With innovation on the minds of many in the defense community, there was good reason to anticipate comments made earlier this month by Air Force General Paul Selva—the current vice chairman of the Joint Chiefs of Staff—at the Center for Strategic and International Studies (CSIS). General Selva, according to a write-up of the event by Colin Clark at Breaking Defense, tried to quell fears that that the military is too hidebound or conservative to innovate.

According to Clark’s report, Selva told the audience, “We grow up assessing risk… We wire that into our leaders from the beginning of their training. It’s not that we’re risk averse; it’s that we’re imminently aware of that risk.” And while it’s unclear why those two things are mutually exclusive—greater awareness of risk seems like an eminently understandable reason for aversion to it—General Selva’s comments get at something very important that Department of Defense officials need to understand in their current efforts to encourage greater innovation in the U.S. military.

Military organizations are risk-averse in large part due to the nature of their organizational purpose. All organizations develop routines or standard operating procedures to minimize complexity and uncertainty in fulfilling their purpose. Thankfully, militaries rarely practice their craft. When they do, the military organization’s purpose is to destroy another similar organization—in the process sacrificing parts of itself. Militaries develop doctrines, as political scientist Barry Posen recently explained, to reduce the uncertainty inherent in this task. Because of the importance of doctrine, military leaders are loath to adopt new technologies that will disrupt it. Military organizations are not so inherently conservative as to reject new technology outright, but they generally approach those new technologies that would require doctrinal changes cautiously.

General Selva may protest, but this caution can and should be called risk aversion. But risk aversion is not necessarily bad. As Posen noted in his seminal work on military innovation three decades ago, innovation is not a priori good. If innovation is disruptive to the military organization, it could lead to diminished effectiveness. If the point of innovation is to improve the performance of a military organization—as Pentagon officials currently hope it will—then caution is appropriate given the chances it could actually lead to battlefield failure.

But diminished organizational effectiveness is not the only risk military leaders will try to avoid. The possibility of programmatic failure is another risk the military services will seek to minimize. Military organizations get resources through the political process and largely in the absence of an organizational output to measure the effect to which those resources are being put. James Q. Wilson argues that these political constraints and contextual goals lead to a great deal of risk aversion. It is much more difficult to specify which investments deterred China from invading Taiwan yesterday, or the Red Army from pouring through the Fulda Gap during the Cold War, than it is to identify costly boondoggles such as the F-35 or the Army’s Future Combat System. Those political constraints mean the military services are more likely to maintain their partnership with traditional defense industry—who can serve as an interface with political overseers in Congress—rather than risk new relationships with Silicon Valley and other high tech commercial providers as Secretary of Defense Ashton Carter hopes.

As the current Pentagon leadership looks to encourage greater innovation in the U.S. military through its Third Offset Strategy and initiatives like DIUx, it has to understand the sources of military risk aversion. General Selva’s comments read like a rebuttal to traditional accusations of military conservatism. However, there are good reasons why the military services are risk-averse. If military innovation is as necessary as defense officials say it is, understanding the sources of that risk aversion is necessary to overcome it.