The Foreign Service Journal, December 2009

D E C E M B E R 2 0 0 9 / F O R E I G N S E R V I C E J O U R N A L 39 were provided to each country for only half of the missiles. In September 1991, the United States judged that the So- viet Union had probably violated the INF Treaty and reaf- firmed that the Soviet Union had negotiated in bad faith. But even the most intrusive verification regime does lit- tle good if policymakers can’t decide if an activity is per- mitted or prohibited. The proposed FissileMaterial Cutoff Treaty provides a good example of this problem. The FMCT would prohibit the production, after a particular date, of fissile material for use in nuclear weapons. Pro- duction of fissile material prior to the cutoff date would not violate the treaty. Neither would production of fissile ma- terial after the cutoff date — if it is for non-weapon pur- poses. Here’s the rub: If, either through national means or in- spection, a cask of fissile material is detected, in order to determine if the material was permitted or prohibited, you have to know whether it was produced before or after the cutoff date, and for what purposes it was produced. Using our best technology, we might be able to determine when the material was last processed, but not always when it was produced. Iranian claims that its uranium enrichment pro- gram is for peaceful purposes offer an example of the chal- lenges of determining the purpose of fissile material production. Similarly, if there were transparency visits or inspections under the Biological Weapons Convention, inspectors might be able to determine that, for example, botulism toxin was being produced. But that would not necessarily, or even in most cases, provide the data to determine whether the production is for weapons purposes or for “peaceful or prophylactic” purposes. In each of these cases, the limited item has to be weighed against the data one can reasonably expect to collect. This enables a determination of the degree of verifiability. This technical assessment must then be weighed against broader questions including the compliance record of each of the parties, the potential significance of violations, and the constraints the agreement places on U.S. programs that might offset noncompliance. On-Site Inspections Specifically, here are some basic questions that must be asked, and answered, early on. What is it that we are try- ing to limit or prohibit? What methods do we have to dis- cover noncompliance? Are prohibited items big, unmovable and hard to hide, or are they small, mobile and easy to hide? And how great a threat do the limited or pro- hibited systems pose to the U.S.? The answers to all these questions should form the basis of developing a verification plan. The methods for discovering noncompliance include satellite intelligence or other technical collection, data dec- larations and exchanges, cooperative measures where the verified party enables our satellite intelligence, and on-site measures. On-site inspections can be critical when satellite observation is unlikely to detect noncompliance. On the other hand, OSIs have significant limitations, starting with the fact that the verified party is a sovereign nation and has to agree to all elements of the inspection. Even when the verified party does agree to an on-site inspection, that is essentially just a snapshot of activities at that particular time and place. The hope was that challenge inspections would remedy some of the limitations of on- site inspections, but they contain their own limitations. No agreement permits mere “fishing expeditions,” so the party requesting an inspection must establish some relationship to the items governed by the treaty. Because on-site inspections are conducted on the terri- tory of a sovereign nation, inspectors must be granted visas to fly to the country and be transported to the inspection site by the inspected party. Depending on what activities are governed by an agreement, there may be miles of ground and hundreds of buildings within a facility that could contain prohibited items or activities. Are inspectors permitted to wander around each and every one, and all parts of all buildings? Not likely. When the United States and the International Atomic Energy Agency were conducting verification activities at North Korea’s Yongbyon site, they were only permitted to visit three of the many buildings at the site. Under the Comprehensive Nuclear Test Ban Treaty, a country re- questing an on-site inspection must produce evidence (some of which might be quite sensitive) and convince 30 of 51 member states of the Executive Council that an in- spection is warranted. Politics often plays a role in this. If the Executive Council does not approve the request, the party that sought the inspectionmay have to pay for the cost of any preparations made for it, may have its right to re- quest inspections suspended, and may even be prohibited from serving on the Executive Council. Given the control the inspected party will have over ac- cess to facilities in its sovereign territory, the inspection team can only be expected to find what the inspected party is willing to have them see (or unable to prevent them from F O C U S