A readers writes in to argue that my post on military innovation misses an important difference between R&D costs in the military and R&D costs as we usually think of them:

As a former civilian Director of the Air Force’s basic research program (now retired), I’d like to make one very important point that is usually lost in the discussions of R&D. The DOD budget for R&D is huge, but that is because it includes all of the product development and testing costs for the final products which the DOD then purchases. In almost every other area of Federal R&D, it is the private sector which foots that bill.

The only part of DOD spending on R&D comparable to the similar spending elsewhere in the government (NIH, NSF, DOE, NIST, etc.) is the so-called S&T [Science and Technology] accounts (specifically called 6.1, 6.2, and 6.3), carried out at DOD labs, universities and some industries. These S&T accounts totaled ~$15 billion of the total DOD R&D of ~$80 billion in fiscal 2011. This means that the real DOD contribution to the governments “innovation” account is not 55 percent, but rather more like 10 percent.

Cutting procurement of future systems may have profound impacts on the product development and testing part of the DOD budget while having little or no impact on the vital S&T.

Your points are well taken, but I believe that to understand the reality of government contribution to innovation one must compare “apples to apples”.