Analysis Paralysis: the Intelligence-Policy Divide, Revisited

The National Intelligence Council’s Global Trends 2030 was released earlier this year (you can find it here).  In that context, it is worth mentioning an important point that Wikistrat‘s Thomas P.M. Barnett made earlier about previous NIC’s forecasts in his 2005 book, The Pentagon’s New Map.  Barnett’s key point in the book for our purposes is that the US Intelligence Community believes that it must only do analysis, and never engage in “advocacy” of any particular policy.  This epistemologically naive point of departure poses a number of problems.

These problems were recognized early.  Back in 1949, the nascent CIA’s famous Kent-Kendall debate was precisely about the so-called “intelligence-policy divide“. Sherman Kent, founding father of the CIA’s analytic arm, advocated a bright and stark line  between the analyst and the policy maker in order for the former to retain analytic objectivity.  Willmoore Kendall, his colleague, argued the contrary view, that only by being closely involved in the policy making process could the analyst really usefully contribute.  But from the distance the argument seems a bit abstract and one of the merits of Barnett’s reminder is to give a concrete example of the problem created by a purely analytic approach to intelligence.

The CIA’s Sherman Kent (L) and Willmoore Kendall (R)

First, such pretended intellectual virginity weakens analysis, because it compels analysts to describe what every other country in the world will do in response to an unfolding series of events while essentially keeping the United States itself static.  Only if analysts are allowed to imagine possible US moves – and this can only be done with policy makers – can a really dynamic approach to analysis be achieved.  Clearly, in the midst of this one could come close to advocacy, Kent school’s greatest heresy.

Your Holiness, call me a heretic, but I cannot analyze without advocacy!

Second, and more subtly, according to Barnett, all things being equal an illusory intelligence-policy divide leads analysts to imagine the future as a straight continuation of the present. But clearly, all other things are not equal, and analysis should not just be about projecting the trends but about evaluating the possible strategic interplay between policymakers and the objects of their policies.

In the end, Barnett advocates a different approach to thinking about the future, one that is more about shaping than predicting. Whereas most analysts define their professional environment as “futures to be avoided,” he prefers to focus on a future “worth creating”.   He concludes “What I want to do is embrace that future and shape it from within”.

This approach is something Milo and I have also long advocated in this blog and in our teaching:  let’s largely move away from prediction – it doesn’t work anyway – and focus on what we can change.  With their deterministic view and impression of ineluctable trends, forecasting exercises really can be the enemy of human action. Far from enlightening us about possible futures, predictions can cloud our thinking and limit our action. The fact that in most human spheres they almost invariably end up horribly off the mark is an additional reason not to even bother reading them. Instead of predicting, put your energy into shaping.  And should you decide to do extensive analysis first, don’t kid yourself that you’re not embedding advocacy in very line.

Barnett’s note can be found here, it is worth reading. The Kent-Kendall debate is described in Jack Davis’ excellent article available here. On forecasting, see our article “We have met the enemy and he is, er, forecasting.”

If you enjoyed this post, why not subscribe to our blog?