News Fighting Proliferation

Chapter 16

Fighting Proliferation with Intelligence

Henry Sokolski

Prior to Operation Desert Storm, US policy toward the proliferation of strategic weapons technology was to delay or prevent it through a policy of nonproliferation—export controls, customs interdictions, end-use checks, diplomatic demarches, nonproliferation pledges, regional arms-control talks, and the safeguarding of sensitive nuclear activities. Desert Storm changed all that. Scud missiles were targeted and intercepted. Coalition forces were inoculated against possible Iraqi use of biological weapons. Iraq’s missile and nuclear, biological, and chemical (NBC) weapons facilities were bombed. Finally, these and related facilities were ferreted out and dismantled as part of the cease-fire plan of the United Nations (UN). In short, with Desert Storm, the United States and its allies moved beyond preventing proliferation to fighting it.

This change from a nonproliferation policy to fighting proliferation is fundamental. Indeed, the US government has yet to comprehend fully what this more combative world requires.l What follows is an examination of how fighting versus merely attempting to prevent proliferation will require policymakers and intelligence officials to work much more closely with one another, not only on the development of new intelligence collection and analysis requirements but on the very definition of proliferation and the strategy used by the United States to combat it.

Instead of pursuing nonproliferation efforts as most governments have—by reacting (often belatedly) to a state’s efforts to acquire the capability to produce strategic weapons—fighting proliferation requires devising a strategy that works backwards from a possible future in which one hypothesizes that these weapons are already employed or might be used against the United States or one of its allies. Such assessments will need to spell out the specific military operational implications of these threats so that policymakers can determine the level of attention each deserves and develop strategies to delay, stop, and—if possible—neutralize them militarily or politically.

All of that assumes that the intelligence community and policymakers can agree on what proliferation is. It also assumes that they have a way to develop relevant threat scenarios that neither requires policymakers to make intelligence determinations nor forces intelligence officers to arbiter everyday policy disputes. Finally, it assumes that the US government understands and can meet the intelligence-collection requirements arising from the new approach.

What Is Proliferation?

Despite (or perhaps because of) the current interest in the issue of proliferation, defining the term has in itself become a topic of debate. Fifteen or 20 years ago, the situation was quite different. Then, the only proliferation that seemed worth worrying about was that related to nuclear weapons. As a result, the definition of proliferation was simple: the spread of unsafeguarded nuclear technology to smaller states.

Because the security implications of additional states going nuclear were so unacceptable and because the technical thresholds that smaller states needed to attain in order to get nuclear weapons seemed so high, US policy was first, second, and last, one of prevention. There was little serious thought given to waging war against a state seeking to acquire nuclear weapons—or much need for such thinking. Instead, the focus was on the near-term efforts of nonproliferation: preventing certain states from getting the wherewithal to go nuclear. That was 15 or 20 years ago. Now the premises of that period no longer pertain.

Consider, for example, how government officials now speak about proliferation. The talk is no longer just about preventing countries from going nuclear but about the need to “counter” a wide variety of weapons technologies. This more ambitious proliferation agenda, especially as currently articulated, is bewildering. In the first six months of the Clinton administration alone, public officials (including those in intelligence) argued that the United States should be concerned about the spread of (1) weapons of mass destruction (NBC munitions); (2) weapons of mass destruction and the means for their delivery; (3) weapons of mass destruction and the missiles needed to deliver them; (4) special weapons; (5) advanced weapons; (6) advanced conventional weapons; (7) destabilizing numbers and types of advanced conventional weapons; (8) conventional weapons; and—to complete the circle—(9) “weapons of proliferation concern.”2

The Confusion of Current Views

Clearly, the intelligence community’s role in fighting proliferation is difficult to pinpoint if what’s being fought is itself left this vague. In particular, although suggestive of something more aggressive than “nonproliferation,” it is not clear what the government’s new desire to “counter” proliferation means. In fact, the dictionary lists six separate entries for the word counter, and only one is a verb. The verb means oppose, offset, or nullify. The question is, In what way? With military countermeasures, counterattacks, counteroffensives, or counterintelligence? All of this sounds exciting. Images of the Israeli air strike against Osirak in 1981 come to mind even though the opportunities for repeating such heroics may be far less than one might wish. The word, however, just as easily appeals to people who want to oppose dangerous proliferation activities through the kind of diplomacy, safeguard inspections, disarmament procedures, export controls, and sanctions that are already in place.3

Although the term counterproliferation may seem useful as new terminology, it is no substitute for clear thinking about what the problem is—especially when one considers whose proliferation the United States is supposed to be countering and what proliferation is.4 A popular view, particularly among many conservatives, is that the United States should simply focus its opposition to proliferation against enemies who have not yet acquired strategic weaponry. This approach was first seized upon at the outset of the cold war, when the US had a nuclear monopoly and a clear fear of the Soviet Union. Today, however, America has few—if any—clear-cut adversaries. There are countries on the US government’s terrorist lists, but the United States and its allies still talk and trade with them. Also, shouldn’t the United States care about states that have strategic weapons systems? What about China’s efforts to significantly upgrade its existing strategic forces with technology from the West and Russia? What of the republics of the former Soviet Union keeping and upgrading the systems the Soviets left behind? Certainly, these developments deserve attention. Do we mean to ignore them?

