An Analysis of the Proposed US NMD System
A version of this paper was published in the Winter 1999-2000 issue of Foreign Policy under the title "National Missile Defense: An Indefensible System"
by George Lewis, Lisbeth Gronlund, and David Wright
In the fall of 2000, President Clinton will decide whether the United States should begin deploying a national missile defense (NMD) system intended to protect the entire country from limited attacks by intercontinental ballistic missiles (ICBMs) armed with nuclear, biological, or chemical warheads. Such attacks, ranging from a few to a few tens of missiles, fall into three categories: a small accidental or unauthorized launch from Russia, a deliberate or unauthorized attack from China, or a deliberate attack from a hostile emerging missile state that might acquire ICBMs. This last threat -- focused on Iran, Iraq, and North Korea -- has emerged as the primary argument for a near-term NMD deployment.
The planned national missile defense is more down-to-earth (literally and figuratively) than former President Ronald Reagan's Strategic Defense Initiative (SDI), which was intended to create a space-based shield against a massive Soviet nuclear attack. In contrast, the current system would use ground-based interceptors to destroy incoming warheads by colliding with them. Indeed, because it is intended to counter a more limited threat than the SDI system, and because it would use existing rather than speculative technology, many assume that, unlike SDI, the NMD system would work against the threat at hand. Moreover, although many observers agree that SDI deployment might well have aggravated the US-Soviet arms race, they assume that, with the end of the cold war and Russia's declining economy, deploying the NMD system will result in few security costs.
However, neither assumption about the security benefits and costs is warranted. Although it is now technically feasible to "hit a bullet with a bullet" on the test range, adversaries would be able to take straightforward steps to defeat this system, not only preventing it from achieving the high levels of effectiveness claimed for it, but also precluding any significant security benefits. Worse still, deploying such a system would open a Pandora's box of problems for the United States, unraveling decades of efforts to reduce US and Russian nuclear stockpiles and to limit proliferation of nuclear weapons and ballistic missiles worldwide.
History of the Planned System
Although the SDI program that began 10 years earlier was effectively dead when Clinton took office in 1993, the idea of deploying a nationwide missile defense lived on within certain quarters of the Republican Party. When the Republicans took control of Congress a year and a half later, it re-emerged as a potent political issue.
In 1995, Congress passed legislation mandating deployment of a limited national missile defense system by 2003. Clinton vetoed the legislation, arguing that there was no threat justifying NMD deployment. However, in the face of continuing congressional pressure, the administration announced in 1996 its "3+3" program, which would first develop and then possibly deploy an NMD system in two phases of three years each. If, after the first three years, no threat justified deployment, then development would continue so that the system would always be three years from deployment with up-to-date technology. The 3+3 program assumed that there would be at least three years' warning of ICBM deployment by an emerging missile state, and for several years the program helped deflect continued Republican pressure to commit to a national missile defense.
However, two events in the summer of 1998 fundamentally altered the missile defense debate. First, in July, the high-level, bipartisan Commission to Assess the Ballistic Missile Threat to the United States (known as the Rumsfeld Commission after its chair, former Secretary of Defense Donald Rumsfeld) released a report concluding that North Korea or Iran could develop an ICBM within five years and with little warning. Then, in August, North Korea launched a three-stage Taepo-Dong I missile over Japan in an attempt to put a satellite into orbit. These developments significantly enhanced the perception of a threat from emerging missile states and undercut the administration's key argument against an immediate decision to deploy an NMD system -- namely, that the United States would have adequate warning before deployment.
In January 1999, the administration announced that it would make a deployment decision in the summer of 2000 based on four criteria: the missile threat, the cost of the system, the strategic and arms control implications of deployment, and the readiness of the technology. (The date of the decision has now slipped to fall 2000.) Acknowledging that the system was being implemented faster than the technology permitted, the administration pushed back the target deployment date from 2003 to 2005.
Nonetheless, Republicans reintroduced legislation requiring deployment as soon as possible. After Clinton indicated that he would not veto such a bill if language on arms reductions and funding were included, the legislation passed by wide margins. The National Missile Defense Act of 1999, signed by Clinton in July, states that it is US policy to deploy a limited NMD system "as soon as technologically feasible."
