Abstract

Many strategists argue that to deter a nuclear attack, states must be certain of their ability to retaliate after a nuclear first strike. China's nuclear posture of uncertain retaliation suggests an alternative logic. Given the catastrophic consequences of a nuclear attack, uncertain retaliation can have a strong deterrent effect, and assured retaliation is not necessary. A simplified nuclear exchange model developed to evaluate China's nuclear retaliatory capabilities against the Soviet Union in 1984 and the United States in 2000 and 2010 shows that China's nuclear retaliatory capability has been and remains far from assured. In its 2010 Nuclear Posture Review Report, the United States promised to maintain strategic stability with China; therefore, the 2010 scenario can be considered as a baseline for China-U.S. strategic stability. Both China and the United States are developing or modernizing their strategic offensive and defensive weapons. The technical competition between China and the United States favors each in different ways. A hypothetical scenario of China versus the United States in 2025 reveals that China-U.S. strategic stability will likely be maintained at no lower than its 2010 level.

Introduction

What levels of cost and risk are required to deter a nuclear attack? Traditional strategists argue that states need to be certain of their ability to retaliate after a first strike, but they disagree on how much damage inflicted by a retaliatory strike is unacceptable.1 The two nuclear superpowers, the United States and Russia, embrace the concept of “assured destruction” by maintaining the capability to destroy the other's society.2 Medium-sized nuclear states, such as France and the United Kingdom, believe that being able to inflict a much lower level of damage is enough.3 This posture of minimum deterrence requires a smaller nuclear arsenal than that required for assured destruction. China's nuclear practice suggests an alternative logic, however; unlike its economy and conventional military, which have been growing exponentially since the late 1970s, China's nuclear program has grown much more slowly. Instead, China has for several decades maintained a small-scale nuclear arsenal that it keeps off alert in peacetime.

Scholars differ in their assessments of China's nuclear deterrent capability. Some argue that its nuclear posture can be labeled as “minimum deterrence” or “assured retaliation,” whereby Chinese nuclear retaliation in response to a nuclear first strike would be guaranteed.4 Others contend that China's nuclear deterrence is based on so-called first-strike uncertainty—that is, creating uncertainty on the other side about whether its first strike would destroy all of China's nuclear weapons.5 Still others recognize the existence of first-strike uncertainty, but believe that it is an insufficient deterrent and that “secure second strike forces have been more difficult to generate than most experts acknowledge.”6

Some scholars have sought to quantitatively evaluate China's nuclear retaliatory capabilities. Charles Glaser and Steve Fetter, for example, argue that ten or more Chinese nuclear warheads would survive a U.S. nuclear first strike and be able to penetrate the United States’ ballistic missile defense (BMD) system. Glaser and Fetter, however, consider only “politically plausible scenarios—the United States attacks once a crisis has become severe or during a conventional war, not by surprise during peacetime.”7 A RAND report released in 2015 modeled counterforce nuclear attacks between China and the United States and concluded, without providing further details of the model, that China's confidence in its secure second-strike capabilities was low before 2010 and medium in 2017. The alert level of China's nuclear forces used in the research for the report was higher than what China maintains in peacetime.8

Academia still lacks the tools, however, to evaluate Chinese nuclear retaliatory capabilities in a meaningful way for two reasons. First, no quantitative analysis of China's nuclear deterrence is publicly available; extant works are either qualitative or based on classified models. Second, the current literature focuses on higher alert levels than what China maintains in peacetime. Although politically plausible, this methodology does not compare China's nuclear posture with that of other nuclear states, and it de-emphasizes the key characteristic of China's nuclear posture—its off-alert status in peacetime.

This study develops a nuclear exchange model involving China and its two main adversaries. Based on this model, it is possible to quantitatively simulate China's nuclear retaliatory capabilities—specifically, the probability that China would be able to launch a successful retaliatory strike after absorbing a first strike from the Soviet Union in 1984 and from the United States in 2000 and 2010.9 Two alert levels of Chinese nuclear forces are considered: day-to-day alert status in peacetime and full-alert status in wartime. The impact of U.S. technological advancements in counterforce and missile defense on China's nuclear deterrent capability as well as a hypothetical scenario—China versus the United States in 2025—are also discussed.

The results of the modeling suggest that China's nuclear retaliatory capability has been and remains far from assured. In 1984, the probability of successful Chinese nuclear retaliation against the European part of the Soviet Union (i.e., Soviet territory west of the Urals) was 38 percent for day-to-day alert status; if Chinese nuclear forces were on full alert, the probability would increase to 64 percent. In 2000, however, the only type of Chinese weapon that could reach the continental United States was the vulnerable, silo-based, intercontinental ballistic missile (ICBM). Therefore, in the face of the United States’ superb counterforce capabilities, China's nuclear deterrent capability was almost neutralized. In 2010, China's deployment of road-mobile ICBMs significantly improved the survivability of its nuclear forces. As a result, the probability of successful Chinese nuclear retaliation against the continental United States was 38 percent for day-to-day alert status, and 90 percent for full alert status.

This study shows that the technological competition between China and the United States favors each in different ways. For example, the United States is able to detect China's fixed nuclear facilities and pursue anti-submarine warfare (ASW). China, meanwhile, has the advantage in being able to hide its mobile missiles. The outcome of the competition in missile defense and penetration remains uncertain. The 2010 China-U.S. scenario can be considered a baseline for China-U.S. strategic stability. Given China's high economic and technological capacity, its level of nuclear deterrence in 2025 will likely be no lower than its 2010 level.

This study also sheds light on the nuclear dynamics between China and the United States. While China is concerned with the development of U.S. BMD, the United States maintains that its BMD system is designed to protect against countries with small nuclear forces such as North Korea and Iran, not China. The model shows, however, that the number of Chinese warheads that would survive a U.S. first strike is so limited that even a small-scale missile defense has enough interceptors to support a four-interceptor salvo firing doctrine. Therefore, the key focus in the missile defense debate should be on target discrimination—that is, the capability to distinguish real warheads from decoys —rather than on the number of interceptors. At the same time, China's modernization and gradual expansion of its nuclear arsenal have raised U.S. concerns. The model in this study shows that the survivability of nuclear delivery platforms is more consequential than the number of nuclear warheads. To avoid an arms race, China and the United States need to restrain the development of their strategic capabilities.

This article is divided into five sections. The first section describes the study's methodology and assumptions of the nuclear exchange model. The second section presents the results of the three modeling scenarios. The impact of future technological advancements and a hypothetical scenario of China versus the United States in 2025 are examined in the third section. The fourth section discusses the theoretical implications of the study, and the conclusion offers policy recommendations for maintaining China-U.S. strategic stability.

Modeling Chinese Nuclear Retaliation

The purpose of this study was to develop a simplified nuclear exchange model and to evaluate the nuclear retaliatory capabilities of China after it has absorbed a first nuclear strike from a major adversary (i.e., the Soviet Union or the United States). This approach should be understood as a way to assess the structural stability of the strategic nuclear relationship between China and its major adversaries; it does not imply that either Moscow or Washington has a first-strike doctrine.

The nuclear exchange model is a tool of strategic analysis developed during the Cold War to simulate “the result of strategic warfare in a 2-player game in which nuclear weapons only are employed.”10 The model can be highly detailed or aggregated.11 The model presented in this article is aggregated. Generally speaking, modeling a nuclear exchange requires knowledge of both sides’ targets, number and capabilities of nuclear weapons, firing doctrine, and formulas for estimating damage.12 Given their nuclear superiority, the Soviet Union and the United States are assumed to allocate their best available weapons against China; both have adequate weapons to target Chinese nuclear assets. The damage resulting from China's retaliatory strikes is not simulated; only the number and corresponding probability of Chinese nuclear retaliatory warheads are calculated. The model is based on publicly available data.

Three scenarios are discussed: China versus the Soviet Union in 1984, China versus the United States in 2000, and China versus the United States in 2010. All three represent turning points in China's nuclear history. In 1984, China's nuclear forces entered combat duty, indicating that Chinese leaders had become confident of their country's nuclear retaliatory capabilities.13 Subsequently, Chinese leaders began to talk publicly about the possibility of Chinese nuclear retaliation.14 The 1990s witnessed a series of events, including the 1995–96 Taiwan Crisis, that redefined China-U.S. relations after the temporary alignment of the two countries during the Cold War; China's return to the target list of the United States’ strategic warfighting plan for the use of nuclear weapons (known as the Single Integrated Operational Plan); and the Cox Report, which ended dialogues between Chinese and U.S. nuclear laboratories and commercial space cooperation.15 After the U.S. bombing of the Chinese embassy in Belgrade, Yugoslavia, in 1999, China began to devote significant resources to defense modernization. In its 2010 Nuclear Posture Review Report, the United States promised to maintain strategic stability with China, which can be considered a baseline for China-U.S. strategic nuclear relations.16

In general, the model relies on conservative assumptions for China. The reliability of Chinese nuclear missiles—the chance that the missile will work as designed—was assumed to be 80 percent. Given the nuclear superiority of the Soviet Union and the United States, both countries could replace failed missiles with missiles held in reserve following a first strike against China. Therefore, the reliability of the adversaries’ missiles was assumed to be 100 percent. In this model, the targets of Chinese retaliation were the adversaries’ heartlands—the European part of Soviet territory and the continental United States.17 In this study, the major indicator of China's deterrence capabilities is the probability of at least one Chinese nuclear warhead surviving a first strike and being used for retaliation.18

The missile units of the Chinese People's Liberation Army (PLA) Rocket Force are organized into six missile bases, each overseeing several brigades;19 a brigade comprises six battalions. For mobile nuclear missile units, a battalion usually has two companies, each operating a missile launcher.20 Therefore, a standard mobile nuclear missile brigade has twelve launchers. There are three levels of missile sites: technical, forward, and launch sites.21 It appears that most of the missiles are kept in technical sites in peacetime and that missiles would be moved to forward sites when alerted. Each forward site is assumed to accommodate one launcher, to which three launch sites are attached.22 The mobile missiles would be unable to launch if all three corresponding launch sites were destroyed.23Figure 1 shows the notional structure of the missile sites of a standard Chinese nuclear mobile missile brigade.

Figure 1.

Notional Structure of Missile Sites of a Mobile Missile Brigade

Figure 1.

Notional Structure of Missile Sites of a Mobile Missile Brigade

The PLA Navy (PLAN) operates China's nuclear ballistic missile submarine (SSBN) forces. China has developed two generations of SSBNs (Type 092 and Type 094) and their affiliated submarine-launched ballistic missiles (SLBMs) (JL-1 and JL-2). The Type 092 has never been deployed on a deterrent patrol. It was reported that the Type 094 conducted its first deterrent patrol in 2015.24 Therefore, China's SSBNs are not considered in the three scenarios, but are considered in the China-U.S. 2025 scenario.

It is believed that China maintains a highly centralized nuclear warhead management system. As noted, Chinese nuclear missiles are kept off alert in peacetime. Also, most nuclear warheads are stored in a special warhead base in Central China, and just a few warheads are kept on missile bases.25 Mark Stokes notes, “[Chinese nuclear] warheads are managed in peacetime through a system that is separate and distinct from Second Artillery missile bases and subordinate launch brigades,” and “warheads are mated with missiles assigned to brigades only in elevated readiness conditions and perhaps on occasion for training purposes.”26

Two alert levels were considered in this study: day-to-day alert status and full-alert status. Under day-to-day alert status, the model assumes that for each Chinese mobile missile brigade, one battalion (two launchers) is dispersed to the forward sites; others are kept in the technical sites; and only those missiles in the forward sites are colocated with—but still separated from—their warheads. All the warheads assigned to the silo-based missiles are kept on the warhead base. It is also assumed that as long as the warhead base could survive a first strike, the nuclear warheads stored on it would always be successfully delivered to the missile bases. In full-alert status, all the mobile missiles are dispersed to the forward sites, and all the warheads are colocated with their missiles in the forward sites or the silos (as shown in table 1).