There is also confusion about which weapons technologies are of proliferation concern. Many officials wish to cling to the convenience of limiting proliferation to only apocalyptic weaponry and related technologies. If proliferation can no longer be limited to nuclear weapons, they argue, it would be best to confine proliferation concerns to weapons of mass destruction.

This approach, however, ignores the very real concern that weapons-delivery vehicles present. Yet, to the extent that trucks, planes, and ships could serve as vehicles, the scope of what is meant by “means of delivery” becomes a problem. In response, two types of solutions have been offered. The first is to limit concerns about delivery systems to missiles and related technology as controlled under the Missile Technology Control Regime (MTCR). The second is to argue that not just missiles but planes should be included as well. That, in turn, has encouraged an even broader approach: any weapon that can inflict military harm against the United States or its friends should be included.

Neither approach is sound. Certainly, to widen proliferation’s focus to cover anything of military concern is to trivialize it. Not just strategic systems but tanks and planes become proliferation worries. The first kind of weaponry, however, is very different from the second. Nuclear weapons can be covertly acquired and employed in small numbers to produce shocking strategic results against the United States or its friends, even in their most defended state. In contrast, years of acquisition and overt training with thousands of tanks or planes are necessary to pose a significant military threat to US forces or US-led coalitions, and—even then—effective military defenses or countermeasures against such weapons are available to the United States. (Thus, current conventional arms-control proposals rarely speak of banning the sale of planes or tanks but talk instead about increasing warning time by requiring states to make such arms sales more public through UN registries and the like.)

Moreover, a broad definition of proliferation would have to include not only the weapons themselves but also the related technology. Designating most conventional arms as weapons of proliferation concern, then, would stretch existing export control, customs, and intelligence collection and analysis efforts beyond any hope of focus. For these reasons, many people in the US intelligence community prefer to limit their attention just to weapons of mass destruction.

Focusing on what has traditionally been of proliferation concern—NBC weapons—however, runs the risk of ignoring new threats that are likely to emerge. After all, since 1945, what is of proliferation concern has itself changed several times, evolving from a worry about the Soviet Union getting nuclear weapons to the current concern regarding smaller states that might get NBC munitions, as well as long-range missiles.

And there is every reason to believe that new worries are on the way. Certainly, congressional interest in the military implications of satellite services (imagery, navigation, and communications) being sold to third world states suggests as much. So, too, does Congress’s recent enactment of a law imposing sanctions against countries selling destabilizing types and numbers of advanced conventional weapons to Iran or Iraq.5 The US Navy is also concerned about Iran’s acquisition of conventional submarines. These vessels will be difficult to find in the confined waters of the Persian Gulf and, if properly armed, could threaten American and allied fleets. In short, our list of concerns is expanding and is likely to include high-leverage technologies and weapons systems that could enable smaller states to threaten war-winning or victory-denying results against the United States or its friends without resorting to weapons of mass destruction.6

Toward a Prescriptive Definition

As difficult as defining proliferation may be, the intelligence and policy communities should make the effort. In fact, they have no choice: Congress in 1992 instructed the executive branch to identify what “types and numbers of advanced conventional weapons [are] destabilizing.”7 There are two schools of thinking on how best to do this. The first is simply to compile new lists of weapons technologies that should be monitored or controlled. Besides being familiar, this approach has the short-term advantage of being responsive to Congress. The difficulty comes in the long term: as already noted, compiling lists of what might be of concern tends to attenuate the government’s already limited ability to maintain a constructive focus.

This legitimate concern with overreach was, in part, what prompted consideration of a second—more prescriptive— approach, which was first presented before the Senate Select Committee on Intelligence in 1990 by Henry S. Rowen—then the assistant secretary of defense for international security affairs.8 Instead of trying to describe what is of proliferation concern by listing specific weapons and related technologies, criteria were established prescribing what was worth worrying about and why. Three criteria were suggested. A weapons or weapons-related technology was of proliferation concern if (1) it enabled another state to inflict high-leverage strategic harm against the United States or its friends; (2) the United States lacked effective defenses or countermeasures against this capability; and (3) its mere acquisition could change other states’ perceptions as to who was the leading power in a given region.