Current administration plans call for a Deployment Readiness Review to be held in July 2000 [now moved to late summer]. This review will be based largely on the results of three intercept tests. The first of these tests took place in early October 1999 and intercepted a mock warhead, although problems with the test that were revealed later call into question how to interpret the hit. The second test, in January 2000, was not successful. If the technology is deemed ready, the president will make a deployment decision, taking into account the other three criteria.
It is possible that this schedule could slip, but it cannot slip far if the 2005 deployment date is to remain an option. According to the administration, to complete deployment by 2005, the United States would need to begin construction activities in spring 2001 that are prohibited by the Anti-Ballistic Missile (ABM) Treaty, signed by the United States and the former Soviet Union in 1972. Because this agreement bans deployment of a nationwide missile defense, the United States has recently opened negotiations with Russia to modify the treaty.
However, Russian public statements have been uniformly critical of changing the treaty, and several meetings between US and Russian authorities have yet to yield any progress. At the same time, administration officials have made it clear that if the United States committed to an NMD system, Russian objections would not delay deployment, and the United States would exercise its option to withdraw from the ABM Treaty. If Russia did not agree to the US-proposed treaty changes, then to meet the 2005 deployment date the United States would need to give its required six months' notice of intent to withdraw in the fall of 2000.
The Proposed System
Although the United States will not make a deployment decision until next year, it now has a specific NMD system on the table. This system would use ground-based interceptors topped with an Exoatmospheric Kill Vehicle (EKV) that is designed to destroy an incoming warhead by colliding with it at high speed.
Existing or new early-warning satellites would first detect the launch of an attacking missile. The NMD system would then use different sensors to detect the missile and any objects it releases, track these objects accurately enough to guide the interceptors, and attempt to discriminate between the real warhead and decoys or other false targets. These sensors include five existing early-warning radars in California, central Alaska, Great Britain, Greenland, and Massachusetts, which would be upgraded to enable them to track targets accurately enough to guide interceptors. In addition, new radars designed specifically for NMD and with much greater discrimination capabilities would be deployed. Roughly 24 missile-tracking satellites in low earth orbit would supplement these ground-based radars.
Using the information from these sensors, the system would launch interceptors and guide them toward their targets. As each interceptor nears its assigned target, it would release the EKV, which would use infrared and visible light sensors to detect the target and attempt to discriminate it from decoys or other false targets, and then use thruster rockets to steer itself into the target. To increase the odds of success, the system would likely fire several interceptors at each target. To conserve interceptors, if time permits, the defense would use a shoot-look-shoot strategy, in which one or more interceptors are initially fired at the target. After observing the results of the intercept attempts, the system would fire additional interceptors if necessary.
Under current plans, the NMD system would be deployed in three phases. The first phase, referred to as Capability 1 (C1), was originally designed to counter a threat composed of a "few" (roughly five) warheads with either "simple" or no countermeasures. However, the administration recently announced that this phase would now involve the deployment of up to 100 interceptors and would be aimed at countering "a few tens of warheads." This phase would upgrade the early-warning radars and deploy one new radar. Because North Korea is perceived to be the most imminent missile threat, the interceptors and radar would be deployed in Alaska.
The second phase, known as C2 and aimed at countering an attack by warheads with more complex countermeasures, would deploy additional radars and enough interceptors to bring the total to 100 or more, as well as the missile-tracking satellite system, if it is ready for deployment. The C3 phase is intended to counter threats consisting of "many, complex" warheads; it would deploy additional radars as well as additional interceptors, including some at a second site, bringing the total to 200 or more. Although the C3 system is the current deployment goal, the system design explicitly permits further expansion and upgrades beyond the C3 level by, for example, deploying additional interceptors, increasing the number of interceptor sites, or deploying space-based weapons. Moreover, a recent Pentagon study concluded that the NMD system could be upgraded by integrating the hundreds of interceptors to be deployed as part of the ship-based Navy Theater Wide missile defense system. These interceptors would be plugged in to the sensor infrastructure of the NMD system.
How much will this system cost? In April 2000, the Pentagon estimated that the cost of building and operating the phase 1 system, which would include 100 interceptors based in Alaska, from 1991 to 2026 would be $30.2 billion. The full C3 system is expected to cost between $50 and 60 billion. As for virtually all major defense programs in such an early phase of development, the actual costs are likely to be considerably higher.
