on september 9, 2013, the Japanese Air Self Defense Force (JASDF) detected an unknown aircraft hovering approximately 200 km northwest of the Senkaku islands in the East China Sea.1 The ownership of these uninhabited rocky islands, only 2.7 square miles in size, has been long disputed by China and Japan. The islands—known as the Diaoyu islands to China—have been under Japanese administrative control since 1972, although China had never relinquished its legal claim that the islands were part of its sovereign territory.2 The scramble for control of these desolate islands had intensified as China’s booming economy increased its appetite for oil and gas supplies worldwide. Reports of vast deposits of oil and gas in the seabed below the islands had made seizing control of them and asserting ownership rights over the surrounding territorial waters a high priority for China. Accordingly, its efforts to challenge Japanese control over the islands had become more insistent. China Maritime Surveillance (CMS) vessels sailed near the islands and manned Chinese aircraft edged around the margins of the island’s airspace.3 But until that September afternoon, China had never attempted a drone incursion over the Japanese-controlled airspace over the islands.
Within minutes of the incursion, the JASDF scrambled F-15 fighter jets to confront the aircraft. They were shocked at what they found: a sophisticated Chinese surveillance drone, believed to be a Harbin BZK-005, a high-altitude, long-range drone employed by the Chinese navy. The Japanese fighter jets shadowed the Chinese drone until it eventually departed the islands’ airspace. While China had deployed drones in the South China Sea before, it never had done so to test Japanese control of the Senkaku Islands. Japan was alarmed and demanded an immediate cessation of drone flights over the islands. China’s response was blunt: “China enjoys freedom of overflight in relevant waters . . . The Chinese military will organize similar routine activities in the future.”4 Under pressure, Japanese Prime Minister Shinzo Abe granted approval to the Ministry of Defense to shoot down any drone that ignored warnings to leave Japanese airspace. An unnamed Japanese defense official remarked that the Chinese drone incursion “was unexpected . . . I fear we are in the position of being one step behind.”5 Weeks later, Japan approved new guidelines for the JASDF to shoot down drones that enter Japan’s airspace if warnings to leave were ignored.6 Geng Yansheng, a spokesperson for China’s Ministry of Defense, said that such an act by Japan would constitute “an act of war” and lead China to strike back with “resolute measures.”7
Alarmed at the escalating rhetoric, the United States jumped into the fray to show its support for Japan. By early November 2013, the Obama administration had agreed to deploy its Global Hawk drones on joint missions with Japanese E-2C early warning aircraft in the airspace over the islands.8 Japan also revised its defense spending plans and ordered the purchase of several US drones for its own fleet. On November 23, 2013, China responded by unilaterally extending its Air Defense Identification Zone (ADIZ) over much of the East China Sea, including the Senkaku islands. This expansion of China’s claims rattled the United States, Japan, Taiwan, and South Korea and increased pressure on Washington for a tougher response. Chinese foreign ministry spokesperson Hong Lei remarked that due to the ADIZ extension, the islands were “an inherent territory of China. Japan’s seizure and occupation of the islands are illegal and invalid.”9 The drone incursion became a forerunner of an expanded Chinese claim over territorial waters in the East China Sea, while the islands themselves had now become “a focal point for Sino-US competition and potential conflict.”10
The Senkaku islands controversy sharpened the lines of division between China and Japan and led both to take measures to prepare themselves for a future war. For both countries, improving their ability to signal, threaten, and fight with drones became a high priority. In 2012, China began the construction of two drone bases to boost its capacity for surveillance over the Senkaku islands.11 By 2014, China was allegedly planning eleven coastal drone bases to allow it to conduct surveillance over an even larger area.12 In June 2015, a leaked People’s Liberation Army (PLA) document revealed that China was considering deploying its Yilong surveillance drones to the airspace over the Senkaku islands.13 Japan took notice of China’s growing interest in drones and responded in kind. In 2014, the Japanese government announced plans to increase its spending on drones by 300%, a three-million-yen investment that would give it substantial surveillance capacity by 2020.14 By 2015, Japan was debating shooting down Chinese drones that intruded into the airspace, but acknowledged that doing so could provoke an international crisis.15 Incursions around the Senkaku islands have continued with Japan reporting another Chinese drone hovering over its territorial waters in May 2017.16 In 2019, the Trump administration decided to up the ante by selling surveillance drones to Malaysia, Indonesia, Philippines, and Vietnam as a way of signaling that it would no longer “tiptoe” around Chinese aggression in the South China Sea.17
In time, the standoff over the Senkaku islands may be remembered as one of the first of the drone age. Today, a growing number of states are using drones as China did: to test the nerves and strategic commitments of their rivals, to chip away at relationships of deterrence, and to gain leverage in crisis bargaining scenarios. While most states are using drones for routine surveillance, emergency response, and other ordinary tasks, governments are introducing military grade drones into long-simmering regional conflicts to gain advantages over their rivals and to see what response, if any, they provoke. This matters because the unmanned nature of drones can change the underlying calculations of risk that lie beneath deterrence and coercion. In most circumstances, when using manned aircraft states are naturally sensitive to the risks of losing crucial military personnel and assets and tend to tread carefully in crisis situations. States also fear that downing another state’s aircraft can damage their reputation and even invite more provocations. As both the United States and Soviet Union acknowledged during the Cold War, attacks on manned aircraft that kill a pilot or crew can precipitate a crisis with a momentum that proves hard to reverse with ordinary diplomacy.
With today’s unmanned aircraft, the calculation is not so straightforward. Almost all of the underlying assumptions about risk, credibility, and the potential for crisis escalation are scrambled when a human being is no longer in the cockpit. With no lives at stake, governments may see drones as cheaper or more expendable and take risks that they would not do with manned aircraft. For example, they may use drones to rattle their opponents and to test the limits of deterrence, under the assumption that an unmanned aircraft will not provoke the same military response that that manned ones do. They may further gamble that shooting down a drone will not precipitate an irreversible international crisis, as killing a pilot in a manned aircraft probably would. Aside from increasing risk-taking in deterrence, drones may also lead to an expansion of a state’s goals and lead them to try to coerce enemies in new and potentially destabilizing ways. In a world awash with drones, strategic moves once thought too dangerous become possible, even attractive, if they can be conducted without physical risk to human pilots. But this perception of limited, controllable risks may be wrong; there is no guarantee that one’s enemies will “see” drone incursions in exactly the same way as their opponent and respond with restraint. A world full of risk-taking with drones, but marked by uncertainty about how strategic moves involving them will be interpreted, is one in which crises may escalate in surprising ways.
The risk of new strategic interactions between states with drones is growing precisely because the technology has spread so quickly, but unevenly, around the world. Less than two decades ago, only a select number of governments with highly capable militaries, such as the United States, Russia, China, Israel, and the United Kingdom, had even small drone fleets. Even these governments tended to see drones as unreliable and expensive and cut those programs first to save other defense programs in times of austerity. As the air wars in Afghanistan and Iraq ramped up, the United States vastly increased the size of its drone fleet to meet reconnaissance and combat demands by ground commanders and to target al Qaeda and associated forces. By contrast, most of the other countries waited until the technology matured and costs decreased to build up their drone fleets in a significant way. When the technology became reliable and the cost curve flattened, drones began to spread to more countries worldwide, depending in part on their technological capacity to develop and run them.18 At present, at least eighty countries have unarmed drones and twenty-four have armed drones.19
Although drone diffusion is a fact of life, not all the world’s drone programs are the same; they range in sophistication from relatively small boutique programs to growing multipurpose drone programs for combat, surveillance, and other uses. For many states, indigenous development of drone technology has moved slowly and remains largely imitative, copying popular US and Israeli models and adapting them to local needs. In some cases, the rationale for developing a local capacity for building drones is essentially one of prestige: developing a token drone industry is a signal that the state is a serious military player in the world. As a status symbol, a drone industry has become a “must have” no matter how strapped defense budgets may otherwise be.20 For this reason, many governments now sponsor small or moderately sized research and development (R&D) programs for drone technology, which are in turn supported by smaller commercial drone firms hoping to corner the local market. This effectively makes the global diffusion of drones look greater than it is. The drone programs in many states have small ambitions. Many are designed only to be effective inside their own borders, and do not aim to match the level of sophistication of US drones like the Reaper or the Global Hawk. But other countries have set their sights higher. Through government-led and private investment, countries such as Russia, China, Turkey, India, Pakistan, and Iran have produced increasingly complex drone models, sometimes by harnessing innovations from their commercial drone industries or by copying US and Israeli technology.
