Saturday, December 26, 2009



Security Certificates: Time for a Plan B
Cross-referencing: National Security Law (NSL), Ch. 10 pp. 410 et seq. and Ch. 14, pp. 569 et seq.

In an October 2009 opinion editorial in the Globe and Mail, I wrote that “[t]he security certificate system as a tool of anti-terrorism is dead.” This op-ed followed the dismissal of the certificate against Adil Charkoui, upon withdrawal of much of the secret information deployed against him. The latter act, in turn, followed a court order to disclose more information than the government (and the Canadian Security Intelligence Service, in particular) believed that it could without prejudicing national security interests. (CSIS director Richard Fadden has since described this decision as follows: “the demands for disclosure in the Charkaoui security certificate case pushed us beyond what we could accept. We were faced with a pretty fundamental dilemma: to disclose information that would have given would-be terrorists a virtual road map to our tradecraft and sources; or to withdraw that information from the case, causing a security certificate to collapse. We chose the path that would cause the least long-term damage to Canada and withdrew the information.”)

In my op-ed, I noted that four other security certificate cases “continue, and the government could ‘win’ a few in the short-term. But even if it demonstrates a reasonable basis for its allegations, the saga will continue – security certificates are supposed to be about deportation. And that prospect seems vanishingly remote because of the risk that the four remaining individuals will be tortured if deported. All of this means that the government will inevitably need a ‘plan B’ for the future.”

It seems useful to expand on my contentions in this blog, in more than the 800 words permitted for the op-ed.

The Objectives of Security Certificates

Security certificates were (and are) supposed to about removal of security threats from Canada. Detention, or other restrictions on liberty, associated with the security certificate are supposed to be temporary, persisting only up to removal. That removal is supposed to be speedier under a security certificate than are regular inadmissibility proceedings in front of immigration adjudicators. A security certificate goes directly to a Federal Court judge (whose decision on the reasonableness of the certificate is final, and amounts to a removal order). Immigration adjudicator decisions are subject to more layers of possible review.

Of course, none of these objectives have been met in relation to the five (now, four, post-Charkaoui) security certificates issued against suspected terrorists/terrorist affiliates. Unlike in past cases involving, e.g., Russian sleeper agents, removal has been vigorously contested by these five, each of whom risks removal to a jurisdiction with (at best) a spotty record on torture. Effectively, the Canadian security certificate labels (or reveals, depending on the truth of the matter) these individuals as members of organizations whose purposes are invariably at variance with the interests of their countries of origin to which they may be removed. In these circumstances, it is to be suspected that authorities in those states will have their own agendas to pursue with these individuals, giving rise to a prospect of torture that does not exist with, e.g., the removal of Russian sleeper agents.

In the result, these individuals have been detained (or subject to strict conditions on their liberty) for very long periods of time as each dimension of the security certificate system is tested up and down the court system. For these individuals, caught between the rock of removal to torture and the hard place of limitations on liberty in Canada, security certificates have become a de facto system of “control orders”, analogous in effect to the system existing in the United Kingdom and Australia.

This scenario seems likely to recur each time a security certificate is used against an individual suspected of being a terrorist/terrorist affiliate, and removal is to a country that a) has a poor record on torture and b) whose own national security interests are engaged by the individual’s suspected acts or affiliations. Moreover, this scenario is likely to recur even if, instead of security certificate, the government resorts to the regular inadmissibility proceedings under the Immigration and Refugee Protection Act (IRPA). This may be a different system, but it creates the same incentive to contest removal, and the same constitutional issues in terms of detention and removal to torture.

The Inadequacies of Immigration Law

This raises the question of whether immigration law can ever be used as an effective tool of anti-terrorism. (There is an argument that, factually, immigration remedies just displace the security risk – fortifying the Canadian border does not reduce the playing field for terrorists elsewhere. Still, I can see how displacement may still be strongly desirable, especially if it disrupts a terrorist agenda.)

Examining the question from the narrower perspective of “legally, can immigration law ever be used as an effective tool of anti-terrorism” my answer would be: Yes, to the extent immigration law can be used to, e.g., pre-screen suspected terrorists before arrival in Canada (for example, security checks and the issuance of visas). Yes, to the extent suspected terrorists/terrorist affiliates who do arrive in Canada are subject to removal to a country that doesn’t torture. No, in other circumstances.

