“It will take all of us working together – government officials, and diplomats, academic experts, and scientists, activists, and organizers – to come up with new and innovative approaches to strengthen transparency and predictability, reduce risk, and forge the next generation of arms control agreements.”
Understanding the Extraordinary Cost Of Missile Defense
December 2000
Ballistic missile defenses, particularly national missile defenses, have always been one of the most divisive issues in national security. From the anti-ballistic missile (ABM) debates of the late 1960s over countering the growing Soviet threat to the current debate over whether to deploy a limited national missile defense (NMD) system to protect the United States from attacks by emerging missile states, the issue has cut to the heart of questions about nuclear deterrence and strategic stability.
Cost has been one factor that has galvanized this debate. It is often the anvil upon which the success or failure of a missile defense scheme (or any other weapons system) is forged. Cost is never the sole reason why a system is deployed or scuttled, nor should it be. But it is a hurdle, a reality check, that any proposed system must pass to survive.
If there is a grave threat to the United States that missile defenses could address effectively and if deploying them would improve U.S. security writ large, a national missile defense system would be deployed even at great expense. But if the threat is not compelling enough or the strategic rationale is not perceived as clearly benefiting national security, the tendrils of congressional oversight and competition for resources within the Pentagon will wrap around the program and gradually squeeze the life out of it.
The 1970s Safeguard ABM system is a good example. The Army successfully developed and then deployed the system in North Dakota at a cost of $23 billion (in year 2000 dollars), excluding the cost of developing and building the nuclear warheads.1 Its mission: to protect an ICBM field from Soviet attack so that its missiles could be used in a retaliatory strike. The system was fully consistent with the consensus strategic concept of the day and even with the ABM Treaty. And yet it survived only four months before the Department of Defense shut it down because it was too expensive to operate, given what became to be perceived as its marginal contribution to U.S. security.
Budget battles have a disciplining effect—a program that is perceived as weak, either because of technical problems or lack of high-level support within the executive branch or Congress, will be tripped up. Budgets will be trimmed or appropriations redirected, slowing the program down until it proves itself to be stronger. Leverage to slow a program comes from one of the ironclad laws of research and development: it takes money to fix technical problems. So even if a program manages to shake off attempts to cut it, it may not get the extra resources it needs to solve the problems and remain on schedule.
The Theater High Altitude Area Defense (THAAD) program is a perfect example of this. Despite the compelling need to protect U.S. forces against an existing theater ballistic missile threat and despite the full weight of the Army and the secretary of defense, a series of test failures led to cuts that delayed the program by several years. The program got "healthy" by hitting two of its targets. Budgets rose and the testing program was quickly ended. THAAD is now enjoying broad support during the early phases of engineering development, but, if it encounters several test failures during the next round of flight tests, it will be slowed again.
Clearly, costs will continue to play an important role in the success or failure of missile defense programs. To date, however, budget hurdles have not been kind to missile defense programs, tripping them up in large part because their costs have continued to rise, seemingly without end. This phenomenon could have a significant impact on the life of whichever national missile defense system the next administration tries to develop, as well as many of the theater missile defense (TMD) systems already under development.
The academic literature on cost growth has shown that costs rise in the vast majority of major acquisition programs (see box).2 In general, the increase has varied by type of system: ships tend to have the lowest rates (roughly 15 percent on average), whereas tactical munitions and vehicles have the highest (roughly 100 percent on average). The average cost growth for other types of systems fall somewhere in between, with most in the 20-30 percent range.
Much of the cost-growth literature is historical, and there are too few data points for missile defenses to have their own category—only Safeguard and the original Patriot system have been included in most academic analyses of cost growth. But costs of strategic missile and space programs, which are similar in some respects to missile defense programs, have only on average risen by 20 percent and 30 percent, respectively. Thus a 20-30 percent growth rate would not seem unreasonable for missile defenses.
But the experience of the last two decades suggests the actual rate will be much higher. Understanding why the costs of missile defense programs seem to grow faster than other types of weapons systems is important because, if any of those systems are to succeed, rising costs must be contained.
