Improving Methods for List Construction and Review
Implementing U.S. export control policy involves balancing conflicting national interests. Achieving that balance requires trade-offs between the potential benefits that can accrue to national security from export controls and the potential costs that such controls can impose on U.S. foreign policy and economic objectives. Principal among U.S. foreign policy interests is encouraging economic liberalization and democratization in the Soviet Union and other countries that are the targets of controls. On the economic side, a principal interest involves building and sustaining U.S. economic strength and competitive position in world markets. If the control system was truly multilateral, the economic cost of controls would be modest, primarily forgone direct sales to the targeted destinations. If the United States continues to implement controls unilaterally, the costs to U.S. economic strength and competitive position in world markets could be substantial.
The process the United States has been using for list construction and review for East-West controls has not struck a balance among national interests in an efficient and satisfactory way. Although the list review undertaken by the Coordinating Committee for Multilateral Export Controls (CoCom) in mid-1990 has changed this situation somewhat, many fundamental problems remain. Historically, there have been five basic difficulties with the U.S. system for list construction and review:
The absence of significant constraints on the defense and intelligence communities with respect to the number of items that can be controlled. As
a result, there is an incentive to list everything that might conceivably have utility to the targeted destinations.
The absence of any mechanism that forces the defense and intelligence communities to reveal the strength of their preferences among the items they propose to control. The result is that it has often been impossible for high-level decision makers to determine the relative degree of criticality of the items proposed for control.
The absence of any mechanism that forces the foreign policy and commercial communities to reveal their views about the relative costs to short-and long-term U.S. foreign policy and trade interests that would result from controlling an item.
The absence of clear decision-making authority to ''balance the national interest." This, combined with bureaucratic political maneuvering, has led to interagency gridlock, which has proved remarkably resilient to legislative "fixes."
The absence of a means by which the system can be easily tuned to respond to changing international political and military circumstances or to changing priorities and judgments of the national political leadership.
This chapter presents an improved method that addresses these five problems. An improved method can lead to more efficient and effective list construction and review, but it is not a magic solution. Consistent, high-level political will can make a variety of methods work; lacking such will, even very good methods may only marginally improve the system.
The process for choosing items for control within any particular control regime should involve the following:
Identification of items of potential concern.
A rank ordering and weighting of items in terms of the national security risks posed by an adversary's acquisition and use of each item, with careful consideration given to the controllability of items.
An approximate rank ordering and weighting of items in terms of the economic and foreign policy costs of restricting trade in each item of concern.
A policy judgment as to how the risks and benefits of control should be balanced.
A comparison of benefits and costs that allows a sorting into controlled and uncontrolled items.
This chapter focuses on trade with the Soviet Union, beginning with how a control list should be constructed. A similar approach can also be applied, however, to other possible CoCom-controlled destinations, such as Eastern
Europe and China, and to the control of technologies important in the proliferation of advanced conventional weapons and weapons of mass destruction. The focus throughout is on developing a general philosophical approach. Many detailed policy and administrative judgments would be required to specify all the operational details of a new system. The ongoing construction of the CoCom core list is methodologically ad hoc. Future list review, as well as possible future reconstructions, would benefit from a more systematic approach.
REDESIGNING LIST CONSTRUCTION AND REVIEW FOR EAST-WEST CONTROLS
Identify Items of Concern
Many dual use items might provide military benefits to the Soviet Union. These can include systems, individual products, critical components, unique or exotic materials, associated test and calibration equipment, software, and technical data and know-how. For simplicity, these have been referred to throughout this report as "items."
If the U.S. objective was to control, without constraints, all items that might conceivably provide some military benefit to the Soviets, list construction would be simple. Listing any item would only require an argument about how, if acquired, that item might contribute to Soviet military capabilities. A balancing of benefits and costs, however, requires a more focused determination. For this, two refinements are required. First, the objective to be served by control must be stated precisely. Second, items to be controlled must be defined with precision so that judgments about relative preferences for control can be made.
Broadly, trade in dual use items can provide military benefits to an adversary in four ways:
Immediate insertion of items into military systems. Examples include substitution of a more reliable Western computer chip in a military system and insertion of a modern telephone switch into an existing military command and control system.*
Direct insertion of items into future military systems. Examples include design of a high-performance Western computer chip into a new weapons system and design of a high-performance array processor into a sonar target acquisition and tracking system.
Short-term enabling of military systems. Examples include acquisition of high-performance machine tools that can make better tank fire-control systems or quieter submarine propellers.