Table 1.

Assumptions on the Alert Rates of Chinese Missiles

Mobile MissilesSilo-Based Missiles
Warhead BaseTechnical SitesForward SitesWarhead BaseSilos
Day-to-Day missiles — 5/6 1/6 — 100% 
Alert warheads 5/6 1/6 100% 
Full missiles — 100% — 100% 
Alert warheads 100% 100% 
Mobile MissilesSilo-Based Missiles
Warhead BaseTechnical SitesForward SitesWarhead BaseSilos
Day-to-Day missiles — 5/6 1/6 — 100% 
Alert warheads 5/6 1/6 100% 
Full missiles — 100% — 100% 
Alert warheads 100% 100% 

To assume day-to-day alert status does not mean that a bolt-from-the-blue, disarming first strike is likely. Discussing peacetime alert status is necessary and meaningful for three reasons, however. First, although unlikely, surprise attacks are not impossible. Maintaining enough retaliatory capability even when caught completely by surprise is necessary for nuclear deterrence. Second, other nuclear states adhere to a doctrine of either assured destruction or assured retaliation under day-to-day alert status. Both Russia and the United States have maintained launch-on-warning capabilities in peacetime, while Britain and France maintain a continuous-at-sea-deterrence posture.27 Therefore, assuming that the target is caught by surprise is typical way when analyzing strategic stability.28 Third, a model that considers China's peacetime, day-to-day alert status may highlight its unique nuclear posture and enable comparison with the postures of other nuclear states.

The most challenging part of this research was determining the probability of detection of Chinese nuclear missiles and facilities. Theoretically, there are technical ways to solve this problem based on radar/optical imaging and target recognition. These technical solutions, however, all require detailed parameters for both sides, but this information is not publicly available. This study does not attempt to give the “true” values of these probabilities. Instead, the relative difficulty of detecting different targets is identified, and then, based on the relative detectability, the detection probabilities for each type of targets are assigned.

The detectability of different nuclear facilities depends on their importance and their level of activity in peacetime. Unsurprisingly, the detection probabilities of China's nuclear warhead base and its ICBM silos are higher than for all other nuclear sites. These two types of targets must be top priorities for adversaries’ military intelligence. The location of China's warhead base can be roughly identified based on publicly available information.29 In the U.S. Department of Defense's annual reports on China's military power, estimates of its number of silo-based DF-5 ICBMs have always been a definite number— rather than a range, as for other types of missiles.30 It can be concluded that China's adversaries are almost certain about the locations of these two types of targets.

The detection probabilities of China's three types of missile sites vary. Technical sites and forward sites are underground facilities (UGFs) and usually take years to build. There are many observable activities along the periphery of China's technical sites during peacetime, whereas its forward sites would be used only in raised readiness conditions. Thus, it is safe to say that the detection probability of the technical sites is higher than that of the forward sites. The launch sites are solid, flat pads. In peacetime, only those launch sites used for training would be employed, and the wartime launch sites should be concealed and camouflaged and activated only before launch. Therefore, the detection probability of the launch sites was the lowest among the three types of missile sites.

Chinese missiles also could be detected while being moved between sites and while being prepared for launch. Liquid-propellant missiles are much easier to detect than solid-propellant missiles for two reasons. First, liquid-fueled missiles usually require more service trucks than do solid-fueled ones. According to a declassified U.S. intelligence document, each liquid-propellant DF-3 missile requires approximately thirty supporting vehicles, whereas a solid-propellant missile requires six vehicles, including a transporter erector launcher (TEL).31 Second, compared to solid-fueled missiles, liquid-fueled missiles need additional time (several hours) to load the propellant. The longer the exposure time is, the higher the probability of detection.32

The next step is to determine the kill probability of Chinese missiles and facilities. There are three types of facilities considered in this study: launch pads, UGFs, and missile silos. If detected, the launch pads would be easily destroyed by nuclear weapons, so their kill probability is assumed to be 100 percent. The UGFs are buried deep underground or deep within mountains. Although directly destroying the deeply buried UGFs would be difficult, there are various methods to disable them.33 For silos, analysts have developed a variety of ways to calculate the kill probability based on silo hardness and warhead parameters.34

All detections, strikes, and intercepts are assumed to be statistically independent events. Although convenient for modeling, these assumptions are probably not accurate in the real world. Detection of one site could lead to detection of a nearby site. Recognizing one camouflaged site might make it easier to identify other sites camouflaged with the same technology. On the one hand, a salvo of missile defenses might all fail in the same mode; on the other hand, the second shoot might perform better, because it can use the information collected by the first one. Although inaccurate, the model can still reflect the structural stability between China and the Soviet Union or between China and the United States.

Evolution of First-Strike Uncertainty

This section presents the calculation results of three nuclear first-strike scenarios: China-Soviet Union, 1984; China-United States, 2000; and China-United States, 2010. Table 2 shows the force structures of the Chinese nuclear arsenal for the three scenarios. As mentioned, only missiles that could target the Soviet Union's or the United States’ heartland were considered. In 1984, China's DF-2 and DF-3 medium-range ballistic missiles (MRBMs) could only target Soviet cities in Siberia and the Far East, so they were not considered. In 2000 and 2010, the DF-3, DF-21 MRBM, DF-4 IRBM, and DF-31 ICBM, which could not reach the continental United States, were not considered.

Table 2.

China's Nuclear Forces

Number of Missiles
Missile TypeRange (kilometers)198420002010
DF-4 4,750 4 (1) — — 
DF-5 13,000 20 20 
DF-31A 11,200 — — 24 (2) 
Number of Missiles
Missile TypeRange (kilometers)198420002010
DF-4 4,750 4 (1) — — 
DF-5 13,000 20 20 
DF-31A 11,200 — — 24 (2) 

SOURCES: International Institute for Strategic Studies (IISS), ed., The Military Balance, 1983– 1984 (London: IISS, 1984), p. 91; IISS, ed., The Military Balance, 1999–2000 (London: IISS, 2000), p. 194; and IISS, The Military Balance, 2010 (London: IISS, 2010), p. 399. See also John Wilson Lewis and Hua Di, “China's Ballistic Missile Programs: Technologies, Strategies, Goals,” International Security, Vol. 17, No. 2 (Fall 1992), pp. 5–40, doi.org/10.2307/2539167; and Office of the Secretary of Defense, Military and Security Developments involving the People's Republic of China, 2010 (Washington, D.C.: U.S. Department of Defense, 2010), p. 66.

NOTE: The numbers of missile brigades are in parentheses. The brigade number in 1984 is the author's estimate.

CHINA VERSUS THE SOVIET UNION IN 1984

In 1984, China deployed DF-4 IRBMs and DF-5 ICBMs that could target the Soviet heartland. Based on the DF-3—the first storable liquid propellant ballistic missile designed by China—the DF-4 added a second stage. The final deployment mode adopted by the DF-4 was “in-cave storage/preparation and out-cave erection/filling/firing.”35 The DF-5 was a silo-based, storable liquid propellant, two-stage ICBM.

The Soviet Union's ballistic missile defense system was not considered in this study, because it can protect only Moscow. Therefore, China could simply target other cities in the Soviet Union. Soviet missiles that could be used to target Chinese missile silos included the SS-20 and the SS-18 IV.36 The kill probabilities of those two missiles against China's DF-5 silos are shown in table 3. The yield of SS-18 IV warheads was bigger, and their accuracy was better, than the SS-20. From China's conservative perspective, it was assumed that the SS-18 IVs could be used against its DF-5 silos.

Table 3.

Soviet Nuclear Weapons against China in 1984

Soviet MissilesYield of Soviet WarheadsCircular Error Probable (meters)DF-5 Silo Hardness (pounds per square inch)Single-Shot Kill Probability3 Warheads against 1 DF-5 Silo
SS-18 IV 10 × 500kt 370 900 64% 95.33% 
SS-20  3 × 150kt 450 900 26% 59% 
Soviet MissilesYield of Soviet WarheadsCircular Error Probable (meters)DF-5 Silo Hardness (pounds per square inch)Single-Shot Kill Probability3 Warheads against 1 DF-5 Silo
SS-18 IV 10 × 500kt 370 900 64% 95.33% 
SS-20  3 × 150kt 450 900 26% 59% 

NOTE: Circular error probable is a measure of a weapon's accuracy; it is defined as the radius of a circle, centered on the planned target, within which half of a missile's projectiles are expected to fall. Single-shot kill probability, which depends on the target's hardness and the offensive weapon's accuracy and yield, is the probability that one offensive warhead destroys the target.

The model shows that in 1984, under day-to-day alert status, the probability that at least one Chinese nuclear warhead survived and was used for retaliation after China absorbed a disarming strike from the Soviet Union was 38 percent, and that the probability would increase to 64 percent if Chinese nuclear forces were on full alert. Some may argue that one retaliatory warhead would not be enough for deterrence, so a Monte Carlo simulation was used to calculate the probabilities that more than one warhead could survive, with one million runs applied. The results show that if the criterion for effective deterrence were set to three, then the probabilities of retaliation with at least three warheads for day-to-day status and full-alert status were 0.7 percent and 5 percent, respectively. Detailed results are shown in table 4.37

Table 4.

Evolution of China's Retaliatory Capabilities

Number of Retaliatory Warheads Required for DeterrenceAlert LevelChina-Soviet Union, 1984China-United States, 2000China-United States, 2010
 day-to-day 38% 0.3% 38% 
one full 64% 1.6% 90% 
two day-to-day 7% 0.002% 11% 
 full 23% 0.01% 65% 
 day-to-day 0.7% 0.0001% 4% 
three full 5% — 37% 
 day-to-day 0.1% — 2% 
four full 0.7% — 17% 
five day-to-day 0.002% — 1% 
 full 0.05% — 6% 
Number of Retaliatory Warheads Required for DeterrenceAlert LevelChina-Soviet Union, 1984China-United States, 2000China-United States, 2010
 day-to-day 38% 0.3% 38% 
one full 64% 1.6% 90% 
two day-to-day 7% 0.002% 11% 
 full 23% 0.01% 65% 
 day-to-day 0.7% 0.0001% 4% 
three full 5% — 37% 
 day-to-day 0.1% — 2% 
four full 0.7% — 17% 
five day-to-day 0.002% — 1% 
 full 0.05% — 6% 

CHINA VERSUS THE UNITED STATES IN 2000

In 2000, China's silo-based DF-5 was the only missile that could target the continental United States. By then, the number of DF-5 silos had increased to twenty. The silos were assumed to be harder than they were in 1984. The W88 warhead atop a U.S. Trident II D5 SLBM with a yield of 455 kilotons and circular error probable of 100 meters is a very good counterforce weapon.38 Against Chinese DF-5 silos, it could realize a very high kill probability with a single shot. The kill probability of three W88 warheads targeted against one DF-5 silo would be almost 100 percent, as shown in table 5.

Table 5.

U.S. Trident II D5 against Chinese DF-5 Silo

Yield of U.S. Warheads (kiloton)Circular Error Probable (meters)DF-5 Silo Hardness (pounds per square inch)Single-Shot Kill Probability3 Warheads against 1 DF-5 Silo
8 × 455 100 2,000 99.95% ~ 100% 
Yield of U.S. Warheads (kiloton)Circular Error Probable (meters)DF-5 Silo Hardness (pounds per square inch)Single-Shot Kill Probability3 Warheads against 1 DF-5 Silo
8 × 455 100 2,000 99.95% ~ 100% 

NOTE: Circular error probable is a measure of a weapon's accuracy; it is defined as the radius of a circle, centered on the planned target, within which half of a missile's projectiles are expected to fall. Single-shot kill probability, which depends on the target's hardness and the offensive weapon's accuracy and yield, is the probability that one offensive warhead destroys the target.