An Alternative Theory
The explanation for this abnormal cost growth appears to be that missile defense programs (at least over the past 20 years or so) are fundamentally different from other development programs, and therefore do not lend themselves to simple projections of cost growth based on historical experience. There are three interrelated and interacting reasons for this:
• Missile defense programs are highly political.
• Missile defense programs respond to a perceived, urgent near-term threat.
• The technical challenges of missile defense are significantly underestimated.
As a result of these factors, the costs of ballistic missile defense programs have been significantly underestimated in almost every case. Other types of weapons programs may encounter one or more of these factors, but few, if any, suffer from all three. This theory has not been subject to rigorous review or statistical analysis, but it seems to explain the most significant causes of the problem and provides a way to gauge how susceptible a missile defense program will be to rising costs. Hopefully, future research will elucidate the causes more systematically and completely.
This article analyzes the causes of cost growth in missile defense programs. It is not a call to increase missile defense budgets or to halt missile defense programs. Nor does it suggest that building defenses will be easy or impossible. These are strategic, political, and technical questions that are beyond the scope of this piece. Rather, this article makes the case that, for a missile defense program to succeed, it must be carefully designed to account for the inevitable problems and technical challenges it will encounter and that proposed budgets must (to the extent possible) accurately reflect the costs of meeting those challenges.
The Political Factor
Missile defense programs tend not to bubble up from below but are instead created in the crucible of ideological combat. There are few security issues as highly charged as missile defense, and people on both sides approach the topic with religious zeal. This polarizes the debate and makes it difficult for a consensus to emerge or for proposals to be carefully considered. In this climate, ideas and programs are not fully conceived or vetted by the Pentagon bureaucracy and the budget process before they are pushed into the spotlight, contributing to poor program design, inaccurate initial cost estimates, and subsequent increases.
Ronald Reagan's claim that the Strategic Defense Initiative would render "nuclear weapons impotent and obsolete" had more to do with a vision than with technical realities. Similar visionary thinking has plagued other missile defense proposals.
A good example is the swirl of claims surrounding the upper-tier theater missile defense known as Navy Theater Wide. The Navy is developing the system to intercept intermediate-range theater missiles. It will be deployed sometime after 2010 on the Navy's existing fleet of Aegis-equipped cruisers and use modified Standard missiles launched from existing vertical launch tubes. The kill vehicle will be lightweight and not very sophisticated at discriminating between warheads and decoys.3 The Aegis radar also has limitations. Its resolution is too coarse for it to discriminate well and its power is too limited.
Despite those shortcomings and before the system's kill vehicle has hit a single target, supporters of Navy Theater Wide have recommended it for everything from a mid-course national missile defense (a direct substitute for the administration's NMD system), to an ascent-phase NMD system, to a boost-phase NMD system. Usually the price tag for one of these capabilities is advertised to be around $2 billion, a figure that neglects many of the requirements for developing an actual system. For the midcourse NMD capability, the Pentagon's own estimate is much higher, $16-$19 billion, in part because it would abandon the Navy Theater Wide's current kill vehicle and instead use the much larger and more capable exoatmospheric kill vehicle that is being developed for the ground-based NMD system. Naturally, this would require that a new, larger missile be developed and that the ships be modified to carry them.4 If history is a guide, prices could be even higher than the Pentagon estimates. No credible cost estimates have been advanced for converting an Aegis-equipped ship to the boost-phase mission.
The Perceived Near-Term Threat
The political warfare surrounding missile defense has been amplified by what are perceived as looming threats to U.S. forces, allies, and even the United States itself from countries in several key regions of the world. The growing capabilities of those countries cannot be denied, although their significance to U.S. security is still an open question. But in the context of this developing threat, new proposals for missile defenses are usually presented in "got to have it as soon as possible" terms. This adds a sense of extreme urgency to the missile defense debate, and the primary effect is that the schedules for missile defense programs are significantly compressed.
The urgency to "solve" the ballistic missile threat has created a "crash program mentality" within parts of the executive branch and Congress, and within the non-governmental groups and industries that push for missile defense. The result is predictable: the proposed defenses are not fully conceived acquisition programs, subjected to the Pentagon's disciplined risk, schedule, and costing analysis. Insufficient attention is paid to testing requirements, developing proper testing and simulation facilities, and the concept of operations.