Skeptics have argued that global drone diffusion will run into its natural limits due to the fact that the technology requires a costly organizational infrastructure, as well as technical base of expertise, to support its development and use.21 It is certainly true that military-grade drones are not an easy thing to construct from whole cloth. Designing drones takes a significant base of technological expertise and a defense manufacturing capacity; using them as the engine of a targeted killing campaign based on information dominance, as the United States has done, takes vastly more resources and is an option unavailable to all other governments at present. 22 Even to deploy drones on a regional basis, as China does, takes a network of military satellites, intelligence capabilities, and data bandwidth capacity that few countries have available. But more to the point, only a very small number of states have a compelling reason to develop drones with that kind of reach. Only governments facing substantial security threats, such as terrorist groups like al Qaeda or regional rivals, have a reason to invest deeply in drone technology in the way that the United States, Israel, and now China have done.23 For the rest of the world, living with relatively peaceful, stable borders, there is no incentive to construct an equivalent of the Reaper. Instead, they will be satisfied with smaller range, less expensive drones that can monitor borders, stop illegal smuggling in drugs and other banned goods, and conduct routine military tasks like target-spotting for their ground forces. This suggests that indigenous drone development will stratify into two tiers: a select number of states who have the need and the organizational and infrastructure base to build high-quality military drones, and a wider array of states engaging in the production of drones which may be only a cut above the commercial models available but are sufficient for routine tasks.
To a degree, this stratification of the world of drones is complicated by two cross-cutting factors. The first is the explosion of the commercial drone market. An estimate by the Teal Group in 2013 suggested that drones will remain a global growth industry for the foreseeable future, with annual sales expected to double from $5.2 billion to $11.6 billion by 2023.24 The Teal Group also estimated that approximately $89 billion would be spent in research and procurement on drones over the next decade.25 In 2011, there were 680 active drone development programs run by governments, companies and research institutions, compared to only 195 in 2005.26 This market has only grown. In 2016, PricewaterhouseCoopers LLP projected that the global market in drones would ultimately be worth $127 billion by 2020.27 According to an estimate by Goldman Sachs, between 2017 and 2021, the majority of spending on drones will be concentrated in a few countries with highly developed programs (United States, Russia, China, United Kingdom, Australia) but smaller programs in countries beyond this set will nevertheless constitute $8 billion in spending.28 Even where the commercial drone industry is small, it can yield benefits for smaller militaries because they can use homegrown drones for routine tasks like reconnaissance and monitoring the perimeters of a base. The global expansion of companies like DJI also allows some cash-strapped militaries to buy “off-the-shelf” drones for these tasks. Beyond the fact that a booming commercial market means more drones are available to buy, expertise also travels between the commercial and military R&D world, especially in countries where there is a permeable border for employment between government and private industry. In such cases, even a small local drone industry will have a spillover benefit in improving the quality of drones available to that country’s military.
The second complication for this picture is the military export market, which is putting more advanced technology into the hands of governments that could not otherwise develop it. The countries that export military-grade drones can be divided into two camps. The first camp consists of the first movers in drone technology, the United States and Israel, who have years of experience developing and using military grade drones and have a sophisticated defense-industrial base with deep expertise and extensive commercial links. Unsurprisingly, the United States and Israel are best positioned to capitalize on the drone export market and have been the leading global exporters so far. The second camp consists of relative newcomers to the drone industry—such as China, Russia, India, South Korea, and Japan—who have poured resources into drone development and are poised to catch up, even imperfectly, with the United States and Israel in the export market. According to an estimate by IHS Janes produced in 2016, these five countries are due to account for $3.4 billion in drone export sales between 2015 and 2024.29 Another new potential exporter is Turkey, which has used its defense manufacturing base and its experience in a long-running war against the Kurds to hone indigenous models and prepare them for sale.30 Of these six countries, China is the best positioned and most eager to capitalize on the export market. This suggests that the export market in the future may be dominated by three large exporters (the United States, Israel, and China), with the other exporters fulfilling highly specialized demands for drones. In part due to the actions of these three actors, the export market will drive global drone diffusion, produce competition that lowers costs, and put sophisticated drones in the hands of governments that could not otherwise develop them.
The vast R&D base for military and commercial drones present in the United States puts it in a strong market position for exports, although its actual record of global exports still fall below those of Israel. According to the Teal Group, the majority of global drone spending—65% of the R&D and 51% of the procurement—will come from the United States between 2013 and 2023. Its long-running investment in drone technology has given the United States a qualitative advantage, and it often develops models years ahead in quality compared to others. This lead in R&D is due in part to deep links between government and private industry. The US commercial drone market—already a world leader, but estimated by the FAA to grow tenfold between 2017 and 2021—is a source of expertise and ideas for innovation in the military drone sector.31 The United States has also developed an array of private and university-based drone labs which are partially supported by Department of Defense funding. This funding allows commercial labs to experiment with drone prototypes and to develop capacities for autonomous flying, swarming, and other types of innovations.
Given its lead in advanced drone technology, the United States is in the best position to take advantage of the opportunity to sell drones on the international market. Yet its actual record of exports is relatively modest. The total direct military commercial sales for US drones was $240 million during the period 2005–2011, with an additional $144 million transferred as part of the foreign military sales program over nearly the same period.32 Most of the early US drone sales went to NATO allies, such as the United Kingdom and Italy, and US officials initially rebuffed calls to sell sophisticated models such as the Reaper or Global Hawk to other long-standing partners. This reluctance was partially due to legal agreements on technology sales and transfer under US law. For the United States, all international arms sales are governed by the Arms Export Control Act (1976), which sets standards for designating the “friendly countries” that the United States can export to, and all US companies are bound by the International Traffic in Arms Regulation (ITAR) law which requires State Department approval in advance of a sale.33 Drone exports were particularly restricted by the Missile Technology Control Regime (MTCR), a voluntary agreement among thirty-four states which limits the sales and export of heavy payload weapons, and by the Wassenaar Agreement, which requires participants to share information on deliveries, especially for dual-use technology.34 The Obama administration proposed changes to the MTCR to enable more drone exports, but was unable to win agreement on the proposed changes from other signatories.35 Further concerns among US officials over the ability of the United States to monitor the uses of drones by other countries also led to caution in implementing an aggressive export policy.
This caution over drone exports incurred the wrath of powerful drone manufacturers who believed that the United States was foolishly sacrificing a chance to become the world’s most prolific drone exporter. With jobs and profits on the line, major companies like General Atomics, Northrup Grumman, DJI, and others sought allies on Capitol Hill and formed unmanned aircraft caucuses in the House and Senate, in part to push for reforms to export policy.36 Drone lobbyists worked to reclassify some drones to exempt them from MCTR and Wassenaar restrictions and to allow manufacturers to tap growing markets in Asia, Europe, and Latin America.37 A 2012 Congressional Research Service report put the pro-export argument succinctly: “Much new business is likely to be generated in the market, and if US companies fail to capture this market share, European, Russian, Israeli, Chinese or South African companies will.”38 Worried that these companies will lose market share to foreign companies, the Obama administration began to consider expanding the range of states that are pre-approved for drone sales.39 The initial draft guidelines proposed by the Pentagon in 2012 suggested that the United States would be permitted to export at least some drones to sixty-six countries around the world.40 But developing a policy ultimately took two years of intense behind-the-scenes negotiations between a number of government agencies that had a stake in drone exports.
