I salute the Arms Control Association … for its keen vision of the goals ahead and for its many efforts to identify and to promote practical measures that are so vitally needed to achieve them. -

– Amb. Nobuyasu Abe
Former UN Undersecretary General for Disarmament Affairs
January 28, 2004
Iraq and the Value of On-Site Inspections

Edward Ifft

In the dramatic speech Secretary of State Colin Powell made to the UN Security Council on Iraq and its alleged weapons of mass destruction on Feburary 5, 2003, a key piece of the evidence presented was images of decontamination trucks at a presumed chemical weapons facility. This was offered as proof that Iraq both possessed chemical weapons and was attempting to deceive inspectors from the UN Monitoring, Verification and Inspection Commission (UNMOVIC), the body charged by the Security Council with carrying out inspections in Iraq. It turned out that the vehicles were likely water trucks—not that unusual in a desert land—that inspectors had earlier seen up close at that facility.

Other pieces of evidence—a “truck caravan” near a facility thought to be related to biological weapons, drone aircraft, suspicious activity at a ballistic missile factory, and so on—were items about which inspectors already had, or could probably have relatively easily obtained convincing explanations.

Indeed, although Powell’s presentation was front-page news around the country and the world, less attention was paid to information that the UN’s chief weapons inspectors, Hans Blix, head of UNMOVIC, and International Atomic Energy Agency (IAEA) Director-General Mohamed ElBaradei provided a week later to the Security Council.

As the country strives to digest the recent Iraq Survey Group (ISG) report by Charles Duelfer and implement the recommendations of the 9/11 Commission with respect to the gathering and use of information related to weapons of mass destruction, part of this effort could usefully be given to gaining a better understanding of the possibilities, limitations, and proper use of on-site inspections (OSI). The public, the media, and government officials must avoid the twin perils of becoming disillusioned by expecting too much from OSI, on the one hand, and gaining false confidence from it, on the other.

Inspections are but one imperfect tool in the arms control and nonproliferation toolbox, but as the Iraq example shows, they can play a singularly useful role in assessing the credibility of intelligence gathered by other means, such as satellites. Policymakers need to pay greater heed to the potential benefits of OSI in promoting confidence that legal controls are working and in promoting global security. As Jacques Baute, the head of the IAEA effort in Iraq wrote recently, such inspection “benefits the international community, which receives the level of assurance it seeks, and also the inspected party, which is given the opportunity to demonstrate the reality of its compliance.”[1]

Over the years, these tools have been applied to a growing number of situations, from stabilizing the arms control competition between the United States and the Soviet Union (and later Russia) to multilateral organizations such as the Organization for the Prohibition of Chemical Weapons (OPCW) to specialized regimes such as the Container Security Initiative. Inspections are now so widespread and routine that teams of Russians and international inspectors visiting U.S. facilities do not merit even cursory notice in local media.

For most of the past half-century, at a time when it was being rejected by the Soviet Union, the United States argued that OSI was the sine qua non for arms control. Now, however, even as OSI is more widely accepted and its tools sharper, U.S. and some foreign officials appear conflicted about the value of such visits. Just before UNMOVIC began its work in Iraq, Vice President Dick Cheney called such inspections “at best useless.” One has the impression that in some countries little weight is given to the efforts and conclusions of inspection regimes if they do not support other political objectives. At the same time, any disruption of an inspection regime, as we have seen when the IAEA was forced out of North Korea, can quickly lead to an international crisis.

The Iraq Experience

The greatest attention has of course been focused on inspections in Iraq. This is a unique case in which inspectors had almost unlimited powers, including the right to immediately destroy illegal objects and facilities. Although not a model likely to be repeated often, it is nevertheless instructive.[2] The post-war ISG arrived at essentially the same result as the IAEA Iraq Action Team and UNMOVIC, even though the ISG had greater resources, greater access, and more time. Thus, it seems clear that the criticism directed at UNMOVIC and the IAEA for their work in Iraq was not justified.[3] Nevertheless, the credibility of the inspection regime appears to have taken a blow.