Given the United States’ superb reconnaissance and counterforce capabilities, the survivability of China's nuclear forces in 2000 was very poor. For day-to-day alert status, the probability of Chinese nuclear retaliation against the continental United States was 0.3 percent; for full-alert status, it was 1.6 percent. The probabilities of more than one warhead surviving are almost zero (see table 4).39 As Patrick Tyler stated in 1999, “For the last thirty years, the Pentagon's strategic planners were confident that China's ICBMs could simply be eliminated with preemptive strikes in any period of hostilities where nuclear war seemed likely.”40

CHINA VERSUS THE UNITED STATES IN 2010

In 2007, China began to deploy road-mobile, solid-propellant DF-31 and DF-31A ICBMs; only the DF-31As could reach the continental United States.41 China still had twenty DF-5 silos, and their hardness in 2010 was assumed to be the same as it was in 2000. The U.S. weapons that could attack Chinese targets had barely changed. The W88 warheads of the Trident II D5 had the same counterforce capability against the DF-5 silos as shown in table 5.

The United States has maintained an operational homeland missile defense, the Ground-based Midcourse Defense (GMD) system, since 2004. The effectiveness of the GMD system, however, seemed problematic. First, target discrimination “[was] not a completely solved problem.”42 The Pentagon's Director, Operational Test and Evaluation fiscal year 2010 annual report stated, “To date, GMD has demonstrated a limited capability against a simple threat.”43 Second, as a result of bad quality management, the GMD's reliability was low.44 In previous intercept flight tests using either no decoy or very simple decoys, the GMD system had a poor test record; in December 2010, only eight of sixteen tests had succeeded.45 In this study, the reliability of U.S. interceptors in 2010 was assumed to be 50 percent, while the target discrimination probability was 10 percent. At the end of 2010, the United States had deployed a total of thirty-three ground-based interceptors (GBIs).46 Assuming the defense fires off four interceptors against one target warhead, the GMD system could handle up to eight retaliatory warheads in 2010.

The survivability of the DF-31As is much better than the DF-5s, thus improving the general survivability of Chinese nuclear forces. The model suggests that in 2010, the probability of successful Chinese nuclear retaliation for the day-to-day alert status and full-alert status was 38 percent and 90 percent, respectively. The Monte Carlo simulation (1 million runs) indicated that if the criterion of nuclear deterrence were increased to three, then the probability of effective retaliation would be reduced to 4 percent for day-to-day alert status and 37 percent for full-alert status, as shown in table 4.47 The simulation also shows that although the scale of the GMD system was limited in 2010, it was big enough to support a four-interceptor salvo firing doctrine against surviving Chinese warheads, and that the probability that Chinese retaliatory warheads could overwhelm U.S. GMD was extremely low and neglectable.

The above analysis demonstrates that although China's deployment of its mobile ICBMs greatly improved its retaliatory capabilities, China still lacked an assured retaliatory capability, as Glaser and Fetter predicted.48 The key difference between their work and this model is the deployment mode of China's land-based mobile missiles. In Glaser and Fetter's article, it is assumed that China alerts its mobile missiles early in a crisis and deploys them in the field over a large area away from any fixed bases or sites.49 As mentioned in previous sections, however, this author believes that China's practice is to keep the dispersed mobile missiles in underground forward sites. From the perspective of survivability, this is not the best practice. Nevertheless, the deployment mode of Chinese mobile missiles could reduce the risk of nuclear accidents and unauthorized launch.

SUMMARY

In 1984, the probability of China's nuclear retaliation against the European part of the Soviet Union following a nuclear first strike was 38 percent for day-to-day alert status and 64 percent for full-alert status. In 2000, regardless of the state of alert, the probability of successful Chinese nuclear retaliation was close to zero. In 2010, China's deployment of the road-mobile DF-31A ICBM significantly improved the survivability of its nuclear forces; as a result, the probability of China's nuclear retaliation for day-to-day status and full-alert status was 38 percent and 90 percent, respectively (see table 4).

A comparison of China versus the Soviet Union in 1984 and China versus the United States in 2010 shows that China's retaliatory capabilities on day-today alert were maintained at roughly the same level, whereas those on full alert improved significantly. The decisive factors for the survivability of China's nuclear forces on day-to-day alert were the off-alert status of its nuclear missiles in peacetime and the poor survivability of its warhead base, which had barely changed from 1984 to 2010. For full-alert status, China's deterrent capabilities were determined by the survivability of its dispersed nuclear missiles; its solid-propellant missiles in 2010 were much more survivable than its liquid-propellant missiles in 1984.

SENSITIVITY ANALYSIS

The study's sensitivity analysis focuses on the 2010 China-U.S. scenario. Figure 2 demonstrates the impact on China's nuclear deterrence of the United States’ capability to destroy Chinese UGFs. The probability of China's nuclear retaliation decreases with the UGF's vulnerability. If the UGFs were invulnerable (kill probability = 0 percent), then even under day-to-day alert, the probability of China's nuclear retaliation would be close to guaranteed (99 percent).

Figure 2.

Impact of U.S. Kill Probablity of Chinese Underground Facilities, 2010 (day-today alert)

Figure 2.

Impact of U.S. Kill Probablity of Chinese Underground Facilities, 2010 (day-today alert)

A conservative perspective suggests that the kill probability of the UGFs is relatively high. In 2010, the U.S. military had various programs on how to identify, characterize, and defeat UGFs.50 Shallow-buried facilities could be directly destroyed by conventional penetrator weapons, whereas more deeply buried ones were “nearly invulnerable to direct attack by conventional means.”51 Regarding a nuclear first strike from the United States, as stated in a report from the U.S. National Academy of Sciences, “Many—but not all— known and/or identified hard and deeply buried targets can be held at risk of destruction by one or a few nuclear weapons.”52 Furthermore, various methods are available to neutralize or disable the UGFs without physically destroying them, such as wiping out the exits/entrances and entombing the UGFs, targeting an environmental control system, attacking an electrical power grid, or using high-powered microwaves to neutralize computer and communication equipment.53 According to one Chinese study, the explosion of a B61-11 nuclear bomb would thoroughly destroy the tunnel exits/entrances of China's underground missile sites.54

Figure 3 indicates the probability of Chinese nuclear retaliation when the alert rate of China's land-based mobile missiles is varied. Clearly, the alert rate has great influence on China's nuclear deterrence capabilities. Two alert levels (day-to-day alert and full alert) discussed previously represent two extreme cases. The exact alert rate in peacetime is unknown, but it is believed that China keeps most of its nuclear missiles in technical sites in peacetime and disperses them in crises.55

Figure 3.

Impact of Alert Rate on China's Nuclear Deterrent Capability

Figure 3.

Impact of Alert Rate on China's Nuclear Deterrent Capability

The Impact of Technological Advancements

Continued improvements in the United States’ counterforce and strategic defense capabilities will eventually undermine China's nuclear deterrence. This section examines China's potential countermeasures and a hypothetical case of China versus the United States in 2025.

IDENTIFYING FIXED FACILITIES

The survivability of fixed-missile facilities relies heavily on their ability to avoid detection. Concealment and camouflage play a very important role in the construction and operation of Chinese nuclear facilities.56 At the same time, how to defeat adversaries’ concealment and camouflage capabilities, the so-called denial and deception problem, is a major focus of U.S. intelligence.57 For several reasons, the competition between concealing and detecting fixed nuclear facilities favors the United States.

First, China requires years to build a missile site, which can then operate for several decades. Concealment and camouflage measures need to be applied carefully and consistently during construction and operation. Any mistake could lead to disclosure. The U.S. military has numerous opportunities to detect a site's existence. For example, in 1984, a DF-5 silo that had begun construction in 1982 was fully exposed, because the camouflage net that was supposed to be suspended over the silo was folded back on support poles.58 The same mistake was made in December 1984.59

Second, the inflexible routines and poorly designed standard operating procedures of states’ military organizations can undermine survivability.60 During the Cold War, Soviet military operating patterns greatly aided U.S. intelligence; these included the SA-2's Star of David deployment pattern, the number of security fences around ICBM silos, and the shape, protrusions, and shipping hooks of crates carrying military equipment.61 U.S. intelligence also benefited from the PLA's standard operating patterns. As declassified U.S. intelligence reports show, the purpose of a Chinese missile site could be determined by the similarities of the geographical layout and camouflage activity of previously observed sites.62

Third, technology developments make concealment and camouflage more difficult. Hyperspectral imaging (HSI) systems, which generate images with hundreds of contiguous narrow spectral bands, pose tremendous challenges for fixed nuclear facilities. All materials emit or reflect eletromagnatic radiation in ways characteristic of their molecular makeup. Therefore, HSI can be used to identify and discriminate among different materials. Concealing objects from detection against hyperspectral imaging is very difficult. The disadvantages of HSI include limited coverage and a requirement for huge amounts of computing. The offense could combine HSI with other sensors, such as synthetic aperture radar (SAR) or panchromatic imaging, to achieve better target characterization and identification.63 Futhermore, technological developments in artificial intelligence, deep learning, and supercomputers will eventually expand the application of HSI technology.64

DETECTION OF MOBILE MISSILES AND ANTI-SUBMARINE WARFARE

Locating mobile missiles is a challenging task. A frequently cited example is U.S. efforts to locate and destroy Iraqi Scud missiles during the 1991 Gulf War; despite 1,460 strikes, there is no evidence that a single Scud TEL was destroyed.65 Since then, U.S. technological developments have provided new options to hunt mobile missiles.66 One began with the Discoverer II program, created in 1998. The program, which was supposed to have both Ground Moving Target Indicator (GMTI) radar and SAR functions, was canceled in 2000, after experiencing difficulties. In 2001, the idea of GMTI/SAR satellites was reborn as the Space-Based Radar program, which was renamed the Space Radar program in 2005 and canceled in 2008.67 By comparison, airborne GMTI/SAR radar, such as the Joint Surveillance and Target Attack Radar System–equipped E-8 series aircrafts and the RQ-4 Global Hawk unmanned aerial vehicle (UAV), has enjoyed wide usage by the U.S. military.

Second, the United States is developing geosynchronous optical imaging technology. For example, the Membrane Optical Imager for Real-Time Exploitation program, sponsored by the Defense Advanced Research Projects Agency, is under development. The concept utilizes transmissive diffractive membrane optic technology, rather than the traditional reflective or refractive optics technology used for large-aperture space telescopes. The satellite would be able to provide 1-meter resolution while focused on a 10-kilometer by 10-kilometer area and provide real-time video at one frame per second.68 One objective of the project is to detect and track Scud-class TELs. The program's original Broad Agency Announcement called for the “probability of detection for a SCUD-class launch of 0.99, with less than one false alarm per month.”69

Third, electronic intelligence also could play an important role in hunting mobile missiles. Radio communications from missile launchers could be intercepted and used for geolocation. The command and control system of mobile missile forces could be hacked, and location data could be extracted. The unattended ground sensors could be deployed to indicate passing missile launchers. The offense could even hack smartphones to track a mobile missile's crew members.70

In contrast to fixed sites, the technological competition between hiding and locating mobile missiles favors China for several reasons. First, the UAV-borne GMTI/SAR might work against a small country such as North Korea, but it cannot cover a country as big as China. Second, the GMTI radar is susceptible to some relatively simple countermeasures such as radar stealth cover.71 Third, target recognition of camouflaged missile launchers would be highly challenging for the GMTI radar. Fourth, prudent operation, such as radio silence, could to some extent mitigate the threat posed by signals intelligence.