Long-term enabling of military capabilities. Examples include almost any item that could accelerate or strengthen the development of the basic capabilities of Soviet high-technology industry, such as large numbers of medium-performance engineering work stations and joint ventures or technical exchanges that result in significant numbers of Soviet citizens learning how to manage manufacturing and research and development operations efficiently.
Any system that is designed to identify militarily critical items for potential control, however, will bog down in conflict and disagreement unless it receives clear presidential policy guidance through a national security directive (NSD) that specifies the objectives it is supposed to address and with what relative importance.
Once a class of items has been identified for potential control, it must be subdivided into smaller groups so that decisions can be made about the relative importance of controlling items with different characteristics. For example, suppose that under the general class of items called "computers" both high-performance engineering work stations and supercomputers are proposed as candidates for control. The security risks posed by trade in these two items are quite different. Both may be candidates for control, but the relative importance of controlling the two items is not simply a technical judgment; it also depends on the policy guidance that has been provided.
In order to determine the relative importance of controlling the two items, it must be possible to differentiate between them. The current system does not provide for such differentiation; rather, current regulations state only that "computers with more than certain specified capabilities are controlled." The result is that if senior decision makers try to make balancing decisions among conflicting national interests, they are unable to determine how much more or less important military and intelligence authorities believe it is to control supercomputers versus advanced work stations. Under the current system, discussion of any such trade-off explodes into a mass of details, for which senior decision makers may not have the requisite expert knowledge or the time to analyze fully.
Given even the few classes of items that are proposed for control on the CoCom core list, the number of potential individual items is enormous. In order to make comparisons feasible, items must be grouped into sets. For convenience, these sets are called item-groups.*
Defining item-groups poses some important technical problems. Details are discussed in Appendix J.
Identify the Relative National Security Benefits of Controls
Once a list of items proposed for control has been assembled and sorted into item-groups, military and intelligence authorities must assess the relative security risks that would be posed by trade with the Soviet Union in each item-group. To make these judgments, the following information is required:
High-level policy guidance, in the form of an NSD, that spells out control objectives. Without such guidance, military and intelligence authorities may have difficulty making judgments about the relative importance of trade in different item-groups because they will not be able place their judgments in the context of U.S. policy objectives.
An explicit assessment of the nature of current and likely future military threats posed by the Soviet Union and an understanding of how Western dual use items might contribute to the evolution of such threats. This is needed so that the relative importance of candidate item-groups can be assessed.
An assessment of the controllability of each item-group. If the objective was simply to deny, without any limits, export of all items that might conceivably provide some military benefit to the Soviet Union, this factor would not matter. Even a very leaky export denial system would, to some limited extent, probably impede acquisition of Western technology. But export denial is not the objective. The objective is to limit the contributions of Western dual use items to the Soviet military while simultaneously balancing foreign policy and trade objectives. (The issue of controllability is discussed at the end of this chapter.)
In light of these three factors, defense authorities, with advice and assistance from the intelligence community, should take the following steps:
Place the entire list of item-groups into a rank ordering, from those in most critical need of control to those that are least in need of control.
Allocate a finite number of points (e.g., 1,000) across the item-groups in proportion to the intensity with which control is desired.
Such a process of ranking and weighting will pose some challenges to defense and intelligence authorities. Appendix J discusses how this might be done. Because it is likely to be sensitive, the weighted list should probably be classified.
Identify the Relative Economic and Foreign Policy Costs of Controls
If the control system was truly multilateral, so that every obstacle to trade that was faced by U.S. suppliers was faced by every other exporter, the
economic costs of controlling exports to targeted nations would be fairly modest. As long as export control is not strictly multilateral, potentially much larger costs in terms of the impact on U.S. competitive advantages, such as early market position and relative reliability of suppliers, must also be considered.
Considerations of foreign policy costs must weigh the potential advantages to U.S. national security of promoting the current processes of restructuring, openness, and democratization that are under way in the Soviet Union. In theory, a full ordering and weighting of item-groups proposed for control could be undertaken to reflect considerations of U.S. national foreign policy and international competitiveness. Although such an exercise may be feasible for the hierarchically organized defense community, it may not be feasible for the heterogeneous groups that represent U.S. national foreign policy and trade interests.