Even after public discussion gradually moved from statements about the failures of the inspection efforts to failures in intelligence, the media have consistently spoken of the failure of inspectors to find weapons of mass destruction. This unfortunate characterization seems to have created in the public mind the impression that the inspectors or the inspection mechanisms themselves were somehow incompetent or corrupt. To be accurate, however, if the inspectors did not find weapons of mass destruction because they were not there or were present only in very small quantities, there was no failure. Indeed, if an inspection can provide reasonable assurance that some suspected objects or activities are absent, it is a success and not a failure—this has long been seen as a major purpose of inspections.

Of course, it would also be incorrect to claim that the inspections proved the absence of these weapons. Here we encounter one of the dilemmas in evaluating OSI. History has shown that inspections can deliver complete confidence that some object has been eliminated or converted according to agreed procedures. Nevertheless, even though inspections can certainly prove that something existed or was disarmed, it can never prove the absence of something. Demanding that it do so will inevitably result in disappointment and confusion. Those who negotiated our numerous OSI regimes understood this quite well. A useful corollary, which Blix noted the United States failed to observe in Iraq, is that “you cannot say that, simply because something is unaccounted for, it exists.”[4]

Another obvious limitation of OSI is that an individual inspection only provides information about a specific location at a particular time. One must be very careful about extrapolations to an entire country. Regardless, an effective inspection can allow for some useful generalizations. For example, it allows one to say with confidence that an identifiable activity did not occur at a particular facility during a certain time period and could not occur for a specified time in the future. Confidence also can be built up using satellites and other national technical means (NTM), interviewing scientists, defectors, and current and former civilian and military officials of the inspected country, as well as conducting a statistically significant number of inspections.

A common problem illustrated by the Iraq situation is the occasional confusion between the effectiveness of inspections themselves and the effectiveness of the response to evidence of noncompliance uncovered by inspections. Undoubtedly, a major problem in arms control today is the reluctance of the UN Security Council, the implementing bodies of various agreements, and individual states-parties themselves, to take effective action when serious violations of legal commitments are discovered.

Of course, the challenge is choosing the action that will effectively solve the problem rather than worsen it. A vivid illustration of this dilemma is found in the differing approaches to dealing with Iran’s nuclear activities. It is certainly possible to point to shortcomings in specific verification systems, although these are generally systemic and not due to careless work by inspectors, but these should not be confused with shortcomings in enforcement mechanisms.

A re-evaluation is needed in cases in which OSI is considered by some to have failed, for example, in Iraq and North Korea. In retrospect, the decade-long work of the United Nations in Iraq (UNMOVIC and its predecessor, the UN Special Commission, or UNSCOM) and the IAEA has been successful. In the cases where inspectors were unable to determine exactly what was going on or there were attempts to deceive them, the inspectors correctly understood that their knowledge was incomplete and deception was being used and reported this to the authorities. If the authorities fail to take appropriate actions in response, it is not a failure of the inspectors or the regimes they are implementing, but of political will.

Apparently, Iraq did not proceed with certain illegal weapons programs because of the fear that they would be caught by inspectors. This is a dramatic illustration of the deterrent effect of inspections. It is widely believed that OSI has a deterrent effect on would-be cheaters, but how strong this is is impossible to quantify. Although the history of OSI is not replete with spectacular discoveries of violations, one reason for this may be that effective regimes can have a powerful deterrent effect.

Lessons Learned, Future Tasks

On-site inspection has become a key component of most arms control agreements and is likely here to stay. Many more examples of the use of inspections could be cited, but those discussed illustrate the variety and effectiveness of the tools available. Although certainly easier to implement in a bilateral context, verification is increasingly becoming a multilateral affair. These mechanisms are put under great stress when states-parties engage in illegal or ambiguous behavior and then allow only grudging cooperation with inspectors (Iraq, Iran, North Korea). Short of war, however, there is little alternative to continuing to use these mechanisms and make them more effective. To accomplish this, individual states need to do a better job of providing timely intelligence information to the inspection systems. It should be possible to assist inspectors in knowing where to look and for what to look while protecting sensitive sources and methods. We will need all of our experience, patience, and creativity to deal successfully with the problem of keeping nuclear, chemical, and biological weapons out of the hands of uncooperative states, as well as stateless terrorist groups.

