Foreign Affairs and Trade, Australia


  Annual Report 1998-99



The strengthened safeguards system being developed today is very different from the ‘classical’ safeguards predominant a decade ago. Classical safeguards were directed at the timely detection of the diversion of declared and safeguarded nuclear material, and relied heavily on nuclear materials accountancy to verify that declared material remained in peaceful use. Essentially they verified the correctness of States’ declarations.

It was recognised that a State might have undeclared facilities separate from its declared and safeguarded activities. But classical safeguards did not require the IAEA to seek out undeclared facilities directly. That was considered more properly the function of national intelligence agencies. Safeguards agreements do provide for special inspections, which allow inspectors to go anywhere, and not just to declared facilities. But strong and location-specific evidence of a safeguards breach is needed for a special inspection and the IAEA never sought one until 1992 in North Korea.

The strengthening of the safeguards system

After the Gulf War it was revealed that Iraq had extensive weapons of mass destruction programs. That highlighted certain shortcomings of classical safeguards:

  • Iraq’s nuclear-weapons program was not based on diversion. It relied upon clandestine facilities and undeclared nuclear material, and so avoided the risk of detection by classical safeguards.
  • Some clandestine activities were on the site of safeguarded facilities. There, Iraq took advantage of restrictions on inspector access under classical safeguards.
  • Information available to governments on Iraq’s interest in nuclear weapons had not been able to be used for safeguards purposes.

So reliance solely on a capability to detect diversion and (indirectly) on national intelligence had proved inadequate. For safeguards to be effective in confidence building, the IAEA needed to be able to detect undeclared nuclear activities. The focus on correctness of declarations had to be broadened to encompass their completeness. It was recognised that the system needed greater transparency of nuclear programs, wider inspection rights, more broadly-based information collection and analysis, and new technology.

The problem of building a stronger system was, and is being, addressed at both technical and institutional levels. Much work has been done on the technical front, but more is needed. At the institutional level, the IAEA Board agreed in 1997 to the text of a Model Protocol additional to existing safeguards agreements. See the outline of the Protocol.

Issues of cost-efficiency

The IAEA needs to make its safeguards more cost-efficient as well as more effective. Some progress in this area is already evident. The IAEA has been able to accommodate a growing safeguards task (a 20% increase in the number of safeguarded facilities between 1990 and 1998) while its overall budget has been subject to zero real growth. The efficiency improvements which have been achieved can be attributed to technical and procedural advances, as well as more efficient working arrangements with Euratom.

Many IAEA Member States want to go further, and are resolved that the cost of new safeguards activities be offset by savings in existing activities so that, after an initial period, strengthened safeguards are accommodated within the existing safeguards budget. This is one factor influencing the concept of ‘integrated safeguards’, discussed below.

Challenges Ahead

The critical challenge for the safeguards community is to establish a credible capability to detect undeclared nuclear activities. Success in achieving that is much harder to measure than for verifying declared material, and the assurance derived will be less certain. While it is important to avoid over-expectation in this regard, care must be taken not to under-estimate what can be achieved.

It is vital that the IAEA is able to present authoritative conclusions about the absence of undeclared activities in a State. If States have no clear conclusions from the IAEA they may act on unsupported suspicions about the perceived proliferation activities of others. If that were to cause some to reconsider their commitment to non-proliferation the consequences for the non-proliferation regime would be severe.

The detection of undeclared nuclear activities will not be easy, but the IAEA will be better equipped than individual States to determine whether or not they exist. It will have at its disposal its own expertise together with substantial information resources, extensive inspector access, and increasingly sophisticated technology. It will be important however that States assist the IAEA where they can, e.g. by providing intelligence information.

Some cultural change will be needed in safeguards practice. Classical safeguards have encouraged a rather mechanistic approach to safeguards. Now inspectors need encouragement to be more inquisitive, but still in a rigorous way so that the international community can have confidence in their findings. The transparency and cooperation of States are further essential elements of the strengthened regime. The international community will view any lack of cooperation, particularly where access is obstructed, very seriously. Overall there is room for optimism that, as strengthened safeguards develop and experience grows, they will make a major contribution to international confidence-building.

Integrated Safeguards

Although implementation of strengthened safeguards is only just beginning, attention is already focusing on merging classical safeguards and safeguards strengthening measures to yield an optimally effective and cost-efficient outcome. This is known as safeguards integration, and has a very high priority within the IAEA. The IAEA’s work program, which was largely defined at an experts’ meeting in Vienna in late 1998, requires conceptual and development tasks to be completed by the end of 2000.