Will It Work?
Before assessing how well the NMD system can achieve its stated objectives, it is useful to review some threats that the system is not designed to counter.
First, ICBMs are not the only -- nor perhaps the earliest, most likely, or most effective -- method available to emerging missile states for delivering weapons of mass destruction to US territory. As the Rumsfeld Commission noted, it would be easier for an emerging missile state to develop shorter-range, ship-launched ballistic missiles than ICBMs. Because such missiles have short flight times and low maximum altitudes, they could not be intercepted by the NMD system. In addition, a nuclear, chemical, or biological weapon could be delivered by cruise missiles, civilian ships, civilian airplanes, or even by being smuggled across the US border.
Second, and more relevant to its stated purpose, the NMD system cannot defend against ICBMs carrying biological or chemical agents packaged in submunitions. To be effective, biological and chemical agents must be dispersed across a large area; for delivery by ICBM this is best accomplished by subdividing the agent into dozens or hundreds of small submunitions that would be released early in the missile's flight. Such submunitions would also render the planned NMD system ineffective because there would be too many targets to intercept.
While these are significant limitations, the planned NMD could still make an important contribution to US security if it could protect against the remaining threat -- a limited number of nuclear weapons (which cannot be divided into large numbers of submunitions) delivered by ICBMs. How effective could a US policy maker expect the NMD system to be against this threat?
Given today's technology, the United States can certainly build a system that could destroy one or several ICBM warheads under controlled conditions, in which the characteristics of the target warhead are known and no serious effort is made to defeat the defensive system. However, the essential question is whether the system will be operationally effective -- that is, whether the defense would be effective in the real world, where the characteristics of the attack would not be completely known in advance and where the attacker would take steps to defeat the defense. The central issue in determining operational effectiveness will be how well the system can deal with countermeasures intended to defeat the system.
The NMD system would employ interceptors that operate above the earth's atmosphere. In the vacuum of space, there is no atmospheric drag, and both heavy objects such as warheads and much lighter objects travel on identical ballistic trajectories. This fact, and other characteristics of the space environment, make possible a wide range of relatively simple but effective countermeasures against the planned NMD system. Such countermeasures could be used alone or in many combinations.
One often discussed countermeasure is lightweight replica decoys. The attacker would release large numbers of these objects along with the real warhead. Unless the defensive system could determine which of the objects was the actual target, it would have no choice but to fire at all of them, which would quickly exhaust its supply of interceptors. Otherwise, it would risk letting the warhead penetrate unchallenged. One way to make replica decoys would be to use warhead-shaped balloons with a thin metal coating on their outer surfaces to reflect radar waves. Such decoys might also contain heaters to simulate the heat emission of the warhead and small weights to give spinning and tumbling motions similar to that of the warhead.
Less discussed, but likely more effective, is a technique known as anti-simulation. Rather than make all the decoys look like the warhead, the attacker would disguise the warhead. For example, the attacker could cover the warhead with irregularly shaped pieces of metallic foil to make it look like a piece of missile debris. Alternatively, the warhead could be released inside a metal-coated balloon, along with large numbers of similar, but empty balloons. These balloons would differ slightly in size -- and by varying their shapes and surface coatings, they would have different equilibrium temperatures. The defense would then face the nearly impossible task of telling which of them was the real target. As an alternative to decoys, the warhead could be enclosed within a thin metallic shroud that would be cooled with a small amount of liquid nitrogen. This technique could reduce the range at which an infrared-homing kill vehicle could detect the warhead by a factor of at least a thousand, in effect blinding the kill vehicle, and thus defeating the defense.
Many NMD proponents argue that countermeasures are hard to develop and implement. In fact, the initial phase 1 deployment is aimed only at a threat that incorporates no effective countermeasures. However, there are many countermeasures that are much easier to build and deploy than either an ICBM or a nuclear warhead small enough to be delivered on an ICBM -- in other words, countermeasures that would clearly be within the means of any emerging missile state that could build an ICBM. As noted by the 1999 National Intelligence Estimate (NIE) on the ballistic missile threat, which represents the consensus view of US intelligence agencies, such countries "probably would rely initially on readily available technology" to develop countermeasures and could do so "by the time they flight test their missiles." The NIE also states that "Russia and China each have developed numerous countermeasures and probably are willing to sell the requisite technologies."