In the anti-submarine warfare domain, the U.S. Navy enjoys both technological and geographical superiority over the PLAN, which is likely to persist for a long time. According to the U.S. Office of Naval Intelligence, China's Type 094 SSBN is quite noisy, making it vulnerable to U.S. ASW forces.72 China would need years to build quiet submarines. The range of China's JL-2 SLBM is reported to be 7,400 kilometers.73 The Type 094 would have to pass through the first island chain to hold the continental United States at risk, further undermining its survivability. The range of China's next-generation SLBM, the JL-3, is reported to be 9,000 kilometers—still insufficient to target the continental United States from China's coastal waters.74

The U.S. military has taken advantage of passive acoustic technology and favorable geography to build undersea surveillance systems. During the Cold War, the United States developed the Sound Surveillance System to detect Soviet submarines. After Soviet submarines became too quiet to detect by long range, passive acoustics, the U.S. military deployed the Fixed Distributed System to critical ocean choke points to detect passing submarines and cue other ASW capabilities.75 China faces a geographic dilemma similar to the Soviet Union's. Since the early 2000s, a more modern system, much like the U.S. Sound Surveillance System—the U.S. Navy's Fish Hook Undersea Defense Line—was commissioned along the first island chain to monitor the movement of Chinese submarines.76 It is believed that the United States also has deployed the Fixed Distributed System in the Western Pacific.77

The PLAN has far to go to overcome these technological and geographical obstacles.

MISSILE DEFENSE

In announcing the release of the 2019 Missile Defense Review, President Donald Trump stated that the goal of the U.S. missile defense system is to “detect and destroy any missile launched against the United States—anywhere, anytime, anyplace.”78 Although the Review maintains the traditional declaratory policy that U.S. missile defense is designed to protect the homeland against North Korea and Iran—not China or Russia—Trump's statement is likely to generate suspicion among China's defense planners. Furthermore, the Review calls for the space sensor layer to track and potentially target hypersonic gliding missiles, asks the Pentagon to explore the possibility of space-based interceptors, and emphasizes the means for a boost-phase intercept.79 If fully implemented, these initiatives would have a serious negative impact on China's nuclear deterrence.

The U.S. military is working to improve its missile defenses’ target discrimination capability. The U.S. Missile Defense Agency awarded Lockheed Martin a $784 million contract to build the Long Range Discrimination Radar (LRDR) in Alaska. Working at S-band, the LRDR was planned to become operational in 2020.80 Forward-based X-band radar, such as the THAAD (Terminal High-Altitude Area Defense) radar deployed in South Korea—which has the capability to detect and track Chinese strategic missiles targeting the United States during their boost phase—also could contribute to target discrimination.81 By viewing the velocity changes of offensive missiles generated by the deployments of light decoys and heavy warheads, the forward-deployed X-band radar can exclude targets with insufficient mass.82

Boost-phase intercept could solve the target discrimination problem by engaging the target missiles before they release their penetration aids. An option is the UAV-borne laser. Compared to its predecessor, the Airborne Laser, the UAV-borne laser operates at a higher altitude, reducing the effects of atmospheric disturbances. The current U.S. plan is to fly a low-power (140–280 kilowatts) laser demonstrator in 2023 to prove its discrimination capability against ICBMs and explore the feasibility of destroying ICBMs.83 Although the UAV-borne laser has great potential, two factors will limit its effectiveness against Chinese strategic missiles. First, the amount of power required for destroying missiles is much higher than the planned 2023 level, probably up to megawatt level.84 Second, all laser weapons have limited range; therefore, the UAV-borne laser cannot cover a country as big as China.85

The U.S. military is also working on solving the GMD system's reliability problem. The Pentagon canceled the Redesigned Kill Vehicle program “due to technical design problems” in August 2019.86 The Redesigned Kill Vehicle was expected to replace the current Exoatmospheric Kill Vehicle on the GBI. At the same time, the Pentagon started seeking bids to develop the next-generation interceptor.87 The new interceptor will be probably more reliable than the GBI. On May 30, 2017, the GBI's first intercept test against an ICBM-class target was successful.88 Another intercept test engaging two GBIs against one ICBM target was successfully conducted on March 25, 2019.89

Figure 4 demonstrates the impact of the U.S. BMD's discrimination capability on China's nuclear deterrence. The simulation assumes that the United States launches a first strike against Chinese nuclear forces and then U.S. missile defenses are used to intercept Chinese retaliatory missiles. Based on the 2010 China-U.S. scenario, it is assumed that the U.S. military has fixed the reliability problem (the interceptor reliability is 90 percent). The figure reveals that a strong U.S. missile defense system would neutralize Chinese nuclear deterrence.

Figure 4.

Improved U.S. BMD and Chinese Nuclear Deterrence as of 2010

Figure 4.

Improved U.S. BMD and Chinese Nuclear Deterrence as of 2010

It should be noted that the United States can compensate for low target discrimination with additional interceptors.90 In general, it is believed that four interceptors would be used against one offensive warhead.91 Given the very limited number of Chinese missiles that would survive a U.S. first strike, however, the U.S. military could have the luxury of allocating more interceptors against one retaliatory Chinese warhead. This scenario would be more likely in the future, when the United States deploys more strategic-capable interceptors and kill vehicles.

China will seek to develop countermeasures to defeat U.S. BMD. It has developed a variety of penetration aids since the beginning of its ICBM program.92 In recent years, China's BMD program has given the country an opportunity to address penetration and target discrimination issues at the same time.93 Some types of countermeasures—such as anti-simulation decoys and a cooled shroud—would be very difficult to defeat, but whether China has adopted these countermeasures for its missiles is unknown.94

Another option for China is to develop boost-glide hypersonic missiles. Boost-glide missiles’ depressed and unpredictable trajectory poses a huge challenge for traditional hit-to-kill missile defenses. It has been reported that China is developing several hypersonic missile programs, including the Wu-14/DF-ZF, Xingkong-2 [Starry Sky 2], and DF-17, and has conducted several flight tests.95 Some sources indicate that those missiles could be used for both conventional and nuclear purposes.96 The longest range reached in the tests, however, was 2,100 kilometers, and the cross-range maneuverability demonstrated in those tests was limited.97 Therefore, to play a role in countering U.S. strategic missile defenses, China's hypersonic glider development program has far to go.

In sum, the United States is seeking to improve its missile defense system, while China is committed to developing countermeasures. The outcome of this technological competition remains uncertain.

CHINA VERSUS THE UNITED STATES IN 2025

This section presents a hypothetical scenario involving China's nuclear retaliation following a U.S. first strike in 2025. The purpose is not to predict the future development of both sides’ strategic capabilities, but rather to provide a nominal force structure to evaluate the prospect of China-U.S. strategic stability.

Table 6 provides a notional Chinese nuclear force structure in 2025. In this scenario, one brigade of land-mobile DF-41 ICBMs is deployed. The DF-41 is assumed to be able to carry three multiple independently targeted reentry vehicles (MIRVs).98 In addition, the DF-41 probably will have better mobility, need a shorter launch preparation time than the DF-31A, and require no preplanned launch sites. China is assumed to have retired all of its DF-31As and deployed the same number of DF-31AGs. There are five Type 094 SSBNs.99 It is assumed, however, that under day-to-day alert status, there is a 50 percent chance that one SSBN is on patrol; under full-alert status, China could put at most three SSBNs to sea. The United States would have a better ability to detect fixed missile sites than it did in 2010. The reliability of the U.S. BMD interceptors is 90 percent, and the target discrimination probability is 30 percent.100

Table 6.

Notional Chinese Nuclear Forces in 2025

Missile TypeNumber of WarheadsRange (kilometers)Number of Missiles
DF-5/A 12,000–13,000 10 
DF-5B 12,000–13,000 10 
DF-31AG 11,200 24 (2) 
DF-41 12,000 12 (1) 
JL-2 7,400 60 (5) 
Missile TypeNumber of WarheadsRange (kilometers)Number of Missiles
DF-5/A 12,000–13,000 10 
DF-5B 12,000–13,000 10 
DF-31AG 11,200 24 (2) 
DF-41 12,000 12 (1) 
JL-2 7,400 60 (5) 

SOURCES: Bill Gertz, “China Tests New Long-Range Missile with Two Guided Warheads,” Washington Free Beacon, August 18, 2015, https://freebeacon.com/national-security/china-tests-new-long-range-missile-with-two-guided-warheads/; and Jane's Sentinel Security Assessment: China and Northeast Asia, March 12, 2018.

NOTE: The numbers of missile brigades or nuclear ballistic missile submarines are in parentheses.

The modeling shows that although both China and the United States had strengthened their strategic capabilities, China's nuclear deterrence level in 2025 is roughly the same as it was in 2010. The probability of successful Chinese nuclear retaliation under day-to-day alert status and full-alert status is 92 percent and 40 percent, respectively. The Monte Carlo simulation (1 million runs) indicated that if the criterion of nuclear deterrence is increased to three, then the probability of effective retaliation would be reduced to 10 percent for the day-to-day alert and 56 percent for full alert, as shown in Figure 5.

Figure 5.

China's Nuclear Deterrence in 2025 versus the United States (notional)

Figure 5.

China's Nuclear Deterrence in 2025 versus the United States (notional)

How Much Uncertainty Is Enough?

The analysis in this study suggests that China's criterion for effective nuclear deterrence is very low. The modeling demonstrates that Chinese nuclear retaliation following a disarming strike from the Soviet Union or the United States has been and remains far from assured. In a crisis or during wartime—in which China's nuclear forces were assumed to be at full alert—the likelihood of successful retaliation would be much higher, but still cannot be guaranteed. Nonetheless, Chinese leaders have felt comfortable with this situation for several decades and have decided not to build a larger nuclear arsenal or raise the alert level in peacetime to pursue assured retaliation.

China's nuclear posture reflects its nuclear philosophy. Chinese leaders believe that nuclear weapons are “paper tigers,” and because of the taboo against the use of nuclear weapons, they are unlikely to be used.101 More realistic is the threat of nuclear coercion or nuclear blackmail.102 Therefore, from the perspective of Chinese leaders, the requirements for nuclear deterrence are not high. As Deng Xiaoping stated, the point of Chinese nuclear weapons is “to show that we have what they have. If they want to destroy us, they themselves will also suffer some retaliation.”103 For Marshal Nie Rongzhen, who once oversaw China's nuclear weapon and missile programs, the purpose of China's nuclear weapons was to have “the rudimentary means of counterstrike” (qima de huanji shouduan) if China came under a nuclear attack.104

This study provides some general insights for nuclear deterrence theory. Given the catastrophic consequences of a nuclear attack, it is hard to imagine that a reasonable world leader would risk losing a city and order a first strike against China. Therefore, uncertain retaliation, like assured retaliation and assured destruction, can have a strong deterrent effect. As former British Secretary of State for Defense Denis Healey once remarked, “It takes only 5 per cent credibility of American retaliation to deter the Russians, but 95 per ent credibility to reassure the Europeans.”105 McGeorge Bundy argued, “Thermonuclear weapons impose a radically new calculus of advantage on anyone seeking to neutralize them: they make it necessary to achieve a kill rate very near 100 percent. Anything less is not good enough for safety—only good enough, at best, for deterrence.”106

China's nuclear deterrent dates back to 1974, when China had deployed liquid-propellant DF-3 missiles, although the PLA could hardly launch them independently.107 At that time, U.S. intelligence admitted that China had acquired “a modest but credible nuclear retaliatory capability,” and that its desire to deter the United States and the Soviet Union was linked to fears for the security of a few bases and cities in the Far East and Siberia.108

This study considers the 2010 China-U.S. scenario as a baseline for stable mutual deterrence, because for the first time, the United States promised to maintain strategic stability with China.109 Whether or not uncertain retaliation is enough to deter nuclear attack or coercion in a specific scenario depends on the level of first-strike uncertainty, the issue at stake, and the other side's perception. Since the 1969 China-Soviet border crisis, China has not been involved in a serious crisis or conflict with a nuclear state, so fortunately, it is impossible to test whether China's nuclear deterrent capability is adequate.

Some scholars have a different view. They argue that secure second-strike capabilities are difficult to develop, and that “the United States stands on the verge of attaining nuclear primacy vis-à-vis its plausible great power adversaries.”110 To some extent, the author agrees that technological developments could enhance counterforce capabilities and erode survivability. The author does not believe that technological competition always favors the offense, however. As analyzed in the previous section, in some areas—such as imaging reconnaissance—the offense has an advantage over the defense, but in other areas—such as mobile missile tracking—the defense could develop relatively simple countermeasures to defeat the offense.