High leverage is not to be confused here with high or advanced technology. Relatively low-technology, nonnuclear submarines in the Gulf, for example, could sink one of the US Navy’s capital ships, preclude the Navy from identifying the perpetrator, and make the political demands to quit the area nearly irresistible. Advanced jet fighters, on the other hand, might incorporate high technology but are low leverage since hundreds of them would be necessary to secure local air superiority against US or allied forces. Even then, US air defenses would prove effective against organized air attacks.

The meaning of strategic harm also requires reflection. During the cold war, this term meant intercontinental, global nuclear conflict with the Warsaw Pact countries. Today, however, wars are more likely to be like Desert Storm than the nuclear wars depicted in popular television movies such as The Day After or novels such as Tom Clancy’s Red Storm Rising. As a result, what constitutes victory-denying harm differs from strategic harm. During the cold war, for example, keeping Soviet nuclear submarines from gaining passage to open seas was critical to US security. Today, however, US and coalition forces potentially could be defeated by an inability to find, identify, and destroy third world conventional submarines laying mines in the Persian Gulf.

Finally, the adequacy of one’s defenses can be determined only in relation to what constitutes war. Certainly, in the 1960s at the height of the cold war, missile defenses against the Soviets did not seem necessary to most US officials because they believed that any Soviet missile strike against the United States or its allies would prompt a massive US nuclear response. That, they reasoned, would deter such attacks. In Desert Storm, however, coalition adhesion was threatened by nonnuclear missile strikes against coalition forces and Israel. Here, even limited missile defenses were understood o be critical.

A key difficulty in getting the type of military threat assessment called for by Rowen is that it requires intelligence and policy officials to cooperate. Working with the policy com- munity on threat assessments (which include the adequacy of US defenses), however, is something the intelligence community has long considered sensitive—if not taboo. The intelligence community fears that working on such assessments will inevitably drag it into policy disputes. It also worries that policymakers will make determinations about intelligence data that they have no business making.

The advantage of this prescriptive approach, however, is that it encourages the kind of communication between intelligence and policy officials which is necessary to anticipate and execute effective diplomatic, commercial, or military operations against proliferation. This interaction is also critical to developing more than merely reactionary or episodic covert-action programs and counterintelligence operations against proliferation. Instead of merely following up tips and heading off particularly dangerous shipments on an ad hoc basis, intelligence operatives could be told what proliferation developments deserve special attention and orchestrate a variety of clandestine efforts in advance. These efforts might include leaking damaging information about projects to the foreign press, introducing faulty software and hardware into programs, and encouraging others to take steps to sabotage them.

If the United States were to pursue this approach, the intelligence community—starting with the intelligence components of the military services—would have to focus on the potential military threats that proliferation developments might pose. The military commands and service staffs, meanwhile, would have to share their views on what they believe high-leverage and strategic weapons systems might be in key scenarios and indicate where their defense preparations might be inadequate. The policy community, finally, would have to work with the military to prioritize the various likely threats and develop strategies in coordination with the military and intelligence communities to contain, reverse, or combat them.9

The Need to Rethink Policy
and Intelligence Reasoning

In addition to working together more closely, fighting proliferation will require the intelligence and policy communities to think about proliferation problems differently. At a minimum, it will require these communities to reconsider their traditional cold war relationship. At the height of the cold war, this relationship was fairly routinized. The United States knew who the adversary was—the Warsaw Pact states; the fear was war. Intelligence was tasked to collect and analyze what Moscow tried to keep secret: its war plans, capabilities, and true foreign policy objectives.

Such clarity about the adversary’s identity and the urgency associated with global war generated a serious effort to learn everything possible about existing Warsaw Pact military capabilities, especially their order of battle. This technical aspect of intelligence, in turn, made it relatively easy to establish a division of labor between the intelligence and policy communities. Intelligence collected secrets that the military needed to help bound the uncertainties associated with war, and policy did all it could to reduce the likelihood of war.

As one recent discussion on intelligence noted, what is most required from the intelligence community is secret information in support of military operations:

[A] former senior official suggested that the touchstone for the intelligence community should be its concentration on secrets. He contrasted a “secret” with an “uncertainty,” which could be studied, and a “mystery,” which is unknowable. A secret, he said, is a valued piece of information that gives its possessor an advantage and would give another who acquired the secret an advantage. The government might want to ponder uncertainties and mysteries, but it should reserve the intelligence service for the pursuit of secrets.10

This makes sense. For purposes of fighting proliferation, however, certain adjustments are necessary.