The NMD system reportedly has a design requirement of 95 percent effectiveness with 95 percent confidence against a small-scale missile attack. Effectiveness is a measure of how well a system works in the real world, which can only be determined through extensive testing or use of the system. The confidence level describes how much trust the user has in the effectiveness of the system, based on prior testing and use. In each case, the stated "95 percent" goal is unrealistic. Even aside from the countermeasure problem, an effectiveness of 95 percent is rarely -- if ever -- achieved by a military weapons system, even after years of use. An additional problem is that an NMD system must work the first time it is actually used. If a nuclear ICBM attack occurs, there will be no opportunity to learn on the job.
Because the real-world operating environment could vary greatly depending on the types and combinations of countermeasures the attacker uses, achieving a 95 percent confidence level would require hundreds of tests conducted under different scenarios, costing billions of additional dollars (the Pentagon reported that the first test cost $100 million). If the tests do not adequately approximate the conditions under which the system would operate, then even a large number of successful tests will provide little meaningful information about the system's operational effectiveness. Worse, such tests could encourage a false sense of confidence.
How does the Pentagon's planned testing program measure up? The three tests that will take place before the Deployment Readiness Review will not even begin to address the question of how well the system would work in the real world. These trials will be limited to demonstrating the basic functioning of the system in a relatively benign test environment.
An additional 16 intercept tests are planned before the target deployment date of 2005. However, according to the FY99 report of the Pentagon's Director of Operational Testing and Evaluation (DOTE), none of these tests will attempt to assess the operational effectiveness of the system against real-world countermeasures. All of these tests will instead test the capability-1 level system. For realistic tests, the countermeasures used should be designed by a truly independent group, and the intercept tests should be conducted without the operators knowing in advance what countermeasures they would face. There is no indication that such a testing program is even under consideration.
Given the extremely demanding operational environment the NMD system will face, and given the need for it to work the first time it is actually used, it is implausible that the system will even approach the high levels of effectiveness claimed for it. Moreover, the inadequate test program means that the United States will not have high confidence in the system prior to its use.
Proposed Missions
Supporters of national missile defense envision three main missions, each of which would impose different requirements on system effectiveness and the level of user confidence in that effectiveness:
Mission One: Preserve US Freedom of Action
Some NMD advocates argue that a nuclear-missile-armed Iran or North Korea might blackmail the United States into not taking some military action it otherwise would, and that deploying an effective NMD system would deter such coercion because the United States could risk an attack if it were able to shoot down incoming ICBMs. In effect, this mission seeks to deny emerging missile states and China a nuclear deterrent against US conventional action.
This mission requires that US policy makers have a very high level of confidence that the NMD system will be highly effective. Would a US president be willing to take an action that could lead to a nuclear ICBM attack on a major US city even if he or she were 95 percent confident that the NMD system would be 95 percent effective? And if even that risk seems unacceptable, consider that 95 percent effectiveness is unrealistically optimistic, because such performance is not achievable in a world with countermeasures. Confidence in the effectiveness of the planned NMD system would not be high enough to increase US freedom of action beyond the level already achieved through deterrence. In fact, deploying an NMD system to preserve US freedom of action might induce the United States to undertake riskier ventures than it otherwise would, thus increasing the risk of a nuclear attack.
Underlying the concern over nuclear blackmail is the notion that emerging missile states such as North Korea or Iraq may be "undeterrable" -- that is, their leadership may not be sufficiently rational to be deterred by the prospect of US nuclear retaliation. There is little, if any, evidence in support of such an argument. However isolated and repressive these regimes may be, the one thing their leaders value highly and have been sufficiently rational to maintain -- their hold on power -- is precisely what would be lost in a US retaliatory attack.
Mission Two: Deter Development and Deployment of ICBMs
Some NMD proponents argue that deploying an NMD system would deter emerging missile states from developing or deploying missiles capable of reaching the United States, because these countries would see the futility of doing so. This logic runs counter to experience in deploying air defenses, which have not deterred deployment of fighter aircraft or bombers.
