Beijing is unavailable to take your call: Why the US-China crisis hotline doesn’t work

During the short thaw of US-China relations over the last six months, the Biden administration has emphasized the importance of crisis communication channels or “hotlines” for stabilizing the relationship between the two powers. After the Biden-Xi Woodside Summit in November 2023, the administration touted an agreement to reopen military-to-military communications—though progress on implementing the agreement has been slow. Ahead of a follow-on call between the two leaders this April, a senior US official explained that “President Biden has made clear that this mil-mil [military-to-military] communication is critical at all times but especially during times of heightened tensions.”

Unlike the United States and Russia, who established the first hotline in 1963 after the near-catastrophe of the Cuban Missile Crisis, the United States and the People’s Republic of China have a much shorter history of crisis communications. The Beijing-Washington hotline was only established in 1998, during President Clinton’s visit to China, followed by the opening of the military-to-military Defense Telephone Link (DTL) in 2008. Since then, US-China hotline relations have been fraught; during several crises between the two countries, Beijing simply has not answered, leading many commentators to call the military hotline “dangerously broken.”

Current hotlines are not up to the task of preventing escalation, in part because the mechanisms are far too slow. The term “hotline” itself may be a misnomer for some of these communication systems. Worse, no one can be confident that the current systems would continue to function in wartime as they are not designed to survive either a direct attack or the indirect effects of a major war. To avoid the possibility that a future crisis spirals into outright conflict and that such a conflict spirals further into an all-out thermonuclear war, the United States and China need to cooperate to establish shared norms and understanding, set up a dedicated channel for faster communications, coordinate the mutual hardening of their communications systems, and take bilateral steps to ensure that hotlines would work when military and political leaders need them most.

History of hotlines. During the Cuban Missile Crisis, as the United States and the Soviet Union nearly crossed the nuclear threshold several times, leaders realized that the atomic age demanded faster and more reliable communication. Soviet Ambassador Anatoly Dobrynin later recalled an attempt to send an urgent message to American counterparts: “We at the embassy could only pray that [the messenger] would take it to the Western Union office without delay and not stop to chat on the way with some girl.” James Acton, co-director of the Nuclear Policy Program at the Carnegie Endowment for International Peace, told me that when Soviet leader Nikita Khrushchev finally decided to back down, he figured radio was the most reliable means of quickly getting the message across to the Americans. But when the messenger rushed to the studio, the elevator got stuck between floors; he had to slip the message through a slit in the elevator’s doors one page at a time so it could be broadcast to the world and thereby to Kennedy.

Something needed to change. Building on ideas by the strategist Thomas Schelling and the advocacy of private citizens, the Cold War superpowers signed a hotline agreement the year after the crisis. Later, this text-based Direct Communications Link (DCL) was upgraded several times because of successive problems with the original cable-based hotline—for example, a farmer in Finland once accidentally severed a hotline cable while plowing his fields and a freighter in Denmark accidentally cut an undersea cable when it ran aground. (Direct communications now run via satellite.) Although it is difficult to establish the extent to which they helped preserve peace, hotlines were used during multiple crises and have been credited with defusing superpower tensions during the Cold War.

Building on this success, the United States initially proposed a hotline to China in 1971, but did not hear back from China’s leadership until after the Cold War. It was only in 1998 that the two countries finally established a presidential hotline and added the military-to-military DTL in 2008. In 2014, President Barack Obama and Xi Jinping signed a memorandum of understanding on notification of major military activities, followed in 2015 by an annex on military crisis notification mechanism. In 2020, the US Defense Department and China’s People’s Liberation Army held a joint working group for crisis communications. Despite new agreements and upgrades, however, China often simply won’t answer in times of crisis. In the words of deputy secretary of state Kurt Campbell, “the hotlines that have been set up have just rung, kind of endlessly, in empty rooms.”

For example, after the accidental bombing by the United States of the Chinese embassy in Belgrade in 1999, President Bill Clinton attempted to contact the Chinese leadership to apologize, but China did not take the call. Then again in 2001, after two US and Chinese aircraft collided mid-air over Hainan Island, US officials frantically tried to reach their Chinese counterparts to avoid an escalating crisis, but China did not answer for over 12 hours. The same problem has continued up until the 2023 “balloon incident” when China once again did not answer over hotlines.

Why China isn’t answering. To explain China’s apparent unwillingness to answer hotlines during crises, some analysts have suggested the lack of a “modern Cuban Missile Crisis” between the two countries. The original Moscow-Washington hotline was born in the aftermath of this crisis, and the United States and Russia share a long history of managing subsequent crises, participating in confidence-building measures, and engaging in arms control talks even when tensions have been high.