One important step that will strengthen the role of inspections in the fight against proliferation will be widespread implementation of the 1997 Model Additional Protocol intended to improve the design of safeguards against the use of civilian nuclear technology for nuclear weapons programs.

This protocol, variants of which have been signed by 86 countries and ratified by 61 of them,[5] will give inspectors access to all parts of a state’s nuclear fuel cycle as well as any other location nuclear material may be present. It also allows access to all buildings on a nuclear site, as well as the collection of environmental samples at locations beyond declared locations.

One idea that has recently surfaced and is apparently favored by the New Agenda Coalition[6] and some other states is to establish and train an international team of versatile inspectors under UN auspices. Such a team, which could evolve from the UNMOVIC apparatus, which still exists, could be sent out on short notice to carry out urgent ad hoc inspections anywhere in the world. Under such a scheme, nuclear issues would presumably generally be left to the IAEA and chemical issues left to the OPCW. The United States would likely wish to retain the option of using its own personnel for certain special inspections when appropriate, but such a system might appeal to many smaller countries

A mixed system might be effective in dealing with Iran and North Korea. The IAEA has considerable experience and expertise in both and should certainly play a key role in any future verification regime. At the same time, some special or ad hoc inspections could be useful. In the case of North Korea, it is even possible that countries could draw upon the work that was done during the 1990s in preparation for implementing the unfulfilled 1992 Joint Declaration of the Denuclearization of the Korean Peninsula.

It is imperative that inspection organizations, especially multilateral ones, have the highest standards of integrity. Thus, these organizations must be kept free of any improper intelligence activities or any failure to properly protect proprietary information unrelated to compliance issues that they may acquire in the course of their duties. A major reason for the reluctance of biotechnology firms to accept an intrusive verification regime for the Biological Weapons Convention is concern that valuable proprietary information will be leaked to competitors. Likewise, some countries, especially those with little experience with inspections and major threats to their security, have been concerned that, as a result of inspections, critical national security or commercial information unrelated to their treaty compliance may find its way to other states. Blix showed that he understood this issue when he declared that, if he found that any members of his staff were working for “other agencies,” he would “throw them out, and I think they will understand why.”[7]

A final, interesting lesson from the Iraq experience concerns the search for records. Some of the information that inspectors received in Iraq came from examining Iraqi records. In some cases, these were actually seized and physically removed. Although such activities have long been an important part of law enforcement investigations, they had not previously played a role in arms control inspections.

Gaining access to paper files and computer records could be a powerful tool in verification, but it also places inspectors and inspected parties in a very difficult position. A widespread search and seizure of records creates a situation in which there may be a significant possibility of gaining access to sensitive information completely unrelated to the compliance issue under investigation. The situation is made even more awkward if the language of the material is unfamiliar to most of the inspectors. Attempts by the inspected party to deny access to records said to be unrelated to the question at hand would clearly raise suspicions to a high level.

On the other hand, if inspectors stumble upon sensitive but irrelevant information, for example, defense plans, financial information, personal leadership matters, etc., how should they deal with the situation, and how can the inspected party be assured that such information will be protected? Some of the more outrageous behavior of the Iraqis during the UNSCOM era might have resulted from just such situations. Although such situations may be rare in the future, giving some thought to the legal and practical issues involved would be useful.

The body of experience we now have shows the negative consequences of overreliance upon any single source of intelligence, be it OSI, NTM, defectors, or declarations by states or institutions. These are all potentially useful sources of information but are most likely to lead to reliable and accurate conclusions when used together in a symbiotic fashion

Still, effective inspection regimes, of which we now have many, can make a major contribution to international peace and security. Coupled with greater transparency and more effective responses to violations, they can help the world find its way to a proper balance between “trust” and “verify.”


1. Jacques Baute, “Timeline Iraq: Challenges and Lessons Learned From Nuclear Inspections,” IAEA Bulletin, June 2004.

2. For an account of inspections in Iraq from 1991 to 1998, see International Institute for Strategic Studies, “Iraq’s Weapons of Mass Destruction: A Net Assessment,” 2002. See also “Iraq: A Chronology of UN Inspections and an Assessment of Their Accomplishments,” Arms Control Today, October 2002, pp. 13-23. For an account of inspections in Iraq in 2002-2003, see Hans Blix, Disarming Iraq (New York: Pantheon Books, 2004).