Efforts to reduce technical risks and integrate myriad system components are also given short shrift. The desire to develop a defense as quickly as possible to address the growing threat from ballistic missiles, coupled with the belief that technology will solve the problem, creates a sense of both urgency and optimism that leads to unrealistic estimates of capability, technical risk, costs, and schedules.
In 1998, a blue-ribbon commission headed by retired General Larry Welch found similar problems in almost every missile defense program it examined. The panel noted that the urgency in theater missile defense programs has led to high levels of risk and "less-than-minimal testing or highly compressed flight testing or both." In its view, the assumptions driving the NMD program have been even more optimistic. Instead of accelerating development times, the panel found that crash programs were leading to program delays. In short, the crash program mentality was causing a "rush to failure."
The crash program mentality is a natural reaction to a serious problem, and, in times of serious crisis, it may be the correct response. But in those situations, special measures must be taken to structure the acquisition program to reduce the technical risks. They do not eliminate the risks and they cost money, often significant sums, but they are essential if an accelerated program is to have a chance to succeed. Most of the NMD and TMD proposals to date do not include anywhere near enough money to fund the robust ground testing, flight testing, and risk reduction programs that would be required for an urgent program to succeed on a compressed schedule. And regardless of how much money you throw at a program, it cannot be accelerated if the technology is not well in hand.
The last decade provides ample evidence of this crash program phenomenon. For example, the Global Protection Against Limited Strikes (GPALS) system was proposed by the Bush administration in January 1991 in response to congressional pressure to focus on protecting U.S. troops from the theater ballistic missiles that Iraq possessed rather than pursue its plans for a large defensive shield to protect the United States from a Soviet attack. The Strategic Defense Initiative Organization developed a concept for a robust system that it planned to deploy in less than a decade to protect the United States (and to some degree its allies) against an accidental launch by Russia of up to 200 warheads accompanied by sophisticated countermeasures. GPALS also included theater missile defenses to protect U.S. troops and allied populations in regional conflicts. It was estimated to cost $53 billion (in year 2000 dollars).
Many of the systems that the United States is developing today were included in the GPALS system or at least had their roots there. Their costs have grown, however, and schedules have slipped.
The national missile defense that is currently being developed includes the X-band radar, the upgraded early-warning radar, the ground-based interceptor, the exoatmospheric kill vehicle, and SBIRS-low missile-tracking satellites, all of which were part of the GPALS architecture, although some of the names have been changed. The estimated price tag for the NMD portion of GPALS was $42 billion when it was first unveiled. Today, the capability-2 configuration of the NMD system is a mere shadow of the original GPALS system, with one-seventh the number of interceptors, two-thirds the number of ground-based radars, fewer than half the number of SBIRS-low satellites, and none of the 1,000 space-based interceptors known as Brilliant Pebbles. Yet that small system is estimated by the Congressional Budget Office to cost more than $30 billion, fully three-quarters of the projected GPALS price, and it will take as long to deploy as was projected for GPALS.5
Many TMD systems that the United States is developing today also were included in GPALS: THAAD and its GBR-T radar, the Patriot Advanced Capability-3 (PAC-3), and some form of the Medium Extended Air Defense System. The GPALS price tag for those systems: $12 billion in today's dollars. The current estimate for just THAAD and PAC-3 is more than $20 billion and is likely to rise further. Likewise, the time it will take to deploy those systems has expanded significantly from the original timelines assumed for them in GPALS.
The Clinton administration's efforts to develop an NMD system present a more recent example of the crash program mentality. Its first foray into NMD was the so-called 2+2 system suggested by Defense Secretary William Perry in 1995—a defense that was intended to be developed within two years and deployed within two if a threat emerged that would warrant it. The idea was that the system would provide some rudimentary protection if a threat emerged more quickly from North Korea, Iraq, or Iran than predicted by intelligence estimates. It would have consisted of 20 interceptors deployed in North Dakota. The system was to use as much off-the-shelf technology as possible and be done quickly and cheaply.