Despite this, some approximate indication of the foreign policy and economic costs of controlling various item-groups would be helpful to decision makers faced with balancing national interests. Two possibilities exist. Costs could be factored in through a qualitative process, perhaps informed by advisory groups and by opportunities for public comment. Alternatively, an independent advisory group appointed by the secretaries of state and commerce could sort item-groups into a small number of very broad categories, such as the following:
Category 1: trade with the Soviet Union in this item is of great importance to meeting U.S. foreign policy and/or trade objectives.
Category 2: trade with the Soviet Union in this item is of importance to meeting U.S. foreign policy and/or trade objectives.
Category 3: trade with the Soviet Union in this item is of limited importance to meeting U.S. foreign policy and/or trade objectives.
The NSD should provide guidance on the factors that should be considered in doing the sorting and weighting of items in the first two categories.* In contrast to the defense/intelligence weights, there is no reason to classify the results of this process. Indeed, making the results public should help to stimulate a full and balanced consideration of the relevant issues.
Compare and Balance Benefits and Costs
After the evaluation of the several national interests, the final step is to strike a balance among those interests and develop a list proposal that will guide the U.S. position in CoCom. Creating this balance requires that the NSD specify two things:
a decision-making process and a responsible decision maker, and
guidelines as to how the balance among interests is to be struck.
Balancing the several national interests involves more than simply looking at the rank ordering provided by the defense and intelligence communities and deciding where to draw a line. When the foreign policy and trade costs of controls are combined with the estimated national security benefits, the actual order of item-groups, in terms of "net benefits," may change.
If the process of enumerating foreign policy and trade costs is qualitative, the final estimation of net benefits and the resulting ordering of the list will similarly have to be qualitative. If foreign policy and trade costs have been sorted into several categories, as outlined above, a somewhat more quantitative sorting may be undertaken to guide and assist decision makers in regularizing the list construction process and enable them to focus their attention on those decisions that most require their powers of qualitative judgment.
In either case, the end result should be a proposed control list that can be presented for multilateral consideration in CoCom. In principle, the identification, prioritizing, ranking, and balancing tasks could be repeated on a multilateral basis within CoCom for final development of the CoCom list. It seems unlikely, however, that many of the CoCom allies would be prepared to adopt so systematic and labor intensive an approach to list construction. If the United States adopts a strategy based on this philosophical framework, however, the multilateral process should be able to share the benefits of the results.
List Review, Sunsetting, and Occasional Reconstruction
It is likely that the new control list would be shorter than previous lists. This should make it possible for CoCom to review the entire list annually. Indeed, review of the entire list as a single process is essential if the balancing of interests is to be achieved.
In order to develop the U.S. position for the annual process of CoCom list review, the internal U.S. evaluation and prioritization effort should be repeated annually. Although the first attempt will be time consuming, future annual cycles should be fairly simple.
In addition to annual review, a more automatic process that regularly removes older items from the CoCom list should be implemented. This might be done in a number of ways. The boxed insert "One Possible Procedure for Sunsetting CoCom List Items" provides a specific example. Another example is in the field of supercomputers. In the absence of periodic reviews, consideration should be given to other review mechanisms, such as indexing
One Possible Procedure for Sunsetting CoCom List Items
The West has, on average, been unable to sustain more than an eight-year technological lead over the Soviet Union. If, as expected, the Soviet Union grows increasingly open over the next few years, and if the perceived and actual threats continue to decline, export controls will probably be unable to maintain even this much of a lead.
These considerations suggest that technologies that are more than eight years old should be removed from the CoCom list. The removal process could be automated if, through its member governments, CoCom maintained a data base to which manufacturers could send information whenever they ship a new product that advances any performance parameter used in specifying a CoCom control. Eight years after the date of the first commercial shipment of that product, items with that level of performance would automatically come off the list. Submissions to the data base would be certified as authentic by participating member governments, subject to challenge through CoCom. The data base would be maintained as an open public record so that manufacturers in any CoCom country could have access.
A few special technologies may warrant control even after their performance specifications are more than eight years old. If CoCom wished to continue to control such a technology, before the end of the seventh year CoCom members could act unanimously to relist the item for up to an additional eight years.
Such a system could be started in either of two ways. Companies could submit shipping records over the past eight years in order to construct the necessary base lines, or the system could be applied only prospectively to new technologies.
(raising the threshold for controls as overall item-group performance capabilities increase).