More importantly, the nuclear primacy argument ignores the effects of uncertainty. As long as the defense maintains some degree of first-strike uncertainty, the offense would have difficulty translating its nuclear superiority into political power.111 To be politically meaningful, the performance of counterforce and strategic defense must be very good, and the offense needs to invest a huge amount of resources. In other words, nuclear stalemate is much harder to escape than technological superiority seems to suggest.

The merits of deterrence through uncertain retaliation are twofold. First, this form of deterrence reduces the risk of an accidental or unauthorized launch. As mentioned, a key feature of China's nuclear posture of uncertain retaliation is that its nuclear forces are off alert in peacetime, with warheads being separated from boosters. In contrast, four other legitimate nuclear weapon states— France, Russia, the United Kingdom, and the United States—deploy nuclear warheads on alert, and the weapons are ready to be launched on relatively short notice.112 Alerting nuclear forces in peacetime would inevitably increase the risk of nuclear accidents.113 Furthermore, U.S. and Russian ICBMs are kept on hair-trigger alert and could be launched before being hit by incoming warheads.114 A strategic missile requires less than thirty minutes to reach its target; state leaders would therefore face strong pressure to make a decision quickly on whether to retaliate. In a false alarm, of which many have occurred, this posture could have catastrophic consequences.115

Second, the posture of uncertain nuclear retaliation could provide a pathway for deep cuts in the number of nuclear weapons in the world. Traditional nuclear doctrines require maintaining strategic nuclear forces of sufficient size to achieve assured destruction or assured retaliation, which constitute a barrier for further disarmament. Henry Kissinger and Brent Scowcroft argued that the United States should not go beyond the levels included in the New Strategic Arms Reduction Treaty, which caps the number of Russian-deployed and U.S.-deployed strategic nuclear warheads at 1,550 each, because “strategic stability is not inherent with low numbers of weapons; indeed, excessively low numbers could lead to a situation in which surprise attacks are conceivable.”116 According to the doctrine followed by France and the United Kingdom—known as continuous-at-sea-deterrence—four SSBNs are required to keep at least one SSBN on deterrent patrol at any time.117 By contrast, there is no minimum requirement for the number of nuclear weapons for deterrence through uncertain retaliation; therefore, each state's nuclear arsenal could be reduced to dozens or perhaps even fewer, if the political environment were conducive and all major nuclear powers joined the disarmament process.

A potential consequence of deterrence through uncertain retaliation is that because its deterrent capability is lower than that of assured retaliation or assured destruction, uncertain retaliation might lead to an arms race, for three reasons. First, uncertain retaliation might be unable to persuade the adversary to acknowledge—or accept—mutual vulnerability. For example, whether the United States should accept mutual vulnerability with China is debatable among U.S. strategists.118 Second, a state's lack of an assured retaliatory capability could encourage adversaries to develop counterforce capabilities and strategic defenses to undermine or neutralize its uncertain retaliatory capabilities. Third, for these two reasons, a country with an uncertain retaliation posture might become skeptical of the effectiveness of its nuclear deterrence and feel pressure to develop “a more calculated strategy of assured retaliation.”119

Conclusion

The analysis presented in this study suggests that China's nuclear deterrence is based on uncertain retaliation, rather than assured retaliation. In 1984, the probability of China's nuclear retaliation against the heartland of the Soviet Union was 38 percent under day-to-day alert status, and 64 percent under full-alert status. Given the vulnerability of its silo-based ICBMs, China's nuclear retaliatory capability against the continental United States in 2000 was almost zero. In 2010, the probability of Chinese nuclear retaliation against the continental United States for day-to-day alert status and full-alert status was 38 percent and 90 percent respectively. Although the United States continues to develop its counterforce and missile defense capabilities, in 2025, China will likely maintain a level of nuclear deterrence no lower than its 2010 level.

Despite rapid progress in counterforce and missile defense technologies, mutual deterrence is more stable than recent technological advancements seem to suggest. On the one hand, countermeasures could be developed and, to some extent, offset the effects of offensive technologies. On the other hand, first-strike uncertainty plays an important role in the offense's decisionmaking. Although in some cases technological advancements could to some extent undermine nuclear deterrence, the offense would need to invest a massive amount of resources to translate its superiority into coercive power.

This study can contribute to the promotion of China-U.S. strategic stability in a few ways. First, the study's modeling identified the roots of the vulnerability of China's nuclear forces and highlighted the direction that China could take to address this vulnerability. China could reduce its reliance on fixed missile facilities and randomly move TELs to simple shelters—such as in a warehouse, under an overpass, or in a tunnel; these locations are soft but difficult to detect. Radio communications and cell phones could be prohibited for dispersed mobile missile units. Missile launchers with off-road capabilities and/or camouflaged as civilian trucks also could improve survivability. A large investment could be devoted to submarine quieting technologies. All in all, to improve its nuclear deterrent capability, as a first step, China should improve the survivability of its nuclear delivery platforms; then, if necessary, it could consider expanding its nuclear arsenal.

Second, the modeling shows that the number of Chinese retaliatory warheads that would survive a U.S. first strike is very limited, or nonexistent; even a small-scale missile defense might be good enough to protect the United States. In previous China-U.S. track 1.5 dialogues, U.S. participants attempted to reassure their Chinese counterparts that the United States’ BMD system was designed to protect against small nuclear forces such as North Korea's, not China's.120 This study demonstrates that the key factor affecting China's nuclear deterrent is not the BMD system's number of interceptors, but its target discrimination capability, which should be the focus of future missile defense dialogues.

Acknowledgments

The author owes special thanks to Daryl Press, who raised the central issue addressed in this article and discussed it extensively with him. For helpful comments, the author also thanks Matthew Bunn, Steve Fetter, Taylor Fravel, Laura Grego, Eric Heginbotham, Li Bin, Martin Malin, David Wright, Hui Zhang, the anonymous reviewers, and participants at numerous seminars where the study's research was presented.

Notes

1. 

Charles L. Glaser, Analyzing Strategic Nuclear Policy (Princeton, N.J.: Princeton University Press, 1990), pp. 19–60; and Lawrence Freedman, The Evolution of Nuclear Strategy (London: Palgrave Macmillan, 1981).

2. 

Alain C. Enthoven and K. Wayne Smith, How Much Is Enough? Shaping the Defense Program, 1961–1969 (Santa Monica, Calif.: RAND Corporation, 2005).

3. 

Kristan Stoddart, “Maintaining the ‘Moscow Criterion': British Strategic Nuclear Targeting, 1974–1979,” Journal of Strategic Studies, Vol. 31, No. 6 (December 2008), pp. 897–924, doi.org/10.1080/01402390802373198; and David S. Yost, France's Deterrent Posture and Security in Europe, Part I: Capabilities and Doctrine, Adelphi Paper No. 194 (London: International Institute for Strategic Studies, 1984).

4. 

M. Taylor Fravel and Evan S. Medeiros, “China's Search for Assured Retaliation: The Evolution of Chinese Nuclear Strategy and Force Structure,” International Security, Vol. 35, No. 2 (Fall 2010), pp. 48–87, doi.org/10.1162/ISEC_a_00016; Fiona S. Cunningham and M. Taylor Fravel, “Assuring Assured Retaliation: China's Nuclear Posture and U.S.-China Strategic Stability,” International Security, Vol. 40, No. 2 (Fall 2015), pp. 7–50, doi.org/10.1162/ISEC_a_00215; Vipin Narang, Nuclear Strategy in the Modern Era: Regional Powers and International Conflict (Princeton, N.J.: Princeton University Press, 2014), pp. 121–152; James C. Mulvenon et al., Chinese Responses to U.S. Military Transformation and Implications for the Department of Defense (Santa Monica, Calif.: RAND Corporation, 2006), p. 97; Michael S. Chase and Evan Medeiros, “China's Evolving Nuclear Calculus: Modernization and Doctrinal Debate,” in James Mulvenon and David Finkelstein, eds., China's Revolution in Doctrinal Affairs: Emerging Trends in the Operational Art of the Chinese People's Liberation Army (Alexandria, Va.: CNA, 2005), pp. 119–154; and Xiangli Sun, “Analysis of China's Nuclear Strategy,” China Security, No. 1 (2005), pp. 23–27.

5. 

Gregory Treverton, “China's Nuclear Forces and the Stability of Soviet-American Deterrence,” in Christoph Bertram, ed., The Future of Strategic Deterrence, Adelphi Paper No. 160 (London: International Institute for Strategic Studies, 1980), pp. 38–44; Avery Goldstein, Deterrence and Security in the 21st Century: China, Britain, France, and the Enduring Legacy of the Nuclear Revolution (Stanford, Calif.: Stanford University Press, 2000), pp. 111–138; Nicola Horsburgh, China and Global Nuclear Order: From Estrangement to Active Engagement (Oxford: Oxford University Press, 2015); and Wu Riqiang, “Certainty of Uncertainty: Nuclear Strategy with Chinese Characteristics,” Journal of Strategic Studies, Vol. 36, No. 4 (August 2013), pp. 579–614, org/10.1080/01402390.2013.772510.

6. 

Austin Long and Brendan Rittenhouse Green, “Stalking the Secure Second Strike: Intelligence, Counterforce, and Nuclear Strategy,” Journal of Strategic Studies, Vol. 38, Nos. 1–2 (January 2015), p. 65, doi.org/10.1080/01402390.2014.958150; Keir A. Lieber and Daryl G. Press, “The End of MAD? The Nuclear Dimension of U.S. Primacy,” International Security, Vol. 30, No. 4 (Spring 2006), pp. 7–44, doi.org/10.1162/isec.2006.30.4.7; Brendan Rittenhouse Green and Austin Long, “The MAD Who Wasn't There: Soviet Reactions to the Late Cold War Nuclear Balance,” Security Studies, Vol. 26, No. 4 (October 2017), pp. 606–641, doi.org/10.1080/09636412.2017.1331639; and Keir A. Lieber and Daryl G. Press, “The New Era of Counterforce: Technological Change and the Future of Nuclear Deterrence,” International Security, Vol. 41, No. 4 (Spring 2017), pp. 9–49, org/10.1162/ISEC_a_00273.

7. 

Charles L. Glaser and Steve Fetter, “Should the United States Reject MAD? Damage Limitation and U.S. Nuclear Strategy toward China,” International Security, Vol. 41, No. 1 (Summer 2016), p. 78, org/10.1162/ISEC_a_00248.

8. 

Eric Heginbotham et al., The U.S.-China Military Scorecard: Forces, Geography, and the Evolving Balance of Power, 1996–2017 (Santa Monica, Calif.: RAND Corporation, 2015), pp. 285–319.

9. 

“Successful” retaliation is defined as at least one Chinese warhead surviving the adversary's first strike, functioning as designed, and penetrating any defenses the other side might deploy.

10. 

Neal H. Hillerman, “The Theoretical Basis of the Code 50 Nuclear Exchange Model” (Arlington, Va.: Center for Naval Analyses, September 1971), p. 1.

11. 

Albert Wohlstetter, “The Delicate Balance of Terror,” Foreign Affairs, Vol. 37, No. 2 (January 1959), pp. 211–234; Glenn A. Kent and David E. Thaler, First-Strike Stability: A Methodology for Evaluating Strategic Forces (Santa Monica, Calif.: RAND Corporation, 1989); Michael M. May, George F. Bing, and John D. Steinbruner, “Strategic Arsenals After START: The Implications of Deep Cuts,” International Security, Vol. 13, No. 1 (Summer 1988), pp. 90–133, doi.org/10.2307/2538897; John D. Steinbruner and Thomas M. Garwin, “Strategic Vulnerability: The Balance between Prudence and Paranoia,” International Security, Vol. 1, No. 1 (Summer 1976), pp. 138–181, doi.org/10.2307/2538581; Dean Wilkening and Kenneth Watman, Strategic Defenses and First-Strike Stability (Santa Monica, Calif.: RAND Corporation, 1986); and Bruce Blair et al., “Smaller and Safer: A New Plan for Nuclear Postures,” Foreign Affairs, Vol. 89, No. 5 (September/October 2010), pp. 9–16, https://www.foreignaffairs.com/articles/russian-federation/2010-09-01/smaller-and-safer.