The 2+2 system never got far enough to be assigned a price tag, but the approach it embodied—get something deployed quickly in response to political pressure from the Congress, but do it as cheaply as possible—set the tone for future administration efforts to develop a national missile defense.
The 2+2 proposal mutated into the 3+3 system, proposed in 1996. It was a more fleshed-out concept that featured 100 interceptors in North Dakota. It would take at least three years to develop, according to the administration, and another three to deploy the first 20 interceptors if deployment became necessary. The price tag for this system was estimated by the Pentagon to be just short of $8 billion. A variant proposed by the Air Force, which could trace its lineage back to Perry's off-the-shelf approach, would have used existing Minuteman boosters and a small kill vehicle that was an improved version of the one being developed for Navy Theater Wide. It also would have used an existing type of radar. The Air Force's price tag: $2.5 billion.
Today, the original 3+3 system has become the first and second phase of the NMD system that the administration is readying for deployment. In the process, the schedule has slipped and the price has risen. According to the current schedule, the system will take at least nine years to deploy and cost at least $20 billion to build, or 2.5 times the price that was originally advertised for the 3+3 system to get the same capability just four years earlier. Much of this huge price rise can be traced to overly optimistic assumptions about technology and costs—an optimism that was created by the crash program mentality combined with a desire to do the job as cheaply as possible. Together, those forces led to an acquisition program that had little of the usual funding for such essential activities as systems engineering, reducing technical risks, and testing. As the Welch panel has pointed out, this was an even more significant omission for such a high-risk and complex program.
It has only been in the last 18 months or so that the NMD program has taken the first steps away from the crash program mentality that has so pervaded missile defense efforts over the past two decades. The program is now beginning to include some of the features that one would expect to see in an acquisition program of this complexity. That does not mean that its cost growth problems are behind it—many aspects of the program need to be made more robust, such as ground testing, flight testing, and reducing technical risk. But at least program managers are beginning to address some of the problems.
The Technical Challenges
Missile defense is a tough challenge, both technically and operationally. It was difficult enough when interceptors carried nuclear weapons and had a kill radius measured in hundreds of meters or even kilometers. But hit-to-kill requires precision that is measured in tens of centimeters and microseconds. It is especially challenging for national missile defense because there is very low tolerance for leakers, warheads that slip through the defense. Nearly everyone underestimates the breadth of the effort that will be required to field effective missile defenses. This does not necessarily mean that the job cannot be done, just that a program must fully account for all the challenges for it to be successful (assuming, of course, that the program is technically feasible to begin with). The technical challenges of missile defense amplify the effects of politically driven proposals and compressed schedules.
Developing the components of an NMD system—hit-to-kill interceptors, seekers, infrared sensors, high-resolution X-band radar, discrimination, and command and control—will require pushing the state of the art in many dimensions. Most other acquisition programs push the state of the art in just a few. Most critically, those subsystems must be integrated into one seamless system of unprecedented scale and complexity that functions with near-perfect reliability. This will require extensive efforts to integrate the system and test it in both laboratory and real-life conditions.
The scale of the task is akin to the one facing the United States in the 1950s when it was developing its entire nuclear deterrent—warheads, missiles, early-warning radars and satellites, launch control centers, communication networks, and a central command and control center—all at the same time. In some critical respects, NMD is even more challenging because the coordination between the pieces must be tighter. Interceptors must be able to distinguish warheads from decoys and maneuver to hit them at closing speeds that may be in excess of 10 kilometers per second. Errors of a few centimeters or microseconds can mean the difference between success and failure. By contrast, ICBMs need only to fly from one point on the ground to another with an accuracy of a few hundred to a few thousand meters.
Similarly, the existing early-warning system that was developed for the nuclear deterrent role needs only to detect an enemy launch and give a general description of the size and character of the attack. The command and control system must reliably transmit that information to the National Command Authority and then transmit any launch orders to ICBM, bomber, and submarine commanders. Solving those challenges was not trivial; it took the United States several decades. But for NMD, the early-warning system must be supplemented with very precise radars and infrared satellites that track the missiles and their warheads and decoys. The battle management and command and control system must then transmit the data to the command center, fuse the data from myriad sensor platforms into a spatially precise picture, discriminate between warheads and decoys, and send instructions to the interceptors so they can fly to the best position to intercept their targets. Theater missile defense systems share many of the same complexities.
