An Uncivil Treatment of a Civil Servant

Jeff Lewis’s article on the reported retirement of Andrew W. Marshall, the Pentagon’s Director of Net Assessment, is mean-spirited, superficial, and inaccurate.

It was mean spirited because Marshall has quietly and loyally served the nation for four decades. He has served presidents as different as Jimmy Carter and Ronald Reagan, George W. Bush and Barack Obama, and secretaries of defense as diverse as Harold Brown and Dick Cheney, Bill Perry and Donald Rumsfeld (twice). He has done so quietly, preferring the wings to center stage and the shadows to the limelight. That is as it should be for one whose job is to advise the Secretary of Defense. Lewis makes it clear that he has never met Marshall, so he instead relies upon several sources that range from the early 1970s to the late 1980s in attempt to characterize (or, more properly, mischaracterize) his views. Lewis goes to heroic lengths of inference and innuendo to portray Marshall as a man who never met an arms race he didn’t like.

Lewis credits Marshall with "only one" idea: the concept of competitive strategies. If that were true, that alone would be remarkable. Those of us who labor in the vineyards of academia (and, one suspects, in the think tank thicket that Lewis inhabits) would derive considerable satisfaction from having developed a single influential idea. Indeed, one is at a loss to name a long-serving (or, for that matter, short-serving) civil servant who has developed "only one" comparable idea. Chuck Spinney comes to mind for his ideas on weapon system affordability, but one runs out of examples after that.

In fact, however, Marshall’s influence has been more far reaching than Lewis admits. Marshall personally and the office that he has led have been responsible for studies on a wide variety of topics, including:

• Bureaucratic politics (read the acknowledgements to Graham Allison’s seminal The Essence of Decision).
• Military power (to include Marshall’s own monograph, Problems of Estimating Military Power, written while he was still at the RAND Corporation).
• Military effectiveness (to include sponsorship of Williamson Murray and Allan Millett’s landmark three-volume study of the same name).
• Military innovation.
• And importing the Soviet concept of Revolutions in Military Affairs.

Indeed, Marshall has helped shape the field of strategic studies as we know it.

And what of the "one" idea that Lewis attributes to Marshall: the notion of competitive strategies? Here Marshall deserves considerable credit, but credit that should be shared with others, including his then senior military assistant (and later Secretary of the Air Force) Jim Roche, as well as Thomas C. Schelling (hardly a proponent of arms races), who was one of the first to write about what would later be called competitive strategies.

I read Lewis’ post with interest because the topic of competitive strategies, arms competitions, and arms races, is one familiar to me. I’ve edited a book on competitive strategies, written a book chapter on arms competitions, and am currently working with two colleagues on a project devoted to arms races in history (all, for the record, sponsored by the Naval War College Foundation; none by the Office of Net Assessment). I am also quite familiar with the documents that Lewis refers to. Unfortunately, much of what he writes about them is incorrect.

Lewis’ central claim is that "competitive strategies was the notion that the long-term strategic competition with the Soviet Union — the nuclear arms race — was inevitable." He goes on to argue that this view dismisses the notion of stability in superpower relations as well as the possibility of cooperation. To bolster his claims, Lewis cites Marshall’s Long-Term Competition with the Soviets: A Framework for Analysis, a monograph written while Marshall was an analyst at the RAND Corporation and before he assumed the position of Director of Net Assessment, first in the White House and then in the Pentagon.

The report deserves to be read cover-to-cover, particularly for those interested in arms races and strategic interaction. It is an insightful piece of work. Unfortunately for Lewis, it doesn’t say what he says it does.

The central finding of the report is that as of the early 1970s strategic interaction between the United States and Soviet Union did not have the characteristics of an action-reaction arms race, as posited by Western arms race scholars. Rather, it was "a much more complex, slower moving process." Nor was it clear that Soviet arms investments were primarily driven by American actions, and vice versa. As Marshall put it, in both the United States and the Soviet Union "the level of military expenditure is probably set not as a reaction to specific military moves of the other side, except in extreme cases, but as part of an independent budgetary process." This view was later corroborated by an exhaustive classified history of the U.S.-Soviet arms race written by Ernest R. May, John D. Steinbruner, and Thomas W. Wolfe that Marshall commissioned after he became Director of Net Assessment.

It is untrue that Long-Term Competition with the Soviets dismissed stability. To the contrary, Marshall lists it as a goal of U.S. strategy (on page vi, for example) — just not the sole goal. Similarly, the monograph doesn’t dismiss the prospect of cooperation with the Soviets — just the notion that cooperation would eliminate entirely competition between the superpowers. As a result, the United States faced the need to plan to compete with the Soviet Union to achieve its objectives. As Marshall wrote, "No available policy can entirely remove competition in strategic arms; therefore, good policy involves competing while attempting to control the stability of the competition process." That seems utterly sensible to me.

Lewis argues, "spending other countries into the ground seems pretty improbable." Marshall would likely agree. In fact, on page vi of Long-Term Competition with the Soviets, he expressed concerns that the United States was pricing itself out of the competition with the Soviets. Indeed part of the original impetus for the idea of competitive strategies in the mid-1970s was that the United States could not successfully pursue its interests merely by throwing money at the problem, but rather needed to think strategically and make choices.

Marshall’s view that Soviet arms decisions were motivated only partially by external developments such as U.S. statements and actions was corroborated by the set of interviews with former Soviet leaders sponsored by the Office of Net Assessment and carried out by John Hines of the BDM Corporation. The interviews showed the Soviet decisions were more driven as much by bureaucratic politics and organizational culture as anything the United States. Lewis agrees that this is a valuable study and credits the Office of Net Assessment with sponsoring it, but then, inexplicably, asserts that "Marshall hated it." Huh?

I would agree with Lewis that strategic competition remains an important subject. Whether interaction between great powers tends to manifest itself in closely coupled action-reaction dynamics or loosely coupled, slower moving interactions has considerable policy relevance. For example, whereas a growing number of analysts resort to talking about an "arms race" in Asia, it is likely that strategic interaction is much more episodic. Marshall’s own work, together with that he sponsored, offers important insights into this subject.

The fact that Jeff Lewis appears to have only a marginal, and distorted, understanding the Office of Net Assessment is fine, in a way. It wasn’t created to serve the Jeff Lewises of the world. Rather, it was put in place to serve the Secretary of Defense and the senior leadership of the Defense Department. This it has done ably for over four decades. One hopes that it will continue to do so after Andy Marshall retires, with the thanks of a grateful nation.