12. 

Lieber and Press, “The End of MAD?”

13. 

Aiping Zhang, Zhongguo Renmin Jiefangjun [The People's Liberation Army of China], Vol. 1 (Beijing: Dangdai zhongguo chubanshe, 1994), p. 113.

14. 

Aiping Zhang, Zhang Aiping junshi wenxuan [Selected military works of Zhang Aiping] (Beijing: Changzheng chubanshe, 1994), pp. 573–584; Qingsheng Guo, “Zhongguo yongyou hefanji nengli: Fang zhongguo zhanlue daodan budui” [China possessed nuclear counterstrike capability: An interview in China's strategic missile force], Liaowang [Outlook], April 22, 1985, pp. 23–25; and Zhiqiang Chen, Su Zhao, and Guanglong Cheng, “Juyou yiding hefanji nengli woguo zhanlue daodan budui chuju guimo” [With some nuclear counterstrike capability our strategic missile force achieves initial scale], Renmin Ribao [People's Daily], January 9, 1988.

15. 

Robert S. Ross, “The 1995–96 Taiwan Strait Confrontation: Coercion, Credibility, and the Use of Force,” International Security, Vol. 25, No. 2 (Fall 2000), pp. 87–123, doi.org/10.1162/016228800 560462; Hans M. Kristensen, Robert S. Norris, and Matthew G. McKinzie, Chinese Nuclear Forces and U.S. Nuclear War Planning (Washington, D.C.: Federation of American Scientists and Natural Resources Defense Council, 2006), pp. 160–164, https://fas.org/pub-reports/chinese-nuclear-forces-u-s-nuclear-war-planning/; and Michael M. May, ed., The Cox Committee Report: An Assessment (Stanford, Calif.: Center for International Security and Cooperation, Stanford University, 1999), https://cisac.fsi.stanford.edu/publications/cox_committee_report_the_an_assessment.

16. 

Secretary of Defense Robert M. Gates, Nuclear Posture Review Report (Washington, D.C.: U.S. Department of Defense [DoD], April 2010), pp. 28–29. The 2018 Nuclear Posture Review did not use the term “strategic stability.” Instead, it stated that the United States seeks “stable relations” with China and Russia. See Secretary of Defense Jim Mattis, Nuclear Posture Review, 2018 (Washington, D.C.: DoD, February 2018), p. 4.

17. 

Only targets in the “heartland” are considered, because this study takes a conservative view of China and prefers to underestimate China's capabilities. The online appendix presents the results considering retaliatory strikes against the adversaries’ whole territory. See the online appendix at org/10.7910/DVN/5EKNJM.

18. 

For formulas used in the model, see the online appendix.

19. 

On December 31, 2015, the Second Artillery was elevated from an independent branch to the fourth military service and renamed as the PLA Rocket Force.

20. 

Mark A. Stokes and Ian Easton, “Evolving Aerospace Trends in the Asia-Pacific Region: Implications for Stability in the Taiwan Strait and Beyond” (Washington, D.C.: Project 2049 Institute, May 2010), p. 10, https://project2049.net/2010/05/27/evolving-aerospace-trends-in-the-asia-pacific-region-implications-for-stability-in-the-taiwan-strait-and-beyond/; and David C. Logan, “Career Paths in the PLA Rocket Force: What They Tell Us,” Asian Security, Vol. 15, No. 2 (2019), pp. 103–121, org/10.1080/14799855.2017.1422089.

21. 

Tao Xie et al., “Changgui daodan zhendi peizhi youhua yanjiu” [Research on optimal deployment of conventional missile positions], Xitong Fangzhen Xuebao [Journal of System Simulation], Vol. 21, No. 6 (March 2009), pp. 1535–1537; and Editorial Committee of China's Strategic Missile Force Encyclopedia, Zhongguo zhanlue daodan budui baike quanshu [China's strategic missile force encyclopedia] (Beijing: Zhongguo dabaike quanshu chubanshe, 2012), p. 89.

22. 

The DF-4s are assumed to have no special launch sites; they can be directly launched outside the missile storage tunnels in the forward sites.

23. 

Zhang, Zhongguo Renmin Jiefangjun, Vol. 1, p. 527; Yong Xian, Longxu Xiao, and Gang Li, “Zuhe zhidao dandao daodan wuyituo kuaisu fashe jishu yanjiu” [Research on offhand fast launch technique for integrated guidance ballistic missile], Yuhang Xuebao [Journal of Astronautics], Vol. 31, No. 8 (August 2010), pp. 1915–1919, doi.org/10.3873/j.issn.1000-1328.2010.08.004. It should be noted that this assumption is conservative. There were some reports that China might be able to launch missiles without a preplanned launch pad. Wenting Yang, “Dongfeng diyi zhi: Ji yici fangwei zuozhan yanxi” [First branch of nuclear forces: A defense operational exercise], in Dier paobing zhengzhibu [Political Department of the Second Artillery], ed., Huihuang niandai: Huigu zai gaige kaifang zhong fazhan qianjin de dier paobing [Golden age: A review of the development and progress of the Second Artillery in reform and opening] (Beijing: Zhongyang wenxian chubanshe, 2008), pp. 107–113; and Zhang, Zhongguo Renmin Jiefangjun [The People's Liberation Army of China], Vol. 2 (Beijing: Dangdai zhongguo chubanshe, 1994), p. 118.

24. 

Bill Gertz, “Pentagon Confirms Patrols of Chinese Nuclear Missile Submarines,” Washington Times, December 9, 2015, http://www.washingtontimes.com/news/2015/dec/9/inside-the-ring-chinas-nuclear-missile-submarine-p/; Andrew Tate, “Satellite Imagery Shows Two Chinese SSBNs in Huludao,” Jane's Defence Weekly, November 28, 2018; and Yongxin Xiong and Shulin Li, “Yongyuan buru qizhi de meiming” [Never disgrace the flag's reputation], Jiefangjun Bao [PLA Daily], January 11, 2016.

25. 

Jixun Yu, ed., Di'er paobing zhanyixue [Science of Second Artillery campaigns] (Beijing: Jiefangjun chubanshe, 2004), p. 202.

26. 

Mark A. Stokes, “China's Nuclear Warhead Storage and Handling System” (Washington, D.C.: Project 2049 Institute, March 2010), pp. 2, 12.

27. 

Gates, Nuclear Posture Review Report, pp. 25–27; Vladimir Isachenkov, “Putin: Russia ‘Ahead of Competition’ with Latest Weapons,” Associated Press, October 18, 2018, https://www.apnews.com/35d5162eb0ff42b2b8dfe4bca2e222ef; National Security Strategy and Strategic Defence and Security Review, 2015: A Secure and Prosperous United Kingdom (London: Her Majesty's Government, November 2015), pp. 34–36, https://www.gov.uk/government/publications/national-security-strategy-and-strategic-defence-and-security-review-2015; and President François Hollande, “Speech on Nuclear Deterrence,” Action des Citoyens pour le Désarmement Nucléaire, February 19, 2015, http://www.nuclearfiles.org/menu/key-issues/nuclear-weapons/issues/policies/President-Hollande-Speech-on_a921.pdf.

28. 

See, for example, May, Bing, and Steinbruner, “Strategic Arsenals after START”; and Paul H. Nitze, “Deterring Our Deterrent,” Foreign Policy, No. 25 (Winter 1976–77), pp. 195–210.

29. 

Stokes, “China's Nuclear Warhead Storage and Handling System.”

30. 

Office of the Secretary of Defense, Military Power of the People's Republic of China, 2010 (Washington, D.C.: DoD, 2010), p. 66.

31. 

U.S. National Photographic Interpretation Center (NPIC), New Mobile Solid-Propellant MRBM Under Development, China (Langley, Va.: Central Intelligence Agency [CIA], November 1983), https://www.cia.gov/library/readingroom/docs/CIA-RDP84T00171R000300410001-4.pdf.

32. 

For the values of the detection probabilities used in this model, see the online appendix.

33. 

Eric M. Sepp, “Deeply Buried Facilities: Implications for Military Operations,” Occasional Paper No. 14 (Maxwell Air Force Base, Ala.: Air War College, May 2000).

34. 

Lynn Etheridge Davis and Warner R. Schilling, “All You Ever Wanted to Know about MIRV and ICBM Calculations but Were Not Cleared to Ask,” Journal of Conflict Resolution, Vol. 17, No. 2 (June 1973), pp. 207–242, doi.org/10.1177%2F002200277301700203; and Samuel Glasstone and Philip J. Dolan eds., The Effects of Nuclear Weapons, 3rd ed. (Washington, D.C.: DoD and Energy and Research Development Administration, 1977), pp. 109–117. The details of the kill probabilities are included in the online appendix.

35. 

John Wilson Lewis and Hua Di, “China's Ballistic Missile Programs: Technologies, Strategies, Goals,” International Security, Vol. 17, No. 2 (Fall 1992), p. 24, doi.org/10.2307/2539167; and Defense Intelligence Agency, First Chinese CSS-3 Rollout Site Is Nearly Operational (Washington, D.C.: DoD, July 16, 1982).

36. 

Robert S. Norris and Hans M. Kristensen, “Nuclear U.S. and Soviet/Russian Intercontinental Ballistic Missiles, 1959–2008,” Bulletin of the Atomic Scientists, Vol. 65, No. 1 (February 2009), pp. 62– 69, doi.org/10.2968%2F065001008; and Pavel Podvig, “The Window of Vulnerability That Wasn't: Soviet Military Buildup in the 1970s–A Research Note,” International Security, Vol. 33, No. 1 (Summer 2008), pp. 118–138, org/10.1162/isec.2008.33.1.118.

37. 

See also the online appendix.

38. 

Hans M. Kristensen and Robert S. Norris, “United States Nuclear Forces, 2017,” Bulletin of the Atomic Scientists, Vol. 73, No. 1 (January 2017), pp. 48–57, doi.org/10.1080/00963402.2016.1264213; and Hans M. Kristensen, Matthew McKinzie, and Theodore A. Postol, “How U.S. Nuclear Force Modernization Is Undermining Strategic Stability: The Burst-Height Compensating Super-Fuze,” Bulletin of the Atomic Scientists, March 1, 2017, https://thebulletin.org/2017/03/how-us-nuclear-force-modernization-is-undermining-strategic-stability-the-burst-height-compensating-super-fuze/.

39. 

See also the online appendix.

40. 

Patrick Tyler, A Great Wall: Six Presidents and China (New York: PublicAffairs, 1999), p. 423.

41. 

Office of the Secretary of Defense, Military Power of the People's Republic of China, 2008 (Washington, D.C.: DoD, 2008), p. 3.

42. 

Science and Technology Issues of Early Intercept Ballistic Missile Defense Feasibility (Washington, D.C.: U.S. Defense Science Board, September 2011), p. 27.

43. 

Director, Operational Test and Evaluation, FY2010 Annual Report (Washington, D.C.: DoD, December 2010), p. 227.

44. 

Defense Acquisitions: Production and Fielding of Missile Defense Components Continue with Less Testing and Validation than Planned (Washington, D.C.: U.S. Government Accountability Office [GAO], March 2009), https://www.gao.gov/products/A85559; Missile Defense: Opportunity Exists to Strengthen Acquisitions by Reducing Concurrency (Washington, D.C.: GAO, April 2012), https://www.gao.gov/assets/600/590277.pdf; Laura Grego, “Missile Defense Oversight: Pulling the Punches,” All Things Nuclear blog, April 10, 2014, https://allthingsnuclear.org/missile-defense-oversight-pulling-the-punches/; and Inspector General, Exoatmospheric Kill Vehicle Quality Assurance and Reliability Assessment—Part A (Washington, D.C.: DoD, September 8, 2014), https://media.defense.gov/2014/Sep/08/2001713396/-1/-1/1/DODIG-2014-111.pdf.