The fact remains that nobody has done most of those things before. Some people doubt that missile defense can ever be done with hit-to-kill, particularly in the midcourse when the vacuum of space allows inexpensive decoys to be effective and makes discrimination extremely difficult. If ballistic missile defense can be done, three key components will be necessary for success: robust risk mitigation, system integration, and testing efforts. Failure to adequately fund any of these will inevitably lead to cost increases and schedule delays.
Risk Mitigation: Risk mitigation is the engineering term for reducing the technical (and by extension, schedule) risks in the program—that is, the chances that the system will not work as hoped. The head of the Ballistic Missile Defense Organization (BMDO) has described the proposed NMD system as very high risk, and most theater defense programs are also high risk.
Program managers try to reduce technical and schedule risks in their program through a variety of methods, including letting alternative technologies or engineering approaches compete with each other so that the best approach will emerge. After selecting one approach, program managers can also continue to develop alternative technologies or systems so that they will be available if the primary one fails. Other approaches include using computer simulations and testing components in facilities that simulate actual combat conditions. All of those approaches cost money up front, and in a tight budget environment that is often caused by external budget pressures or internal technical problems, program managers are often forced to cut back on risk reduction efforts to keep the program on schedule. This happens over and over again.
Take the SBIRS-low program, a constellation of 24 space-based infrared satellites (previously known as the Space and Missile Tracking System and Brilliant Eyes) designed to track warheads and decoys. Two teams were initially selected to build prototype satellites and fly them in space, but for budget reasons, one was dropped. The winning contractor was tasked to build a single satellite for a flight demonstration system. But after program costs began to grow out of sight, a second contractor was added to develop an alternative sensor. Finally, costs got so high that both programs were cancelled. To keep costs down, the Air Force has now decided that it does not need to demonstrate key high-risk technologies in space before it deploys the first phase of the system.
The THAAD program presents another example of how a more robustly funded risk mitigation program could have helped avoid problems. THAAD initially suffered well-publicized problems including four straight flight-test failures. The program was delayed after each failure and costs went up. But THAAD was not killed because the Army viewed it as essential for protecting its soldiers from longer-range theater ballistic missiles. If a second contractor had been funded to develop another prototype, two things might have happened: the first contractor might have tried harder knowing that it might lose the competition, or the contractor could have been fired if it could not correct the problems. Either way, the Army would have had an alternative for addressing its pressing need for force protection.
Reducing technical risk is essential, particularly for ballistic missile defense programs that push the state of the art in many dimensions. When a program starts cutting back on planned risk reduction efforts, it is usually an indication that costs are likely to rise—a case of being penny wise and pound foolish.
System Integration: Making sure that the individual components of a missile defense system are carefully designed to work together and thoroughly tested together is known as system integration and can be the most difficult part of an acquisition program. Much of the system integration rubber for missile defense meets the road in the battle management/command and control system. The challenges are evident from the discussion in the previous section. But how well the components of a defense have been integrated—and thus how effective a defense will be—is revealed only during combat, or in test conditions that mimic combat as realistically as possible.
System integration for a complex missile defense program can cost billions of dollars. For example, the Pentagon estimates that it will cost $5.4 billion for system integration of the first phase of the NMD system, or nearly 30 percent of the $18.6 billion it expects the system to cost. The more complex the system integration task, the more weight is placed on the testing program to prove the system's effectiveness.
Testing: Proper testing is complicated, expensive, and to date has not been included in most missile defense programs. Responding to criticisms from the three Welch panel reports, more money for testing has been added to the theater and national missile defense programs, something that should have been done from the start if the development programs had been properly conceived and designed and the technical challenges were not understated. But despite the Welch reports and the responses from the programs, it is not clear that the fundamental problems have been solved, and this raises the specter of further cost growth.
