National Security Advisor John Bolton’s recent speech condemning the ICC has generated significant criticism for going beyond his traditional attacks on the court to targeting its judges and prosecutors in a uniquely personal way.  But far from his plan to “let the ICC die on its own,” as he put it, he is actively trying to get other states to undermine it by reviving the Article 98(2) agreements that the administration of George W. Bush gave up on 12 years ago.

Most discussion has focused on Bolton’s threat to target ICC judges and prosecutors, given its dubious legal foundation and the potential to give cover to autocrats who wish to retaliate against or marginalize either the ICC or their own judicial institutions.  His denunciation echoed previous Trump administration statements denigrating “so-called judges” and criticizing attorneys representing immigrants in removal proceedings.

But his threats of individual retribution extended to personnel from “any company or state that assists an ICC investigation of Americans.” Many of these could well be officials from traditional allies, potentially including NATO soldiers serving alongside U.S. troops in Afghanistan. Of the twenty-nine members of NATO, twenty-seven are States Parties to the ICC; only the U.S. and Turkey are not. If the Afghanistan investigation is approved, it will extend to the activities at black sites in Poland, Lithuania, and Romania, all members of NATO.

He also asserted that “[w]e will negotiate even more binding, bilateral agreements to prohibit nations from surrendering U.S. persons to the ICC.”  This threat may sound obscure to those who began following international criminal law only in the last decade, but it refers to the “bilateral non-surrender agreements” or “bilateral immunity agreements” (BIAs) the U.S. developed and promoted in the early 2000s. Given the threat to revive these agreements, officials in foreign governments and civil society interested in preserving the authority and efficacy of the ICC may need to dust off the legal analyses and political playbooks of 15 years ago.

What are the agreements, and where did they come from? 

The bilateral non-surrender agreements that Bolton referred to are largely what they sound like – agreements between the U.S. and another state indicating that the counterpart will not surrender any U.S. national to the ICC without U.S. consent.  While there are slight variations on whether any given agreement protects only U.S. persons or is mutual, and on whether it prevents surrender only to the ICC or to any international tribunal not established by the Security Council, the key operative provision is:

“U.S. persons present in the territory of [the state] shall not, absent the express consent of the Government of the United States of America, (a) be surrendered or transferred by any means to the International Criminal Court for any purpose, …”

“U.S. persons,” in turn, are typically defined very broadly in the agreements:  not just officials and military personnel, but all U.S. nationals, plus all current and former employees and contractors.

The agreements were a key part of the Bush administration’s 2002 strategy to undermine the ICC by denying it legitimacy or the chance to be effective: Bolton’s “unsigning” the ICC statute, the U.S. vetoes of peacekeeping missions that did not exclude ICC jurisdiction, and the passage of the American Servicemembers Protection Act (ASPA).

And the U.S. was not simply asking nicely if countries would like to sign a BIA – it carried a big stick when it did so, as ASPA contained a presumption that any state refusing to enter into a BIA would be denied U.S. military aid, though this could be waived.  The U.S. was also demanding these BIAs from a position of strength – at a time that it still benefited from substantial international sympathy and goodwill following the attacks of Sept. 11, 2001. At the same time, the U.S. was also treating countries according to Bush’s view that they were either “with us or you are with the terrorists.”

However, even from this position of relative strength, attacking an independent judicial institution designed to prosecute the most serious international crimes where domestic systems had failed to do so was a bridge too far for many traditional U.S. allies. None of the “five eyes” allies (the intelligence partners of the U.K., Canada, Australia and New Zealand) entered into a BIA.  Nor did most European states:  the European Union took a position against such agreements, though not before Romania became the first state to sign an agreement (this seemed less surprising when it came to light that Romania subsequently hosted a CIA rendition black site).  And while a number of African and Middle Eastern countries signed BIAs, support in the Americas was largely limited to the smaller Caribbean and Central American nations.

Ultimately, the U.S. signed about 100 agreements, almost all between July 2002 and January 2005.  The last agreement was signed in June 2006. Later that year, however, the Bush administration changed course and attempted to undo the damage caused by its attacks on the ICC, by ceasing to pursue BIAs and entering a period of constructive (if cautious) engagement with the court.

How do the agreements relate to the ICC statute?

That all depends on the other state.  If the other state is not a party to the ICC Statute, it has no obligation to surrender anyone to the ICC. Under the circumstances, the BIA is likely to be effective (unless cooperation obligations are imposed as part of a referral from the United Nations Security Council to the court, which would likely override the BIA under Chapter VII of the U.N. Charter).

But about half of the states that entered BIAs with the U.S. are parties to the ICC statute, and those cases are more complicated.  Ordinarily, those states are obliged to surrender a person when the court requests (Article 89).  However, the U.S. claimed that these agreements were consistent with the ICC statute — in particular, that they were allowed under Article 98(2) (hence their colloquial name of “Article 98(2) agreements”), and therefore superseded any more general cooperation obligations under the Statute.

How would this work?  Article 98(2) provides:

“The Court may not proceed with a request for surrender which would require the requested State to act inconsistently with its obligations under international agreements pursuant to which the consent of a sending State is required to surrender a person of that State to the Court, unless the Court can first obtain the cooperation of the sending State for the giving of consent for the surrender.”