3. A summary of criticisms is given in Joseph Cirincione, Jessica T. Mathews, and George Perkovich, “WMD in Iraq: Evidence and Implications,” 2004, pp. 45-46. See Charles Duelfer, “The Inevitable Failure of Inspections in Iraq,” Arms Control Today, September 2002.

4. “Getting It Right the Next Time: An Interview With Hans Blix,” Arms Control Today, July/August 2004, pp. 14-17.

5. International Atomic Energy Agency, “Strengthened Safeguards System: Status of Additional Protocols,” October 13, 2004.

6. The New Agenda Coalition was launched in 1998 by the foreign ministers of eight non-nuclear nations—Ireland, Brazil, Egypt, Mexico, New Zealand, South Africa, Slovenia, and Sweden—with the purpose of pressuring the nuclear weapons states to fulfill the obligation they undertook in Article VI of the nuclear Nonproliferation Treaty to eliminate nuclear arsenals. The coalition officially consists of seven non-nuclear nations: Ireland, Brazil, Egypt, Mexico, New Zealand, South Africa, and Sweden.

7. “Anticipating Inspections: UNMOVIC Readies Itself for Iraq,” Arms Control Today, July/August 2000, p. 5.



The Universe of On-Site Inspections

As on-site inspection (OSI) has gradually gained acceptance around the world, thousands of inspections have been successfully implemented in a wide variety of circumstances and under many different arms control regimes. Nonetheless, certain practices have emerged.

Formal OSI regimes generally contain important provisions on the status of inspectors, which generally draw upon the Vienna Convention on Diplomatic Relations. They also specify a host of details, such as certification of inspectors, entry and exit points, equipment to be allowed and procedures for its use, timelines, procedures for escort by nationals of the host country, communications, provisions for lodging and meals, financial matters, and resolution of disputes.

Formal regimes also generally establish bodies to deal with implementation issues related to OSI or other matters. For example, the Strategic Arms Reduction Treaty (START I) includes the Joint Compliance and Inspection Commission, which normally meets in Geneva twice a year. Other agreements led to the formation of similar bodies, which help remove pressure from inspectors in the field to settle all controversial issues on the spot.

National governments also generally institute or designate an agency responsible for conducting inspections and providing escorts for inspections on its own territory. In the United States, the On-Site Inspection Agency was created in 1988 to implement the OSI portions of the Intermediate-Range Nuclear Forces (INF) Treaty and gradually was assigned similar responsibilities for later agreements. It was incorporated into the Defense Threat Reduction Agency (DTRA) in 1998, and today, the On-Site Inspection Directorate of DTRA has more than 600 employees. Other countries have similar organizations, but none approach the size of the DTRA effort.

Inspections can either be routine (conducted according to an agreed schedule or quota) or challenge (conducted to clarify some suspicious or ambiguous situation). Some inspections such as those that monitor conversion or elimination activities, are scheduled well in advance. Others are conducted on short notice, to minimize the possibility of hiding illegal activities.

If an inspection regime is to preserve and enhance confidence, it must be designed in such a way that it protects the legitimate rights of the inspecting and inspected parties. Specifically, the inspecting party needs to have sufficient access to provide confidence that the agreed obligations are being fulfilled. At the same time, the inspected party should be protected from unwarranted intelligence activities or unreasonable interference with its normal activities. Arms control regimes are generally successful when they find this balance.

Although inspections are now commonplace in the countries of NATO and the former Warsaw Pact, other parts of the world remain skeptical of the benefits. For example, some countries argue that they have “few secrets” and therefore must guard them vigorously and cannot become more transparent or allow intrusive inspections. Treaty compliance, however, can be demonstrated without leaving oneself “as naked as Adam,” to use Nikita Khrushchev’s colorful phrase.

OSI Regimes

Much of our current expertise about inspections derives from the arms control treaties painstakingly crafted by the United States and the Soviet Union during the Cold War.