In the face of significant technical challenges, the current NMD program is using a testing program that is based on a relatively new approach—one that has yet to be fully validated for a program this complex. Most TMD programs are also using this approach. In the traditional approach, large numbers of flight tests are conducted during development and after the system is deployed. This allows more time and opportunities to resolve any technical issues that arise.
By contrast, the new approach relies on extensive simulations and ground tests to complement a historically small number of flight tests. Table 1 illustrates this trend. It shows that 40-60 flight tests were conducted during research and development programs for first-of-their-kind ICBM and SLBM programs. That number fell into the 20s in more recent years, in part because of the rising sophistication of simulation and ground-test facilities. Long-range ballistic missiles are similar to missile defenses because very high reliability and effectiveness are demanded of both. But in important ways ballistic missile programs are different: they only need to fly from one well-known point on the earth to another. By contrast, missile defense interceptors must fly to a precise point in space to meet a dynamic target that could come from almost infinite directions and angles and deploy a wide range of countermeasures against the defense. In this regard, the missile defense problem is more like air defense.
Air defenses have been subjected to far more flight tests than ballistic missiles: Block I and II of the Navy's Standard missile had 88, the original air-defense version of the Patriot had 114, and Advanced Medium Range Air-to-Air Missile had 111. The testing program for Safeguard, the only NMD system that the United States has deployed, was consistent with these air-defense programs—165 tests were conducted.
By comparison, 19 flight tests have been planned for the current NMD system, about the same number for THAAD, and 10 tests with 16 missiles for PAC-3. In several important ways, though, missile defense is even tougher than air defense, in large part because of the very short time lines, high speeds, and the challenges of overcoming decoys.
The primary motivation for adopting this new approach has been cost. For the current NMD system, a flight test that pits a single interceptor against a single missile costs $80-100 million, including all the pretest and post-test activities and analysis. Even the modest flight-test program planned for the NMD system will require $1.5-2 billion—a sum that would make any program manager look hard for alternatives. Computer simulations, hardware-in-the-loop tests, and ground tests in facilities that simulate a few aspects of real operational environments are reasonable and well-established ways to supplement flight tests, and BMDO is making extensive use of all three.
But it is not clear that those tools can substitute for flight tests rather than just supplement them, particularly for a program as challenging as missile defense where reliability and effectiveness must be so high. First, it is not clear that the programs have the proper simulation and ground-test facilities. Although the Pentagon has built several significant simulation and ground-test facilities over the past 20 years for TMD and NMD systems, the 1998 Welch report argued that THAAD, Navy Theater Wide, NMD, and to some extent Patriot did not have properly designed ground-testing programs. Second, the few flight tests planned are not demanding enough. The NMD system is designed to defend against a few dozen warheads and by shooting four or five interceptors at each, but no tests have been planned in which the system will face more than one warhead or in which the system will target the warhead with more than one interceptor. And questions have also been raised about how realistic the targets and decoys are. Unless both these issues are addressed, costs will rise.
Lessons for the Future
What lessons can the new administration and the new Congress draw from the experience of the past two decades as they wade into the thicket of missile defense? First and foremost, they must recognize that the same three forces that have caused costs to rise in the past will continue to plague ongoing programs, particularly NMD, unless those programs are subjected to careful scrutiny before revised price tags are attached. This is also true for the alternatives that have been suggested to current programs, including new land, sea, and space-based boost-phase systems. Because the success of missile defense programs will depend in part on how well the acquisition programs have been designed and how complete the resulting cost estimates will be, the next administration and Congress should keep the following points in mind as they review the ballistic missile defense programs that they have inherited and consider new approaches:
• To the greatest extent possible, minimize the political fighting and seek consensus. The challenge of this task may seem huge, but the dividends of consensus could be larger. They include a stable program with predictable funding and a higher tolerance for the inevitable flight test failures and other problems, making planning and executing the program much easier. In fact, it may be impossible to deploy a system as complex as NMD without a broad consensus.
• Resist the temptation to compress schedules. Compressing the development and testing time has actually slowed progress on missile defense rather than accelerated it, according to the Welch panel. If the threat is so compelling that deployment must be accelerated, extraordinary efforts should be made to include robust risk reduction programs for every critical technology and system component. Properly funding this effort will be very expensive and will not guarantee success, but it will reduce the chances and significance of delays and failure.