Of course, Chinese policymakers and scholars also study the Cuban Missile Crisis. “You do have a number of Chinese strategists [who] have read the literature about the Cuban Missile Crisis and do see value in those mechanisms [of confidence-building],” David Santoro, a nuclear strategy expert who convened several US-China track-2 (unofficial) diplomatic talks on strategic nuclear issues, explained to me. For example, Chinese senior colonel Zhou Bo has argued that “well-established hotlines for emergency communication” helped the United States and the Soviet Union avoid nuclear war during the Cold War. Similarly, The Science of Military Strategy, a key Chinese strategic document translated by US Air Force’s China Aerospace Studies Institute, highlights the benefits of “simple and efficient crisis communication and control channels.”

Beyond the available literature, the Chinese party-state may lack the institutional memory of just how uncontrollable military crises can become. This may lead to overconfidence in the possibility of “crisis prevention”—a term many Chinese strategists favor over “crisis management.” Tong Zhao, a senior fellow at the Carnegie Endowment for International Peace and an expert of Chinese security policy, explains that because of the lack of the lived experience of a nuclear crisis, Chinese political leaders are relatively more confident than their US or Russian counterparts “about their capacity to maintain good control of the situation on the battlefield.” Chinese officials are therefore “less appreciative of the risks of the fog of war, and things getting out of control because of unforeseeable developments on the front line,” he added.

Lyle Morris, a senior fellow with the Asia Society Policy Institute and the country director for China in the US Defense Department from 2019 to 2021, saw firsthand the challenges of crisis communications with China. Fundamentally, China views crisis communications “through a political lens” and as a tool “to come out on top of a conflict,” Morris told me. Zhao expressed a similar view: In traditional Chinese military thinking, “not to respond to the other side’s request to communicate at a time of crisis is a useful tool to impose pressure on the other side, to express one’s displeasure, and to compel the other side to change its own behavior.”

Chinese strategic documents seem to support this view by encouraging the use of “various means to control and guide military crises to develop in a direction that is beneficial to them.” Chinese officials may selectively answer or ignore messages from their adversaries to play the crisis on their own terms. Similarly, according to Acton, “one could imagine the Chinese are worried that if hotlines exist and they’re used regularly, the United States may behave more aggressively because it thinks it can get out of crises without escalation.” Zhao, too, explained that some Chinese leaders blame the United States as the sole source of tensions in the region and worry that functioning hotlines “could actually embolden perceived American military provocations,” especially “close range surveillance activities near China.”

This creates uncertainty in the minds of many US officials on whether Chinese leaders would take calls in a crisis. But some experts are more hopeful that higher stakes could change the calculus. For Acton, “the Chinese actually would answer the phone in a more intense crisis.” Similarly, Zhao believes that “if the crisis has reached the level that makes Chinese leadership really concerned about the consequences of miscommunication, yes, I think Beijing thinks it has the option to communicate with the other side more directly and more timely.”

And indeed, China did use the DTL when tensions were high.

During the summer of 2020, some officials in China’s bureaucracy and military apparently genuinely believed that the United States would attack China in the South China Sea as a kind of “October surprise”—referring to both how the Cuban Missile Crisis started one morning of October 1962 and how a news event in October can influence the outcome of US presidential elections held in November. Santoro told me that Chinese officials “thought [President] Donald Trump would create a crisis” and that they used the DTL and other channels “to get reassurance from the United States that they would not create a crisis.” Concerns raised via track-2 backchannels made their way up to US defense leadership, and two top generals, Mark Milley and Li Zuocheng, spoke twice via the DTL about the issue. According to Washington Post’s Bob Woodward and Robert Costa, who interviewed Milley for their book “Peril,” the US general used the channel to dispel these fears. “General Li, you and I have known each other for now five years. If we’re going to attack, I’m going to call you ahead of time. It’s not going to be a surprise,” Milley remembers telling his Chinese counterpart in October 2020.

Bureaucratic obstacles. Beyond Chinese attitudes, a more practical reason as to why hotlines don’t work with China may have to do with the structure of the communications links themselves.

First, hotline problems are in part the result of bureaucratic and organizational obstacles in the Chinese party-state. Sometimes those political officials who are “authorized” to communicate via the DTL do not have clear visibility into the movements of Chinese military forces, and those military officials who have knowledge of the situation on the battlefield are not authorized to speak. Despite attempts to coordinate national security decision-making, the China’s People’s Liberation Army still needs to get the Communist Party’s approval on the content of crisis communications messages. And the centralization of the Chinese party-state under Xi and ongoing purges of the military likely worsen these problems. As Zhao explained to me, “[Chinese] military officials feel even less comfortable picking up the phone and talking with their American counterparts without first consulting the political leadership and receiving explicit approval and guidance.”