What emerged was a cautious, hedged export policy that shied away from blanket approval of drone sales, particularly of the armed variety. In February 2015, the State Department announced a final policy for the export of military and commercial drones. According to their guidelines, the United States could export armed military drones to other governments under strict conditions, among which was that the recipient government must agree to end-use assurances to ensure US technology was not deployed for illegal purposes, such as repression of the local population. The new export policy did not have a formal list of countries approved for drone exports, as some in the Pentagon originally envisioned, but rather processed each request for a sale on a case-by-case basis.41The policy had a “strong presumption of denial” for licenses to export lethal drones and required four specific commitments from those that received US drones: (1) that the drone system would be used in accordance with international law, including international humanitarian and international human rights law; (2) that the drones would be deployed “only when there is a lawful basis for the use of force under international law, such as national self-defense”; (3) that the drones would not be used to “conduct unlawful surveillance or use unlawful force against their domestic population”; and (4) that drone users would have to agree to require technical and doctrinal training to “reduce the risk of unintended injury or damage.”42 These standards are relatively high and, as one critic noted, it is not clear that the United States would meet them for some of its own military operations, especially targeted killings.43 The Obama administration’s policy was designed to be consistent with the MTCR and to restrict the transfer of high-payload drones.44 Under the new export policy, medium- or high-altitude drones would rarely be exported and then only to NATO allies who could use them for counterterrorism and other US supported missions. In many of these cases, they would be used for operations in which the United States shared a military command with its NATO allies, such as airstrikes in Libya and Iraq. Even with these restrictions, US officials nevertheless argued that this export policy would vastly improve the capability of partner nations and improve interoperability with US military forces, especially for intelligence, reconnaissance, and surveillance missions.45
The Obama administration’s export policy on drones was designed to resolve some of the uneasiness around unleashing drone sales worldwide, but it left a number of questions unanswered. The unclassified parts of the export policy did not make clear what substantive penalties, if any, a state would incur for violating any of these end-use assurances. The United States could use its existing Blue Lantern end-use monitoring system, administered by the State Department, to check compliance with these conditions. In general, a failure to meet end-use assurances or issues involving misuse or misdirection of US technology leads to a revocation of a license and a ban on future sales. Such a program would presumably allow the United States to use its market power to punish those who misuse US technology, but monitoring compliance in this way would be limited by the high levels of secrecy that many countries apply to their drone programs. It is also not clear if excluding countries from future US drone sales would be an effective deterrent, given the growing number of other exporters in the current drone market. To do so would require the United States to induce its allies who are drone exporters, such as Israel, to refuse to sell to those who misuse drone technology. Even if Israel and NATO allies like the United Kingdom were cooperative, the United States has relatively little leverage to force other less friendly drone exporters like Russia or China to do the same; on the contrary, in a world awash with drones, every instance of US restraint opens up a business opportunity for someone else.
This new policy enabled the United States to export non-lethal drones to a larger range of non-NATO allies, including Panama, Burundi, Lebanon, Egypt, and Pakistan.46 As the export policy was being shaped, the Obama administration also sought to build an international consensus on how exported drones might be used. In October 2016, with the new presidential election looming, the Obama administration issued a joint declaration on the export and use of armed drones.47 The policy, supported by forty governments, called for compliance with international law, transparency, and a voluntary code of standards for the sale and use of drones. While some powerful NATO drone users signed up to this agreement, many US rivals for the export market—Israel, China, Russia, among others—did not agree to the declaration for both financial and political reasons. At a minimum, such a code of conduct would forbid the sales of armed drones by those states to some lucrative markets and push them into direct competition with the United States for the few “approved” markets left. But some also noted that the call for voluntary standards was ironic because the Obama administration could not even confirm that its own drone usage, particularly for targeted killings, met these standards. There were few political reasons to say “yes” to a lame duck Obama administration and potential costs to incur for doing so if the Republicans won. Beyond that, this proposal seemed like slamming the stable door shut after the horses had already bolted.
The administration of President Donald Trump was initially slow to move on drone exports but eventually broke with the caution that characterized the approach of the Obama administration. In April 2018, the Trump administration announced a series of policy changes designed to loosen the restrictions on drone sales.48 It changed the process by which sales had been approved and eliminated some of the bureaucratic barriers to direct sales by companies to foreign governments. It also reclassified drones with strike-enabled technology, like laser target designators, as unarmed, thus making them easier to sell. The Trump administration also signaled a willingness to renegotiate the MCTR to exclude drones and some of their related components. Part of the administration’s objection to the MCTR was that the United States was losing not just customers but control, as its allies were increasingly buying drones from China. When this happened, the United States essentially lost the ability to influence how those drones were used.49 The goal of the Trump administration was to increase the speed and volume of US exports and allow US manufacturers to compete more effectively with Israel and China for business with non-NATO countries. But this was not an overwhelming success: slow approval processes for foreign exports inside the Department of Defense delayed the sales and the United States still found itself losing business, especially in the Middle East, to China.50
One of the most notable states that rejected the Obama administration’s call for voluntary export standards was Israel. Israel plays a complex role in the drone export game, producing and selling more than virtually any other player in the export market, but calibrating its export policy in a way to avoid a direct confrontation with the United States. For Israel, the commercial imperative to sell drones is balanced against the geopolitical imperative of maintaining its alliance with the United States. In many cases, however, the commercial imperative prevails. Israel’s powerful defense manufacturers and commercial aviation industry have pushed the government to capitalize on the advantages that the country possesses in the drone market. A number of Israeli companies, such as Israel Aerospace Industries, Elbit Systems, Rafael, and Aeronautics Defense Systems, are world leaders in drone technology R&D. Historically, Israel has been the world’s most significant drone exporter, selling 60.7% of all drones sold worldwide between 1985 and 2015. By some estimates, Israel has sent drone technology to at least fifty countries since it began exporting in the mid-1980s.51 One report on Israeli drone production and sales concludes that “the reality is that if you scratch any military drone and you will likely find Israeli technology underneath.”52
In recent years, Israeli drone exports have increased in volume. Between 2005 and 2013, Israeli drones exports have totaled $4.6 billion in sales to a wide variety of countries, including Australia, Singapore, Turkey, Thailand, Azerbaijan, Nigeria, Uganda, and Indonesia.53 According to the Stockholm International Peace Research Institute, Israel accounted for 41% of all drone exports between 2001 and 2011.54 By 2014, drones accounted for 10% of all Israeli military exports.55 Israel is projected to export $500 million per year of drone-related technology (including both anti-drone technology, sensors, etc.) with an annual increase of 5%–10% between 2015 and 2020.56 Although Israel sells to a wide range of countries, its biggest clients are countries with sizable militaries, such as India, Brazil, and the United Kingdom. Israel’s export destinations include Europe (50.2%), Asia Pacific (33.3%), South America (11.2%), and North America (3.9%).57
Israel has managed to make drones attractive to a wide range of countries with very different needs. For example, its Heron model, widely used in Gaza, has become a default surveillance drone for many states without indigenous production capabilities. One reason that Israel has been successful is that it has emphasized the production of drones which are adaptable both to a wide-range of purposes and to very specific niche needs. Such adaptability enables Israeli companies to aggressively advertise their products to a wide range of potential consumers inside and outside the military.58 To some extent, Israeli drones also benefit from US restraint. Since they are not subject to the same foreign military sales legal restrictions that the United States imposes, Israeli companies are free to sell to a greater range of countries than their US competitors.59 In general, Israel is also less concerned about selling to countries that have a long history of human rights abuses.60
Israel’s export strategy is also tailored to boosting its ability to sell. Israeli companies sometimes set up a network of subsidiary companies in countries where there is a strong preference for domestic production.61 This allows them to sidestep regulations that insist on domestic production for military assets. Israel also sometimes sells through third country suppliers.62 As a result of these moves, Israel has now become the main supplier of drones for many states in Europe, Asia, South America, and Africa. The chief limitation that Israel faces in exploiting the drone market lies in the hostility that it experiences in the Middle East. Most neighboring Arab states will not purchase from Israel, which in turn leaves an opening in that regional market that other countries like China and Russia will seek to exploit. The sales of Chinese drones to the United Arab Emirates, Jordan, and other states show how this gap in the market might be filled in the future.