First, as noted before, adversaries of the United States are far less apparent than they were during the cold war. The United States must plan for a security environment in which there may well be a shifting set of allies and adversaries. Moreover, even if Washington had perfect clarity about who its future enemies might be, that knowledge would not exhaust the parties the United States would be interested in monitoring for proliferation purposes. For instance, the United States is obviously concerned about what North Korea is doing. But precisely because the United States is worried about having to go to war in Korea and having to cope with NBC weapons and long-range missiles, it is also keenly interested in what South Korea might be doing to acquire or use such weapons itself. One of the last things the United States wants is to be drawn into a war with (or by) a friend and then have its joint defense plans undermined by this ally’s unilateral employment of proscribed strategic weaponry.

Also, the United States is interested in stemming the proliferation of strategic weaponry in general. Given the broad uncertainties of the future security interests of the  United States and, therewith, the uncertainty of who its friends and competitors will be, a prudent expenditure of effort in curtailing proliferation today can pay big dividends in the future. Thus, the proliferation activities of a country like Indonesia, which at present is neither a formal ally nor an adversary, are of interest to the US government particularly if, with a modicum of effort, the United States and its friends can persuade that country to forgo the acquisition of strategic weapons.

What fighting proliferation requires, then, is something more than what became routine in the cold war. It requires not just the collection of secrets from known adversaries but military threat assessments that consider current and possible future threats along with analysis that will enable the United States to slow or prevent proliferation in a much more entrepreneurial fashion.

This kind of forward-looking assessment will require intelligence and policy officials to examine a wide set of uncertainties about the future and to use a variety of analytic methods to build likely proliferation scenarios. Instead of merely trying to determine which countries have which weapons capabilities, policy and intelligence officials will need to determine how these capabilities might be employed against the United States or its allies, spell out what the military operational implications of these employments would be in the most probable war settings, and pinpoint what military vulnerabilities these employments are most likely to create.

A New Paradigm

Working together on such assessments will not be easy for the policy and intelligence communities. That is understandable. Collecting and analyzing proliferation secrets—information that other people are trying to conceal from the United States to prevent it or its friends from taking appropriate action—is what intelligence officials who have been working with proliferation issues are most comfortable with. Policymakers, on the other hand, are generally most at ease developing policy justifications concerning proliferation—arguments supporting or opposing positions or actions the government might take.

Clearly, if the intelligence and policy communities are to focus on the uncertainties that fighting proliferation requires, a new paradigm governing their relationship on these issues needs to be established. One model that suggests itself is the operations research and gaming activities the US military has done for years in order to anticipate what it will need to meet military contingencies. The following sections borrow from that experience. They comprise a suggested hierarchy of policy and intelligence reasoning that would encourage the kind of analysis and operations research that a true fight against proliferation would require. This hierarchy consists of four kinds of proliferation knowledge: proliferation secrets, uncertainties, excursions, and policy judgments.

Proliferation Secrets

Proliferation secrets are information that other people are trying to conceal to prevent a significant US or allied response. This information is collected and analyzed for policymakers by the intelligence community. Proliferation secrets tend to be straightforward. An example of a proliferation secret might be information as to whether or not the People’s Republic of China (PRC) has transferred M-11 ballistic missiles or missile technology to Pakistan in contravention of the guidelines of the MTCR. The PRC and Pakistan would want to keep this information from the United States since the transfer of such missiles could trigger trade sanctions and would be diplomatically embarrassing.

Proliferation Uncertainties

Proliferation uncertainties are less clear-cut. They can only be known imperfectly since they concern possible future technical, economic, political, or military developments—that is, educated guesses about the future. An example of a proliferation uncertainty might be projections as to when Pakistan would be able to produce and operate M-11 ballistic missile systems on its own. Getting a handle on such uncertainties in the Pakistani case would be most useful to policymakers since it would tell them how much time they had to head off or prepare for such activities.

Intelligence estimates concerning proliferation should highlight such uncertainties but rarely do. They do not analyze what the government does and does not know, the variety of futures that might happen, their probabilities, and the independent variables or determinants of each likely future. Instead, intelligence estimates have tended to be fixed, oracular “determinations” that simply reflect the consensus of the moment—that is, what most intelligence analysts agree is the best single guess of what the truth may be. Generally left unaddressed are the gaps in the government’s knowledge, which, admittedly, can be embarrassing. That is perhaps why these proliferation estimates are sometimes treated as sensitive secrets rather than the speculative analyses that they often are.

Proliferation Excursions

Proliferation excursions take uncertainties one step further. They involve the use of information and uncertainty analysis (that is, projections) to divine the likely relation or possible operational implications of known or possible proliferation developments. Operations research, scenario building, and gaming would all fit under this heading. An example might use operations-research techniques to determine the military implications of Pakistan’s employing M-11 ballistic missiles against Indian forces. This information would be useful to understand just how serious a problem the spread of this technology might be and allow the United States and its allies to prepare militarily for the consequences.