The 1987 INF Treaty, with its meticulous set of rules and procedures for OSI, became the model for later agreements, both bilateral and multilateral.[1] That pact required the United States and the Soviet Union to eliminate and permanently forswear all of their nuclear and conventional ground-launched ballistic and cruise missiles with ranges of 500-5,500 kilometers. It called for five types of intrusive inspections and permitted perimeter and portal continuous monitoring (PPCM) at one site in each country. At Russia’s Votkinsk site, U.S. inspectors routinely took X-ray images of missile canisters containing Russian ICBMs. The images allowed U.S. officials to determine if permitted missiles, rather than banned SS-20s, were leaving the factory. Russian officials were able to monitor a similar U.S. plant near Salt Lake City that had been involved in producing the banned Pershing II missile.[2]

Although all the INF systems were dismantled or converted by 1991 in accordance with the terms of the treaty, inspections continued until 2001 in what was probably a case of verification overkill. During the 13 years of INF inspections, a total of 851 inspections were carried out. Roughly 60 percent of these were carried out by U.S. inspectors at 130 sites in Russia, Ukraine, Belarus, and Kazakhstan. Roughly 40 percent were carried out by the other parties at 31 sites in the United States and the five western European basing countries (the United Kingdom, Germany, Belgium, the Netherlands, and Italy).

The 1994 START I, in addition to reducing deployed strategic nuclear delivery vehicles (ICBM and submarine-launched ballistic missile launchers and heavy bombers), also succeeded in reducing the warheads on these delivery vehicles. Thus “accountable warheads” (multiple independently targetable re-entry vehicles, single re-entry vehicles, heavy bomber armament) for the United States and four former Soviet republics (Russia, Ukraine, Belarus, and Kazakhstan) were reduced by more than 40 percent, to 6,000 for each side. Ballistic missile throw-weight (lifting power), in which the Soviet Union had a 3-to-1 advantage, also was reduced to 3,600 metric tons for each side.

START provides for 12 types of OSI, and its inspection protocol runs well over 100 pages. There is also a massive data exchange to aid both inspectors and national technical means. This Memorandum of Understanding on data is updated every six months. It is supplemented by a significant number of notifications regarding numbers, locations, movements, and other aspects of the sides’ strategic forces, which are exchanged on a day-to-day basis among the Nuclear Risk Reduction Centers of the parties. Inspections are continuing successfully under START at the rate of about two per month. These consist primarily of data update inspections at declared facilities to monitor the status of treaty-limited items and re-entry vehicle inspections to verify that the number of re-entry vehicles on deployed ballistic missiles does not exceed the number allowed for that type of missile, plus occasional inspections of other types.[3]

Inspections under the 1990 Conventional Armed Forces in Europe (CFE) Treaty have verified the successful elimination of more than 70,000 pieces of military equipment. An interesting feature of its inspection process has been the gradual evolution of procedures that allow an inspector from one country to head a team including inspectors from other countries. For example, all inspections led by the United States have the participation of inspectors from other NATO allies and even non-NATO countries. Under the CFE Treaty, DTRA had conducted 264 inspections and 168 reduction inspections and escorted 130 inspections of U.S. facilities as of January 2004.[4]

Multilateral Regimes

The pioneering multilateral arms control inspection regime was established by the 1968 nuclear Nonproliferation Treaty (NPT), under which the five nuclear-weapon states commit to pursue general and complete disarmament, while the non-nuclear-weapon states agree to forgo developing or acquiring nuclear weapons. The NPT incorporated, under Article III, a system of safeguards applied by the International Atomic Energy Agency (IAEA) to prevent diversion of nuclear energy from peaceful uses to nuclear weapons. The IAEA continues to play a crucial role in this field, although until recently it has been handicapped by the fact that its monitoring activities were only applied to declared facilities.

The 1997 Model Additional Protocol, once it is widely accepted, should help to close this unfortunate loophole, which has been exploited by certain countries and thus weakened the nuclear nonproliferation regime in dangerous ways. This addition to a country’s safeguard agreement gives inspectors access to all parts of a state’s nuclear fuel cycle, as well as any other location where nuclear material may be present. It also allows access to all buildings on a nuclear site, as well as collection of environmental samples at locations beyond declared locations.

