Readings

Readings: The Diffusion of Drone Warfare: Industrial, Infrastructural and Organizational Constraints by Andrea Gilli and Mauro Gilli

By Kenneth Anderson
Tuesday, June 10, 2014, 7:27 PM

Political science graduate students Andrea Gilli (European Union Institute, Florence) and Mauro Gilli (Northwestern University, Evanston) have posted a new and provocative paper to SSRN--"The Diffusion of Drone Warfare: Industrial, Infrastructural and Organizational Constraints."

I read this paper when first posted to SSRN some weeks back, but I waited to discuss it in a Readings post until I had talked through its themes with a few friends expert in this area. Its thesis can--and certainly should--be vigorously debated on some of its factual premises.  But I think it makes a quite plausible case on important questions of weapons diffusion and technological innovation.  (The article refers in the title and abstract to "drones," but neither the argument nor its examples is narrowly limited to "drones" as they currently exist.  Indeed the argument extends (as the authors note) to certain other weapon systems, such as autonomous weapons, that have similar or interrelated issues of infrastructure, expertise, and organization.)  Abstract:

Many scholars and policy-makers believe that drones will spread quickly because of their low price, their reliance on cheap commercial components, and their relative unsophistication. According to this view, this process will redistribute military power at the global level and, possibly, promote international instability. The literature in international relations and on globalization almost unanimously supports these concerns. Drawing from the scholarship in management, we show that such consensus is unwarranted. Specifically, even if we assume that advanced components are cheaply and easily available, drone warfare casts two major challenges. First, the production of combat-effective drones require advanced competences and industrial capabilities that are generally difficult, expensive and lengthy to develop. Second, the employment of drones calls for expensive and burdensome organizational and infrastructural support that, often, only few countries can afford. We test our claims by focusing on three types of military relevant drones – loitering attack munitions (LAMs), unmanned combat aerial vehicles (UCAVs) and ground and airborne surveillance drones (ISR and AEW&C). Our analysis shows that drone warfare poses significantly more daunting challenges than the current debate acknowledges. 

The most important and most persuasive part of the paper is that it is much, much harder to get--and keep--drone aircraft in the air than many appear to believe, and that to run a drones network of a kind that the US government does today, with a communications network via satellite, is not going to be easily duplicable, save by a quite limited set of states.  All of this has important implications, of course, for strategic studies, international relations, and defense policies that make assumptions about the availability of these technologies, or the speed of their availability and diffusion across the international arena, that might be quite mistaken.

This is also an analysis with many practical policy implications.  For one thing, the US and a handful of others, if they correctly assess the entry barriers for the practical military use of these technologies if they have to be developed from scratch, will have to consider whether to try to keep the advantages close to the chest, so to speak, or whether to commodify them to the extent of exporting them to close allies such as NATO partners. The authors urge in the conclusion that it makes better sense for the US to sell them to its allies--and I certainly agree--but an implication of this analysis is that the basic drone hardware won't mean much to those allies unless the US is also willing to give access to its communications and network infrastructure, both hardware and software.  That dependence, of course, means that one is in the hands of the US regarding actual use of the system.

But it seems unlikely that very many countries will be able or willing to develop the whole technological infrastructure from scratch.  In that case, they won't so much be buying a weapon system as doing what we all do with Apple TV or similar devices: buy a relatively modest piece of hardware in order to "lease" access to the networks.  This is not likely to make even close allies overwhelmingly happy with the arrangement.  In that regard, then, one of the most interesting bits of the discussion walks through the failed attempts by states to do so to date.

Conversely, however, high barriers to entry to getting one's own drone system in place and operational does not mean that it is equally difficult to attack such a network.  On the contrary, the complexity of the system means that it will tend to be fragile.  The communications links will certainly be places to attack.  This, in turn, will push for "hardening" of the communications systems over time and (I would add, drawing on Matthew Waxman's and my work) will gradually create pressures to automate and finally perhaps make autonomous certain parts of the system--to reduce vulnerabilities while it is being operated remotely and in real-time.  As one might imagine, all of these technological difficulties and more only increase as the ambition becomes the production of more and more highly automated jet aircraft--especially insofar as their competition is seen to be manned jet aircraft.

This is not an argument that the technological barriers will not be surmounted. Instead it's an argument that the complexity of technology, and its economic concentration in the hands of a small number of advanced states, means two things. First, all but a small number of states will seek to purchase such technologies if they want them, rather than starting from scratch.  Second, only a (perhaps surprisingly) small number of states will have the technical sophistication to be able to deploy these systems effectively--or even to keep them operational. Perhaps the automation of systems will lead them to be more end-user friendly over time, but that will be a ways off.  Meanwhile, crucial parts of a purchased system will depend on being able to "lease" the network necessary to use it.  My personal guess is that several leading industrial powers will conclude that the stakes are not merely security stakes, but global economic ones in terms of technological innovation--and perhaps that will be enough to induce the necessary investment, but it doesn't seem likely.  No doubt China will proceed to steal and reverse engineer the technology, but to judge by some of the evidence in this paper, many of the barriers to entry involve infrastructure and organization that cannot be immediately overcome by stealing IP.

