tilts the wrong way.
The most
problematic recommendation, however, is that counterinsurgent forces should
simultaneously implement as many good practices as they can. This argument ignores the fact that the operational problems of
COIN are inexorably wrapped up with the strategic problems of state-building,
which is a protracted and bloody process. 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.
For
these reasons, policymakers should be skeptical about RAND's findings. The
utility of this study is, sadly, only marginal for the Afghanistan debate.
Joshua Rovner and Tim Hoyt are assistant professor and professor, respectively, of strategy and policy at the U.S. Naval War College. The
opinions expressed here are those of the authors alone, and do not
necessarily reflect the views of the Naval War College, the U.S. Navy, or any
other entity of the U.S. government.
(6)
HIDE COMMENTS LOGIN OR REGISTER REPORT ABUSE