The 1993 Chemical Weapons Convention established the Organization for the Prohibition of Chemical Weapons (OPCW), which plays a key role in assuring that existing chemical weapons are being destroyed according to agreed procedures and that new chemical weapons are not being created. The OPCW has conducted more than 1,300 inspections in more than 50 countries. Like the IAEA, the OPCW has a professional staff of inspectors who are responsible to the organization and not their own governments. In both organizations, raw data obtained on inspections is carefully guarded to give confidence to the parties that sensitive national security and proprietary information will be protected. Still, the OPCW has undergone some controversy and turmoil in the past, including a budget crisis in 2001, and has not yet been tested by any requests for challenge inspections to clarify ambiguous situations.

The Comprehensive Test Ban Treaty (CTBT), while not yet in force, establishes a unique worldwide verification system of 337 facilities, many of which are already operating. (See ACT, October 2004.) Less well known is the fact that the CTBT also provides for an ambitious challenge inspection system. This is being worked out by the Comprehensive Test Ban Treaty Organization (CTBTO) in Vienna. The work is proceeding slowly in some cases because of the uncertainty regarding the CTBT’s entry into force, the caution of countries with little experience with OSI, and the fact that the United States no longer provides support for the organization’s OSI work.

Still, if the CTBT comes into force, inspectors would employ a variety of sophisticated types of equipment to detect the seismic, radiological, and other indications of a nuclear explosion. The CTBTO is procuring equipment, preparing operational manuals, and training prospective inspectors from many countries, although the treaty’s entry into force is uncertain. One complication is that, because all inspections will essentially be of the challenge variety, it is difficult to predict how frequent they would be and thus how many trained inspectors and how much equipment will be needed.

Other Regimes

As the acceptance of OSI has increased, there has been a proliferation of relatively short-notice, ad hoc inspections to deal with specific situations. These generally do not have the benefit of detailed legal arrangements, but many have been quite useful in providing information to clarify ambiguous events or to diffuse tense situations. The United Nations has organized a number of these, for example, in Asia to investigate “yellow rain,” in Iran to investigate the use of chemical weapons in the Iran-Iraq War, and in the Caucasus to investigate charges of the use of chemical weapons in Nagorno-Karabakh. In the late 1980s, the United States and the Soviet Union successfully used ad hoc OSI visits to clarify ambiguous situations that had arisen in connection with arms control agreements in force or under negotiation. Each side visited certain radar installations of the other side, and a U.S. team visited suspected “submarine tunnels” in the USSR. In 1992 the United States, the United Kingdom, and Russia agreed to use OSI to resolve some uncertainties regarding compliance by Russia with the Biological Weapons Convention. U.S. teams have also gone to North Korea to see underground facilities that were suspected of being associated with illegal nuclear activities.[5]

The war on terrorism has recently led the United States to create what might be termed a specialized inspection regime: the Container Security Initiative, a series of measures designed to move the screening of shipping containers further back in the supply chain than traditional customs inspections. With more than 90 percent of the world’s international cargo transported in these ubiquitous containers, the danger that terrorists might use them to infiltrate nuclear, chemical, or biological weapons into the United States have become obvious. The new initiative has rapidly resulted in agreements with some 20 of the world’s largest ports and is leading to a layered defense of better documentation; passive detectors; active detectors; and, when necessary, the opening of the containers themselves.[6]

The line between OSI and peacekeeping can be rather blurred. When peacekeepers get involved in activities such as inspecting military activities, searching for and destroying weapons, and so on, what they are doing could fairly be described as OSI. One long-running operation that makes an interesting case study is the Multinational Force and Observers (MFO), which was created to help implement the 1979 peace treaty between Egypt and Israel. The MFO is a unique model for peacekeeping completely outside the UN system, reporting directly to the treaty parties and including the participation of 11 countries as varied as the United States, Colombia, and Fiji. Its tasks are to operate checkpoints, observation posts, and reconnaissance patrols in the Sinai peninsula; carry out periodic verification of the agreement, with additional verification within 48 hours after a request from either party; and to assure freedom of navigation through the Strait of Tiran. Although much of the Middle East has been in turmoil, the Sinai has remained peaceful.[7]


1. Edward Ifft, “Verifying Nuclear Arms Control and Disarmament,” in Verification Yearbook 2001, eds. Trevor Findlay and Oliver Meier (London: Corporate & Commercial Printing, Ltd., 2001), pp. 26-27.