There are a number of issues raised in the paper that might be reasonably disputed, particularly some of the business model assumptions that the authors believe are already shaping economic concentration of research, development, and innovation in the field, as well as some of the assumptions about the paths of technology.  I also think that among the cognoscenti - among, for example, military experts in the US and Israel with whom I informally discussed these questions over the last few weeks - this will seem much less new and startling than these conclusions might seem to outsiders relying on second-hand understandings of the technologies and the underlying industrial and military structures for these new systems.  But the misperception that Gilli and Gilli identify is, as they say, widely held--particularly among political scientists, international relations policy experts, and international lawyers and academics--and their article offers a needed corrective.  Below the fold is an excerpt from the article's conclusion.

From the Conclusion (notes omitted):

[D]rone warfare poses more challenges than generally acknowledged. By analyzing three types of drones with relevant military capabilities, we have shown in fact that UAVs are far from easy and cheap to develop and to employ. Future revolutionary breakthroughs, making unmanned systems cheaper and easier to build, are certainly possible. However, the history of warfare and our empirical analysis warn some skepticism. Similarly, since even wealthy European countries with an advanced defense industrial base struggled, and in few instances even failed, to develop the UAVs we have considered, there is little reason to believe that other countries currently lacking the necessary industrial capabilities will fare any better. Additionally, UAVs’ employment cast major organizational and infrastructural challenges even for the US, Israel and the biggest European countries. For other nations, the adoption of drone warfare will then likely be extremely challenging and possibly even overwhelming. We conclude that drone warfare is unlikely to spread quickly and easily and hence to affect the distribution of military power at the global level and to promote international stability at least in the short term.

Our analysis has important implications both for theory and practice. Our article contributes to the literature on the diffusion of military innovations in few important ways. First of all, by looking not at the single platform but at the ecosystem required for the successful employment of military technology (the tripartite challenge related to the platform, component and complement challenges), our analysis complements the existing scholarship as it sheds light on aspects that have so far received scant attention – like the infrastructural requirements at the complement level. Second, our analysis of the platform challenge casts strong doubts on the widely held view in the field of IR that imitating military technology (the hardware) is easy. Consistent with the works of Eugene Gholz and Jonathan Caverley, our article shows that the very opposite is true: even for relatively unsophisticated drones, the platform challenge can prove extremely daunting. Third, our analysis supports the works of Emily Goldman and Michael Horowitz about the organizational challenges new military technologies trigger in terms of bureaucratic structures, codes and doctrines. Finally, our findings also speak to the literature on globalization and national security. Specifically, our analysis suggests that the availability of advanced components is far from sufficient to develop first-class military capabilities.

Our work has also relevant policy implications. In contrast to the consensus about drones, it provides a more reassuring view. From our analysis we conclude in fact that the concerns about autonomous and unmanned weapons and their potential destabilizing effects on the international system are likely exaggerated as such capabilities will not spread easily and quickly – because of technological, organizational and infrastructural constraints. Some important policy implications follow. First, when we put into the equation also the role of future anti-drone systems, the US technological advantage is in fact likely to last and, possibly, even to widen further. This means that the only real prescription for the United States is to continue – and even to increase – its investments in unmanned technologies and in potential counter-systems so as to preserve its hedge over both friends and foes. Second, this means that a new legal and normative framework, including amendments to the Missile Control Regime Treaty, is likely going to be both insufficient and unnecessary. On the one hand, despite the clarity of the treaty’s provisions, France and the UK have exported in the past their joint air-launched cruise missile program, the Scalp EG/Storm Shadow, to Saudi Arabia. Any amendment to the treaty will likely fail as long as commercial considerations push in the opposite direction. On the other hand, because of their organizational and infrastructural requirements – let alone the difficulties of sourcing specific components (like appropriate air-launched missiles) – many UAVs will not immediately and widely endow countries with advanced military capabilities. Finally, when Israel struck a deal with China for upgrading its Harpy LAM, the US did not need any treaty, it could force its ally to backtrack by leveraging its economic, diplomatic and military role.

This brings us to the third aspect: the role the United States can actively play in the age of drone warfare. Some suggest that the US policy should be aimed at limiting the sale of armed drones, even to its own allies. We are unconvinced. This policy may in fact backfire and provide US allies with an incentive to develop their drone platforms autonomously. This would not only be counter-productive but it would also harm the US’s long-term interests. In the age of budgetary austerity within NATO, the US should promote efficient spending among its allies, not provide them with further incentives to duplicate capabilities. A more liberal export policy is then likely the most suitable, especially if coupled with different forms of subsidies to favor the transfer of US unmanned technologies. On the one hand, because of the ecosystem’s challenges discussed in this article, the active promotion of drones is unlikely to trigger major effects on international security. On the other, the US could prevent the emergence of potential commercial competitors while locking-in its allies into its weapon systems – de facto strengthening their security commitment.