However, organizational hurdles to crisis communications exist not just on the Chinese side.

US-China crisis communications channels have structural features that simply make them unsuitable as “true” hotlines. According to Morris, “mechanically, there is a misperception out there, about there being a direct line” between the two countries that is analogous to the DCL between the United States and Russia. But current US-China military-to-military links may be structured closer to a game of “phone tag” in which the two sides have to communicate back and forth asynchronously before even getting on the phone.

When someone at the Pentagon wants to communicate with their Chinese counterparts about an incident, they first must use the DTL to let those counterparts know “there’s an incident and we’d like to talk,” Morris told me. This initial contact can include a variety of formal questions, such as “Which level of official do you expect to talk to?” or “What’s the alternative official that you expect to talk with if your primary official is not available?” Zhao added. Second, the Chinese side must send a message to agree to discuss the incident. Third, both parties must schedule a time to talk. Then, finally, the parties can discuss the crisis—though, of course, not without the Chinese military leadership first running everything by the party bureaucracy.

Unfortunately, no part of this process is fast enough for rapid de-escalation in a crisis. The fact that the two countries technically have to give each other 48 hours to respond to a request for a call is no match with how quickly a nuclear crisis could unfurl—in mere minutes. Morris explained that the same onerous process applies to the Beijing-Washington presidential hotline. Even though these links are generously called “hotlines,” in practice, they are more similar to communication through the slower and more routine National and Nuclear Risk Reduction Center—a communications center for exchanging notifications about arms control and missile launches—than to a true crisis hotline like the Moscow-Washington hotline.

Hotline survivability. In a major crisis, even if US and Chinese military leaders were to somehow manage to establish communication, the channels themselves may fail if war were to break out.

Over the years, some defense analysts have pointed out the glaring omission of state-to-state communication systems in the effort to harden nuclear command, control, and communications infrastructure. In 1987, Ash Carter, then a Harvard professor, explained that “some thought needs to be given to communications between the superpowers, since terminating a nuclear war before it escalates to all-out exchanges is a goal of US strategy.” Carter concluded that, “[f]or the existing Washington-Moscow hotline to have a chance at functioning in a nuclear war, both sides would have to withhold attack on national capitals.” Others have publicly agreed that the hotline “is not a survivable link and will not be available for general nuclear war termination communications” and that “despite the system’s critical importance for war termination, no special measures have been taken to protect it from the collateral effects of a nuclear exchange.”

Of course, a hotline’s most valuable function resides in its availability, even—or especially—after hostilities have already broken out.

Chief commanders and officials need to be able to talk to each other in order to limit a war and find routes to quickly end the war and negotiate a peace—before the war escalates further to a suicidal all-out thermonuclear exchange. As Acton put it, “perhaps the even more important application for hotlines [is] being able to communicate and try and find a formula for ending the hostilities.”

At least some elements of hotline infrastructure may be able to survive a war. For example, DCL terminals are known to exist not only at the Pentagon and the White House, but also at the National Military Command Center located inside the Raven Rock Mountain Complex, a massive underground nuclear shelter in Pennsylvania from which surviving elements of the US government would conduct a nuclear war. Experts, however, still believe that the government is neglecting hotline survivability. A 2020 report by the California-based Institute for Security and Technology stated that “adversaries may not have a trustworthy means to communicate to avoid nuclear cataclysm.” The institute’s chief executive officer, Philip Reiner, who served on the National Security Council and in the Defense Department, explained that nuclear command, control, and communications survivability has predominantly been first to “make sure that the EAM [the Emergency Action Message that directs nuclear attack] gets through.” Communicating with an adversary, on the other hand, is “always an afterthought.”

The problem of hotline survivability—a highly technical topic that can fall into the bureaucratic cracks between defense and diplomacy—might not even cross the minds of most decision-makers. For Acton, “it would not for one second occur to national leaders to worry about the fragility of the hotline.” When asked about the survivability of the US-China DTL, Morris told me that this question “quickly gets into classified, sensitive information,” adding that after a nuclear detonation, state-to-state communications may become extremely challenging. After this point, he said, “you really can’t say with certainty that we would be able to talk. All bets are off.”

Despite the lack of publicly-available information on nuclear communications infrastructure, there are reasons to suspect current hotlines would not survive long in a conflict.