Note that if there is a conflicting obligation within the scope of Article 98(2), it acts as a pre-emptive bar to the court even requesting surrender – it doesn’t resolve a conflict, but rather prevents one arising in the first place.  But Article 98(2) also prescribes the type of agreements that are covered and envisions “a sending State” as the type of state that can benefit from protection.

Why include such a convoluted provision?  When states drafted the statute, they knew that those who ratified the ICC Statute would be obliged to detain and surrender persons when the court requested.  But they also realized that sometimes a state might have other obligations that it would breach by doing so.

To avoid a “damned if you do, damned if you don’t” situation, states built in mechanisms to manage these conflicts.  Article 98(1) covered immunities under general international law (state and diplomatic).  And Article 98(2) covered agreements under which one state has sent its personnel (the “sending state”) to another country, and the host state has agreed that it will not exercise jurisdiction over those personnel unless the sending state consents.

Such provisions are a standard component of Status of Forces Agreements (SOFAs), under which military forces are deployed abroad. The U.S. alone has over 100 SOFAs.  Indeed, the language of “sending state” comes directly from these SOFAs; and it is to these types of agreements, where one state has “sent” its personnel to another state for a specific purpose, that Article 98(2) applies.

The difficulty faced by the U.S. with its BIAs is that, by attempting to cover private U.S. citizens, or former government employees and contractors years after they retired, the BIAs go well beyond any plausible interpretation of Article 98(2).  The U.S. can’t claim to be “a sending state” for Joe Average, who is escaping a Wisconsin winter for a week of sunning himself on the beach in Belize.  As one foreign government lawyer quipped, “’sending state’ doesn’t include the state’s travel agents.”  These BIAs simply do not fall within the language of Article 98(2), at least where they are applied beyond deployed military personnel or current officials on government business.

In reality, the difficulty is faced not by the U.S. but by the other state.  Since the agreements do not fall within the scope of Article 98(2), there is nothing to stop the court from requesting surrender of a person, and the state party is obliged to comply.  But this does not invalidate the BIA.  It just puts the state in the very position that Article 98(2) was trying to avoid, of being forced to choose which obligation to violate:  its obligation to surrender under the ICC Statute; or its obligation not to surrender under the BIA.  Damned if they do; damned if they don’t.

So, what now, Mr. Bolton?

Bolton said he wants to make a renewed push to execute more BIAs. What are the prospects for and implications of this?  On both fronts, they are limited.

The agreements are most effective when signed with states that bear no obligations to cooperate with the court, where they would prevent a state from voluntarily surrendering a person to the ICC. But they do not absolve an ICC state party from its obligation to surrender an ordinary U.S. citizen or former official to the court. And deployed personnel are likely already protected by the U.S.’s broad network of conventional SOFAs (while other officials on government business may be covered by SOMAs,  analogous agreements for missions with substantial civilian components which also likely fall within Article 98(2)).

What is more, as has been noted, the U.S. already has over 100 of these BIAs.  Most states that wanted to sign a BIA have already done so; and those that were reluctant to enter into such an agreement 15 years ago are unlikely to do so now.  There may be a handful of states that would sign a BIA to show their new-found disdain for the court.  But I can’t see the U.S. signing an anti-ICC partnership with Myanmar (regardless of state media’s attempts to make common cause with Bolton on this issue).

That said, the court is not protected by the same enthusiasm that it enjoyed in its early years:  Alex Whiting has highlighted the current weakness of the court, compared with the groundswell of support in 2002-2004.  Yet active criticism of the court is primarily from African states parties, many of whom have already signed BIAs.  Many traditional allies of the court (and of the U.S.) have already come out in support of it following Bolton’s threat.  Indeed, doing so is a cheap, easy, and visible way for those states to take a stand against a U.S. administration suffering from substantially diminished stature on the world stage.  And there is little reason to believe that the silence of others will translate into a willingness to actively undermine the court by signing a BIA.

In short, looking at the geopolitical map, it is difficult to identify many states that (1) are sufficiently disenchanted with the Court that they will assist the U.S. in undermining it, (2) the U.S. would be willing to enter into an agreement with, and (3) have not already done so.  Might the U.S. pick off another couple of states?  Perhaps.  But absent Hungary and Poland deciding that this is an opportunistic way to stick it to the rest of Europe, adding a handful more BIAs is unlikely to damage the court in any real way.  The impact is likely to be greater for the U.S., which again looks hyperbolic and detached from current reality in its dealings with international institutions – using the last decade’s weapons to fight the last decade’s fights.

The fact that they are unlikely to be effective doesn’t mean that Bolton’s threat to pursue such agreements is an empty one, however.  If the U.S. does pressure states to sign new BIAs, officials and civil society in those countries should remember that there are options other than outright rejection or simply accepting the U.S. template (each of which may be politically unpalatable).  States, especially ICC states parties, can and should insist that the terms of any agreement match the scope of Article 98(2), i.e. that they cover only deployed military personnel and current officials on government business.  Agreeing to the blanket immunity from ICC jurisdiction for all nationals is not only bad policy; it would place any state party at risk of the very legal jeopardy that they were trying to avoid when they drafted Article 98(2) in the first place.

Image: Ugandan troops in Central African Republic exhume the remains of top Lord’s Resistance Army commander Okot Odhiambo, once a Joseph Kony deputy, in March 2015, under International Criminal Court protocols. Odhiambo was among the LRA leadership indicted by the ICC for alleged war crimes. (Photo by Brent Stirton/Getty Images)