Policy Judgments

Policy judgments are opinions or conclusions about what positions or actions the government should take toward particular proliferation developments. These judgments are based on arguments that, one would hope, reflect the best information and analysis of the situation, its possible implications, and full consideration of the ramifications of whatever action or position is decided upon. An example here would be the US government’s current judgment that M-11 missile proliferation to Pakistan jeopardizes continued peace in Southwest Asia and should, therefore, be opposed.

As noted before, the hierarchy of policy reasoning and the hierarchy of intelligence certitude concerning proliferation issues are opposites. Whereas policy officials are most comfortable with developing and arguing over policy justifications concerning proliferation, the intelligence community is most at home collecting and analyzing proliferation secrets. As for proliferation uncertainties and excursions, neither enjoy much favor in either community since they seem either too complex or technical for busy policymakers or too close in their implications to policy-making for intelligence officers.

To the extent that it was less necessary during the cold war to have policy and intelligence cooperate in doing uncertainty analyses and excursions on the Soviet threat, this division between policy and intelligence was relatively clear, and it generally worked. With proliferation issues, however, such a marked division of labor quickly becomes dysfunctional. Whereas policy officials are eager to use intelligence to issue demarches to US allies about specific exports, imports, or other activities, intelligence agencies are naturally worried about jeopardizing sources and methods. Also, knowing the military implications of specific proliferation developments is critical to policymakers to gauge the importance of particular developments and to come up with the appropriate responses. Yet, the military is uncomfortable discussing possible US force vulnerabilities. Finally, policymakers need to know the full range of what technical, political, and military outcomes are possible for a particular proliferation development. Intelligence analysts, on the other hand, are typically leery of making any but the most conservative projections for fear of being accused of being wrong somewhere down the road.

Each of these dysfunctions has caused friction between the intelligence and policy communities and is part of the reason why the director of Central Intelligence (DCI) created the Nonproliferation Center. Designed to coordinate all of the government’s intelligence efforts related to proliferation and to serve as a single point of contact within the community that policymakers could turn to, the center has gone a long way to improve relations between policy and intelligence officials working on proliferation. Such liaison work, however, is no substitute for rethinking the policy and intelligence suppositions that are the cause of the friction between these groups.

Certainly, officials who are fighting proliferation need to stop thinking of their goal as simply one of buying time—a static objective; instead, they need to view their efforts as a campaign, which is dynamic. Progress in this struggle would not be measured in terms of how many states have signed up to agreements such as the Nuclear Nonproliferation Treaty but by how well an agreed strategy to prevent, delay, and combat proliferation has been implemented.

Once policy and intelligence officials begin to think of our efforts against proliferation in this light—as a kind of warfare—it will be more natural for them to become interested in delving into the uncertainties and excursions normally associated with military planning. In fact, in wartime, intelligence officials’ cooperation with policymakers is expected (as, for example, in the Office of Strategic Services during the Second World War). Operations research, gaming, and predictive analysis would all be needed in fighting proliferation in order to identify technical, political, and economic opportunities to slow or prevent proliferation and to disinform and sabotage the efforts of those states seeking to acquire nuclear weapons. They would also be needed to identify future proliferation-induced military vulnerabilities, both of American forces (along with those of US allies) and of potential adversaries, and ways of mitigating them.

New Intelligence Requirements

The DCI’s Nonproliferation Center has rightly focused on developing an intelligence strategy to tackle the issue of proliferation. However, development of a basic strategy document has been under way for nearly two years. In part, that is because the security environment is changing, and an intelligence strategy—like any strategy—must take account of those changes. Yet the center’s development of a strategy has been—and will continue to be—hindered by the absence of a coherent, agreed-upon, prescriptive definition of proliferation.


Of course, the intelligence community cannot be expected to develop a prescriptive definition on its own. Nevertheless, it makes sense for the community to work on such a definition for two very practical reasons. The first is that the intelligence community retains an expertise in proliferation that should not be ignored out of some overly fastidious concern about the line between policy and intelligence. Thus, the intelligence community often generates (and rightly) reports on proliferation matters without specific requests from policymakers. These reports are intended to alert policy officials to issues and trends that policy has not addressed and presume a prescriptive definition of what proliferation is. One can ignore this behavior or reflect upon it and use the expertise exhibited in the best of these reports to fashion a realistic and coherent definition of proliferation.

A second, more practical, reason for the intelligence community to help define proliferation is that without a sound, prescriptive definition, intelligence officials and agencies will be hard pressed to do their job or measure their performance. How much of a failure would it be, for example, if US intelligence failed to anticipate North Korea’s acquisition of several night-vision goggles as compared to failing to anticipate its development of crude, unmanned air vehicles that could penetrate US air defenses? Knowing which is more important is possible only with a prescriptive definition that identifies what is of greater concern and why.