Would this dynamic be different in the case of ICBMs? As the 1999 National Intelligence Estimate concludes: "Countries with emerging ICBM capabilities are likely to view their relatively few ICBMs more as weapons of deterrence and coercive diplomacy than as weapons of war . . . ." Thus, an emerging missile state is unlikely to impose stringent requirements on the effectiveness of its ICBMs, because deterrence would not require the certainty of a successful attack -- only the real possibility. (Emerging missile states are already willing to deploy relatively unreliable missiles, as indicated by the very limited development and testing programs identified by the Rumsfeld Commission.) The deployment of countermeasures, either developed indigenously or purchased abroad, would introduce enough uncertainty about the effectiveness of the defense to accomplish this objective. In fact, rather than discouraging proliferation, NMD deployment may have the opposite effect: Emerging missile states might deploy more ICBMs to maintain their deterrent and coercive potential by making it clear that they can penetrate the defense.
Mission Three: Limit Damage
Finally, some NMD supporters, pointing to the catastrophic damage that a single nuclear warhead explosion could cause, argue that even a partially effective NMD system might prevent or reduce damage in the event of an attack. In other words, when many thousands of lives could be at stake, any defense is better than none.
Although it is true that even a defense of unknown or low effectiveness could limit damage, this benefit would likely be more than offset by the increased risk of an ICBM attack that an NMD deployment may produce. The NMD system would provoke deep suspicion in Russia and China, likely increase the risk of a Russian accidental or unauthorized launch, and undermine existing nonproliferation efforts worldwide.
Security Costs of Deployment
Advocates of the NMD system focus on its possible security benefits. Building such a system, however, will also incur far-reaching security costs.
Despite the much-trumpeted end of the cold war, the United States and Russia continue to rely on nuclear deterrent policies based on deploying large numbers of nuclear-armed missiles ready for immediate launch. If the United States and Russia should ever decide to abandon mutual deterrence, then their concerns about national missile defenses and the ABM Treaty would become anachronistic and irrelevant. But until this happens -- which appears highly unlikely in the near future -- US deployment of an NMD system that Russia believes could undermine its deterrent will almost certainly provoke a reaction that will undermine US security.
If the proposed system is likely to be ineffective, why should Russia feel threatened? Although Russian scientists understand that countermeasures would defeat the system, Russian policy makers may not have such confidence and are instead likely to wonder why the United States would pour billions of dollars into an ineffective system. In addition, Russian leaders will realize that US actions will be based not on the system's actual effectiveness, but on US perceptions of its effectiveness. Finally, a US NMD system would create uncertainties against which Russian military planners would want to hedge.
Many influential military and political leaders in Russia strongly oppose US deployment, and their position has hardened since Clinton signed the National Missile Defense Act, which they interpret as a decision to deploy regardless of Russian concerns. They warn that the United States' disregard of Russian strategic interests seriously undermines relations between the two countries, especially following the expansion of nato and the bombing of Kosovo -- cases in which Russia also believes its concerns were ignored. Some warn that US NMD plans could become an important issue in the June 2000 presidential election in Russia and lead to the election of a hard-liner.
Many Russian officials find unconvincing US reassurances that the planned system is not aimed at Russia. Even as the United States builds its missile defense, it will retain large numbers of highly accurate ICBMs and submarine-launched ballistic missiles (SLBMs), which have considerable first-strike capabilities against Russia's nuclear forces. Of particular concern are the highly accurate Trident II SLBMs, which could potentially exploit gaps in the crumbling Russian early-warning system to attack and destroy even heavily hardened targets with little or no warning. US nuclear-powered attack submarines continue to operate near Russian ballistic missile submarine bases, posing a direct threat to the few missile submarines Russia is able to maintain at sea at any given time. In the event of a future deterioration of US-Russian relations, Russian planners could find themselves with relatively small numbers of warheads capable of surviving a US attack. From this perspective, the planned C3 phase of the NMD, with its 200 to 250 interceptors, could appear as a direct threat to Russia's retaliatory capability.
Moreover, although the United States insists its NMD system will be limited, the large set of sensors planned would lay the base for a much more extensive system, which could be expanded relatively rapidly by increasing the number of ground-based interceptors or by incorporating sea-based interceptors, such as those in the Navy Theater Wide system.