No Removal to Torture, Period

To expand on this last statement: the IRPA at present does permit removal where national security interests are substantial enough, even where there is a serious prospect of torture. But international law does not. As discussed in NSL, Article 3 of the Torture Convention is unequivocal: no removal, no matter what, “where there are substantial grounds for believing that he would be in danger of being subjected to torture.” Article 3 further specifies: “For the purpose of determining whether there are such grounds, the competent authorities shall take into account all relevant considerations including, where applicable, the existence in the State concerned of a consistent pattern of gross, flagrant or mass violations of human rights.” (It is worth noting that the UN Human Rights Committee has construed the International Covenant on Civil and Political Rights, Art. 7, as containing an equivalent restriction).

Accordingly, any application of the IRPA (or the Charter of Rights and Freedoms) that would allow removal in the face of these “substantial grounds for believing” would violate our international obligations. It really is that simple.

Much ado has been made about the Supreme Court’s 2002 decision in Suresh that seemed to open the door a crack under the Charter to removal to torture where the extenuating circumstances were dire enough. I think, personally, that that crack no longer exists. Suresh wasn’t dealing squarely with the issue – the comment was obiter. Since 2002, the courts seem to have become much less sympathetic to government national security claims. Most importantly, as I argue at p.577 of NSL, the Supreme Court has also since reaffirmed its commitment to interpret the Charter in keeping with Canada’s international obligations. And, as noted, those international obligations are unambiguous.

No Guarantees from Assurances

A frequently cited solution to the conundrum of removal to torture is the concept of “diplomatic assurances”; essentially, supplementary promises from the receiving state that it won’t abuse the person upon his or her return. If credible, these assurances would vitiate the “substantial grounds to believe” the person will be tortured. As discussed in NSL, p.576, assurances are regarded as doubtful guarantees by the human rights community and attracted negative commentary by the Supreme Court in Suresh (and in subsequent Federal Court cases).

The assurance concept has been pursued most earnestly in the United Kingdom. Human Rights Watch has prepared an overview of this policy. This approach earned a conditional blessing from the House of Lords in its 2009 decision of RB v. Secretary of State. Essentially, whether the assurance displaces the reasonable grounds to believe threshold is a question of fact, dependent (among other things) on the state in question and the nature of the assurance. (See in particular, the reasoning of Lord Phillips in paragraphs 106-126, Lord Hoffmann in paragraphs 182-194 and Lord Hope in paragraphs 235-242).

However, I find it hard to believe that those facts ever add up to a viable assurance in the Canadian context. The problem with assurances, of course, is that they constitute a promise not to do something that is already illegal in conventional and customary international law and, invariably, the law (and perhaps even the constitution) of the torturing country. It strains credulity that when these robust sources of obligation prove ineffectual, some supplementary memorandum of understanding will prove effective, especially if that understanding includes no on-going monitoring by the deporting country of the returnee’s status. The following are, for me, a sobering and damning indictment of assurances:

First, the assurance is not enforceable in international law. It is not a treaty obligation. The deporting state would have no standing to enforce it, even assuming that there was an international venue competent to assume jurisdiction over the dispute.

Second, a diplomatic assurance issued by one branch of government – the foreign ministry, for example – may be done completely in good faith and have no impact whatsoever over the actions of the security services. As CSIS director Fadden commented in his recent speech, “in some countries the intelligence agencies are more influential than their foreign ministries”. Whether a security service will feel bound by a diplomatic assurance entered into by some other branch of government is no sure bet.

Third, even if the security service itself agrees to the assurance, compliance depends entirely on a cost-benefit analysis: what is the cost of non-compliance against the benefits of violating the agreement. A country that views its key national security interests imperiled by the actions of the terrorist entity with which the returnee is implicated and which takes the view that extreme interrogation/intimidation is in its interests may have strong incentives to violate the agreement. The countervail is the disrepute that country would then be held in by the deporting country (assuming the latter even found out about the abuse, an uncertain possibility). That disrepute might imperil future returns and other forms of intelligence cooperation.

This may be a reasonably potent countervail where the assurance is given to, e.g., the United States or the United Kingdom – states with whom the torturing state may have important relationships. Canada’s place in the hierarchy of global power is very different, and the countervail is accordingly much weaker. Moreover, for a long time, the government has repeatedly urged (in submissions on the need to protect foreign-provided intelligence from disclosure and elsewhere) that Canada is a practical supplicant in the world of information-sharing. All told, the government’s evidence in Canada Evidence Act and security certificate cases is that we receive more intelligence from our partners than we give – we need them more than they need us. In Mr. Fadden’s words: “just as we have diplomatic links with countries with poor human rights records, so must there be intelligence links. If Canada’s only intelligence sharing took place with countries that had pristine human rights records, there would be little we could do to track threats across a turbulent planet.” It seems very unlikely indeed that we would truly imperil or curtail a relationship with a foreign state’s intelligence services over the treatment of someone we ourselves regard as a security threat.