While eager to exploit the drone export market, Israel is also generally careful to pay attention to US sensitivities when approving exports to Washington’s enemies and less trusted friends. In a few cases, Israel has brokered co-production or service agreements, rather than outright sales, in an effort to mollify the United States. Nevertheless, Israel has shown itself willing to break with the United States over drone sales to big markets like China and Russia. In the 1990s, Israel sold China 100 Harpy armed drones.63 When Israel considered selling an anti-radar attack drone to China in 2005, the Pentagon temporarily shut Israel out of the F-35 aircraft program.64 Similarly, in 2009, the United States objected to an Israeli sale of drones to Russia, according to documents released by WikiLeaks.65 With Russia, Israel continues to do a particularly delicate dance. It has sold drones to Russia and to its enemies, including Georgia, leading to the unusual situations of both combatants in the same war using Israeli technology.66 The Israeli Foreign Ministry has been willing to use drone sales as a bargaining chip to get Russia not to sell weapons systems to Iran, although it has to steer clear of transferring to Russia any technology jointly developed by US manufacturers for fear of violating US law and intellectual property standards.67 Israeli manufacturers have also courted controversy in other ways. In 2017, one Israeli drone manufacturer, Aeronautics Defense Systems, was accused by the Israeli government of using the Orbiter 1K drone in a kamikaze attack on Armenian soldiers in an attempt to sell Azerbaijan the technology.68
Of the newcomers to the export market, China is in the best position for selling in the years ahead. The Chinese drone program began in the early 1960s, when the Soviet Union transferred its La-17 target drone to the PLA for target practice. This drone was reverse engineered and rebranded as the Chang Kong 1 target drone. Similarly, the PLA’s recovery of a US Firebee drone provided the underlying technology for what became the Chinese Wu-Zhen 5.69 In the 1980s, China began to invest in indigenous production and to develop a manufacturing base across government, private industry, and academic centers.70 While many of China’s drones still lag in quality behind the top US and Israeli models, the scale of the investment across these sectors in China suggests that it will soon become a global competitor with the United States and Israel. A 2012 assessment from the US Defense Study Board concluded that China could “easily outmatch or outpace US spending on unmanned systems, rapidly closing the technology gaps and become a formidable global competitor in unmanned systems.”71
While many drone newcomers build relatively small or specialized drone programs, China has done the opposite, investing heavily in an array of R&D centers around the country and diversifying its production of drone models. Some of its leading universities, such as Beijing University and Nanjing University, have already yielded drone models which have been deployed by PLA units.72 China has strong existing capabilities in avionics, propulsion, and flight control systems that can be rapidly adapted to drones.73 In addition, virtually every major Chinese arms manufacturer has a research center on drones, and dozens of different models are in development.74 The Aviation Industry Corporation of China (AVIC), state-owned and one of the leading commercial aviation companies in the country, also has a number of subsidiary companies which are heavily involved in drone R&D. Chinese intelligence is also interested in hacking US companies for information on their programs.75 Many of China’s drones are designed for target practice, ground support, and reconnaissance, but China is beginning to invest in long-range drones that could extend its ability to engage in surveillance.76 China is also investing heavily in unmanned combat aerial vehicles (UCAVs) and mini- and micro-drones that might be used to overwhelm opponents. By 2020, China was expected to match overall US spending on drones.77 While China’s drones are not generally equivalent in quality to US models, they are often seen as good enough in quality and cheaper, thus drawing interest from countries that are barred from buying US-made drones.
China’s drone program is also designed for a different type of battle than much of the US fleet. Most of China’s drones are flexible platforms, used for surveillance, reconnaissance, and combat operations, and some of these are designed to counter US power projection capabilities in places like the South China Sea. A 2012 report from the Defense Science Board noted that China’s move into unmanned systems was “alarming,” especially given China’s evident interest in using drones to swarm US vessels in the event of a naval conflict.78 An estimate of China’s drone program from Project 2049 noted that unlike the United States China’s goal was to have drones that were successful in denied or contested airspace.79 China has also been pursuing small, agile, and autonomous drones that could disable radar on a naval vessel, overwhelm it, and enable other, more conventional attacks. Another option that China has explored is the use of small, disposable drones that could fly directly into targets. Along these lines, China has been investing heavily in electronic warfare packages on its drones to jam GPS and other communication devices, to disable early warning systems and some defense systems, and to confuse the detection of incoming threats to US vessels.80
As of 2012, approximately 93% of China’s produced drones were tactical in orientation.81 But China is not content to leave the export of medium- and high-altitude drones to the United States alone. China developed its own iterations of the medium-altitude Reaper model, known as the Wing Loong, which can conduct reconnaissance missions but also fire missiles at targets. It has also developed a sophisticated high-altitude stealth drone, called the Lijian (or Sharp Sword),82 a UCAV specifically designed for reconnaissance and combat operations in contested environments. Broadly similar to the US X-47B drone, the Sharp Sword has two internal bomb bays, equipped for holding a payload of 4,400 pounds of explosives and is designed to obscure its exhaust fumes to avoid detection by enemy radar.83 While the United States abandoned the X-47B due to cost concerns, China pursued the model by mimicking US technology and now has the lead in the export market for those looking for stealth combat drones.84 China has also developed the Divine Eagle drone, a highly secretive, high-altitude drone designed to provide a comprehensive picture of a specified battlespace, including tracking targets like US warships and enabling attacks through other means.85 In contrast to the low- and medium-altitude drones that China intends to sell, China keeps these high-altitude drones, designed for a quiet but deadly serious geopolitical competition with the United States, cloaked in secrecy.
This penchant for secrecy regarding drones has extended to China’s export policy as well. The desire to tap the export market has been among China’s most important motivations for rapidly catching up in drone technology. Unencumbered by the MTCR and the export restrictions faced by the United States, China sees an extraordinary opening for exporting surveillance and attack drones to governments to which the United States cannot sell. Zhang Qiaoliang, a representative of the Chengdu Aircraft Design and Research Institute, remarked that, “The United States doesn’t export many attack drones, so we’re taking advantage of that hole in the market. The main reason is the amazing demand in the market for drones after 9/11.”86
In 2011, China began offering for export the Wing Loong drone, which is capable of holding laser guided missiles and is comparable to the Predator in flight range. Unlike the Predator, it costs only $1–2 million.87 In September 2013, the China Daily reported that China was making substantial gains into the drones market, and was negotiating with at least three governments on Wing Loong sales.88 China eventually exported the Wing Loong I and II models to a range of countries, including Egypt, Saudi Arabia, Nigeria, United Arab Emirates, and Kazakhstan. In 2016, China reported making its “biggest overseas purchase order” in the sale of Wing Loong II drones. The details were kept quiet, but most reports suggested that the buyer was Saudi Arabia.89 China also allegedly struck a co-production agreement with Saudi Arabia. The sale to Saudi Arabia—rumored to be for as many as 300 drones—suggests that the Middle East offers a significant opportunity for China’s drones due to the export restrictions facing both the United States and Israel in that region. China has also sold drones to Iraq and the United Arab Emirates, which in both cases, have been used for strikes against militants.90
But China has also capitalized on selling capable drones that can be purchased by countries without deep pockets for defense spending. One of its most successful lines has been the Caihong (Rainbow) drone models. The Rainbow models, originally designed for reconnaissance and short-range surveillance, have now been equipped with air-to-ground missiles and rockets that can be used to attack targets at short range. Among the most popular is the Caihong-4 model, which can be equipped with laser-guided missiles capable of reliably finding targets and even piercing through 40 inches of armor.91 While there are persistent doubts among experts about the technological capabilities and reliability of the Caihong models, the Caihong system is widely seen as good enough by a growing number of governments. Although the Caihong-4 can only carry two missiles and remain in the air for six hours, it is sufficient for fighting enemies at close range, especially if those enemies are relatively poorly armed.92 As one expert remarked, these drones are “fast becoming the Kalashnikovs of the drone world—entry-level alternatives for countries eager to achieve a basic unmanned strike capability quickly and cheaply.”93 The Caihong drones have been allegedly purchased by Iraq, Nigeria, Pakistan, Egypt, the United Arab Emirates, and Saudi Arabia. There is already evidence that Pakistan, Iraq, and perhaps Nigeria have used them to kill militants in their territories. At the moment, the range of these drones is limited, but future iterations of the Caihong model will be capable of operating at a distance of over 4,000 miles.94 The base model Caihong drones are also significantly cheaper than comparable US or Israeli models and come with none of the concerns over end-use assurances that US exports do. For this reason, sales are swifter and less encumbered by burdensome regulations. By some estimates, China’s drones cost as little as a quarter of the comparable US models.95But the chief attraction is that China does not disclose details of its sales. In the words of Ian Easton of the Project 2049 Institute, China’s drone export strategy revolves around a three-pronged offer of “price, privacy, and product.”96 This ask-no-questions approach is allowing China to capture a significant portion of the market for small- and medium-sized drones, especially among countries that are not authorized to buy from the United States and Israel. Over time, China could be a powerful force for the global proliferation of drones, as it has little compunction about selling drones to countries with poor human rights records and can exploit the market in selling to governments in the Middle East and Africa that are forbidden from accessing US technology.97
Aside from the three main exporters, there are a number of latecomers to drone development and export that could play an important role in the future. What distinguishes the latecomers is their capacity: all have a substantial military infrastructure and robust commercial defense sector that could be used to exploit the drone market with sufficient investment. One of the most important latecomers is Russia. Long one of the world’s biggest arms exporters, Russia has lagged behind the United States, Israel, and China in the production of drones and is now racing to catch up to their capabilities. In 2013, the Russian government announced a $13 billion, multi-year effort to develop military drones by the end of the decade.98 It has also closed deals with Israel and the United Arab Emirates for the purchase of drones, while in August 2013 the Russian Defense Ministry ordered that the speed of drone development must be doubled over the next decade to capture its share of the export market.99 For years, Russia relied on Israeli-provided technology, but it has found its access to some Israeli models blocked by US pressure. It discovered to its chagrin that it lacked enough drones to dominate the battlefield in Georgia in 2008 and was determined to ensure that this did not happen again.100 But by 2015, Russia was estimated to have approximately 800 drones, with the greatest emphasis placed on small tactical drones for use by its own armed forces.101 In 2016, it deployed up to sixteen different “low, small” drones in eastern Ukraine and has used these to achieve air superiority, to identify targets, and to rattle its opponents.102 It has also used drones in Syria to spot targets, to conduct reconnaissance of the battlefield, and even disseminate propaganda.
