Correspondence: Military-Technological Imitation and Rising Powers

To the Editors (Michael C. Horowitz writes):

Andrea Gilli and Mauro Gilli should be lauded for making clear how the growing complexity of capital-intensive military platforms such as fighter jets hinders states, such as China, seeking to mimic the United States.¹ Gilli and Gilli join a long line of thinkers, myself included, who argue that military technology does not always diffuse easily and that the characteristics of technologies matter in driving how those technologies spread and influence international politics.² Although there is much to like about Gilli and Gilli's article, their analysis has some theoretical limitations with implications for policymaking.

First, because Gilli and Gilli evaluate only military technology adoption, they miss the broader ways that human capital, tacit knowledge, and organizational practices shape military power. Gilli and Gilli's unit of analysis is military-technological superiority (p. 145), suggesting a technologically determinist view of military power. Technology, however, is only a subset of how states generate military power. More important for victory and defeat, on average, is how states develop and employ their capabilities on the battlefield.³

Gilli and Gilli's ignoring of the organizational component of military power is relevant because another reason why a country such as China might struggle to adopt some of today's key military capabilities involves a lack of tacit organizational and operational knowledge. Moreover, explaining how countries use their military equipment, not just their military technologies, requires understanding the varying organizational capacity requirements for those technologies.⁴

Second, Gilli and Gilli's key conceptual advance is to show how the growing complexity of military technologies slows technology diffusion. However, they incorrectly bundle multiple things into their concept of complexity. The level of difficulty in copying technology changes depending on its absolute complexity—the number of intersecting components, challenges in machining them, systems integration, and the like—and on the number of players with the capability to produce those systems.

My adoption capacity theory argues, for example, that the resource and organizational requirements of a particular innovation influence the number of players who can adopt a technology, in addition to broader innovation adoption rates.⁵ Technology diffusion slows as unit costs for the applicable systems increase and when the relevant technology has exclusively military applications.⁶ These factors help explain, for example, why stealth (part of Gilli and Gilli's F-22 case) has spread so slowly. Stealth technology is expensive and has only military purposes, which also causes it to spread slower than less expensive technology based on commercial products.⁷ When the underlying basis of a technology is commercially oriented, more actors get involved, for reasons involving market forces, and technology can spread via more pathways.

When Gilli and Gilli argue that more complicated technologies are inherently harder to copy, they also imply that the diffusion rate for all complicated technologies should be the same—whether an exclusively military technology such an F-22 or a dual-use technology such as a smartphone. Smartphones have spread around the world, however; F-22s have not. Moreover, contrary to their argument, some complex military technologies, including cruise missiles and armed uninhabited vehicles, have spread widely.

Finally, I feel compelled to correct Gilli and Gilli's repeated misrepresentation of my research. Gilli and Gilli selectively quote my work to imply that I argue that all military technology diffuses easily.⁸ As described above, this is incorrect. For example, they quote me as saying "it is [not] difficult to copy . . . specific technologies" (p. 188). Here is the quote in context: "Organizational capital is a critical element of firms' success, helping businesses survive during periods of change by giving them the institutional capacity to shift practices without being blocked by bureaucratic obstacles. Because it is difficult to copy business processes, as opposed to specific technologies, organizational assets are difficult to duplicate. Darby and Zucker find a ten- to fifteen-year gap between the beginning of the biotechnology revolution and widespread knowledge of the practices necessary to take advantage of the evolution for business purposes...