Also of note, our record in terms of forestalling torture appears to be spotty, even where we do (presumably) have influence. The record of treatment of battlefield detainees transferred from the Canadian Forces to Afghan authorities in Afghanistan is now a matter of some controversy, but it seems very likely that some number of them have been maltreated. This is true even with the prisoner transfer arrangements concluded between Canada and Afghanistan – arrangements that while imperfect, are probably more demanding than anything likely to be found in a diplomatic assurance.

In sum, diplomatic assurances are an unsustainable solution to the inadequacies of immigration law as a tool of anti-terrorism.

Plan B

So the solution must come from a completely different body of the law. RCMP Commissioner Elliot championed a criminal law-led approach in an October 2009 speech, and it is a view I also expressed in my op-ed.

To be clear, the criminal law is no solution to the dilemmas posed by the five anti-terrorism security certificates that have proven so contentious (that is, those of Almrei, Charkaoui, Harkat, Jaballah, and Mahjoub). The alleged behaviour that has been used to implicate these individuals pre-dates the anti-terrorism provisions in the Criminal Code and Security of Information Act added by Bill C-36 in 2001. This criminal law cannot, for constitutional reasons, be applied retroactively.

Still, these individuals have been incarcerated and subject to restraints on liberty for length periods of time, and courts have now universally agreed that that fact attenuates the security threat they are alleged to pose. Put another way, if they were dangerous, they are now less dangerous and that danger can be contained by measures short of detention and even the initially very strict limitations on liberty imposed on their initial release. Plan B should not, therefore, be driven by a preoccupation with these five individuals – it risks contorting policy to fit the peculiarities of a handful of cases.

The Criminal Law’s Reach

Critically, the criminal law now reaches very far in terms of terrorist activity. As discussed in NSL chapter 7, it has a strong preemptive dimension, encompassing behaviour that once would have been unassailable in criminal law because it was “merely preparatory” to an actual offence. Indeed, it captures much – if not all – of the sorts of actions apparently at issue in the security certificate cases. Note, for example, the provisions included (rather incongruously) in the Security of Information Act:

20. (1) Every person commits an offence who, at the direction of, for the benefit of or in association with a foreign entity or a terrorist group, induces or attempts to induce, by threat, accusation, menace or violence, any person to do anything or to cause anything to be done

(a) that is for the purpose of increasing the capacity of a foreign entity or a terrorist group to harm Canadian interests [a very broadly defined concept, set out in section 3]; or

(b) that is reasonably likely to harm Canadian interests.

(2) A person commits an offence under subsection (1) whether or not the threat, accusation, menace or violence occurred in Canada.

(3) Every person who commits an offence under subsection (1) is guilty of an indictable offence and is liable to imprisonment for life.

22. (1) Every person commits an offence who, for the purpose of committing an offence under … 20(1), does anything that is specifically directed towards or specifically done in preparation of the commission of the offence, including

(a) entering Canada at the direction of or for the benefit of a foreign entity, a terrorist group or a foreign economic entity; …

(2) Every person who commits an offence under subsection (1) is guilty of an indictable offence and is liable to imprisonment for a term of not more than two years.

Note also this provision from the Criminal Code:

83.18 (1) Every one who knowingly participates in or contributes to, directly or indirectly, any activity of a terrorist group for the purpose of enhancing the ability of any terrorist group to facilitate or carry out a terrorist activity is guilty of an indictable offence and liable to imprisonment for a term not exceeding ten years.

(3) Participating in or contributing to an activity of a terrorist group includes … (d) entering or remaining in any country for the benefit of, at the direction of or in association with a terrorist group …

Under the latter provision, entering Canada to serve as a “sleeper” for a terrorist group is almost certainly a crime. Under the former provision, entering Canada for the purpose of extorting, e.g., Tamils to provide financial support to the Tamil Tigers (a terrorism financing offence) would likely be guilty of a crime.

Proving Crimes

Proving criminal offences is another story, and is the true disincentive to pursuing this Plan B. Criminal proceedings are open – much more so than the security certificate proceeding – and disclosure rules are robust (although the difference between the two bodies of rules has been narrowed by the Supreme Court’s Charkaoui II ruling).