But Russia is not content to ignore medium- and high-altitude drones or new technology. After years of relying on Israeli medium-range drone models, Russia has recently begun to spearhead the production of a new long- and medium-range attack drones similar to the US-made Reaper drone.103 Some of these models were expected to be available in the 2018–2020 time frame.104 It is also developing heavy strike UAVs, such as the Korsar and the Sukhoi’s Okhotnik UCAV, which also bears some similarities to the cancelled US X-47B.105 Russia is investing in swarming drone technology, with semi-autonomous function, with a view toward using this in future battles.106 One of its planned exports will be an exploding suicide drone produced by the Kalashnikov Group—the same company that produced the ubiquitous, eponymous rifle. As with the rifle, the aim is for the new Kalashnikov drone to be sold widely and cheaply across the world.107
Russia has a number of obstacles before it develops a strong drone export capacity. Its R&D base still lags behind that of the United States, Israel, and China, and it has not developed attack drones to match its ambitions. What the Russian industry has developed does not match the country’s needs for ISR, particularly for patrolling its vast territory. Yet once its drone technology catches up, Russia will be in a good position to take advantage of the growing global demand for drones and the gap in the Middle Eastern market. It is already working on joint projects with Iran.108 Like China, Russia is less encumbered by human rights considerations over its sales of drones and has an opportunity to sell to states (such as Iran, Syria, and others) that are not authorized for drone sales by the United States or Israel. It is not yet a leading global exporter, but if the record of Russia’s arms sales are anything to go by it is just a matter of time before it achieves this status.
Three other latecomers also have the capacity to play an important role in the export market. In Asia, both Japan and South Korea have the organizational and defense industrial capacity to become major players in the drone export market. While Japan has lagged behind China in domestic production of drones, it has invested heavily in building its domestic capacity for drones and may reach its desired fleet size even earlier than the projected date of 2023. By some estimates, it now has the fastest growing drone fleet in the world.109Using both US Global Hawks and its own models, Japan envisions a future with expanded ISR capabilities in its airspace and immediate region, including the South China Sea and the Korean peninsula.110 Similarly, South Korea is developing its own set of strategic UAVs designed to provide deterrence against North Korea and to expand its own ISR capabilities. Like Japan, South Korea is temporarily relying on US drone technology, but plans to develop an indigenous medium-altitude, long-endurance drone for ISR missions over North Korea by 2020.111 Over time, given their strong technological capacities, Japan and South Korea could become important exporters of drones. Finally, Turkey has also been exploring the export market and is now trying to sell unarmed and armed drones to Saudi Arabia and Qatar, while also exploring a joint production agreement with an Indonesian company.112
Finally, a number of European countries are possible, future drone exporters. Despite having very sophisticated, powerful defense industries, many European countries have lagged far behind in drone domestic production. While European companies are pouring millions into R&D, the hard fact is that their governments are mainly takers, rather than makers, of drones.113 As a result they are ceding the export market to others. As the European Commission noted in 2012:
[UAS] technologies are a source of important spin-off to civil aviation and a key element of the future aeronautics sector. Presently, the US and Israel dominate the sector although also other non-European countries show great potential to become strong competitors. The European aeronautics industry is still lagging behind and must quickly catch up to be able to compete on this global emerging market.114
The degree to which Europe has fallen behind in drone production means that even countries with significant defense export industries, like the United Kingdom, Germany, and Italy, have become reliant on US and Israeli drones. One estimate by the Guardian newspaper in 2015 suggested that the United Kingdom is the world’s biggest importer of (largely US) drones.115 Others, like Germany, have been importing Israeli Heron drones and US Global Hawk drones for tactical and strategic purposes.116 Under the auspices of NATO, a number of European countries are also pooling contributions to field five Global Hawk drones as part of the Alliance Ground Surveillance (AGS) system,117 designed to provide NATO allies with ISR assets. It also illustrates how dependent these countries are on US drone technology, at least until their domestic UAV industries are further developed.
There have been a number of multilateral efforts by European countries to pool resources in order to build drones which might be competitive with popular US and Israeli models. Among these is the Future Combat Air System, led by France and Germany, which aims to produce a mix of manned aircraft and UCAVs that should be able to conduct surveillance, target selection, intelligence gathering, and targeted strikes in combat situations.118 It is due to be rolled out for demonstration by 2025 and be the basis for a new concept of operations by 2030. Another joint endeavor between France, Germany, Spain, and Italy is the European MALE Remotely Piloted Air System, nicknamed the Eurodrone. This is a medium-altitude drone broadly comparable to the Reaper.119 This project, in the works with Airbus and other European companies for a decade, carries the promise of delivering a share of the global export market for medium-altitude drones to a consortium of European companies. It aims to be in the air by the mid-2020s. Finally, there is the multinational effort, led by Dassault Aviation, behind the nEUROn, a UCAV designed to be stealthy and work in contested environments. All of these models are in the testing stage, but may eventually go into production and possibly export in the coming decades.
The combination of the expansion of domestic drone production around the world, the rapid development of commercial drones that can be retrofitted for a military purpose, and the booming export market together suggest that drones will soon be in the hands of most militaries worldwide. But will this matter? For governments blessed with peaceful neighborhoods, few military rivals, and stable borders, drones are likely to remain just another useful tool in the arsenal, an efficient way to do routine tasks more cheaply or with less risk to personnel. For the majority of new users, it is unlikely that drones will alter their strategic position in any fundamental way. But there is a subset of cases in which the introduction of drones may produce a very different effect. Today, drones are being inserted into a series of active conflict zones where they have not previously been deployed. This marks an important shift. For the last two decades, drones have been used over battlefields in the undergoverned spaces of the world (for example, the tribal areas in Pakistan and Yemen, or Iraq and Syria) in which the United States and its allies were pitted against non-state actors like al Qaeda or the Islamic State. In these circumstances, the United States typically has full air superiority at high altitudes, meaning that it has destroyed enemy air forces (if they exist) and faces no real risk that its aircraft will be shot down. These conflicts can be best described as asymmetric, pitting a well-armed state against a weaker non-state actor in an environment where the drone user effectively “owned” the skies. In these conflicts, drones were not required to be nimble or to rapidly respond to threats in their environment, but could instead loiter safely to identify their targets and strike them at will. Even Israel’s drone use follows this broad asymmetric template. Israel deploys its loitering drones against Palestinian militant groups in Gaza and the West Bank for surveillance and aerial policing, and faces no real air-to-air risk from drones owned by Hamas and other militant groups. With medium-altitude drones concentrated exclusively in the hands of the United States, Israel, and a select few NATO allies and deployed in asymmetric contexts, few states have had to imagine how things might change if the principal enemy were another state with its own equally capable drones.
Today, drones are being introduced into that exact situation; they are now in play in interstate rivalries in which long-standing enemies eye each other warily and monitor each new technological advance to see if it will confer any advantage. In these circumstances, the introduction of drones may give one side an advantage or, at a minimum, sharpen the conflict between them. The degree to which drones can change the equation can be seen in the US and Chinese drone flights in the South China Sea, and in other locations as well. Both North Korea and South Korea are deploying drones along the demilitarized zone (DMZ), and India and Pakistan are flying drones over each other’s territory along the Line of Control (LoC) in Kashmir. In the war in Iraq and Syria, US and Russian drones glide through the same airspace on opposing sides, a situation that would have been considered almost unthinkable less than a decade ago. In all of these cases, governments are deploying drones in the context of what the scholar Thomas Schelling once called “dirty bargaining”: a contentious negotiating relationship in which the threat of violence is ever present.120 This in itself is not a new development. States have always intermixed threats of violence and diplomatic pressure in crisis zones, and nuclear deterrence during the Cold War often depended upon careful aerial surveillance such as is seen today with drones. But drones may alter the underlying calculation behind “dirty bargaining” because, as unmanned platforms, they so dramatically change the risks and costs involved.