2. Joseph P. Harahan, “On-Site Inspections Under the INF Treaty,” OSIA Treaty History Series, 1993.

3. David B. Thomson, A Guide to the Nuclear Arms Control Treaties (Los Alamos, NM: Los Alamos Historical Society, 2001), chap. 7.

4. Joseph P. Harahan and John C. Kuhn III, “On-Site Inspections Under the CFE Treaty,” OSIA Treaty History Series, 1996.

5. Edward Ifft, “The Use of On-Site Inspection in the Avoidance and Settlement of Arms Control Disputes,” in Avoidance and Settlement of Arms Control Disputes: Vol. II, Arms Control and Disarmament Law, ed. Julie Dahlitz (New York: United Nations, 1994), pp. 16-19. See George L. Rueckert, On-Site Inspection in Theory and Practice (Westport, CN: Praeger, 1998).

6. For a discussion of the Container Security Initiative and the concerns which led to it, see Stephen Flynn, America the Vulnerable (New York: HarperCollins, 2004), chap. 5.

7. The Multinational Force and Observers, Rome, 1999; “Annual Report of the Director General, Multinational Force and Observers,” January 2002, found at http://www.mfo.org/files/Annual%20Report%202002%20-%20web.pdf.




The Long Road to On-Site Inspections

The effort to achieve broad acceptance of on-site inspection (OSI) is a story of high diplomacy, creativity, and determination through several U.S. administrations. The ambitious but somewhat naïve Baruch Plan, put forward by the United States in 1946 as a result of the Acheson-Lilienthal Report, was the first great effort to control nuclear weapons.[1]

The Baruch Plan called for unlimited inspections of all atomic energy-related facilities by personnel from an international inspection organization, as well as complete international ownership and control over fissionable materials. It was rejected by Soviet leaders, whose progress in developing nuclear weapons was not fully understood at the time. Soviet leaders soon developed the rationale that OSI could be used to monitor disarmament, but not armament. In 1948, Jacob Malik, the Soviet Union’s permanent representative to the United Nations, stated a view that would guide Soviet policy for many years: “It would be folly for the USSR to disclose everything and then have others invent conditions as a pretext for dropping the whole question of armaments after they had found out everything they wanted to know.”[2]

The Soviet obsession with secrecy, combined with the fact that the United States was sometimes not above using proposals for OSI to embarrass the Soviets, made progress almost impossible for years.[3] W. Averell Harriman, a long-time U.S. interlocutor with Moscow, recalled that Soviet Premier Nikita Khrushchev met his arguments for OSI to verify a comprehensive test ban with one of his typical earthy analogies, saying, “You remind me of a cat saying that he would only eat mice and not the bacon lying in the room.” He added that he would not trust such a cat, “as it would undoubtedly snatch the bacon when no one was in the room...he knew what cats were like.”[4] It is interesting that Khrushchev himself later regretted Soviet intransigence on OSI during this period.[5]

During the 1960s, the development of satellites and other forms of national technical means (NTM) of verification gradually made OSI less essential for monitoring certain kinds of constraints. President Lyndon B. Johnson, trying to convince the Soviets that negotiating an end to the most dangerous forms of military competition would be in the interest of both sides, informed General Secretary Leonid Brezhnev that the United States would be prepared to place “maximum reliance” upon NTM, that is, the decades-long insistence upon OSI could be set aside, at least for certain kinds of constraints. This was probably a key factor in the Soviet agreement to begin the Strategic Arms Limitations Talks (SALT) negotiations in 1969, during the Nixon administration. For the next 10 years, difficult but productive SALT I and II negotiations proceeded without recourse to OSI.