Accordingly, if the evidence to be deployed against an individual comes from an allied security service, extreme sensitivity over its disclosure would likely terminate a prosecution – as discussed in NSL, chapter 10, the “third party rule” and “originator control” make disclosure of foreign-provided intelligence a damning sin in intelligence sharing circles. Likewise, domestic security services are extremely wary of their own sources and techniques being dragged into open court. The Canada Evidence Act does provide a means for protecting this information from disclosure. It also allows a trial judge to toss criminal charges if non-disclosure would produce an unfair trial.

This reality makes criminal trials possible in the following circumstances:

1. The information supplying state consents to the disclosure of the key inculpatory information; or,

2. In the absence of that consent, a security-sanitized summary of the evidence can be prepared that accommodates the national security preoccupation while not materially impairing the fairness of the trial; or

3. The Canadian security services have enough inculpatory evidence of their own which does not prejudice their own sources and techniques to support a conviction (note that the common law has long recognized informer privilege in criminal trials).

(There is, of course, a fourth possibility: Canada ignores the third party rule, and uses foreign-provided intelligence in criminal prosecutions without consent. Such a course of action would almost certainly destroy Canada’s relationship with foreign intelligence services, putting in great peril further information sharing of a sort that may be absolutely critical to Canada. I do not, therefore, believe that this fourth possibility merits discussion).

Option 3 appears to be in play in the “Toronto 18” proceedings. Option 2, in effect, arose in the Khawaja case (through the Canada Evidence Act). Option 1 is the most thorny issue. The bottom line is that every state faces the dilemma of the third party rule in terrorism cases – modern terrorism straddles borders and implicates a whole web of information sharing networks. To the extent each state treats the third party rule as absolute – and refuses to consent to disclosure in the court proceedings of another state – each pursues a policy that may, applied against it, imperil its own prosecutions. The result is a sort of intelligence “beggar thy neighbour” approach.

In these circumstances, states must consider requests for disclosure of their shared intelligence carefully (and not dismiss them out of hand, or demand secrecy for even the most banal information). Indeed, it seems wise to develop protocols on how and where information can be disclosed in court proceedings, perhaps concluded initially between states with similar legal systems. These protocols would vitiate CSIS’s complaint in some of the security certificate cases that even asking for permission to disclose foreign provided intelligence creates the impression that the agency is soft on secrecy, and not to be trusted.

(As an aside, I would not support the introduction of IRPA-style special advocates in criminal cases. It departs too radically from our criminal law tradition to exclude the accused from his or her own trial – I can’t see it being desirable or constitutional or (given recent successful prosecutions in Canada) necessary. That said, I think there is a role for special advocates to represent the interests of the accused in the collateral Canada Evidence Act proceedings where decisions on the disclosure of secret evidence to the accused are made).

Preventive Detention?

Objections to a Plan B focused on the criminal law also often urge that there will be instances where actions needs to be taken immediately, perhaps even precipitously, to disrupt a terrorist plot before evidence of a quantity or calibre to support a conviction beyond a reasonable doubt has been collected.

As noted, the post-2001 terrorism provisions reach very far into once “merely preparatory” behaviour. It stands to reason that these urgent scenarios arise less often if behaviour several steps removed from actual terrorist violence is criminalized. But of course, urgent situations where the criminal law is not yet engaged will still arise, even if more rarely.

In these circumstances, the state does need a tool. As discussed in NSL, chapter 14, the United Kingdom and Australia both have tools of preventive detention and so-called “control orders” that remove or limit the liberty of action of terrorist suspects. Comparative systems of preemptive detention – or detention without charge – are described in a report by the UK human rights organization, Liberty.

Canada’s system of so-called “preventive detention” – sunsetted several years ago – is modest in comparison. The bill currently before Parliament would restore this same system. As I will describe in a supplemental blog, I believe that this system is appropriate (although the bill deserves some tightening up). I think it is enough, in the absence of a very compelling case grounded in the actual facts of the Canadian security situation that something more draconian is necessary.


In sum, I believe that as a legal matter, the tools for an effective system of anti-terrorism are available, even though the back of the immigration approach is broken. I defer to RCMP Commissioner Elliot on whether the sort of plan B that I support – a back to basics criminal law approach – is adequately financed. His view is that law enforcement has been the poor cousin in the post-9/11 terrorism strategy. His opinions on this point are certainly consistent with what I’ve heard from others. All of this leads to my final (and perhaps banal) conclusion: Rather than throwing more money at security certificates, we should now be better financing criminal investigations.