This raises the question as to whether and how drones will be used in dirty bargaining. Some experts have argued that drones themselves do not fundamentally change the strategic logic of bargaining in a serious way.121 According to this line of thinking, current generation drones are a tool, like any other, and are not as effective as manned aircraft for deterrence or coercion because of their technological limits. It is certainly true that current generation drones like the Reapers or Global Hawks are poorly suited for contested airspace and would be more vulnerable than manned aircraft to being shot down. For example, unlike manned aircraft, most medium-altitude drones are not quickly adaptable and have few, if any, air-to-air defenses.122 Drones are also subject to other potential disruptions, such as jamming, hacking, or spoofing, which can lead to data links being disrupted and drones crashing or falling into an enemy’s hands. This is why one US Air Force general concluded in 2013 that, “Predators and Reapers are useless in a contested environment.”123
Since they are slower and less agile, current generation drones like the Reaper might then not be the best choice for a government hoping to deter or coerce another government in a crisis situation. Even if that were not the case, skeptics argue, it is not clear that the introduction of current generation drones into conflict zones would necessarily make crisis points more unstable. Surveillance drones may have the beneficial effect of reinforcing deterrent relationships between states by improving the flow of information and reducing the risk of miscalculation.124 If, for example, a state is able to track the movement of troops along a border and determine that the movement is a training exercise rather than an invasion, this may reduce the risk of accidental war. For example, India is deploying forty-nine hand-held mini-drones along its borders with China and Pakistan to watch troop movements in high-altitude regions.125 Such an effort, while fallible, might provide better intelligence to Indian commanders and reduce the risk of misperception and conflict. Similarly, the United States is conducting drone surveillance operations over Iran from its bases in Qatar and the United Arab Emirates, thus providing fine-grained details on Iran’s nuclear programs that may help verify its claims about the purpose of its facilities.126 The use of surveillance drones, and the quality of the information that they provide on a rival’s activities, may reduce uncertainty, enhance decision-making, and make accidental war less likely.127
On balance, more information is a good thing, and no state would want to know less about a crisis that it faces. Yet there are three reasons to be less sanguine that the improved quality of information from drones will always be a force for stability in crisis zones. First, it is not clear that more information will always be a net positive in crisis bargaining situations. Information alone is not useful unless the receiving government is able to collect and analyze the data efficiently. Even the United States has struggled on this point, with senior Pentagon officials noting that they cannot analyze all of the data that they collect from their vast drone surveillance apparatus.128 Whether a government can actually use the new information provided by drones is likely to vary based on its organizational characteristics and intelligence capabilities. Some militaries that lack the deep pockets and extensive intelligence collection and analysis capacity of the United States, China, and Israel may find that additional imagery produced by drones will not be processed quickly or efficiently enough to be useful in crisis-management scenarios. Given that most new drone users will fall into that category, it is not clear that the rapid provision of rich, detailed information will always be as beneficial to others in crisis situations as it might be to the United States.
Second, the provision of vivid imagery from drones may not always be a net positive for decision-making. Vivid information has been shown to draw some types of decision makers away from other salient or more useful information.129 In a crisis situation, vivid drone imagery of a potential incursion could exaggerate cognitive, affective, or psychological biases that decision makers already have. For example, such imagery may heighten their sense of urgency, exacerbate pathologies in the decision-making process, or truncate their deliberations by forcing a decision too quickly. In crisis bargaining situations, some decision makers will profit from the ability to visually track an enemy’s movements and reassure themselves of its non-threatening nature; others may panic and rush to decisions based on the same drone imagery.
Information can also cut in different ways depending on whether it is released into the public domain or used as leverage. The danger of vivid aerial imagery inflaming a crisis, charging public opinion, and giving a new character to routine or transactional violence was clear in a crisis around the LoC in Kashmir in 2016. In September of that year, India claimed to have conducted a series of “surgical strikes” on militants in Kashmir.130 Although Pakistan denied that this happened and claimed the strikes were an “illusion,” India essentially backed the Pakistani leadership into a corner by letting it be known that the entire operation was filmed with drones and hinting that footage could be publicly released.131 The result was an even stronger public condemnation and retaliatory actions by Pakistan, which then shot down an Indian drone on November 20.132 The prospect of drone imagery and the public reaction to it may have produced more pressure on Islamabad to respond more forcefully to an incursion or violation of deterrence than it otherwise would have and further ratcheted up tensions between the countries.
Third, it is unlikely that drones will continue forever to be able to provide the kind of information that improves decision-making in crisis situations in contested environments. This is largely because states in those situations have strong incentives to take countermeasures to knock drones out of the sky and to reduce what they can see. As drone technology improves the flow of information about what other actors are doing, there will be pressure for more aggressive (and risky) countermeasures designed to block the gaze of drones, as well as calls for more aggressive operational security about controversial activities, such as building missile sites. Aside from investing heavily in counter-drone technology, states are considering new steps that they can take to shoot drones from the sky if they are detected in their airspace. Russia has developed a new model of its Tor M2 surface-to-air missile defense system, designed deliberately to shoot down drones.133 Another response in the future might be the development of counter-UAV operations in which drones are developed specifically for the purpose of detecting and destroying other drones.134 This may lead to a dangerous sequence of UAV operations and counter-UAV strikes by adversaries, thus setting the stage for conflict spirals and dangerous accidents.
Even if a war of drone versus drone does not develop, states will not accept the widespread use of surveillance drones, and the corresponding loss of secrecy, without a response. One way that they may respond is to go further to ground to conceal their activities and to adopt stronger countermeasures to block their detection. Russia has recently fielded a Krasukha 4 radar system designed to block surveillance of ground targets and emphasized that the system was capable of blocking both Global Hawk and Reaper drones.135 China has also developed a homemade laser defense system capable of shooting down US drones.136 It is just a matter of time before states in regional conflict rivalries—North and South Korea, or India and Pakistan—also begin to invest in anti-drone technology in this way. They may also likely build even more underground nuclear and military facilities to avoid drones’ gaze. In the long run, the spread of surveillance drones—and the corresponding conclusion that one must assume everything is being watched from the skies—may paradoxically lead these states to become more opaque, not less, due to aggressive countermeasures and improved operational security over military bases and sensitive locations.