• Never underestimate the difficulty of missile defense. Despite many optimistic statements to the contrary, the technology for missile defense is not yet well in hand. Expect problems, delays, and, most of all, test failures. A program should be designed to withstand them. It must have a robust risk reduction program. Every component that pushes the state of the art should have one or more backup components in development. Also, do not overlook system integration. This may be the most difficult activity of all, so it must be well designed and fully funded.
• Keep an eye on the full testing program. The new testing paradigm that the Pentagon is using for missile defense relies heavily on computer simulation, anchored by extensive ground testing and occasional flight testing. Data from flight tests will be used to validate simulations and ground-test facilities. In turn, validated simulations and test facilities make it possible, the argument goes, to run very lean (by historical standards) follow-on flight tests after the system has been deployed. Remember that this approach is driven by cost, and a number of questions need to be kept in mind. Does it strike the right balance between flight tests and the others? Given the importance of each flight test, does the schedule allow enough time to prepare for the test and incorporate the results of the test into the program before the next test? Will all the necessary ground-test facilities be built and fully exploited? More fundamentally, will this approach prove itself?
• Don't forget the extras. Developing and deploying a missile defense is not the end of the story, but rather the beginning of a long process that will span decades. Budgets should include the costs of operating the system (including follow-on flight testing to ensure the system remains reliable and effective), sustaining the system (especially for space-based systems where satellites must be replaced every 5-10 years), and continuing research and development (missile defense is a dynamic process, and research must continue to keep ahead of what potential adversaries are doing to undermine the defense). These activities usually represent 50 percent or more of the lifetime costs for a system.
• Watch for inconsistent cost assumptions. Make sure that the assumptions are consistent when comparing the price tags on competing missile defense proposals. Does the price include development, procurement, and operating costs? Are the estimates presented in constant dollars (adjusted to eliminate the effects of inflation) or inflated dollars, which make price comparisons more difficult.
Costs in any complicated acquisition program are likely to grow—20 percent to 30 percent or so on average. But if a missile defense program has not been carefully thought through and well designed; if it does not include a robust effort to reduce technical risk for all critical components; if it short-changes system integration; or if it cuts corners on simulation, ground testing and ground-test facilities, and realistic flight tests, costs will spiral out of sight and delays will be rampant. Unfortunately, that has been the experience of the past two decades. Only if all those factors are properly addressed will missile defense programs experience more normal levels of cost growth and avoid some of the cost-related pitfalls that lay ahead.
In short, there is nothing cheap about missile defense except talk. There are no shortcuts to building a complex system. And missile defense, particularly national missile defense, may be the most complex system that the United States has ever attempted. Based on the experience of the past two decades, it is clear that, if the United States is to deploy the missile defense systems now under development, it will have to spend more, probably significantly more, than current estimates suggest.
NOTES
1. Stephen I. Schwartz, ed., Atomic Audit: The Costs and Consequences of U.S. Nuclear Weapons Since 1940 (Washington, D.C.: Brookings Institution Press, 1998), p. 289.
2. J.A. Drezner, et al., An Analysis of Weapon System Cost Growth (Santa Monica, CA: RAND, 1993) and Karen W. Tyson, et al., The Effects of Management Initiatives on the Costs and Schedules of Defense Acquisition Programs (Alexandria, VA: Institute for Defense Analyses, 1992).
3. Ballistic Missile Defense Organization, Summary of the Report to Congress on Utility of Sea-Based Assets to National Missile Defense, June 1, 1999, p. 10-11.
4. Rodney W. Jones, Taking National Missile Defense to Sea: A Critique of Sea-Based and Boost-Phase Proposals, (Washington, D.C.: Council for a Livable World Education Fund, October 2000).
5. Congressional Budget Office, "Budgetary and Technical Implications of the Administration's Plan for National Missile Defense," April 2000.
David E. Mosher, who worked at the Congressional Budget Office for 10 years analyzing nuclear weapons, arms control, and missile defense policies and programs, is currently at RAND. The views expressed here are his own.