Budgetary considerations, and the resolve of many Member States that strengthened safeguards do not increase the cost of safeguards, are important factors driving safeguards integration. It would be incorrect, however, to see integration as purely as response to budget pressures. There is also an important in-principle issue of efficiency. Enhanced efficiency in integrated safeguards derives from redundancy between the old and the new safeguards measures, where proliferation strategies are detectable by both (e.g. diversion of safeguarded spent fuel and reprocessing in a clandestine plant). Some strategies are detectable only by classical safeguards (i.e. diversion of weapons-useable material), and some can be detected only by safeguards-strengthening measures (i.e. totally clandestine fuel-cycles). Where there is redundancy there is room for rationalisation to enhance cost-effectiveness.

There is a link between safeguards designed to detect diversion and those designed to detect undeclared facilities, because most nuclear materials found in safeguarded activities need enrichment or reprocessing to make them weapons-useable. If we have credible assurance that a State has no undeclared enrichment or reprocessing facility, then diversion of those materials becomes a lesser concern.

It will be some time before the technical means and procedures necessary to demonstrate a high degree of assurance of the absence of undeclared enrichment and reprocessing will be established. An appropriate level of assurance of non-diversion will be required from classical safeguards for the foreseeable future. That level will be high for materials that require minimal downstream processing (e.g. highly enriched uranium, separated plutonium), and thus offer little opportunity for detecting downstream facilities. But for less sensitive materials the appropriate level can be determined in the light of progress with strengthened safeguards.

Further improvements in efficiency

The challenge in integrated safeguards is to achieve effectiveness and efficiency without compromising the overall assurance derived. Historically, changes have been made in classical safeguards without reducing effectiveness, and there is no reason to believe further improvements are not possible.

Safeguards have already benefited considerably from the introduction of new technology. The IAEA is now in a position to deploy remote monitoring, which enhances safeguards effectiveness by removing the delay in the acquisition of safeguards information associated with the gap between inspections. If needed, safeguards information can be accessed in real time. That has profound implications for the concept of timeliness in safeguards. Remote monitoring also gives an opportunity to save inspection effort by reducing inspection frequencies. Of course inspectors’ under-the-roof observations are important for safeguards, but remote monitoring gives the IAEA an opportunity to prioritise and send inspectors when and where they are most needed.

Under classical safeguards hitherto the need for non-discrimination has meant that safeguards have been implemented in an identical way in each State. That results in concentration of inspection effort in the States with the biggest civil fuel-cycles. At one point 70% of IAEA inspection effort was devoted to just three States (this is now nearer to 50%, which is still very high). While complex fuel-cycles will plainly attract significant inspection effort, it is not clear that they need such a very high proportion of the total.

In fact, NPT safeguards agreements do not prescribe a uniform approach. Rather they provide for flexibility to meet States’ circumstances and include criteria for determining the intensity of safeguards (e.g. as expressed by inspection frequency). The factors to be taken into account include the characteristics of the State’s nuclear fuel-cycle, its international interdependence, and the effectiveness of the State’s system for accounting for and control of nuclear material (SSAC).

Given that integrated safeguards are to be applied to the State as a whole, and that no two States will have identical circumstances, a basis is needed for determining how safeguards should be applied in each State. For optimal effectiveness and cost-efficiency, differences between States need to be taken into account. Provided this is done in a transparent way using objective criteria, it need not be discriminatory. It is worth repeating this is not a new idea, it is provided for in existing safeguards agreements.

It is also open to the international community to change certain underlying safeguards parameters such as timeliness goals, detection probabilities sought and how nuclear materials are categorised. They might be changed for States where it is found that current verification intensities are no longer needed under strengthened safeguards, or they could be changed for individual States to take account of State-specific circumstances. In addition, concepts drawn from the IAEA’s new safeguards arrangements with Euratom might prove to be applicable elsewhere, allowing further flexibility.


The IAEA safeguards system is undergoing the greatest period of development and change since the introduction of NPT safeguards agreements in the early 1970s. We are moving from a system characterised by:

  • heavy reliance on the verification of nuclear materials accountancy, using containment and surveillance only as complementary measures; and
  • a focus, inherited from pre-NPT safeguards, on the concept of diversion as comprising the removal of nuclear material from declared facilities or locations;

towards a system which:

  • is shifting the focus from declared inventories and flows of nuclear material at individual facilities, towards safeguards approaches based on evaluation of the State as a whole;
  • seeks to provide credible assurance of the absence of undeclared nuclear material and activities in the State; and
  • diversifies the methods of detection, introduction methods based upon quite different principles, resulting in a system which is expected to be more robust.

Return to the ASNO Annual Report Index

This page last modified: Wednesday, 07-Apr-2004 10:54:07 EST

Local Date: Tuesday, 27-Jan-2015 07:33:11 EST