Aside from these arguments, there is a more fundamental reason to believe that crisis zones populated with drones may be more dangerous than those with manned aircraft. If those using drones are more risk-taking because a pilot is not in danger, a world full of drones might be one in which crises are more, rather than less, likely to escalate into a war. The core of this claim—that risk-taking in war grows in dangerous ways if decision makers and warfighters are insulated from its costs—has a long intellectual lineage from Immanuel Kant to the early twentieth-century anti-war movement.137 Although many forms of military technology reduce or even remove elements of risk during wartime, drones go further than most by abolishing entirely any physical risk to personnel on their own side: their uniqueness comes from their ability to “allow you to project power without projecting vulnerability,” in the words of Lt. General David Deptula.138 As David Hastings Dunn has pointed out, drones appear as a disembodied threat, which appears to “enable their use with domestic political impunity, minimal international response and low political risk and cost.”139 This does matter if the absence of risk does not affect strategic choice, but evidence from the US targeted killing program suggests that it does. Even President Barack Obama acknowledged this, saying that his use of unmanned aircraft had edged closer to being a “cure all” for terrorism and eroded some of the decision-making barriers on the use of force during his term in office.140 The central characteristic of drones—that they are unmanned and therefore wholly detached from the risks associated with losing personnel—may induce leaders and even the public to roll the dice and use force in more risk-taking ways than they might with manned aircraft.141 Although there is not enough evidence to say that drones alter strategic choice in every case, there is some early evidence that operational patterns and norms evolve when drones are put into play. An analysis by the Royal United Services Institute (RUSI) in London found that three of six drone users in the Middle East altered their operating norms, and engaged in more extra-territorial strikes, as a result of having access to the technology.142
If drones change governments’ calculation of risk and their ensuing behavior, they will begin to affect relationships of deterrence between long-standing rivals. One way that this might happen is that states begin to use drones to test the nerves and strategic commitments of their enemies. States may calculate that small drones will be able to evade detection in contested space or at the least be able to evade being shot down. Drones are particularly conducive to what Alexander L. George referred to as “salami tactics”—that is, small steps used to probe or test deterrent relationships by violating commitments in a small but measurable way.143 The crisis in the Senkaku islands in September 2013 provides an obvious example, but there are others in Kashmir and the Korean peninsula. In August 2013, India flew a drone from Kashmir into Pakistani airspace. Pakistan scrambled its aircraft in response and demanded an explanation of the violation.144 In July 2015, Pakistan shot down an Indian drone taking aerial photographs near Jammu and Kashmir, but India’s appetite for drones was undimmed, and it continued to fly even more drones near the LoC.145 India also pressed the United States to sell it Predator drones for surveillance, while Pakistan increased its efforts to boost domestic production of surveillance and armed drones.146 By 2019, Pakistan was allegedly flying its own surveillance and armed drones over the LoC in an attempt to rattle India, while at the same time seeking to buy Wing Loong II drones from China.147
The same dynamic is emerging on the Korean peninsula. In spring 2014, three crude drones allegedly from North Korea crashed in different locations in South Korea, one near the city of Peju, another close to the DMZ, and another on Baengnyeong island in the Yellow Sea.148 They were reported to have been equipped with cameras that contained images of the Blue House, the South Korea’s presidential residence. In January 2016, it was estimated that at least six North Korean drones had crashed in South Korea, and at least one of them contained images of South Korea’s nuclear reactors, presumably to allow them to be sabotaged.149 While the drones themselves were poor quality compared to those South Korea has, this carried its own advantages. As Van Jackson put it, “it’s the low performance qualities of North Korea’s drones that enable them to evade South Korea’s defenses, which are optimized for more traditional threats from bigger, faster high-altitude aircraft.”150 These efforts culminated in South Korean forces shooting at a North Korean drone over the DMZ in January 2016.151 Yet another North Korean drone escalated a missile crisis in May 2017, with South Korean forces firing ninety shots at it before it disappeared.152 If North Korea develops more capable armed drones, the tempo and nature of these tests of South Korea’s deterrence posture may also change.153 Although the details of North Korea’s drone arsenal are hard to confirm due to the regime’s secrecy, a South Korean think tank estimated that North Korea had approximately 1,000 drones which it could deploy for surveillance and for terror attacks with crude chemical and biological weapons.154
There is also some evidence that even powerful states can find their posture of deterrence challenged by weaker states with drones. In June 2019, Iran shot down a US Global Hawk drone over the Strait of Hormuz in the midst of a crisis that had been escalating since the Trump administration pulled out of the Joint Comprehensive Plan of Action—the international deal to review and contain Iran’s nuclear program. There was some precedent for Iran’s gambit, as it had hijacked an RQ-170 Sentinel drone by “spoofing” it in 2011.155 But this was different. The drone was shot down, not hijacked, the Global Hawk was a much more expensive, potent symbol of US power, and the shoot-down occurred in the middle of a crisis. Would Iran have been just as willing to test the United States by shooting down a manned aircraft, especially in the middle of a crisis? On balance, Iran has tolerated the limits of US deterrence in the Middle East and refrained from open attacks, although it has conducted a covert campaign of subversion and support for terrorism to harm US interests. But with drones, the calculation appeared to change. Iran gambled that the United States would be more willing to let them destroy a drone without a military response. This calculation was proven correct: President Trump said that he decided against a US military response to the destruction of the Global Hawk at the last minute because the response would cost 150 Iranian lives, thus making it disproportionate.156 This crude moral calculation—that one drone was not worth that many lives—ultimately rebounded to Iran’s benefit. But the entire events are marked by uncertainty: Iran had no guarantee that the Trump administration would respond as it did or that the limits of US-Iran rivalry in the Middle East would remain intact. Yet it was still willing to take a shot against a much more powerful opponent because, in the end, the target was only a drone. This interaction may have emboldened Iran even further. In July 2019, one or more Iranian drones approached the USS Boxer in the Strait of Hormuz and were disabled by the United States, with no Iranian response.157 Although this crisis appears to have been averted, there is also no guarantee that a future drone interaction between the United States and Iran—or indeed between any two other states—would remain limited in this way. Much depends on the nature of the drone incursions and even the whims and personalities of the leaders involved.
The response might also depend on whether the details around the drone incursion are clear and widely accepted, as in the drone attack on Saudi Aramco oil facilities in Abqaiq, Saudi Arabia, in September 2019. Initial reports suggested that Houthi rebels in Yemen launched a coordinated attack with at least ten drones on the oil-processing facilities as revenge for aerial bombings conducted by the Saudi military in Yemen’s civil war.158 But the United States, Saudi Arabia, and others soon doubted that the Houthis were responsible and pointed a finger at Iran, a longtime backer of the Houthis. The government of Saudi Arabia presented evidence of Iranian cruise missiles in the wreckage and suggested that some of the damage was due to missiles rather than drones.159 While the Houthis continued to claim responsibility, Iran denied any involvement in the attack and cast doubt on the evidence that Saudi Arabia presented. Although US and European leaders eventually agreed that Iran was responsible, confusion over what caused the damage and who was responsible undercut momentum for a response, at least over the short term. The lesson was that violations of deterrence with drones might be possible, even attractive, if the incursion itself can be cloaked in uncertainty.
As these incidents show, among the dangers of using drones to test deterrence is that their low cost might make such violations of deterrence more routine, even normal, and sustainable over the long run. This is in part because low-cost violations of deterrence are more politically sustainable than those in which pilots are at risk. Democracies are naturally cost sensitive, and their publics—as well as their leaders—will resort to options which minimize costs, especially for attacks that are going to be frequent. As Amy Zegart has argued:
Precisely because drones are lower-cost options to fulfill a threat, they are more likely to be initiated, sustained and supported by a domestic public. With drones, the low-cost threat becomes credible for the first time: “I can send drones at you, all day long, with no risk to me” becomes plausible, sustainable, attractive, and true. Indeed, the drone threat could become more credible than higher-cost signals to put boots on the ground. Why? Because high-cost signals may show a willingness to initiate a course of action but not to sustain it. In the long run, domestic audiences are more likely to maintain support for options that do not run risk of American lives; presidents, legislators, and adversaries should know this at the outset. 160
By making these threats to deterrence not just possible but routine, drones may begin to shake the agreed limits of previously stable deterrent relationships and make them bearable for cost-sensitive governments, effectively giving them a free hand to test deterrence far more often than they might with manned aircraft. Given how cost effective drone attacks can be—the Saudi Aramco attack knocked out 5% of the world’s oil supply at the cost of only a few drones and cruise missiles—it is not hard to see how tempting it might be for other governments to test deterrence with drone incursions.