It is for this reason that deployment of the planned NMD system is fundamentally inconsistent with the purpose of the ABM Treaty and will require much more than the minor treaty changes the Clinton administration claims are needed. The treaty, a centerpiece of the US-Russian security relationship for over 25 years, provides two essential security guarantees: that a nationwide defense would not be deployed and that an infrastructure that could be used to rapidly deploy or expand such a nationwide defense would not be put in place. Not only is the NMD system designed to provide nationwide coverage, but it would deploy up to 10 new radars around the world, in addition to the five upgraded early-warning radars, that would constitute the infrastructure for a much larger defense. Thus, even if Russia agrees to modify the treaty, it will be a document with little strategic value.
How might Russia respond? Russian officials have indicated that if the United States deployed an NMD system, Russia would continue to field land-based missiles with multiple warheads -- rejecting a key provision of the second Strategic Arms Reduction Treaty, (START II) -- and would deploy countermeasures on its missiles to ensure that those missiles could penetrate the NMD system. This step would also nullify any capability the system would have to defend against a small accidental or unauthorized Russian missile launch.
But this is unlikely to be Russia's only response. To maximize its number of survivable missiles, Russia would probably resist "de-alerting" measures to reduce the rapid-launch status of its nuclear forces and might even increase the fraction of its missiles on high-alert status. Such steps would have the side effect of increasing the risk of an accidental or unauthorized Russian missile launch, which arguably presents the greatest threat to US security.
Moreover, Russian officials say that they continue to see a link between offensive and defensive weapons. So while Russia's nuclear arsenal is expected to decrease as a result of economic constraints, a US NMD deployment would likely block cuts to levels below about 1,000 warheads, raising the possibility that the United States and Russia would become locked into unnecessarily high levels of nuclear weapons for the indefinite future. Russia might also emphasize shorter-range tactical nuclear weapons, which could be used to threaten US allies or forward-deployed US forces, reversing the progress made over the last decade in withdrawing these weapons from deployment. And Russian officials have already suggested that even the existing transparency measures and inspection arrangements with the United States could fall victim to a more antagonistic relationship between the two countries.
China may be even more alarmed than Russia. China's small deterrent force of roughly 20 ICBMs would be directly threatened by even the first phases of the NMD system. How then might Beijing respond?
Recent Chinese missile tests have included countermeasures, which China is likely to deploy in the future. But the uncertainties raised by the deployment of a US missile defense and the increased strain on US-Chinese relations would likely strengthen the position of those in China who want to build up its nuclear forces. China clearly is capable of expanding its offensive forces in the next few decades by building more missiles and deploying multiple-warhead missiles to overwhelm an NMD. While some modernization of China's missile forces is likely in any event, deployment of a US missile defense will influence both the timing and scope of those efforts. To preserve its option to enlarge its nuclear arsenal, China might also refuse to participate in a fissile material production cutoff.
The unavoidable deterioration of relations with Russia and China will also undermine US nonproliferation goals. Progress on nonproliferation requires cooperation with both countries, which have economic incentives to sell military technology and provide technical assistance abroad. However, both countries will be less willing to participate in efforts to limit proliferation of ballistic missiles if they feel that the United States is ignoring their security concerns.
Russian and Chinese responses to a US NMD deployment could have a significant ripple effect worldwide. For example, a Chinese nuclear buildup could trigger buildups in India and Pakistan. More generally, the demise of the ABM Treaty, the halting of US-Russian nuclear reductions, and the deterioration of the US-Russian and US-Chinese security relationships cannot help but threaten the viability of the Nuclear Non-Proliferation Treaty (NPT), which is, at least in principle, predicated on continuing arms reductions by the nuclear weapons states. (In this regard, the US Senate's recent rejection of the Comprehensive Test Ban Treaty, if not ultimately reversed, will only compound problems for the NPT). European leaders have warned that a US decision to deploy a national missile defense system would undermine the concept of "shared risk" that has held together the nato alliance for five decades.
In the medium to long run, the price of a national missile defense system deployed by the United States may well be a world with more ICBMs and weapons of mass destruction. Compared with these large and nearly certain security costs, the benefits the planned NMD system would provide are both too small and too uncertain to justify its deployment.
George Lewis is associate director of the Security Studies Program at the Massachusetts Institute of Technology (MIT). Lisbeth Gronlund and David Wright are senior staff scientists at the Union of Concerned Scientists and research fellows at MIT's Security Studies Program.
