APPENDIX
J

Some Details on the Proposed Method for List Construction and Review

Chapter 10 outlined a general framework for approaching the problem of list construction and review. It included the following basic elements:

  1. Identification of items of potential concern.

  2. A rank ordering and weighting of items in terms of the security risks posed by trade in each item, with careful consideration given to the controllability of items.

  3. 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.

  4. A policy judgment as to how risks and benefits should be balanced.

  5. A comparison of benefits and costs and a sorting into controlled and uncontrolled items.

The panel did not consider it appropriate to make the detailed policy and administrative judgments that would be required to spell out all the operational details of such a system. This appendix, however, provides brief additional discussion of three issues:

  1. defining an item-group;

  2. the feasibility and necessity of rank ordering item-groups; and

  3. the use of quantitative analysis to assist in list construction.

DEFINING AN ITEM-GROUP

To facilitate judgments about the need to control different items within a broad class of items (an internationally accepted standard category, for in-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment APPENDIX J Some Details on the Proposed Method for List Construction and Review Chapter 10 outlined a general framework for approaching the problem of list construction and review. It included the following basic elements: Identification of items of potential concern. A rank ordering and weighting of items in terms of the security risks posed by trade in 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 risks and benefits should be balanced. A comparison of benefits and costs and a sorting into controlled and uncontrolled items. The panel did not consider it appropriate to make the detailed policy and administrative judgments that would be required to spell out all the operational details of such a system. This appendix, however, provides brief additional discussion of three issues: defining an item-group; the feasibility and necessity of rank ordering item-groups; and the use of quantitative analysis to assist in list construction. DEFINING AN ITEM-GROUP To facilitate judgments about the need to control different items within a broad class of items (an internationally accepted standard category, for in-

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment FIGURE J-1. Hypothetical parameter space for control of a particular class of items. In this example, the highest performance item of this class that is made has p1 = P1max and p2 = P2max. Items with P1c > p1 > P1max and P2c < p2 < P2max are controlled. Items with p1 < P1c and p2 < P2c are not controlled. stance, Harmonized System category 8471.91—processing units for computers), items should be sorted into item-groups. In the current list, classes of items are controlled by specifying a set of performance parameters, such as bandwidth and operating frequency. Typically, items that have performance parameters above those thresholds are controlled, and all those lying below those thresholds are decontrolled. Suppose, for example, that for a particular class of items, I, two parameters are specified as important in making control decisions, p1 and p2. Items are controlled if p1 > P1C or if p2 > P2C. Suppose that the highest performance version of I has p1 = P1max and p2 = P2max. The situation is shown graphically in Figure J-1. The shaded portion of this figure represents the subset of items of type I that are subject to export control. The objective in defining item-groups is to cluster items that lie in the shaded region into a finite number of groups. For example, if the controlled range of both p1 and p2 was divided into two equal intervals, the result would be the creation of the eight item-groups labeled I1 through I8 in Figure J-2. To make the rank ordering and weighting process feasible, enough item-groups must be defined to allow distinctions to be made, but without over-

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment FIGURE J-2. Hypothetical division of the parameter space into 8 item-groups, I1 through I8 whelming the sorting process. Also, items must be spread across groups with enough uniformity that ''all the action" does not end up in a single item-group. Item-groups could be defined in several ways: With fixed increments in the parameter values (e.g., every factor-of-10 increase in pj defines a new item increment). By dividing according to natural "technological generations." By sorting specific products. Specific definitions should be left to the technical working groups that possess this expertise. However, the procedure should lead to a sorting that has approximately the following characteristics: Each Harmonized Custom code category used in the list should be divided into not fewer than 5 and not more than 20 item-groups. Each item-group defined should contain at least one actual product on the market. No item-group should contain more than 20 percent of the products or 20 percent of the annual sales by U.S. manufacturers in a custom code category.