When drone incursions happen, the responding state faces a number of important questions. Is a drone incursion a sign of serious intent to harm, or is it just a trivial foray which suggests nothing has changed? Does a drone incursion present a credible threat, or can it be dismissed? Answering these questions is crucial for deciding what kind of response is appropriate. Judging this accurately is hard because the rules of engagement for violations of deterrence with drones are unclear and may evolve in very different ways from manned aircraft.161 It is possible that there will be an implicit understanding between states that destroying drones is not a big deal, or that it is not worth paying for that destruction in human life. If that is the case, a drone can be shot down, or the responding state may respond with its own drones in a tit-for-tat exchange. But it is equally possible that states will have varying interpretations of what is and is not appropriate to do to an unmanned aircraft, producing disagreements about whether there was a violation of deterrence and what the appropriate response may be. This is particularly problematic in deterrence relationships because they rely on what Schelling referred to as a “tacit bargain”—that is, an unspoken agreement about what is and is not permissible between two actors.162 No one yet knows what the tacit bargains surrounding drones will be. The absence of an agreed interpretation of a destroyed drone—that is, whether it should be treated as an aircraft and interpreted as a major provocation or dismissed as easily as a downed weather balloon—may not be a problem in some cases, but it may suddenly become a serious one in others. Although some experts are confident that all countries appear to know that their opponents will value drones differently than manned aircraft, there is not yet enough evidence to accept that conclusion given the limited number of drone-to-drone interactions in crisis zones so far.163
Armed drones might also be used to coerce other states into changing their behavior. In theory, drones are attractive for use in acts of coercion because they radically lower the costs of coercion for the attacker, effectively allowing power projection at lower levels of risk.164 Because drones are also generally cheap, drone coercion could be sustained for a long time, especially against a comparatively weaker enemy.165 As Zegart has argued, armed drones can keep the weaker party trapped in a “constant state of ambush,” fearing attack if they defy the stronger party.166 Especially in cases of stark asymmetry, drones can increase the certainty of punishment for the weaker party and shift more vulnerability onto them, as the coercer can just keep promising that more drones will keep coming at them until they comply. Following this logic, it is possible to imagine a scenario in which governments would threaten their enemies with repeated drone attacks in an attempt to coerce them over an issue of importance. A government might threaten important military assets and personnel, or even other heads of government and high-ranking officials, to get them to change their behavior. In a different form, coercion is what the United States has been doing with non-state actors like al Qaeda and the Islamic State. This model—called “targeted hurting”—may be the way in which coercion by drone occurs in the future.167
Understanding whether coercion would actually work with drones—an often cheap, risk-limiting technology which presumably conveys a weaker signal of resolve—is hard to do at present. It is worth noting that the United States has only been able to coerce with drones given clear air superiority and a much weaker non-state opponent. Coercion might not work the same way between rival states. For example, if North Korea threatened repeated drone attacks against South Korea to coerce it for some purpose, it is not clear that South Korea would be sufficiently alarmed to do what was demanded. It might choose to simply bear the cost of North Korea’s attacks, to respond in kind with drones, or even to develop its anti-drone defenses in order to nullify the potential threat. There is not enough evidence of state practice with drones to know definitively how coercion would operate when there is a rough strategic balance between the two government rivals. Because current generation drones are largely designed for uncontested airspace, few governments in this situation have used them for coercion for fear that they would be quickly shot down. But as drones become more capable of flying in contested environments and using stealth to cloak their activities, more governments will have an incentive to take the risk of using drones for coercion, if only to see how effective such a low-cost tool would be at changing an enemy’s behavior.168 This incentive is likely to increase as next generation UCAVs, capable of moving fast, using stealth, and striking deep into an enemy’s territory despite air defenses, come into the hands of more governments worldwide.169
How coercion would play out when one state attacks another to force it to scale back its support of non-state actors is also an open question. For example, Pakistan has long supported militant groups in Kashmir to contest Indian control over the territory. India has responded with strikes against militants in Kashmir, but both India and Pakistan have observed tacit bargains in their responses to ensure that the situation in Kashmir does not generate an irreversible momentum for war. Drones, however, may be changing this calculation. In 2018, one Indian general announced that, “the Indian Army is capable of using drones to attack hostile targets inside Jammu and Kashmir and across the Line of Control, and sees no problem in using them provided the nation is willing to accept mistakes and collateral damage.” The chief concern was not whether Pakistan would respond but whether the “international community [would] get after us” over civilian casualties from the strikes.170 This suggests that at least some in India’s military believe that the risks associated with a coercive drone strike against Pakistani-backed militants across the LoC are manageable and that restraint would prevail in Islamabad. But it is not clear that the Pakistani government would show such restraint in the face of an armed drone strike from India. Would Pakistan simply accept it, perhaps with some protest as it does with the United States, or would the fact that its bitter regional rival India launched the attack change its calculations? Pakistan might calculate that the reputational costs of being coerced by a drone attack from India are high and might respond with equal or greater force. This calculation might also change depending on whether the strikes are a one-off event or a semi-regular occurrence. Drone coercion is a seductive idea, but the chief contrast surrounding it—that it is so costless for the attacker, but implies so much about the weakness of the attacked if they give in—suggests that the underlying logic might be different than it is with other forms of coercion.
The biggest problem with using drones for coercion concerns credibility. It is not self-evident that drones will convey the same kind of message that manned aircraft do. The absence of risk to a pilot suggests that their use may be seen as a less credible threat by opponents and hence less useful to states as a signal than a manned aircraft. For coercion to work, a state must be able to send a message that it is serious about the issue and willing to pay the costs to get an enemy to change its behavior. If all a state is willing to send is a drone, it may not be taken seriously. Moreover, if routine incursions by drones become almost normalized between enemies over time, it will undercut the urgency of a threat and render achieving goals through coercion more difficult.171 The result of making drone incursions routine is that the underlying calculations behind coercion—for example, how many drones does one send to change an enemy’s behavior, and what response will follow—will be harder to make. If governments decide to try coercion with drones, they may need to be prepared to escalate the level of threat beyond what would be ordinarily needed with manned aircraft in order to convey their resolve. It may not be sufficient to send a single drone; instead, a government might have to send dozens to make it look like a real threat. Because drones are detached from risk to pilots, their natural lack of credibility as a threat may have the perverse effect of making governments more likely to ramp up the intensity or scale of the threatened violence.
A final reason to be concerned about the growing drone arms race is the danger of accidents and the conflict spirals that can subsequently arise. While drones are becoming more sophisticated, they are still prone to frequent accidents. According to an estimate in 2010, the United States has experienced at least seventy-nine drone accidents costing at least $1 million each, as well thirty-eight Predator and Reaper drone crashes during combat missions in Afghanistan and Iraq.172 Unmanned platforms such as the Pioneer and Shadow drones have even higher rates of accidents.173 A later estimate in 2014 put the total number of major drone crashes at over 400 since 2001.174 Although it is estimated that some of these accidents are caused by human error, and that the accident rate is declining for military aircraft, these rates are still higher than comparable manned aircraft.175 Among the problems noted in military drone crashes were a drone’s limited ability to detect other aircraft or bad weather, pilot error, mechanical defects, and unreliable communication links.176 It is also possible that less sophisticated models sold by China and other new suppliers will have a higher rate of accident than the more robust US models. Simply as a matter of probability, drone accidents will become more commonplace as more military drones take to the skies in the future.
Some military drone accidents may have political costs that are hard to estimate, especially if they cause a substantial loss of life or are interpreted as intentional. The possibility of a military drone colliding with a civilian airliner, while improbable, is not impossible. In 2004, a German UAV nearly crashed into an Ariana Airlines Airbus A300 carrying 100 people in the skies over Kabul.177 Over the last fifteen years, drones have been equipped with anti-collision software designed to avert such crashes, but dangers remain. One estimate in 2012 found that at least seven US Predator or Reaper drones have crashed overseas in the vicinity of civilian airports.178 In September 2013, the United States was forced to move its drone operations from Camp Lemonnier in Djibouti due to concerns that drones would crash into passenger planes from a nearby airport.179 Whether a state views an accident as negligent or intentional will matter. If a US drone struck an Iranian passenger airliner, for example, it is not hard to imagine the incident causing a serious international crisis, along the same lines as following the accidental US downing of Iran Air Flight 655 in 1988. The risk of a conflict spiral from a drone accident between India and Pakistan, or Israel and one of its neighbors, should not be ignored. Similarly, a collision between a Chinese drone and a Japanese civilian aircraft in the East China Sea could produce disastrous consequences. It is further possible that governments or insurgents will try to hijack drones and cause international incidents with crashes or accidents. Hijacked drones are a particularly attractive way to test an enemy as they are hard to trace and can shield the perpetrator with a degree of plausible deniability. As drones wind up in the hands of more unscrupulous actors, and less reliable drones sold by China and others flood the market, governments around the world will face a vastly increased risk of a conflict spiral from drone misuse, hijacking, or collision with a civilian aircraft.
The world will soon be full of drones. Given the explosive growth in the government-led, commercial, and export markets, it is impossible to stop the spread of drone technology or to convince states to not use them. For most countries, there is no reason to do so, especially if drones are only going to be put to routine purposes. But for a select number of states, embedded in deeply contested rivalries, the introduction of drones will have important strategic consequences. From the Strait of Hormuz to the South China Sea, we are moving into an age in which drones may tilt the strategic balance between adversaries in some of today’s most dangerous flash points. In these, drones will not be a mere substitute for manned aircraft, replicating the same calculations of risk and opportunity that have governed how these crises unfolded in the past. Because they are unmanned, drones are fundamentally different and the strategic calculations around their use will be marked by uncertainty rather than the tacit bargains that have prevailed for decades.
At present, we do not know how this will turn out because only a few years have passed since many of these states have acquired drones. But there is evidence that unmanned platforms will make governments become more risk-taking and expand their ambitions in ways to test their enemies’ mettle. This is potentially dangerous. As a technology that promises to make violence antiseptic and carefully controlled, drones may convince governments that a potential violation of deterrence is negligible, or that it can be managed. Alternatively, it could convince them that coercion will be easier or more sustainable than it actually is. It may also convince them to be careless with the risks around drone use and to proceed while unaware of how conflict spirals could emerge even following simple accidents and other misunderstandings. All of this flows from the fact that the unmanned nature of drones detaches risk from the act of flying in another’s airspace and inflates the confidence of users to set new, perhaps ambitious goals. This is why it is so crucial for drones to operate with some code of conduct and set of norms, as the Obama administration envisioned, rather than let the technology proliferate without clear legal and moral understanding of how it may be used. Such an effort would be imperfect and marked by defections, but the alternative—a world of drones with no rules—might be worse. In that world, we may find that the biggest strategic effect of drones would be to exaggerate the hubris of the governments who have them and to make the path to war different but no less likely.