One way to develop this definition without suffering the dilution of 1,001 coordinations would be for the intelligence community to work with people in the military who are most interested in proliferation problems—that is, members of command staffs charged with actually having to worry about fighting a war in their region. That has the immediate advantage of engaging people most likely to be affected by weapons proliferation and whose views are critical to giving descriptive details to any prescriptive definition once it is in place. Once a working definition has been developed, the other policy elements within the Departments of State and Defense can be brought on board. With a prescriptive definition in hand, both policy and intelligence could establish a process for gauging proliferation threats, prioritizing them, and establishing an overall strategy of diplomatic, economic, and military activities and goals for fighting proliferation.


Proliferation-related intelligence collection to date has tended to be technocentric. Tremendous amounts of attention continue to be paid to imagery and its interpretation, as well as to the collection of signals. There is also a heavy emphasis on technical matters in the way proliferation issues are handled and discussed. Generally, far more time is spent on the question of whether a particular country has a particular capability and the specific technical facts associated with its acquisition than on what such acquisition might mean economically, politically, or militarily.

To some extent, this emphasis is unavoidable: tracking nuclear weapons and missile acquisitions or developments requires rocket scientists and nuclear engineers to make sense of things. Yet, relying too heavily on these experts and focusing exclusively on their collection requirements comes at the cost of collecting the kinds of political and economic information that is needed to curtail or prevent proliferation in the first place.

In this regard, what might help most would be a greater understanding, in policy and intelligence circles, of “opportunity analysis.”11 As one proponent of this type of analysis has written, we need much more of the “kind of analysis [that] illuminates for the policy maker opportunities for advancing U.S. objectives and interests through diplomacy, military and economic moves, cultural activities, and other political action.”12 This type of analysis requires that intelligence be collected that will point to “opportunities and vulnerabilities the United States can exploit to advance a policy as well as to the dangers that could undermine a policy.”13 It is obviously important, for example, to know the technical details of Argentina’s Condor 2 missile program but no more so than knowing which elements within that country’s military are secretly opposed to the missile’s development. The former is useful to know if the effort to stem the missile’s proliferation fails, while the latter is critical to diplomatic and covert maneuvering to end the program altogether. Both kinds of information are secret, yet only one is especially relevant to policymakers looking for opportunities to defeat proliferation.

Further, in support of policymakers, intelligence collectors (and analysts) need to pay even more attention than they have in making sure they have as much unclassified material as possible on the matters about which they are collecting classified materials. That, again, may not seem a task worthy of agencies known primarily for their work with secrets, but frequently the government cannot make a demarche or inform allies of developments that require their assistance unless the information can be used in an unclassified forum.

Finally, even more needs to be done to collect information relevant to war. How and why other states might use the strategic weapons systems they are developing or acquiring is at least as important as specifics about the weapons themselves. What kinds of exploitable problems might the employment of these capabilities produce within the military and political leadership? What efforts are smaller states making to hide development or employment of strategic capabilities from American intelligence collectors? Are US and allied efforts succeeding in forcing them to change their acquisition or employment plans? How are countries that are seeking to acquire nuclear weapons reacting to efforts to stop or hinder their programs? What sectors of those governments or their populations are opposed to acquiring strategic systems? To what extent are these doubts related to American or allied actions aimed against these programs?

If intelligence officers are collecting the answers to these questions, these answers should be reflected by changes in Washington’s strategy against targeted states seeking to acquire nuclear weapons and also by changes in US collection requirements. Indeed, if collection requirements stay the same, it is a sign that the intelligence community is not collecting what it should or that the policy community is failing to implement an effective strategy.

Analytic Requirements

The effort to fight proliferation will require more than just new targets for collection; intelligence analysis will also have to change. In particular, to support policymakers looking for opportunities to disrupt, slow, or stop a proliferation program of another country, an increased need will exist for analysis that lays out the uncertainties and variables connected with a particular program. That is true not only in the technical arena (for example, what engineering bottlenecks remain for country X to complete project Y; and what are the range of possibilities for country X to master them?) but in political and economic affairs as well.

Recently, we have seen how nuclear and missile programs in countries such as Argentina, South Africa, Taiwan, and Brazil have been either terminated or suspended because of political considerations or economic factors. The United States and its friends could have done more to reinforce some of these forces earlier, had more collection and analysis existed on what the various domestic constituencies for and against these programs were.