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment New item-groups should be defined as needed to accommodate new higher performance items that cannot be appropriately included in the highest existing group. In doing this, boundaries around existing groups should not be changed unless required to remain consistent with the constraints outlined above. As item-groups are decontrolled, boundaries around remaining groups should not be changed unless required to remain consistent with the constraints outlined above. THE FEASIBILITY AND NECESSITY OF RANK ORDERING ITEM-GROUPS In Chapter 10, the panel proposed that defense decision makers, with advice and assistance from the intelligence community, begin the process in the following manner: Place the entire list of item-groups proposed for control into rank order, from those item-groups judged to be in most critical need of control to those least in need of control. Allocate a finite number of points (e.g., 1,000) across the item-groups in proportion to the control desired. Some may argue that such a rank ordering is not possible. The panel's judgment is that it is not only possible but necessary if a balancing of national interests is to be achieved. Such a weighted ordering might be done in a variety of ways. First, several broad categories might be constructed, based on the threat assessment, high-level policy guidance, and judgments of controllability. If it seems to be appropriate, a similar strategy can be adopted to achieve further refinements within categories, especially those that are most critical. Once the point is reached at which there is no clear preference between item-groups being compared, the process has gone far enough. Details, such as selecting the kinds of analysis that could best support the necessary decision making, are best left to those who oversee the task. However, an intelligence-based analysis of Soviet weapons systems development could be useful. USE OF QUANTITATIVE ANALYSIS TO ASSIST IN LIST CONSTRUCTION In the list construction method proposed in Chapter 10, defense and intelligence decision makers identify dual use items of potential concern and then rank order them and assign weights to each item-group in terms of the security risks posed by trade to the targeted country. For simple illustration, suppose that only five item-groups are of concern and the process of ranking and weighting has produced the following:

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment Item-group F 470 points Item-group B 336 Item-group R 113 Item-group A 53 Item-group C 28   1,000 The weights indicate an assessment of the relative cost to national security that would result from unlicensed trade with the targeted country in each of the item-groups. As discussed in Chapter 10, a full weighted rank ordering of foreign policy and economic costs is not feasible. For purposes of illustration, however, suppose that it is. The communities concerned with foreign policy and international trade might produce a weighted list of the following sort to indicate their judgment of the relative costs of restricting trade in each item-group: Item-group R 470 points Item-group C 336 Item-group F 109 Item-group A 53 Item-group B 32   1,000 If guidance provided in the national security directive on export controls led the decision makers doing the balancing to weight foreign policy and trade considerations with half the weight of national security considerations, the approximate net benefit of controlling each item-group could then be computed as follows: Item-group F 470-(0.5*109) = 416 points Item-group B 336-(0.5*32) = 320 Item-group A 53-(0.5*53) = 26.5 Item-group R 113-(0.5*470) =-122 Item-group C 28-(0.5*336) =-140 Comparing the order of this list (F, B, A, R, C) with the order of the original defense list (F, B, R, A, C), one of the important consequences of doing the full net-benefit calculation is that it can rearrange the order of items on the list. Thus, computing net benefit is not the same thing as simply deciding where to draw the line in the defense-ordered list. Under the hypothetical circumstances illustrated here, item-groups F and B should be controlled, since the benefits of control clearly outweigh the costs. Item- groups R and C should be decontrolled, since the costs clearly outweigh the benefits. Item-group A should be looked at with greater care

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment before deciding whether to control or decontrol, since costs and benefits are fairly close. Because the heterogeneity of foreign policy and trade interests prevents the costs of control ever being estimated with the sort of precision used in this example, such a complete benefit-cost calculation is not possible. Suppose, however, that it is possible to sort the foreign policy and trade costs of controlling item-groups into a small number of categories, such as the following (introduced in Chapter 10): 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; or, Category 3: trade with the Soviet Union in this item is of limited importance to meeting U.S. foreign policy and/or trade objectives. Then, foreign policy and trade decision makers might produce the following partly ordered list: Item-group R Category 1 Item-group C Category 1 Item-group F Category 2 Item-group A Category 3 Item-group B Category 3 Suppose again, for illustration, that guidance provided in the national security directive leads decision makers to assign no weight to any item-group ranked as category 3, and twice the weight to items in category 1 as items in 2. Continuing to weight foreign policy and trade considerations with half the weight of national security considerations, there are 500 points to be allocated across two item-groups in category 1 and one item-group in category 2. Under these conditions, each item-group in category 1 should be given a weight of 200, and in category 2 a weight of 100. A crude calculation of net benefit can then be performed as follows: Item-group F 470-100 = 370 Item-group B 336-0 = 336 Item-group A 53-0 = 53 Item-group R 113-200 =-87 Item-group C 28-200 =-172 Once again, item-groups F and B clearly should be controlled, and item-group C should clearly be decontrolled. Because of the cruder nature of the estimate, the details of the case for item-groups A and R should probably be looked at more carefully before a final decision is made.

OCR for page 349
Finding Common Ground: U.S. Export Controls in a Changed Global Environment One would never want to make decisions slavishly on the basis of such quantitative evaluations of the net benefits of proposed control. However, as a "decision aid" to assist decision makers to regularize the list construction process and focus their attention on the decisions for which their powers of qualitative judgment are most needed, a quantitative approach, of the sort briefly illustrated here, could be very useful in supporting the implementation of a specific decision process within the general philosophical framework proposed in Chapter 10.