James ActonLessons from the RSA

A reader, Paul Carroll, made an important point in commenting on a previous post of mine about verifying the North Korean declaration:

Obviously, when dealing with North Korea and the risk of diversion or concealment we want the uncertainty to be as low as possible. But I would urge us to caution against demanding levels of certainty that may be unattainable, and that could undermine progress on the other aspects of the deal.

Verification in North Korea is tricky because it involves ‘proving a negative’, a perennial problem that is frequently discussed on these pages. It boils down to the following question: How can North Korea demonstrate that it does not have a clandestine Pu stockpile or a uranium enrichment programme?

The IAEA faces this problem on a routine basis when trying to draw its broader conclusion about the absence of undeclared nuclear activities in a state with an additional protocol in force. John Carlson, former chairman of the Agency’s Standing Advisory Group on Safeguards Implementation, put it thus

The revolutionary aspect of safeguards development is that judgment is coming to the fore in drawing safeguards conclusions. Conclusions about the absence of something undeclared activities—can never be as definitive as conclusions based on quantitative methods applied to a finite problem—the verification of a declared inventory.

This problem was faced in verifying disarmament in South Africa—and the lessons from this case are worth reflecting on in light of the upcoming verification challenges in North Korea.

South Africa is the only state to have developed nuclear weapons indigenously and then abandoned them. When South Africa acceded to the NPT in 1991, the IAEA was tasked with verifying its initial declaration. Significantly, the results of verification were inconclusive.

To cut a long story short, there was a discrepancy between the amount of HEU South Africa had apparently produced and the amount that was in its stockpile. Such a discrepancy could have had two causes: either it was the result of errors in estimating past production (a tricky business, especially for enrichment plants) or it was indicative of the existence of a clandestine stockpile of HEU. It was impossible to tell on purely ‘technical’ grounds which explanation was correct.

In practice, of course, no-one (or at least very few people) seriously believed that South Africa had a clandestine fissile material stockpile and this raises an interesting question: How did South Africa convince the international community of the veracity of its declaration even though the results of IAEA verification activities were inconclusive?

Part of it, of course, was the broader political context. South Africa was in the process of transitioning from the apartheid system to genuine democracy. It had no security concerns that would have driven it to keep nuclear weapons and was making a concerted effort to reintegrate itself into the international community.

However, I believe that part of it was the way that South Africa built the trust of the inspectors on the ground: namely, by being exceptionally open and transparent. South Africa offered inspectors almost carte blanche to go anywhere, have access to any documents and speak to anyone—and it honoured this promise.

This transparency did not produce new information which explained away all discrepancies in the fissile material accountancy. What it did do was build the confidence of inspectors—and through them the international community—that South Africa had nothing to hide. It was a visceral thing—but no less effective for it.

Perhaps understandably, this is an aspect of the verification process in South Africa that the official history merely hints at rather than dwells upon. But, I think it’s how verification often works in practice.

The lessons from South Africa are particularly salient for North Korea. It seems that North Korea is likely to submit its declaration soon prompting the start of the verification process. If North Korea wants to demonstrate that it does not have an undeclared enrichment programme or a clandestine plutonium stockpile, transparency may be key. To convince the inspectors on the ground, active cooperation is needed; the odd concession given grudgingly within the context of an on-going battle for access won’t suffice.

However, there are also lessons here for the rest of the international community, the US in particular: namely that the results of verification can be ambiguous in a state that has nothing to hide.

Based on the objective results of verification, it might be effectively impossible for the US either to prove North Korea’s declaration is correct or to show that it has lied. If the results are indeed ambiguous, it will be very tempting for the Bush Administration to accuse the North Koreans of cheating in order to avoid a fight with Congress, which has already made noises about the Administration being weak in this area.

The only solution is to pre-empt this potential problem. The US needs to discuss these ‘what ifs’ with both North Korea and the US Congress (if it is not already doing so). It is vital that all parties have a realistic idea of what verification can—and can’t achieve—before it gets going. Otherwise the process may be doomed to fail before it’s started.


  1. Gridlock (History)

    Does the close military relationship between Israel and SA not suggest a possible residing place for the anomalous HEU?

  2. Bruce Klingner (History)

    The key is to follow the precedent of previous arms control treaties (INF, START, CFE) and have an agreed-upon verification protocol which allows for, amongst other things, short-notice challenge inspections of non-declared facilities. The quota doesn’t need to be ridiculously large (perhaps 10 the first year of an agreement and fewer in subsequent years). By having the ability to go anywhere in the country, it puts any illicit program at risk of discovery. If North Korea doesn’t allow entry to a suspect site, then there is a problem that needs to be resolved and it would lead to a slowdown or cessation in benefits. If the inspectors successfully gain access to site after site and nothing illicit is discovered, then monitoring confidence increases (even if suspicion remains) which allows for continued progress/negotiations/benefits.

    Initial data exchanges with the Soviet Union and Warsaw Pact showed great discrepancies between US estimates and Soviet declarations. But inspections of declared and non-declared sites allowed for continual revision of US estimates which reduced the discrepancy, e.g. the US might have assessed 31 tanks in a storage shed but an inspection showed fewer tanks since it was not a full TO&E tank battalion or that the unit had departed years earlier.

    But this requires the US demand such provisions and North Korea agree to a sufficiently rigorous inspection regime.

  3. MarkoB

    One problem here might be something akin to the UNSCOM experience. The DPRK might see an instrusive inspection process as a cover for intelligence operations.

    The RSA was easy because the RSA didn’t really face a security threat from the outside but it is the very existience of a perceived outside security threat that lead to a DPRK bomb programme in the first place.

    My understanding of the RSA bomb was that it was a programme more motivated to maximise leverage…although the DPRK bomb has elements of that too.

    I doubt whether the DPRK would grant free and unlimited inspections given the context and the politics of verification can easily be used by hawks to undermine arms control processes.

    We have a pretty good idea about how much Pu the DPRK produced from its known nuclear facilities. We know that the uranium enrichment “plant” story is highy suspect.

    Of course unlimited inspections would be best but as pointed out here even they would not be perfect. Good arms control need not be perfect arms control.

    We also know that the DPRK has not received that much fuel oil either. The DPRK provided a low Pu declaration partly because of this.

    It takes two to tango.

  4. Lao Tao Ren (History)

    This is somewhat tangential to the discussion about whether South Africa really “came clean” about their nuclear program.

    It is well known that at least a few people involved in the bio-warfare programs cached away their custom bugs, etc. and then years later, attempted to sell it to the US, only to be turned away because the stuff is pretty primitive by US standards. These miscreants then “went public” about it….

    That certainly cast some doubts as to how “clean” South Africa came in other areas.

  5. Cameron (History)

    The people involved were holding on to the info on their bio-weapons program on a personal level. If anything I’d suggest it casts doubts on SA’s security measures.