In the area of multilateral negotiations, however, OSI gained fairly rapid acceptance. The Antarctic Treaty of 1959, the Outer Space Treaty of 1967, the nuclear Nonproliferation Treaty of 1968, the 1975 Helsinki Final Act, and the more ambitious Stockholm Accord of 1986 all permitted some form of inspections. In addition, the U.S.-USSR Peaceful Nuclear Explosions Treaty of 1976 and its subsequent protocols provided for quite intrusive inspections under certain circumstances. Nevertheless, none of these agreements could be said to impinge upon vital interests or allow inspections of the most important weapons systems.

The Reagan administration from its early days placed much greater emphasis on verification, changing the stated U.S. goal from “adequate” to “effective” verification, although the distinction was never made clear.

Mikhail Gorbachev came to power in Moscow in the mid 1980s and he soon began to match Reagan’s rhetoric about effective verification. In January 1986, Gorbachev made his proposal for the complete elimination of nuclear weapons by 2000, accompanied by optimistic words about verification. He also addressed the old problem of verifying a ban on nuclear testing by declaring, “Any reference to verification as an obstacle to the establishment of a moratorium on nuclear explosions is totally groundless. We declare unequivocally that verification is no problem as far as we are concerned.”[6]

After showing some flexibility on OSI at the dramatic and contentious Reykjavik summit in October 1986,[7] Gorbachev moved the ball forward significantly again in his famous “European Common Home” speech. Speaking in Prague in April 1987, the Soviet leader said that “appropriate verification measures, including on-site inspections, must encompass the missiles and launchers remaining after the reductions, including those of combat duty and at other facilities—i.e., testing grounds, manufacturing plants, training centers, etc. Inspectors must be guaranteed access to the other side’s military bases on the territory of third countries. This is necessary to ensure absolute certainty that the agreement is really being fully observed.”[8] It was at last clear that OSI could be applied to operational systems and not used simply to verify the process of destruction. It was not long until the Soviet negotiators in Geneva were actually advocating forms of OSI the United States could not accept.

Events moved rapidly following this breakthrough. The Intermediate-Range Nuclear Forces (INF) Treaty was signed in December 1987. It incorporated much of the verification regime proposed by the United States and became the model for later agreements. The Strategic Arms Reduction Treaty (START), signed in 1991, is even more intrusive and contains the most extensive OSI regime ever negotiated.

By 2000 it appeared inevitable that virtually all future arms control agreements would incorporate some form of OSI, until the United States pulled back in 2001 from efforts to incorporate a verification regime into the Biological Weapons Convention. The events of the last three years indicate that future efforts to use OSI will be examined and debated carefully on a case-by-case basis.


1. This discussion is adapted from a presentation the author prepared for the American Association for the Advancement of Slavic Studies.

2. Timothy Pounds, “Proposals for On-Site Inspection Over the Years,” in Arms Control Verification and the New Role of On-Site Inspection, eds. Lewis A. Dunn and Amy E. Gordon (Lexington: Lexington Books, 1990), p.73.

3. For an account of this period, see Thomas Schmalberger, In Pursuit of a Nuclear Test Ban Treaty: A Guide to the Debate in the Conference on Disarmament (New York: United Nations, 1991), pp. 18-24. See also Anatoly Dobrynin, In Confidence: Moscow’s Ambassador to America’s Six Cold War Presidents (New York: Crown, 1995), p. 46.

4. W. Averell Harriman, America and Russia in a Changng World (New York: Doubleday, 1971), p. 94.

5. Strobe Talbott, Khrushchev Remembers (New York: Little, Brown & Co., 1974), pp. 537-538.

6. U.S. Arms Control and Disarmament Agency, “Documents on Disarmament,” 1986, p. 12.

7. Mikhail Gorbachev, “Statement at the Press Conference in Reykjavik,” in “Reykjavik Documents and Materials,” Novosti, 1987, pp. 6-9.

8. Mikhail Gorbachev, “For a Common European Home, for a New Way of Thinking” (speech, Prague, April 1987).


Edward Ifft is a retired State Department official who served as Deputy Director of the On-Site Inspection Agency and Senior Advisor to the Defense Threat Reduction Agency. The views expressed in this paper are those of the author and do not necessarily reflect the policy of the U.S. Government. For more information on on-site inspections see the author's research paper on the subject.