Beyond annual list review, it may, from time to time, become appropriate to undertake additional reconstructions of the CoCom list. There are two reasons why such periodic reconstructions might be useful:
The direction of the CoCom unanimity rule is reversed so that unanimity becomes necessary to put something on the list rather than take it off. This could help to reduce the accumulation of low-end items, a problem that has plagued the CoCom list management process in the past.
The promise of list reconstruction, tied to accomplishment of certain specified milestones in arms control, might provide a useful additional incentive to the Soviet Union to be more forthcoming in strategic arms negotiations.
GENERALIZATION TO OTHER CONTROL REGIMES
The philosophy of list construction for controls on the Soviet Union can be generalized to other countries and other export control regimes. The same need to balance national interests applies in the context of other CoCom-controlled destinations, such as the People's Republic of China, and to destinations of proliferation concern. Because U.S. national security, foreign policy, and economic interests are significantly different in these contexts, however, separate prioritizations are likely to be necessary, although they should benefit greatly from the analysis that is done in support of controls on exports to the Soviet Union.
In the long run, export controls are a defensible policy tool only if they are applied to items that are, in fact, controllable. Two kinds of controllability are important in East-West dual use export control: export controllability and end-use controllability. The first refers to the feasibility of denying or significantly limiting exports from the West to the targeted countries. The second refers to the extent to which, once exported to a targeted country, a dual use item can be prevented from being used to direct military advantage.
Effective East-West export controls require that the items to be controlled have the following properties:
Manufactured and/or sold by only a modest number of suppliers whose actions can be controlled.
Consumed or used by only a modest number of consumers whose export actions can be controlled.
Individually traceable or not easily concealed or disguised.
Other, rather different properties may apply in the context of proliferation controls.
Modest numbers of suppliers and consumers are required to make policing possible. If the number of either becomes large, it becomes impossible to keep track of everyone, and the odds of undetected diversions increase rapidly. Similarly, if individual items cannot be accounted for, they can be easily concealed or disguised and/or can easily get "lost" and be diverted. The current system of foreign availability addresses only the first of these requirements. If items of concern are manufactured only in CoCom countries,
or in countries subscribing to CoCom rules, the first requirement above can probably be met. If items are also manufactured in countries that are unwilling to abide by CoCom rules, however, this requirement cannot be met. It is for this reason that Congress implemented the foreign availability rules, which require the decontrol of items that are available in substantial quantities from countries that will not agree to abide by CoCom rules.
Foreign availability constitutes one dimension of controllability. Incorporating foreign availability considerations into the determination of controllability as an integral component of annual list review will obviate the need for a separate and independent foreign availability assessment process. Several other kinds of uncontrollable items can also be defined, three of which are listed below:
Items that, through economies of mass production and mass marketing, have evolved into the status of "commodities."
Many kinds of software (some but not all of which also have achieved commodity status).
Many kinds of technical know-how.
Personal computers and countless other once exotic, high-priced technologies like them have become commodities. Because they violate all three of the attributes for a controllable item listed above, technologies that have become commodities are not controllable.
Although many people "know a commodity when they see one," defining a commodity poses some challenges. One possible definition is provided in a boxed insert. Some computer software has achieved commodity status. Other software that is not yet a commodity is nevertheless uncontrollable because it is widely available and easily accessed or transferred over public and quasi-public computer networks.
Effective end-use controls require that items to be controlled have the following properties:
Used or consumed by only a modest number of entities.
Used or consumed in an environment in which access can be limited and/or in which users can be positively identified.
Used or consumed by people and organizations that have the authority, means, and will to limit access.
Under the system proposed in Chapter 8, determination of end-use controllability would shift from national decision makers to a multilateral basis within CoCom.
In order to make such a system work, the United States will have to persuade CoCom to develop* collectively, and regularly promulgate, general
A Definition of Commodities
By nature, a workable definition of a commodity must be somewhat arbitrary. The following is one possible approach to identifying item-groups that are commodities.
The item-group must involve the following:
policy guidelines on the attributes of an acceptable end-use control system. In addition, if end-use arrangements are approved for specific classes of items (e.g., software and data), CoCom will have to develop** descriptions of specific strategies that have proved to be acceptable. Although following a previously successful strategy in an application for approval of an end-use arrangement might speed CoCom approval, it should be explicitly stated that following such a strategy is not a requirement for approval. Exporters should always be free to innovate and to propose alternative strategies. To ensure that CoCom's criteria for judging the acceptability of proposed end-use controls do not become outdated by rapidly evolving political realities, the United States should urge CoCom to reconsider and revise its end-use control guidelines on a regular basis.