Taiwan’s decision in 1991 not to develop a space-launch vehicle is a case in point. After considerable internal debate, Taiwan decided to focus its development efforts in the area of satellite technology rather than rocketry. Knowing who supported what, for what reasons, and what political and economic costs they were willing to pay to pursue their aims was critical for policymakers who wanted to move Taiwan along a more benign path of satellite development.

Beyond this, it would also be helpful to have analyses of how each of these countries could better meet the peaceful goals they claimed they were pursuing by investing in these projects (for example, civilian nuclear and space-launch vehicles, large mainframe supercomputers, and so forth). This analysis is especially useful if one is to mount effective public diplomacy efforts against proliferation developments. Also, one must do counterintelligence analysis on how incipient nuclear states are likely to hide their activities or how they plan to get around weaknesses in existing safeguard and inspection regimes and how they plan to avoid intelligence-collection efforts. Such analysis goes beyond uncertainty analysis into the realm of excursions. This work need not be done by intelligence analysts alone; it can be contracted out. In any case, it is work that should be jointly managed by intelligence and policy officials who have a clear grasp of the facts and know what kind of analysis is needed.

This requirement for gaming, economic analysis, and operations research is even clearer in the case of developing military threat analyses. Here, the involvement of the intelligence agencies, particularly those of the military services, is imperative. Without their involvement, no threat assessment—no matter how correct—is likely to alter defense requirements in the areas of weapons acquisition and development or service training or doctrine. Again, such involvement does not come naturally. As one Defense Intelligence Agency analyst explained to me, “We don’t do excursions; they are too hypothetical.” When analysts in the defense intelligence services attempt to do such work, moreover, it is often heavily edited and reduced to banality out of concern that it might upset the military services, who have a large stake invested in their five-year spending plans. They do not need or want any second-guessing—implicit or otherwise—generated by intelligence officials. Unfortunately, such a perspective can be literally self-defeating.

One way to change that is to sponsor threat assessments by analysts from outside the government and arrange for the military commands or service staffs to participate in their production. The money is likely to be there: for fiscal year 1996 alone, the policy arm of the Department of Defense spent several million dollars on proliferation studies, and the Energy Department and the Central Intelligence Agency together spent orders of magnitude more on this same set of issues.

Given this spending, it is important that a concerted effort be made to focus and manage the government’s study efforts. The National Intelligence Council or the Intelligence Community Management Staff might play a useful role in seeing that this money is used to develop the right kind of analyses. They have a solid bureaucratic interest in doing so, since these studies should have a direct impact on intelligence-collection requirements. At a minimum, some effort is needed to keep track of what is being done by the government as a whole. Without such monitoring, matching analysis on proliferation to the govern- ment’s nonproliferation strategy will be nearly impossible.


Clearly, if the US government wants to do more than just react to the proliferation of strategic weapons capabilities, the role of intelligence in this fight will have to change along with policy. If the United States is to anticipate developments in the proliferation of strategic weapons; slow or reverse them through diplomatic, political, or economic appeals; or develop military options for coping with their employment, the government will have to commit itself to a long-term strategy of competition not unlike what it did during the cold war.14 The heart of such a strategy is to match US strengths against an adversary’s weaknesses in an effort to force it into less threatening areas of competition.

Instead of engaging in one major competition—as the United States did with the Soviets—the United States will have to engage in and manage a varied number of competitions against several suppliers and acquiring states. As with the earlier competition with the Soviets, the US will have to anticipate each of these entities’ reactions to US and allied moves to fight their actions with regard to the proliferation of strategic weapons and be able to maintain relative advantage in defeating or mitigating these moves. Yet, as with the Soviets, the endgame is the same: America’s goal should be to contain the threat until each government trying to acquire strategic weapons is defeated by its own internal contradictions and gives way to a new, more peaceful regime.

Iran, for example, despite severe economic difficulties, continues to pursue costly nuclear programs and conventional military systems in an effort to dominate the Persian Gulf and its neighbors. Although the United States must worry about the military implications of Iran’s acquiring these weapons and defending against them should they be employed, a competitive strategy might also attempt to check Tehran’s proliferation activities by challenging Iran in areas where it is especially vulnerable and where the United States and its allies hold key advantages. Specifically, in conjunction with allies, Washington and its allies could take additional measures to lessen Iran’s access to Western economic assistance and assets as long as it pursues programs dangerous to US and allied interests in the region. A second element in a competitive strategy might be to encourage Tehran to spend more on defensive weaponry by regularly demonstrating the decisive nature of US and allied air superiority in the region and the potential cost to Iran of ignoring its air defenses. A concerted effort to bring this point home to the leadership in Tehran by a vigorous enforcement of UN resolutions against Iraq might result in the Iranian government’s spending its limited resources in a more benign area (air defense) rather than in a more dangerous one (long-range missiles). Finally, the US and its allies should be more alert to Iranian political and domestic opposition that might divert Iran from its current, hostile course.15