45. 

Ballistic Missile Defense Intercept Flight Test Record (Fort Belvoir, Va.: U.S. Missile Defense Agency, September 2019), https://www.mda.mil/global/documents/pdf/testrecord.pdf.

46. 

James D. Syring, “Homeland Defense,” presentation at the 2014 Space and Missile Defense Conference, Huntsville, Alabama, August 13, 2014.

47. 

See also the online appendix.

48. 

Glaser and Fetter, “Should the United States Reject MAD?”

49. 

Ibid.

50. 

Report to Congress on the Defeat of Hard and Deeply Buried Targets (Washington, D.C.: DoD and Department of Energy, July 2001), https://nsarchive2.gwu.edu/NSAEBB/NSAEBB372/docs/Underground-DeeplyBuried.pdf.

51. 

Office of the Secretary of Defense, Proliferation: Threat and Response (Washington, D.C.: DoD, January 2001), p. 90, https://fas.org/irp/threat/prolif00.pdf.

52. 

Committee on the Effects of Nuclear Earth-Penetrator and Other Weapons, Effects of Nuclear Earth-Penetrator and Other Weapons (Washington, D.C.: National Academies Press, 2005), p. 1.

53. 

Sepp, “Deeply Buried Facilities.”

54. 

Chunmei Deng, Jinyu Xu, and Liujun Shen, “B61-11 hezhuandidan qinche baozha xia kengdao koubu pohuai xiaoying fenxi” [Analysis of damage effect of tunnel portal under B61-11 penetration and explosion], Danjian yu Zhidao Xuebao [Journal of Projectiles, Rockets, Missiles, and Guidance], Vol. 28, No. 2 (April 2008), pp. 97–100, org/10.3969/j.issn.1673-9728.2008.02.030.

55. 

China's National Defense in 2008 (Beijing: Information Office of the State Council of the People's Republic of China, January 2009), p. 40.

56. 

Linsuo Si, Tao Wang, and Junhong Zhao, Daodan zhendi anquan guanli yu anquan jishu [Security management and technology of missile sites] (Xi'an, China: Shanxi kexue jishu chubanshe, 2007).

57. 

Jeffrey T. Richelson, America's Secret Eyes in Space: The U.S. Keyhole Spy Satellite Program (New York: Harper and Row, 1990), pp. 190–198.

58. 

NPIC, CSS-4 ICBM Silo Construction, Jingxian Probable SSM Launch Site 1, China (Langley, Va.: CIA, January 1984), https://www.cia.gov/library/readingroom/docs/CIA-RDP84T00171R000300950001-5.pdf.

59. 

NPIC, CSS-4 Upper Silo Configuration, Lushi SSM Launch Site 3, China (Langley, Va.: CIA, December 1984), https://www.cia.gov/library/readingroom/docs/CIA-RDP85T00060R000300610001-4.pdf.

60. 

Scott D. Sagan, “The Perils of Proliferation: Organization Theory, Deterrence Theory, and the Spread of Nuclear Weapons,” International Security, Vol. 18, No. 4 (Spring 1994), pp. 66–107, doi.org/10.2307/2539178; and Scott D. Sagan and Kenneth N. Waltz, The Spread of Nuclear Weapons: An Enduring Debate, 3rd ed. (New York: W.W. Norton, 2013), pp. 41–81.

61. 

Dino A. Brugioni, “The Art and Science of Photoreconnaissance,” Scientific American, March 1996, pp. 78–85, doi.org/10.1038/scientificamerican0396-78; and Dino A. Brugioni, Eyes in the Sky: Eisenhower, the CIA, and Cold War Aerial Espionage (Annapolis: Naval Institute Press, 2010), p. 326.

62. 

NPIC, New Possible CSS-3 Rollout SSM Launch Site Construction, China (Langley, Va.: CIA, January 1981), https://www.cia.gov/library/readingroom/docs/CIA-RDP81T00380R000100840001-0.pdf; and NPIC, New Probable CSS-3 ICBM Development Area, Lushi, China (Langley, Va.: CIA, August 25, 1982), https://www.cia.gov/library/readingroom/docs/CIA-RDP90T00784R000200340003-4.pdf.

63. 

S.M. Hsu and H.K. Burke, “Potential Payoff of Fusion between HSI and Other Sensors,” presentation at the 19th Space Control Conference, Hanscom Air Force Base, Massachusetts, April 4, 2001.

64. 

Richard A. Marcum et al., “Rapid Broad Area Search and Detection of Chinese Surface-to-Air Missile Sites Using Deep Convolutional Neural Networks,” Journal of Applied Remote Sensing, Vol. 11, No. 4 (October–December 2017), pp. 1–31, doi.org/10.1117/1.JRS.11.042614; Jeremy Hsu, “Wanted: AI That Can Spy,” IEEE Spectrum, Vol. 54, No. 12 (December 2017), pp. 12–13; and George Melillos et al., “Detection of Underground Structures Using UAV and Field Spectroscopy for Defence and Security in Cyprus,” Proceedings of SPIE, October 2017, pp. 1–10, org/10.1117/12.2279131.

65. 

Eliot A. Cohen, Gulf War Air Power Survey, Vol. 2, Part 2: Effects and Effectiveness (Washington, D.C.: Office of the Secretary of the Air Force, 1993), pp. 331–332.

66. 

Alan J. Vick et al., Aerospace Operations against Elusive Ground Targets (Santa Monica, Calif.: RAND Corporation, 2001), https://www.rand.org/pubs/monograph_reports/MR1398.html.

67. 

Dwayne A. Day, “Radar Love: The Tortured History of American Space Radar Programs,” Space Review, January 22, 2007, http://www.thespacereview.com/article/790/1; Joseph A. Post and Michael J. Bennett, Alternatives for Military Space Radar (Washington, D.C.: Congressional Budget Office, January 2007); Defense Acquisitions: Space-Based Radar Effort Needs Additional Knowledge before Starting Development (Washington, D.C.: GAO, July 2004), https://www.gao.gov/products/GAO-04-759; DOD Is Making Progress in Adopting Best Practices for the Transformational Satellite Communications System and Space Radar but Still Faces Challenges (Washington, D.C.: GAO, August 2007), https://www.gao.gov/products/GAO-07-1029R; and Defense Acquisitions: Assessments of Selected Weapon Programs (Washington, D.C.: GAO, March 2008), pp. 159–160, https://www.gao.gov/products/GAO-08-467SP.

68. 

“First Folding Space Telescope Aims to ‘Break the Glass Ceiling’ of Traditional Designs” (Arlington, Va.: U.S. Defense Advanced Research Projects Agency [DARPA], December 5, 2013), https://www.darpa.mil/news-events/2013-12-05.

69. 

Broad Agency Announcement: Membrane Optic Imager Real-Time Exploitation (MOIRE) (Arlington, Va.: DARPA, April 16, 2010), p. 5.

70. 

Paul Bracken, “The Intersection of Cyber and Nuclear War,” Strategy Bridge, January 17, 2017, https://thestrategybridge.org/the-bridge/2017/1/17/the-intersection-of-cyber-and-nuclear-war; Long and Green, “Stalking the Secure Second Strike”; and Noah Shachtman, “This Rock Could Spy on You for Decades,” Wired, May 29, 2012, https://www.wired.com/2012/05/spy-rock/.

71. 

Li Bin, “Tracking Chinese Strategic Mobile Missiles,” Science & Global Security, Vol. 15, No. 1 (May 2007), pp. 1–30, org/10.1080/08929880701350197.

72. 

The People's Liberation Army Navy: A Modern Navy with Chinese Characteristics (Washington, D.C.: U.S. Office of Naval Intelligence, August 2009), p. 22; and Wu Riqiang, “Survivability of China's Sea-Based Nuclear Forces,” Science & Global Security, Vol. 19, No. 2 (May 2011), pp. 91–120, org/10.1080/08929882.2011.586312.

73. 

Office of the Secretary of Defense, Military and Security Developments Involving the People's Republic of China, 2014 (Washington, D.C.: DoD, 2014), p. 8.

74. 

Ankit Panda, “China Conducts First Test of New JL-3 Submarine-Launched Ballistic Missile,” Diplomat, December 20, 2018, https://thediplomat.com/2018/12/china-conducts-first-test-of-new-jl-3-submarine-launched-ballistic-missile/.

75. 

Owen R. Cote Jr., “Invisible Nuclear-Armed Submarines, or Transparent Oceans? Are Ballistic Missile Submarines Still the Best Deterrent for the United States?” Bulletin of the Atomic Scientists, Vol. 75, No. 1 (January 2019), pp. 30–35, doi.org/10.1080/00963402.2019.1555998; and Owen R. Cote Jr., “Assessing the Undersea Balance between the U.S. and China,” SSP Working Paper (Cambridge, Mass.: Security Studies Program, Massachusetts Institute of Technology [MIT], February 2011).

76. 

Desmond Ball and Richard Tanter, The Tools of Owatatsumi: Japan's Ocean Surveillance and Coastal Defence Capabilities (Canberra, Australia: ANU Press, 2015), pp. 51–54.

77. 

Cote, “Invisible Nuclear-Armed Submarines, or Transparent Oceans?”

78. 

“Remarks by President Trump and Vice President Pence Announcing the Missile Defense Review” (Washington, D.C.: White House, January 17, 2019), https://www.whitehouse.gov/briefings-statements/remarks-president-trump-vice-president-pence-announcing-missile-defense-review/.

79. 

Office of the Secretary of Defense, 2019 Missile Defense Review (Washington, D.C.: DoD, January 2019).

80. 

Jen Judson, “Lockheed Wins $784M Long-Range Radar Contract,” Defense News, October 22, 2015, https://www.defensenews.com/2015/10/22/lockheed-wins-784m-long-range-radar-contract/.

81. 

Wu Riqiang, “South Korea's THAAD: Impact on China's Nuclear Deterrent,” RSIS Commentary No. 192 (Singapore: S. Rajaratnam School of International Studies, Nanyang Technological University, July 26, 2016), https://www.rsis.edu.sg/rsis-publication/rsis/co16192-south-koreas-thaad-impact-on-chinas-nuclear-deterrent/#.W4QR7S2ZP2J.

82. 

Committee on an Assessment of Concepts and Systems for U.S. Boost-Phase Missile Defense in Comparison to Other Alternatives, Making Sense of Ballistic Missile Defense: An Assessment of Concepts and Systems for U.S. Boost-Phase Missile Defense in Comparison to Other Alternatives (Washington, D.C.: National Academies Press, 2012), pp. 5–6.

83. 

James D. Syring, “Ballistic Missile Defense System Update,” presentation at the Center for Strategic and International Studies, Washington, D.C., January 20, 2016; Gareth Jennings, “MDA Seeks Laser-Armed Hale UAV for Counter-ICBM Role,” IHS Jane's Missiles and Rockets, June 14, 2017; and Office of the Secretary of Defense, 2019 Missile Defense Review, p. 57.

84. 

The beam power of the Airborne Laser was believed to be 3 MW, according to an American Physical Society study. David K. Barton et al., Report of the American Physical Society Study Group on Boost-Phase Intercept Systems for National Missile Defense: Scientific and Technical Issues (College Park, Md.: American Physical Society, 2003), p. 133.

85. 

The APS study concluded that the maximum useful range of the Airborne Laser is 600 kilometers against liquid-propellant ICBMs and 300 kilometers against solid-propellant ICBMs. Ibid., p. 133.

86. 

Jen Judson, “Pentagon Terminates Program for Redesigned Kill Vehicle, Preps for New Competition,” Defense News, August 21, 2019, https://www.defensenews.com/pentagon/2019/08/21/dod-tanks-redesigned-kill-vehicle-program-for-homeland-defense-interceptor/.