First, parties to a war may target each other’s military communications infrastructure—on the ground, under sea, and over the electromagnetic spectrum—to gain a tactical advantage on the battlefield. But in doing so, they might disable hotlines that serve for strategic stability and that rely on the same infrastructure or simply happen to be nearby. Reiner explained that in the first phase of a conflict, China would likely “knock out as much of [US] communications capability as possible.” Second, the effects of nuclear weapons (like electromagnetic pulse damage) could outright disable communications, especially for high-bandwidth communications like the videoconference capability that the DTL currently relies on. (A text-only channel may be easier to protect against such damage.). Third, war between the United States and China or Russia would likely include space operations, which might either directly disable or destroy satellites necessary for hotlines or create massive amounts of space debris that could also disable the required infrastructure.

Moreover, even if the US side were fully hardened, it takes two to talk.

For Zhao, hotline survivability does not seem to be much in the minds of Chinese leadership either: “China has been paying much more attention to maintaining effective communication and command and control of its military forces, rather than maintaining effective diplomatic communication channels. I will not rule out the possibility that China has indeed not systematically thought through the maintenance of crisis communication with enemies in a serious major conflict.”

No expert I spoke to was able to confirm whether hotlines would work even in a non-nuclear war or a limited nuclear war. Many believed the system would fail at this critical moment for civilization—after hostilities have broken out but before a conflict has escalated into a total thermonuclear war and therefore could potentially still be avoided. To negotiate an end to a war, parties need a way to communicate their intentions—one that ideally does not involve more bombs. But more likely than not, hotlines may fail just when we need them most. In fact, based on all publicly available evidence, the odds that hotlines would fail should a major war break out might well be greater than 50 percent.

Cooperate to communicate. Fortunately, the problems with hotlines are all tractable, and leaders of both sides have signaled their willingness to discuss crisis communications. In May, for example, Zhou Bo, now retired from the Chinese military, argued that “trust but talk” should be the motto of the US-China relationship—just as “trust but verify” was the motto of the US-Soviet relationship during the Cold War—and pointed to the importance of resuming military-to-military communications.

Policymakers need to acknowledge that the issue is not merely one of cultural differences or strategic behavior; the problem involves the very organization of the US-China crisis communications infrastructure itself, with its 48-hour response allowance and arcane bureaucratic procedures. The simple commitment to “reopen” existing channels is not enough. The overarching problem with hotlines needs practical solutions.

First, setting up a true US-China hotline modeled after the US-Russia DCL, and designing it for speed, would solve half of the issue (hotline slowness). Like the US-Russia DCL, a true hotline could be text-based, and it could be limited to nuclear issues (or it could also include other high-consequence technologies, like artificial intelligence).

An improved system should also include a mechanism for acknowledging receipt of information without having to formally issue a reply—a suggestion some analysts already made. As Morris explained to me, “just being able to have the [Chinese military] say ‘confirm receipt,’ is very important.” There may be situations when even knowing unilateral technical information has made it to the other side can help reduce tensions. Such information could include messages designed to reduce warhead ambiguity, such as whether a given weapons system is armed with a conventional or a nuclear explosive.

The second half of the issue (hotline survivability) may be more challenging to address, as it involves many sensitive questions about nuclear weapons effects, nuclear command, control and communications resilience, continuity of government planning, operations in space and cyberspace, and more. But working with hotlines already requires a certain level of trust. As Morris explained to me, there is a sealed Chinese encryption box in the Pentagon and a US encryption device on the other end of the DTL in China, which serve to ensure the security of communications.

Conducting discussions of hotline survivability would not necessarily require disclosure of sensitive information. For example, agreeing on a text-based system (rather than videoconference) would facilitate low-bandwidth communications, if a war or natural disaster degrades communications infrastructure. Reaching an agreement on hotline survivability does not need to include technical discussions of the related disabling events. Similarly, each side could agree that it would be prudent to unilaterally harden their terminals, clearly mark hotline-related infrastructure, and avoid targeting certain communications systems even in a war—a step that nuclear arms control advocates have recommended since the Cold War.

However difficult, bilateral discussions on these issues between the United States and China would also function as a confidence-building tool and help both sides better understand the causes of their disagreements. In the latest round of US-China track-2 dialogues earlier this spring, Santoro recalls witnessing a slight change in tone among some Chinese participants compared to previous sessions: Chinese strategists now “do see value in crisis management. They do think it needs to happen. They do want better channels.”

Editor’s note: In his capacity as fund manager of the Global Catastrophic Risks Fund, the author recommended a grant to CEIP for a new project “Averting Armageddon,” which will fund work by James Acton and Tong Zhao (among others), including on hotline survivability, and a grant to Pacific Forum for the track-2 dialogues that David Santoro organizes.