Again, developing and implementing such a competitive strategy will require several changes in the way intelligence and policy officials currently address proliferation problems. As I have argued, first, it will require that they at least agree on a prescriptive definition of what it is they are fighting. Second, it will require that they reconsider the basic relationship between the policy and intelligence realms that currently makes fighting proliferation so difficult—if not impossible. It simply is impractical for policy and intelligence officials to continue to avoid cooperating on the kinds of operations research and uncertainty analyses needed to gauge and give priority to the potential proliferation threats the United States faces. To assure any meaningful level of success in this area, the role of the military—both the service staffs and the commands, relative to the current actors in proliferation-related policy and intelligence—will have to grow.

Developing such a strategy against proliferation will not be easy. It will take more than a month or even several years. But Washington should be in no rush to get it wrong. Certainly, our experience in Iraq has made that much clear.


1. That is not to say that the shift was not anticipated. For the creation of the US Department of Defense Proliferation Countermeasures Working Group, see, for example, Barbara Starr, “DoD to Track TBM Proliferation,” Jane’s Defence Weekly, 10 February 1990; and idem, “Third World SSM Threat Studied,” Jane’s Defence Weekly, 16 November 1991.

2. See, for example, Senate Committee on Governmental Affairs, testimony of Director of Intelligence James Woolsey, hearing on “Proliferation Threats of the 1990s,” 102d Cong., 2d sess., 24 February 1993; and House, Department of Defense Appropriations Act for FY 1993 (PL 102-396), 102d Cong., 1st sess., 1992, H.R. 5504, sec. 1607.

3. For example, see agenda, “Counter Proliferation” conference, Arms Control and Disarmament Agency, Washington, D.C., 24–25 May 1993, which emphasizes technology transfer, export controls, verification technologies, and information management.

4. See Les Aspin, secretary of defense, remarks before the National Academy of Sciences Committee on International Security and Arms Control, 7 December 1993.

5. See House, Department of Defense Appropriations Act for FY 1993, sec. 1607; and House, The National Defense Authorization Act for FY 1993, 102d Cong., 2d sess., 1992, H.R. 2100, sec. 1336.

6. For a discussion of the potential strategic significance of submarines in closed seas; accurate conventional missiles; and command, control, communications, and intelligence (C3I) technologies, see Henry Sokolski, “Nonapocalyptic Proliferation: A New Strategic Threat?” The Washington Quarterly, Spring 1994, 115–27.

7. See House, Department of Defense Appropriations Act for FY 1993, sec. 1607.

8. An unclassified presentation of this approach can be found in Senate Committee on Governmental Affairs, Proliferation and Regional Security in the 1990’s, 101st Cong., 2d sess., 9 October 1990, 49–54. The approach has also been considered within the Department of Defense’s Defense Intelligence Agency. See Dan Spohn, “Proliferation Production,” Staff Working Paper of the Office of the Defense Intelligence Officer for Research and Development, October 1992.

9. This might seem to be what counterproliferation is all about. In fact, it is something different. Indeed, it is impossible to neutralize weapons of proliferation concern since by definition, they are systems for which we lack adequate military countermeasures. We can limit the damage such weapons might inflict through active and passive defenses and offenses. Neutralizing these threats with effective military countermeasures in the way we do with electronic countermeasures against enemy radars, though, is not yet possible. If it were, these weapons, by definition, would no longer be of proliferation concern but merely of military concern: they would not be capable of inflicting strategic harm, assuming they were properly opposed.

10. Abram Shulsky and Jennifer Sims, What Is Intelligence? (Washington, D.C.: Consortium for the Study of Intelligence, 29 April 1992), 32.

11. See Jack Davis, “The Challenge of Opportunity Analysis: An Intelligence Monograph” (Langley, Va.: Center for the Study of Intelligence, Central Intelligence Agency, 1992).

12. Kenneth deGraffenreid, “Intelligence and the Oval Office,” in Intelligence Requirements for the 1980’s: Intelligence and Policy, ed. Roy Godson (Lexington, Mass.: Lexington Books, 1986), 28.

13. Davis, 7.

14. See Andrew W. Marshall, Long-Term Competition with the Soviets: A Framework for Strategic Analysis (Santa Monica, Calif.: RAND, 1972); and idem, “Competitive Strategies: History and Background” (Washington, D.C.: Office of Net Assessment, Department of Defense, 3 March 1988).

15. See Ken Timmerman’s “Opportunities for Change in Iran”—chapter 12 of this volume—and Patrick Clawson, Iran’s Challenge to the West: How, When, Why (Washington, D.C.: Washington Institute for Near East Policy, 1993).