87. 

Robert Burns, “Pentagon Cancels Billion-Dollar Missile Defense Project,” Associated Press, August 21, 2019, https://apnews.com/b6d01e01f2a44031a86c7780c835d60a.

88. 

“Homeland Missile Defense System Successfully Intercepts ICBM Target,” press release (Fort Belvoir, Va.: U.S. Missile Defense Agency, May 30, 2017), https://mda.mil/news/17news0003.html.

89. 

“Homeland Missile Defense System Successfully Intercepts ICBM Target,” press release (Fort Belvoir, Va.: U.S. Missile Defense Agency, March 25, 2019), https://mda.mil/news/19news0003.html.

90. 

Stephen D. Weiner, “Ballistic Missile Defense System Analysis,” in William P. Delaney, ed., Perspectives on Defense Systems Analysis: The What, the Why, and the Who, but Mostly the How of Broad Defense Systems Analysis (Cambridge, Mass.: MIT Press, 2015), pp. 145–194.

91. 

Ankit Panda and Vipin Narang, “Deadly Overconfidence: Trump Thinks Missile Defenses Work against North Korea, and That Should Scare You,” War on the Rocks, October 16, 2017, https://warontherocks.com/2017/10/deadly-overconfidence-trump-thinks-missile-defenses-work-against-north-korea-and-that-should-scare-you/; and Glenn Kessler, “Trump's Claim That a U.S. Interceptor Can Knock Out ICBMs ‘97 Percent of the Time,'” Washington Post, October 13, 2017, https://www.washingtonpost.com/news/fact-checker/wp/2017/10/13/trumps-claim-that-u-s-interceptors-can-knock-out-icmbs-97-percent-of-the-time/.

92. 

Office of the Secretary of Defense, Military and Security Developments Involving the People's Republic of China, 2017 (Washington, D.C.: DoD, 2017), p. 60; U.S. Ballistic Missile Defense Organization, “Countermeasure Integration Program: Country Profiles, China” (Washington, D.C.: DoD, April 1995); and Ming Yang, Youlian Chu, and Qi Chai, “Dantou baoluoqiu zai hongwai tufang zhong de yingyong yanjiu” [Warhead enveloping-ball and its application on penetration in infrared wavebands], Hangtian Dianzi Duikang [Aerospace Electronic Warfare], Vol. 25, No. 4 (August 2009), pp. 5–7, org/10.3969/j.issn.1673-2421.2009.04.002.

93. 

Wanxing Zhou, Dandao daodan leida mubiao shibie jishu [BMD radar target recognition technology] (Beijing: Dianzi gongye chubanshe, 2011); and Dejun Feng, “Dandao zhongduan mubiao leida shibie yu pinggu yanjiu” [Study on target recognition and its evaluation in ballistic midcourse], Ph.D. dissertation, National University of Defense Technology, 2006.

94. 

Andrew M. Sessler et al., Countermeasures: A Technical Evaluation of the Operational Effectiveness of the Planned U.S. National Missile Defense System (Cambridge, Mass.: Union of Concerned Scientists and MIT Security Studies Program, April 2000).

95. 

Ankit Panda, “Introducing the DF-17: China's Newly Tested Ballistic Missile Armed with a Hypersonic Glide Vehicle,” Diplomat, December 28, 2017, https://thediplomat.com/2017/12/introducing-the-df-17-chinas-newly-tested-ballistic-missile-armed-with-a-hypersonic-glide-vehicle/; Zhen Liu, “China Takes Step towards Precision Warheads for Unstoppable Nuclear Weapon, State Media Says,” South China Morning Post, September 29, 2018, https://www.scmp.com/news/china/military/article/2166298/china-takes-step-towards-precision-warheads-unstoppable-nuclear; and Bill Gertz, “China Successfully Tests Hypersonic Missile,” Washington Free Beacon, April 27, 2016, https://freebeacon.com/national-security/china-successfully-tests-hypersonic-missile/.

96. 

Kristin Huang, “China's Hypersonic DF-17 Missile Threatens Regional Stability, Analyst Warns,” South China Morning Post, August 23, 2019, https://www.scmp.com/news/china/military/article/3023972/chinas-hypersonic-df-17-missile-threatens-regional-stability.

97. 

James M. Acton, “China's Advanced Weapons,” testimony before the U.S.-China Economic and Security Review Commission, https://carnegieendowment.org/2017/02/23/china-s-advanced-weapons-pub-68095.

98. 

The DF-41 flight tests had demonstrated the use of two MIRVed warheads. Bill Gertz, “China Tests New Long-Range Missile with Two Guided Warheads,” Washington Free Beacon, August 18, 2015, https://freebeacon.com/national-security/china-tests-new-long-range-missile-with-two-guided-warheads/; and Bill Gertz, “China Flight Tests Multi-Warhead Missile,” Washington Free Beacon, December 11, 2015, https://freebeacon.com/national-security/china-flight-tests-multi-warhead-missile/.

99. 

Catherine Dill, “Counting Type 094 Jin-Class SSBNs with Planet Imagery,” Arms Control Wonk, November 21, 2018, https://www.armscontrolwonk.com/archive/1206320/counting-type-094-jin-class-ssbns-with-planet-imagery/; and Patrick Tucker, “China Has More Nuclear Subs than the West Believed,” Defense One, November 20, 2018, https://www.defenseone.com/technology/2018/11/china-has-more-nuclear-subs-west-believed/152984/.

100. 

For detailed parameters of the model, see the online appendix.

101. 

Zedong Mao, Mao Zedong waijiao wenxuan [Mao Zedong's selected works on diplomacy] (Beijing: Zhongyang wenxian chubanshe, 1994), pp. 57–62.

102. 

Li Bin, “China's Potential to Contribute to Multilateral Nuclear Disarmament,” Arms Control Today, March 3, 2011, https://www.armscontrol.org/act/2011_03/LiBin.

103. 

Zhongyang junwei bangongting, ed., Deng Xiaoping guanyu xin shiqi jundui jianshe lunshu xuanbian [Deng Xiaoping's essay on military construction in the new era] (Beijing: Bayi chubanshe, 1993), pp. 44–45.

104. 

Nie Rongzhen, Nie Rongzhen Yuanshuai huiyilu [Marshal Nie Rongzhen's memoirs] (Beijing: Jiefangjun chubanshe, 2005), p. 645.

105. 

Denis Healey, The Time of My Life (Harmondsworth, U.K.: Penguin, 1990), p. 243.

106. 

McGeorge Bundy, “The Bishops and the Bomb,” New York Review of Books, June 16, 1983, https://www.nybooks.com/articles/1983/06/16/the-bishops-and-the-bomb/.

107. 

Wu, “Certainty of Uncertainty.”

108. 

U.S. Director of Central Intelligence, National Intelligence Estimate: China's Strategic Attack Programs, NIE 13-8-74 (Langley, Va.: CIA, June 13, 1974).

109. 

Gates, Nuclear Posture Review Report, pp. 28–29. It should be noted that this is a conservative assumption for China; the criterion of first-strike uncertainty for stable deterrence could be even lower. As mentioned, only those Chinese missiles that could target the adversaries’ heartland are considered in this study. The online appendix shows the results of the calculation considering retaliatory strikes against the whole territory of the Soviet Union and the United States, including the eastern part of the Soviet Union, Hawaii, Alaska, and Guam. Understandably, the probability of retaliation is higher than that in previous analysis. In 2000, China also had dozens of MRBMs (DF-21 and DF-3) that could target the U.S. oversea military bases in the Western Pacific. The extent to which those retaliatory strikes could deter nuclear attack or coercion is an open question. See the online appendix.

110. 

Long and Green, “Stalking the Secure Second Strike”; Green and Long, “The MAD Who Wasn't There”; Lieber and Press, “The New Era of Counterforce”; and Lieber and Press, “The End of MAD?” p. 7.

111. 

For a summary of the debate on nuclear superiority and nuclear coercion, see Matthew Kroenig, “Nuclear Superiority and the Balance of Resolve: Explaining Nuclear Crisis Outcomes,” International Organization, Vol. 67, No. 1 (Winter 2013), pp. 141–171, doi.org/10.1017/S002081831 2000367; and Todd S. Sechser and Matthew Fuhrmann, “Crisis Bargaining and Nuclear Blackmail,” International Organization, Vol. 67, No. 1 (Winter 2013), pp. 173–195, org/10.1017/S0020 818312000392.

112. 

Hans M. Kristensen, “Alert Status of Nuclear Weapons (Version 2),” presentation at the Elliot School of International Affairs, George Washington University (Washington D.C.: Federation of American Scientists, April 21, 2017), https://fas.org/wp-content/uploads/2014/05/Brief2017_GWU_2s.pdf.

113. 

Caroline Gammell and Thomas Harding, “British and French Nuclear Submarine Collision ‘as Serious as Sinking of Kursk,'” Telegraph, February 16, 2009, https://www.telegraph.co.uk/news/uknews/defence/4640673/British-and-French-nuclear-submarine-collision-as-serious-as-sinking-of-Kursk.html; and Scott D. Sagan, The Limits of Safety: Organizations, Accidents, and Nuclear Weapons (Princeton, N.J.: Princeton University Press, 1993), pp. 156–203.

114. 

Isachenkov, “Putin: Russia ‘Ahead of Competition’ with Latest Weapons”; and Jeffrey Lewis, “Is Launch Under Attack Feasible?” (Washington, D.C.: Nuclear Threat Initiative, August 24, 2017), https://www.nti.org/analysis/articles/launch-under-attack-feasible/.

115. 

Geoffrey Forden, Pavel Podvig, and Theodore A. Postol, “False Alarm, Nuclear Danger,” IEEE Spectrum, Vol. 37, No. 3 (March 2000), pp. 31–39, doi.org/10.1109/6.825657; and Robert Kazel, “Ex-Chief of Nuclear Forces General Lee Butler Still Dismayed by Deterrence Theory and Missiles on Hair-Trigger Alert” (Santa Barbara, Calif.: Nuclear Age Peace Foundation, May 27, 2015), https://www.wagingpeace.org/general-lee-butler/.

116. 

Henry A. Kissinger and Brent Scowcroft, “Nuclear Weapon Reductions Must Be Part of Strategic Analysis,” Washington Post, April 22, 2012, https://www.washingtonpost.com/opinions/nuclear-weapon-reductions-must-be-part-of-strategic-analysis/2012/04/22/gIQAKG4iaT_story.html.

117. 

National Security Strategy and Strategic Defence and Security Review, 2015, pp. 34–36.

118. 

William J. Perry and Brent Scowcroft, U.S. Nuclear Weapons Policy, Independent Task Force Report No. 62 (New York: Council on Foreign Relations, 2009); Robert L. Pfaltzgraff Jr., “China-U.S. Strategic Stability,” presentation at the Carnegie Endowment for International Peace conference, Washington, D.C., April 6–7, 2009; and PONI Working Group on U.S.-China Nuclear Dynamics, Nuclear Weapons and U.S.-China Relations: A Way Forward (Washington, D.C.: Center for Strategic and International Studies, March 2013).

119. 

Eric Heginbotham et al., China's Evolving Nuclear Deterrent: Major Drivers and Issues for the United States (Santa Monica, Calif.: RAND Corporation, 2017), p. xi, https://www.rand.org/pubs/research_reports/RR1628.html. See also Li Bin and Hu Gaochen, “Meiguo shiyu zhong de zhongguo heweishe youxiaoxing” [The effectiveness of China's nuclear deterrence in the U.S. perspective], Waijiao Pinglun [Foreign Affairs Review], No. 5 (2018), pp. 21–41, org/10.13569/j.cnki.far.2018.05.021.

120. 

Michael Glosny and Christopher Twomey, U.S.-China Strategic Dialogue, Phase V, 2010 (Monterey, Calif.: Naval Postgraduate School, 2010).